For many people, listening to music elicits such an emotional response that the idea of dredging it for statistics and structure can seem odd or even misguided. But knowing these patterns can give one a deeper more fundamental sense for how music works; for me this makes listening to music a lot more interesting. Of course, if you play an instrument or want to write songs, being aware of these things is obviously of great practical importance.
In this article, we’ll look at the statistics gathered from 1300 choruses, verses, etc. of popular songs to discover the answer to a few basic questions. First we’ll look at the relative popularity of different chords based on the frequency that they appear in the chord progressions of popular music. Then we’ll begin to look at the relationship that different chords have with one another. For example, if a chord is found in a song, what can we say about the probability for what the next chord will be that comes after it?
The Database
To make quantitative statements about music you need to have data; lots of it. Guitar tab websites have tons of information about the chord progressions that songs use, but the quality is not very high. Just as important, the information is not in a format suitable for gathering statistics. So, over the past 2 years we’ve been slowly and painstakingly building up a database of songs taken mainly from the billboard 100 and analyzing them 1 at a time. At the moment the database of songs has over 1300 entries indexed. The genre and where they are taken from is important. This is an analysis of mainly “popular” music, not jazz or classical, so the results are not meant to be treated as universal. If you’re interested, you can check out the database here. The entries contain raw information about the chords and melody, while throwing out information about the arrangement and instrumentation.
We can use the information in the song database to answer all sorts of questions. In this introductory post, I’ll look at a few interesting preliminary results, but we invite you to propose your own questions in the comments at the end of the article.
Let’s get started.
Are some chords more commonly used than others?
This seems like such a basic question, but the answer doesn’t actually tell us much because songs are written in different keys. A song written in the key of C♯ will have lots of C# chords in it, while a song written in G will probably have lots of G‘s. That G chords are more popular than C♯ chords is likely only a reflection of the fact that it’s easier to play on the guitar and piano. So instead of answering this meaningless question, I’ll answer the slightly more interesting one of, what keys are most popular for the songs in the database?
C (and its relative minor, A) are the most common by far. After that there is a general trend favoring key signatures with less sharps and flats but this is not universal. E♭ with three flats, for instance, is slightly (though not statistically significantly) more common than F with only one flat. B♭ only has two flats but is way at the end of the popularity scale with only 4% of songs using that as the key.
What are the most common chords? Part 2
It’s much more interesting to look at songs written in a single common key. That way direct comparisons are possible and more illuminating. We transposed every song in the database to be in the key of C to make them directly comparable. Then we looked at the number of chord progressions that contained a given chord. Below we’ve plotted the relative frequency that different chords occurred in descending order.
As expected, C major is a very common chord for songs written in C (it’s the I chord in Roman numeral or Nashville Number notation), but F major and G major (the IV and V respectively) are used just as often. Interestingly, F and G actually show up in more chord progressions than C! C major is the tonal center and one might expect it to be ubiquitous, but it turns out to be pretty common to omit this chord in some sections of a song for effect. “My Heart Will Go On” by Celine Dion is one of many examples in the database that exhibit this behavior. Clicking on the above link will take you to the song’s entry in the database and show you that of the two sections that were analyzed (the chorus, and the verse), only one contains a C.
The A minor chord is the next most popular, but after that there is a significant drop off in use. If you’ve ever heard someone complain about the “four chord pop song”, this is what they are talking about.
Is there a reasonable explanation for the relative popularity of certain chords?
Why are A minor chords so popular but A major chords practically non existent? There won’t always be easy answers, but in this case these results can easily be explained with some basic music theory. A discussion of this is out of the scope of this post, but we’ll definitely explore the music theory behind this in future articles.
Even if you don’t know the music theory behind this yet, there is a lot of practical information to take away. If your song is written in C and you want it to sound good, you probably shouldn’t use any A major chords unless you really know what you’re doing. Better stick with A minor, for example.
The team over at Apple, Inc. evidently know their music theory. Their latest version of GarageBand lets you play with “Smart Instruments” that “make you sound like an expert musician… even if you’ve never player a note before.”
I’m skeptical of their claims, but look at the chords they’ve chosen for these “Smart Instruments”:
Don’t those chords look familiar? Based on what our database is showing, I might suggest some small changes.
In particular, Bdim, while diatonic in C, is much less common than some other chords, like D, and E. Perhaps in the next version of garageband, Apple will fix this (they really should).
However, overall Apple is making good choices for the chords that the average “garage band musician” might want to start with.
If a song happens to use a particular chord, what chord is most likely to come next?
The previous question took an overall look at the relative popularity of different chords, but we can also look at the relationship that different chords have to one another. For example, a great question to ask is, if a song happens to use a particular chord, what chord is most likely to come next? Is it random, or will certain chords sound better than others and thus be more likely to show up in the popular songs that make up our database?
There are a lot of relationships to analyze, but we’ll start it off by looking at just one for now: For songs written in C, what chords are most likely to come after an E minor chord? The relative popularity of what the next chord will be is shown below:
This result is striking. If you write a song in C with an E minor in it, you should probably think very hard if you want to put a chord that is anything other than A minor or F major after the E minor. For the songs in the database, 93% of the time one of these two chords came next!
There are lot of interesting questions to ask, and we want to know what is most interesting to you. Let us know in the comments below.
Other posts in the 1300 song series
Comments
516 responses to “I analyzed the chords of 1300 popular songs for patterns. This is what I found.”
This is really awesome. Must have been a ton of work
Very interesting! Although, for most common chords chart, I think that instead of simply counting the number of times a chord appears, you should have also considered the duration of that chord in the song.
Didn’t Pandora radio did the same analysis for its recommendation engine?
That’s a good recommendation. I have a feeling it wouldn’t change the results much for pop songs, but it’s something to try!
I know Pandora has done some analysis like this for their database, but I thought it was limited to things like major or minor tonality, upbeat tempo, etc. and didn’t delve as much into the nitty gritty harmony. One reason for this might be that these patterns are so universal (spanning lots of genres), that it might not be too helpful for determining what types of music people like. I could be wrong about this though.
Hi Dave, sure lots of work sure on “Pop” music but I’m really lamenting the ever disappearing “music” on modern youth as generations go by. The study probably encourages the easy way out for nowdays would-be-guitarists.bouncing off each other and believing they are making good music that might be remembered for 6 months if they’re lucky. Unless “accompanying” video clips are spectacular their “music” doesnt climb the charts. Pity the genre of Andrew Lloyd Webber, that supposed fool Charlie Chaplin “Smile” 1936, “You Needed Me” 1978, “Embracable you” 1928 and many more doesnt get someones attention because of the commercial push as in “pop”. Did you see the revealing documentary involving “Sting” on how the musical brain is only stimulated by a limited range of chord progressions? It crystalliized for me why modern music is rubbish. Did you know that probably the best music last century was born from the hard times during two world wars and the great depression? Diminished 7ths and augmented 5ths didn’t get a mention in your chord chart reflecting perhaps and equating to the lack of emotional depth in “pop” and its associated non-musical youth. Oh to see what your study would reveal as applied to last century music and even centuries before that. Take “masters” such as Beethoven and Bach now they KNEW their chord sequences. I wish I could honestly say keep up the good work Dave and that “Pop” music was really going somewhere musically but our “musical” brains trust is becoming more extinct, generation by generation.
What do you think about doing an n-gram type analysis for melody? Like what are common patterns, whats the progression like (in a normalized set).
Adrian,
we’ve analyzed the melody in many of the songs in the database and are definitely planning to do some analysis in this area. Stay tuned!
“stay tuned”. yes. One should always do that.
😉
You seem to know some theory but one of the odd things here is that you lightly touch on is key. I would say the frequency of commonly related chords (relative to the key) is more important than what chords are most common. Apple choose those chords as they are the relative chords in a major C scale (or A minor scale) which is a completely logical way to approach what they are doing.
So for folks not familiar with music theory a melody could be written over chord changes that go from G major to C major. That same melody could be transposed to the key of A and where the chords are now A major to D major. It’s the chords position relative (and sometimes the lack of a chords relativity) to the key that’s important.
Key is probably mostly a function of where the singer can sing the melody successfully with most singers choosing to record the key highest in their range as it’s assumed people with voices pitched higher will be more successful on the charts although that’s somewhat a function of genre as well as other factors.
Sorry, didn’t think this posted. Got an error back after submitting.
Also key depends on the instrument that it’s written on/for. A guitarist is far more likely to write in E, a pianist in C. Music written for brass instruments are often F or B flat. Those are the easiest keys to play for those instruments.
I disagree C is the easiest key on the piano, while it may seem that way from a smarting pupil, its not, the more black keys in the key, the easier it is.
In C, because you have no context of feeling (no black keys), you tend to look at the keyboard more to see which white key you’re hitting. In say, Db, the black keys provide context to your fingers without looking.
Completely agree. I also started in C. Once I started playing scales in B major I quickly came to wonder why they didn’t start with B.
you guys are speaking from a highly technical point of view. When you take into account only the ease of viewing a keyboard and not which keys are most “comfortable” or “fast” for your fingers, most beginners in fact prefer to start without learning about what the black keys are for.
E followed by A are the two most common keys on the electric guitar. Whether the singer is female or male has an effect on which one is more likely. On the acoustic G is the most common key due to the open chords available.
Sure didn’t see a lot of 9ths, 11’s or 13ths in that list. No surprise. This just backs up my contention that copy right on most music is insane. Performance copyrights excluded.
This is sadly a very, very common misconception about music, which comes mostly from a culture of amateur musicianship (even at professional levels).
This is neither an innate characteristic of all music nor one in which we should base actual laws.
Music made by people without much knowledge of their instrument will most likely play things that are easy to play but listen to more refined types of music (note I do not think they are “better” though!”) like jazz or classical music (specifically anything after the 17th century) and those preferences go away.
Music is not solely, not even mainly, about the notes. Similar chord progressions and melodies can be sung to evoke teenage swagger or intense grief; slightly differences can make huge changes in the way it’s perceived; in any case, be it a pop song or a 12 hour opera, a personal work that took effort and creativity (no matter the amount) should be the creator’s property.
Your argument would be comparable to someone saying that since all books have similar grammar then copyright should not exist. Insane indeed.
Wrong. As a brass player for over 20 years, I can safely say that the easiest key to play in for trumpet is by far E major (concert D). Just because the instrument is in Bb does not mean it is the easiest key for the instrument.
“A guitarist is far more likely to write in E….. Those are the easiest keys to play for those instruments.”
That’s why God invented the capo. :-))
For a guitarist, the number of flats and sharps is irrelevant, since changing key is mostly a matter of moving up and down the fretboard with or without a capo.
Also note that many guitarists tune down a half step, resulting in songs in Eb rather than in E.
I don’t find a discussion of key very interesting at all. Most people don’t have perfect pitch, so they don’t know a C from a B when they hear it. I would write out all the chord progressions Nashville style, using numbers to represent the scale tones do-re-mi, etc, and then see what patterns develop. Then you’d actually see the patterns. For best results, start minor songs from 6, to keep diatonic.
I’d guess popularity would go something like: 1,5,4,6,2,3,b7,7
But 6 might come higher depending on how many minor songs you chose…..
Actually, this is exactly what they did, only using a different approach.
By transposing all the music to be in the key of C, you have C/F/G representing your “typical” I/IV/V rock song, with C/Am/F/G being your “typical” I/vi/V/VI pop song. These are directly comparable regardless of what key the song really is played in.
It’s just a notation difference; you could read the popular chords chart as 4,5,1,6m,2m,3m,3,2,b7,6 if you prefer.
That’s I/vi/V/IV for pop, of course; Roman numeral fail.
Sure, but doesn’t it worry you that they had to reinvent this particular music-theory wheel, instead of just knowing that they could use roman-numeral notation? It does make me wonder what else they might have missed…
OK, guess I should have read further in the comments (the subtle trap of nested comments/replies!)
Isn’t it normal what you found? Given the rules of classical harmony? The only interesting (but very predictable) part was that most of the songs are written in the (for non transposed instruments) most simple key!
Yeah. These rules are very simple and are expected from classical harmony. I thought I’d start with the basic stuff first. I think a lot of people don’t realize how prevalent these classical harmony “rules” are in popular music.
The most common chords are dictated by basic harmony, but the most common sequences – now you have some really interesting data there, please publish more! also, it might make more sense to indicate I ii iii IV V vi vii in addition to C G F etc as the chord names are just coincidental (except in the initial analysis of the most popular keys). I’d love to see the graph for each chord (we’ve got iii, let’s see what follows each chord in the scale)
Here’s a great video ‘Determining Chord Progressions’:
http://www.youtube.com/watch?v=fXIEmMDwc7E&feature=relmfu
My theory lessons too:
http://www.youtube.com/playlist?list=PLC315FC2AF8D6BA56&feature=plcp
Karen Cuneo Ramirez is one of my favorite person when it comes to practical music theory
this is common knowledge.CGFAm – these chords make music sound ‘pretty’…… so all pop songs have them.
Actually, the chord progression that you’re probably thinking of is C-G-Am-F, I-V-vi-IV.
I thought about writing a song entitled I-V-vi-IV (of course using I-V-vi-IV), paying “homage” to the progression, capped with a bridge medley of some of the most well-known examples…
Yep — the “Don’t Stop Believin’” progression.
You seem to know theory but it seems odd that you are looking at actual chords and their frequency and not the chords relationship to the scale which would seem to me to be more relevant.
By way of example I would say analyzing songs and finding that popular songs are frequent with a I chord and a minor IV chord verse is of more value than saying there are a lot of hit’s with C to A minor verse progressions.
I would think the actual key is more a component of the singer’s range than anything about the songwriting process as any song can be easily transposed. Although it would be interesting if hit songs in the key of C were prevalent especially if the singers tended to sing in a particular range.
By range I mean octave.
And I meant the root chord to a minor 6 chord not root to minor 4 chord.
Hey Chuck,
this is actually exactly what we are doing. I didn’t want to use Roman Numerals to refer to the chords because a lot of people aren’t familiar with this terminology. This is why I transposed all the songs to the key of C.
a real markov model of chord moves would be nice! I’ve been wanting to do that for years but was too busy or lazy 🙂 although a friend of mine did this manually for a musician he liked, then used the most popular moves on songs he wrote. he’s not very successful, but I don’t blame the chord moves 🙂
Dave, thanks for the work you’re doing – it’s very interesting.
I do disagree about common understanding of Roman Numerals for chords though. It’s really only rote players that don’t know or need them. For all jazz, blues, rock and folk players and for all composers Roman numerals for chord are standard. Just google up “I-IV-V” 😉
Thanks again,
Dave
“93% of the time…”
Yes, but which 93%? If it happens that all of the best music is in the 7% that doesn’t, this doesn’t help me!
True, as a songwriter I would also be more interested in writing a song that lasts. We still listen to The Beatles because they wrote great chord changes. Take a look at the tabs for “In My Life” or “No Reply”. Just super musical and one of the reasons they’re one of the few boy bands to last.
I’d argue that the reason the songs are great is not because they wrote great chord changes, but because they combined those particular chord changes with a particular melody, a certain rhythm and a deliberate instrumentation/production, the combination from which emerged a great song.
That is to say that focusing only on harmony in an attempt to decode what makes for good music is akin to saying that what most accounts for great literature is the frequency of letters.
That’s a disingenuous comparison. Letters in words have no inherent meaning unto themselves; words themselves will be largely meaningless without the context of a sentence. Analyzing frequency of letters would be a surface analysis at best.
Harmonic structure is not the surface of a piece of music; it creates an underlying backbone for the melodic and rhythmic elements to work from. It is structural, not decorative.
I have no qualms using this as a starting point for deeper analysis, and am excited to see this series continue.
Right. Perhaps a better comparison would be between harmonic progression and gramatical structure.
Goovitude, it is true that a lot of guitar and piano players, when composing, find a chord (harmonic) progression that they like and then locate a melody within that structure. However, historically, melody has been the king of music composition. Most of our modern concepts of chord progressions (harmony) are based on Bach’s rules of counterpoint, which were all about how to support/harmonize the melody.
When trying to convey the basic essence of a song, someone is far more likely to sing/play the key melodic motif than they are to strum/play the chords or hum/play the chord roots. You can rearrange the chords of a song but keep the melody and everyone will still recognize the song. But if you keep the chords and change the melody, the original song would be unidentifiable.
But of course chord progressions are, as you say, very important.
I look forward to seeing what you’ve found… but I am a bit curious. “Popular” songs covers a lot of ground–while I’ve not done this sort of study, for example, I’d suspect that dance tunes have less variation in harmony. Could you be a bit more specific about your choices of songs to study? Also, since you mention “My Heart Will Go On”, am I correct in guessing that your transposition to C was done in a way that kept modulation in a song from making any difference?
Good catch on the modulation in My Heart Will Go On. The database entries small individual chord progression so each is analyzed with respect to the key it is currently using.
You can find most of the songs we’re using in the song analysis link at the top of the page (and linked at the beginning of the article).
At some point I’d like to hear some examples of songs you’ve run across that are a) popular and b) use atypical chord progressions or similar musical devices.
Very interesting analysis. Thanks for sharing your findings!
I’d be very interested to know if you observed patterns or deviations from the norm when comparing between music genres and perhaps even eras / decades.
Great stuff.
Yes, I think you should do an analysis relative to the key, e.g., roman numeral analysis: http://en.wikipedia.org/wiki/Roman_numeral_analysis
This would give a good concept of the organization within the key and also a good basis for chord progression analysis…
Thanks for doing this!
Very interesting. Compare your results to the circle of fifths and the chords for each key. Apple use the first seven chords in the key and add an eighth which is interesting. Is your data base available? Do you use the Chord Pro format? If not, what format do you use? What process did you use to “clean” your data base?
Eric,
most of the songs we used are can be found by clicking the “Song Analysis” link at the top of the page. We created the format ourselves, and all the entries were added using our “Music Editor” which is also linked from the top of the page.
You can configure what chords you play in GarageBand for iOS using the Smart Instruments. If you want to make changes, go right ahead!
Thanks for the tip Danny. I’ve haven’t played much with smart instruments myself. It’s good that Apple let’s you have that flexibility.
I’d love to see the most common 4-chord progressions, specifically the relative progressions regardless of Key (I’m sure C, G, Am, F is very common (this would be C Major), and I’d like to match up the songs with the same progression across keys)). This will essentially tell us which songs have the same melodies. I’m tired of people saying electronic music is unoriginal – I’d like to demonstrate just how unoriginal many rock progressions are as well!
Hey Taylor, we totally agree. In fact, all of the songs in the Hooktheory database use relative notation, (e.g., I V vi IV) to represent C G am F in the key of C, so that you can compare songs that have different keys. We also have an algorithm that figures out how similar two songs are, which is used to generate the “songs with similar chords” section. We’ll write an article soon that goes over the most common chord progression in rock/pop songs.
This is pretty awesome what you’re doing here!
Cryptography has recently piqued my interest and in the course of learning about it I have written some simple programs to analyze text for frequency of n-grams, frequency of the first letters of words, etc and I can see a bit of a parallel between written text analysis and the analysis of parts of song! Music really is sort of it’s own language isn’t it?
Is there any chance you would publish your database, so that we can write our own queries?
Yes we are working on this. Currently the public database entry for a song returns entries for songs with “similar” chord progressions. We’re looking at making it possible to do more advanced queries as well.
Don’t know if I skipped the part where you mention it, but how did you analyze the songs?Parsing midi files?
Also I’d love to see you talk more about the uncommon song structures, the ones that stand out and are still top of the pops never the less (off the top of my head, bohemian rhapsody??)
What happened to the key of B major? I don’t see it in your “most popular keys” chart.
Apparently B Major is a less common key in songs contained in our database. Howie Day’s Collide, and Jason Mraz’s I’m Yours are two songs that use B Major in our database. One possible explanation is that it’s a little more difficult to play in B on the piano. Guitarists can capo up to B; both Collide and I’m Yours use guitar as the main instrument in the rhythm section.
I don’t understand your complaint about the chords in the key of C for GarageBand. D and E are indeed there, but in the key of C they are the ii and iii chords, which are minor. A is also minor in the key of C being the vi chord.
I’d like to add to Enc and ask for a copy of the database. I mean, we could crawl http://www.hooktheory.com/analysis but that wouldn’t be as awesome as a dump from you guys.
Hmm interesting but his post is a bit pointless to me. The data eludes to the most common key which is C and is very well known to most musicians. The reason being that this is all the white keys on the piano (no flats) and easiest to play in, which is why it would be the most common and popular.
For the chords it doesn’t quite make sense because it is very biased because those are the 3 primary chords of the C major scale. So of course the most popular scale’s primary chords will be the most popular he encountered. Roman numeral analysis would have been much more useful here, as it is not dependent on the actual chord letter or key, but its overall function. For example C-F-G is generically describing a I-IV-V movement and describes all major scales in any key not just C major. Functionally this is known as root-predominate-dominate progression and is the most popular movement in all modern music. It’s cool he did his own study though, but a bit pointless in its final information since it is extremely biased to the most popular scale C major.
I find roman chord transitions much more useful like shown below, because they can use any key, and they show the most common movements in music. Below is his entire study summed up in a simple state diagram using roman notation that works for any key. These aren’t technically all those choices, just the most common ones. Using C major in the below diagram would be: I = C, ii=d, iii=e, IV=F, V=G, vi=a, vii=b(diminished) where a capitol letter is a Major chord and lowercase is minor chord.
http://www.musictheory.net/lessons/57
We’re definitely going to start using the language of Roman Numerals in future posts. A lot of folks aren’t familiar with this type of notation and we wanted to start off simple.
When you think about it, this type of analysis is a really good motivation for why thinking about music in Roman Numerals is important and useful. We’ll have more to say about this shortly.
I should also note that all the analyses in the database were done originally in Roman Numerals. If you go to the raw database, all the entries are analyzed using Roman Numerals, not chord names.
That would be excellent. Keep up the good work!
I’ve found that people like to learn the Roman numeral system but, they first need a simple step-by-step process of introduction. I put this together on my Youtube channel and, I’ve had many people sending messages of thanks (as well as some insults!).
Here’s the link:
http://www.youtube.com/playlist?list=PLC315FC2AF8D6BA56&feature=plcp
Also, Garage Band is a very good way to introduce people to the creative side of music. As for Transcribing and theory, it seems to be accepted gradually.
Justin Sandercoe gives great help, as does Marty Swartz.
I hope this helps.
Tony 😉
All very obvious results to be expected by anyone knowing basic music theory. I think it’s a nice analysis, in illustrating well known aspects of the formulaicness of music generally, I just wouldn’t call any of the results at all ‘striking’ or ‘suprising’ to any half-decent muso as you do.
To explain for you one of the results that you found ‘suprising’, it’s obvious why ‘F’ and ‘G’ ‘suprisingly’ came out ahead of ‘C’ in songs like ‘My Heart will Go On’ and thus slightly overall on average – that song is in a minor key (i.e. ‘Am’ in your terminology), so ‘C’ is _not_ really the ‘tonal center’ (contrary to your claim) in such songs, and thus it’s no great suprise that minor songs in the key of ‘Am’ wouldn’t be expected to have ‘C’ as the most common chord, and thus lumping the averages for major+minor songs in one overall average as you did (which is perfectly reasonable in itself, just you seem to have forgotten this at the point that you call ‘C’ the ‘tonal center’ there) might lead to such averages overall.
If you were able to split the results for major+minor keys (which would be a tricky and somewhat arbitrary distinction somewhat defined by precisely the averages you’re deriving, so I’m not saying you should), you could for example expect to find ‘F’ + ‘G’ + ‘C’ as generally the most common in major key songs and ‘F’ + ‘G’ + ‘Am’ in minor key ones (With ‘Am’ still being quite common in major key ones and ‘C’ quite common in minor key ones and generally of course the distinction is pretty minor between the major+minor keys).
I think the tonal center for most of My Heart Will Go On is in the major mode for the simple reason that there are no authentic cadences to A minor. You’d expect to see some E major chords.
Pop songs tend to flirt with the parallel minor a lot, however, and this is one reason why a chord progression would omit the I chord (C major).
I’m interested about songs that change key in the middle. Eg, Dream a Little Dream of Me changes from A minor to A major for the bridge (or C to A if you like). Or more popular example, Karma Police is C/Am in the verse, G/Em in the chorus, and D/Bm in the bridge/outro. Do you have to hand-annotate these changes?
I’m also curious about how you selected which songs to diagram. Are they popular songs? Classic songs? etc.
The other interesting thing about chord changes in popular music is you have two different dominant approaches. The “Germanic” approach which is what is being discussed here for the most part and the Blues/R&B/Funk approach which have a lot of atypical chord relationships where jarring juxtapositions of chords is seen as a benefit.
what do you mean with “two different dominant approaches”? Could you point me to some references?
I actually really like using A in a C major song (or using non-minor VI chord in general), though. It works wonders for many choruses, in a sequence like this:
F G/F Em (or Em7b5) A (or A7) Dm7 G7 C C7
F G/F Em (or Em7b5) A (or A7) Dm7 G7 C
That A just adds much more tension to be resolved by the Dm7 that follows…
The first chart has only 11 keys, not 12 – it’s missing B major. Were there none in your sample, or is this a glitch?
good catch. It’s very uncommon but it somehow didn’t find it’s way into the graphic. I’ll update this as soon as I can. It’s going to be a very small number of songs though.
Yeah. We definitely did not want to say you can’t use these other chords. They’re just less common and have very specific uses. In your case the A is functioning as a V of the ii chord. Do you notice that when you use it, you usually have to play that Dm chord right after or it sounds weird?
That’s what I would suspect in general. We definitely want to talk more about how these statistics relate to basic (and in this case, more advanced) harmony soon.
Allow me to make a little commercial break..
I have analyzed 12.600 songs ( mostly pop & jazz ) and made the statistics available in a app.
http://itunes.apple.com/dk/app/pro-chords-instant-inspiration/id335317698?mt=8
The result here is slightly different, but I may be because of the Jazz songs and the fact that I have used sheets where you get the complete harmony.
So the top list looks like this:
reads like( times / half notes from tonica / harmony )
1120 0
732 2 m7
727 5
719 9 m7
701 7 7
570 9 m
445 7
428 0 maj7
328 4 m7
320 0 7
275 2 m
258 4 7
In jazz & blues I vi ii V I is quite common in harmony we call it the anatole
http://en.wikipedia.org/wiki/Rhythm_changes
okay, your chart show a difference on the ponderation, but we are close. And if your analysis made false recognition, we have it 🙂
Congrats for the work. I hope you give us more interesting analysis though, once counting the chords is as useful as counting the letters of a text. What matters in harmony (guess you know) is the relationship between them and between them and the whole form and movements…
You should be focusing more on syntax of chord progressions and less on lexical stuff.
Understand the analysis but a basic music theory class would explain all of this and the fact that you leave out the music history concept (people generally write music that is similar to things they’ve heard / are ears are trained to like certain sounds and progressions relative to others)
This is great work you’ve done here. Thanks for taking the time to compile all this useful information.
As a small (haha) favor, would you analyze a raft of Tom Waits’ tunes for us? 🙂
(maybe stick to the ones that no sheet music seems available for?)
Wow, this was really cool. I will definitely be coming back to read more.
Cut the guy a break people, this is the first entry it’s going to start out at a basic level.
You might end up calculating the “average song” of those 1200 songs 🙂
Old thoughts, old patterns of behavior, old dreams, prevent us from having new ones.
I’m sure this is cool to non-musicians, but this is fundamental stuff to a musician. I don’t mean to be a wet blanket, but this just music theory 101. (Actually, it’s the first week of Music Theory 101.)
If your study helps non-musicians have a greater appreciation for music, cool. But it’s not exactly ground-breaking work.
Definitely agree that these results were expected from a very basic music theory perspective.
I do think it’s interesting to see how the statistics actually break down though, and rest assured, we’ll be going deeper.
What you said.
Can the chords progression (which chords comes one after another) be a simple by-product of how these chords can easily be chained on piano and/or guitar for example, moving a single or two fingers only toget them. Thus making them easier to play in sequence. For instance c and am, on the piano you onle move one finger from f to a. Ond so on…
It’s turns out that it’s more fundamental than that. The rules governing how chords work with one another is actually very well studied (it’s called the study of Harmony).
We’ll definitely be looking at where these rules come from in a future article.
Philibert, what you are referencing is often referred to as “voice leading,” and is definitely a significant consideration in composition. Ever notice how altos in choir often sing just 2-3 notes for the entire song? With good voice leading between traditional 3-note chords (and depending on where the melody is within the chord) you’ll often see one note remain the same, one note move along the cycle of 4ths (5ths), and one note move by a step – preferably a half-step.
Have you used McGill’s Billboard dataset?
http://ddmal.music.mcgill.ca/billboard
Thanks for putting in the work to do this. Now you have a good base line (no pun intended) for how “pop” songs work. I think what would now be REALLY interesting is to analyze the Rolling Stones 500 best songs of all time, and compare the groups. As an experienced song-writer, my guess is those songs would be way more diverse than the ones analyzed here. Some would be very simple, others rather complex. And all kinds of different chords would get used a little bit. Sure, C, G, F, Am would still get used a lot, but the Beatles and other greats knew how to throw in just the ONE out-of-key, weird chord that would make all the difference.
I’ve always been curious about this kind of data but never gained enough knowledge in music to consider doing this research. Bookmarked, I like where this is going. Keep it up.
Would love to see some further analysis and real depth to the thinking that goes into the analysis of what you’ve indexed.
What has been presented here could have been told to you by a first or second year music student.
Definitely interesting but I agree with one of the comments above. None of this is all too surprising and seems to fall almost exactly in line with basic music theory.
Also, when you were taking the number of times a chord occurs, and what comes after, is that measure by measure? So let’s say there is a I chord for four measures, does that mean in that case a I chord followed by a I chord followed by a I chord followed by a I chord? or is it blindly by the change in chord barring repetition over multiple bars?
What about passing tones, chord variations, and chord variations that create passing tones? A great song, even popular, can be created with such and can break the mold found in this data. Were there some that u found were creative and unique? I want to write like those. I think the data shown is proof of commercialized laziness in the industry. Arrangers abandon thick chords for simple ones and lose the accompaniment that the catchy melodies so desire.
Cool stuff, I think this data should both influence us to look towards trends but also be revolutionary in our writing.
For future posts, I think the Roman numeral system could of been explained and grasped easy enough in this post. I think leaving it in C could motivate the uneducated to stay that way. Let’s challenge one another towards a better understanding and learn from one another as well. #EndMeaninglessMusic #brainymusicpleez
*Were there any songs that you found that we’re both pop and unique?
This is great stuff. Thanks for sharing.
Did you see any differences between the songs of rock bands (e.g. Oasis) and pop groups (e.g. Backstreet Boys)?
My impression is that the latter type is more likey to have the more “unusual” chords (both in absolute and relative terms), due to the significant musical training of the contributing songwriters and backing band.
Also, regarding your comment above about the duration of the chord not making a difference, my amateur impression is that major chords are generally played longer than minor chords. (I can’t think of many songs that play a VIm for a prolonged duration).
Yeah. Some of the Boy Bands have some really interesting harmony.
When we analyze this stuff, sometimes we like to play the game of “did this person right this song, or did they get someone trained to do it for them”.
The minor major duration thing would be an interesting thing to look into. I might think songs that play around with the relative minor (the vi) might be likely to play it for a prolonged period of time.
[…] hooktheory.com通过分析1300首热门歌曲的和铉，发现了一些有趣的数据，如果你想自己写歌，这些数据对你或许有帮助。 […]
Hi author. Please post the TOP TWO percentage 4-chord progressions. I have a STRONG HUNCH that they are:
1. I V vi IV (e.g. C G a F)
2. i VI III bVII (e.g. a F C G)
thanks in advance
max loh, I agree with your #1. For #2, my vote is:
I – IV – V – IV
These two have current nicknames:
1. Axis of Awesome 4-chord progression
2. “Sensitive Female” Chord Progression
Somewhat related, but if anyone reading this post has never listened to ‘Four Chords’ from Axis of Awesome, it’s definitely worth a listen. 🙂
http://www.youtube.com/watch?v=oOlDewpCfZQ
Niiice!
Actually, I’m sorta surprised nobody’s mentioned the Pachelbel Canon (near as I can tell) in all this. Well, perhaps not entirely surprising since the article presumably(?) focuses on modern pop music rather than music that’s popular… period — I didn’t actually check the song list to see if it included any highly popular classical music, but nobody’s given me any indication that it does (so far).
Anyway, since you shared that video, I’ll share this older one that I love in similar vein… 😉
http://www.youtube.com/watch?v=JdxkVQy7QLM
Maybe this study series will eventually make reference to the Pachelbel Canon…
Cheers!
I remember a while back reading an article about a piece of software that the music industry uses to asses new songs. It compares it to a database of previous top songs and gives a given song a score. The artists then have to tweak their music until it breaks a certain threshold of points. Maybe you are stumbling into their secrets.
This is the least insightful article I’ve ever read.
I am reminded, reading this, that pop tunes can also be hugely successful by using very unusual chord progressions.
See, for example, this analysis of The Eagles’ “Hotel California”: http://www.icce.rug.nl/~soundscapes/VOLUME09/Locked_into_the_Hotel_California.shtml
The verse of Hotel California is |i|V|VII|IV|VI|III|iv|V – which sounds pleasant but isn’t close to any traditional chord progression!
This is an awesome analysis! Just one nit-pick:
“In particular, Bdim, while diatonic in C, is much less common than some other chords, like D, and E. Perhaps in the next version of garageband, Apple will fix this (they really should).”
In my opinion, Bdim can added in between the progression between G and Am in almost all cases and it gives a really good effect. In fact I suspect in that many songs at least the bass hits B when going from G to Am. Which brings me to the next question – did you guys consider bass chords like C/E, G/B, etc? Those bass chords add a nice touch to many songs and it would be nice to see how popularly they are used.
Goods news. We did analyze chord inversions (C/E, etc.) in these songs and will be talking about their uses in a future article.
Excellent. I was going to ask the same question.
Brilliant! I applaud your ability to data mine, however many of your questions and fatal flaws in your logic could have been answer by researching the basics of major key harmony. “Music Theory” is not a law, but an explanation of how western music uses chords in a key, and what closely or distantly related keys a song may modulate to. These can often have genre specific explanations. Take some harmony classes and then figure those “patterns” into your research and you may find some answers.
Is there a way to download the database in a nice format (preferably csv)?
Hi Jeremy,
Not at the moment. We are planning to open it up to people on a case-by-case basis depending on what they want to use if for. Pretty much any interesting / innovative use of the data would be something we’d be happy to support.
Brilliant! I really enjoyed some of the figures you came up with and your enthusiasm for the subject. I applaud your efforts and ability to mine data, however many of your questions ( and fatal flaws in your results ) could be answered by learning the basics of major key harmony. “Music Theory” is not a law but an explanation for why things sound good or bad to our ears. Very commonly chords exist in one key together. When they don’t, they are often a quick modulation to a closely related key. Sometimes chord changes can genre specific. It is all very interesting. Take some courses in harmony and plug some of the “rules” into you equation, I bet you’d get even more interesting results. For the record, there is no b diminished chord in the key of c. The chord is b half diminished or b-7b5. They are very different chords with very different functions and implications.
Hi Mike,
we’ve definitely taken our fair share of college level music theory coursework in harmony. I’d definitely be interested to know where you think the “fatal flaws” lie in our results.
We’re doing our best to make this stuff accessible to folks that haven’t necessarily seen this stuff before, and apologize if we’ve oversimplified some concepts for you.
A point of clarification: Bdim is in fact diatonic in the key of C. The triad built on the 7th scale degree of the major scale is a diminished chord.
If you make this triad a 7 chord it does become a half diminished 7 chord as you said.
I haven’t read all the comments, but this seems as good a place as any to bring up the point that many scales are not simply alternate modes of a major scale. Even the humble harmonic minor doesn’t fit. How can you account for that by simply translating to the key of C? Consider Cmaj7, C#m7b5, Dm7, Ddim, just as a simple example, and note that the scale on that last chord could readily be a mixture of a E major chord (dominant 7th) and a D minor, basically taking us into a A harmonic minor, but it wants to resolve (in context) back to the Cmaj7, not an Am. That’s a very quick example. Maybe I’m getting a few things mixed up here, not being an expert, but it just occurs to me that your approach is similar to the music bots I’ve been creating with Analog Box — and their primary flaw is that there is no easy way to swtich to an alternate scale to achieve the harmony movement desired. To do that requires some representation of scales beyond just the modes of a major scale. This is extremely important to being able to create “interesting” music programmatically, IMHO, and therefore is important for this sort of analysis as well.
I am primarily a classical musician (I play a little bass guitar) so this might seem like a dumb question, but are there no popular songs in minor keys? Presumably you will deal with minor key chord progressions in a later article? Nevertheless I love your first article since I am constantly thinking along these lines in the music I make. Without awareness of the changing chords none of it makes sense. I eagerly wait for the next article.
great question Anne. We will be dealing with this more in later posts.
One of the interesting things about popular music is that songs often flirt between their relative major and minor modes all over the place (even in the same progression) and it is often a very subjective notion as to whether a song should be analyzed in the minor or major mode.
When a song is flirting with its minor mode you can think of it as tonicizing the vi chord.
Should a song that goes
vi IV I V
actually be analyzed in the minor mode because it is centered around vi, or is it really just the famous I V vi IV progression shifted over?
We’ll talk about this more soon
Some time ago, I looked into the idea of using Markov chains to generate music. This kind of data would be ideal!
Here is the blog post describing the deal: http://talesofcode.tumblr.com/post/23377011585/stochastic-generation
Is there a way I can access the data in a program-friendly format?
You just discovered the circle of fifths progression.. which has been around for almost 1000 years. There is nothing new about C major resolving to F major.. that is just V going to I. I can appreciate what you’re doing, but there is a much simpler way to do this.. Take a basic music theory course at your local community college. 🙂
I think you’re sort of missing the point of the analysis. The point is that the circle of fifths is pretty much all that is required in pop music and this is proved by the results. I swear these guys know more about music theory than you and certainly more about the scientific process.
This is so amazing!!! Please keep going!!! You guys have put in some incredible work!!!!!! 😀
Wow, so much feedback already. Great idea, I wish I had the technical capabilities to play with the data a bit. However, I feel if you are shooting for a popular audience here then the kinds of questions you might probe the database with will likely be boring. Questions that I might ask like: what are the most common harmonic rhythms for verses? What percentage of songs start the chorus on a chord that is not the tonic? How many songs use sections of odd numbered measures? May only be useful for songwriters…
I sure hope that the database grows to include non billboard stuff. I transcribe a lot of indie rock and alt pop stuff that is a bit more adventerous and feel like having another data set would provide better content.
Interesting post… but I agree. Considering the way music works (harmony as someone mentioned) this is very much expected.
But good to show newbies that music isn’t completely random.
There’s a mistake in your first graph. The relative minor of A is F#min, not Fmin.
I am glad this jumped off the page for someone else.
Really cool stuff! … I’m personally interested in using this kind of info to build a general music composer some day … would be really helpful to see a full table of the data of chord progressions for each chord, like you did with Em, … perhaps some data on the most used patterns of 4 -or even more- chords after each other..
And what I would find really cool as well is to analyse the lyrics for song theme’s and see what kind of themes get what kind of chords / chord-patterns … but I guess that would involve quite a bit of semantic research as well 🙂
This seems like a fascinating project, I’ll definitely check back for the followup posts!
Thought about moderating the feedback? A hundred comments saying the exact same thing, and all after this being posted to the musictheory reddit…
maybe you should just use chords that sound good …
What really surprises me is not that C chords are popular (esp on keyboarded instruments), but that E doesn’t rule the roost. It’s pretty much THE default key for guitarists.
it would be interesting to see if there was a cumulation of songs in the key of e at the height of the guitar craze in western popular music.
when would that have been? the seventies?
.~.
Freddy, I was also surprised at how low E major was on the list for the same reason.
amazing..
[…] "CRITEO-300×250", 300, 250); 1 meneos Cual es el acorde mas usado en las canciones mas populares? blog.hooktheory.com/2012/06/06/i-analyzed-the-chords-of-1… por kastanedowski hace […]
Wow, very interesting post! Awesome use of data to dive into one of my favorite topics, music.
Now, I have to preface my comment by saying that I have been around musical instruments for all of my life, and that I have studied the theory of popular music for quite some time, albeit on my own. I have come to many conclusions during this time, but have never had the data to really back it up. That is why this post from you guys is so exciting to me! So first of all, let me thank you for putting your time into this.
Some comments from things that I noticed while reading the blog post:
1. Most popular keys:
Interestingly enough, this pattern presented in your bar chart lines up nicely with the circle of fifths, and follows 2 general principles that I think hold true. One is the avoidance of too many sharps or flats relative to C Major, and the other is preference of sharps over flats. It has long been speculated that sharps give an uplifting feel to the music, while flats bring about a sort of depression, or depressive character to the music.
Here is a picture I have made showing this trend: http://i.imgur.com/FSnBC.jpg
Also, beyond just sharps and flats, great musicians that have long since passed have quite often commented on the nature of the different keys, and the colors and emotions associated with them.
Here is a brief overview: http://www.rollingball.com/A01c.htm
Note that all of this supports your data thus far.
There is one anomaly however, the prevalence of the Eb key. I understand it is not statistically more relevant than say F, but it still requires a comment to understand it’s placing among the other keys. You’ll note that in the rollingball link above, even though the key contains flats, it is still given a good description that seems to fit with the feelings of popular music. But I do not think that is why it is so prevalent.
An interesting thought is that when you detune a guitar by a half-step, you have put it in the key of Eb. Many bands have done this to separate their sounds from others, or to help the singer hit the high notes. Most notably, Jimmi Hendrix used this tuning: http://www.ehow.com/how_4798600_tune-guitar-like-jimi-hendrix.html
I believe that most beginning guitarists would know this key or tuning due to its prevalence.
The next two charts on this blog post are also very illuminating. There are hints that the music theory will be explained in further posts, so i’ll hold comments on those charts, since i’m sure you guys will cover the main points of chord families, tonic resolution (and why that is bad in a lot of cases), etc.
Again, thank you very much for putting this post up!
As an aside: I personally have recently started to learn how to code (I found this blog through hacker news), and am very interested in data/big data, and finding trends in data that bring new insight to the world. Because of that, I would love to see the data behind these findings. I remember spending nights upon nights in my room listening to the same song endlessly on loop so that I could analyze the progression and find patterns. Unfortunately, I didn’t know about programming or databases at the time.
I can already think of a million things I would do with the data, though I do not have the programming skill to carry it out yet. In either case, I will definitely stay tuned to future posts from this blog. Thanks again.
Very cool post. Instead of just seeing what chord follows when the first chord starts with C. You could see what kind of Chord Progression (I, IV, V, etc) follows. This way, it could use all your data and provide a more complete view. It may make sense as well to separate the songs that start on Major and minor chords. Great post regardless, and thanks for keeping the internet interesting.
Nice work, really looking forward to your next posts…I’ve always wanted to have some scientific evidence to prove that my friends have bad taste in music 🙂
This is certainly interesting data, but I’d be hesitant to recommend certain chords, melodies, or rhythms just because they happen to be popular. For me, it’s definitely not the common chords that make a song interesting. It’s the unexpected ones. Music is all about the balance between expectation and surprise—too much of the expected and you end up with “Mary Had a Little Lamb.”
Look at “Yesterday,” whose first two chords are F and Em7. I’d be surprised if those two chords were back-to-back more than 1% of the time, and yet that initial drop is one of the things that makes that song so heartbreaking and amazing.
Following the “rules” may get you a tune that simple and easy to sing, but it probably won’t have any of the magic that makes for a truly great song. Rather than using a formula when writing a chord progression, or even starting with chords at all, I find it much better to begin with a melody. Then I come up with chords that work with the melody, which on the first pass usually end up being rather predictable and boring (and likely follow the results above). The third, and arguably most important, step is to take those chords and swap a few of them with more interesting and less predictable chords until you find something that really shines. Yes, it’s a labor-intensive process, but nobody said writing great music would be easy!
To those of you complaining that classical music theory makes this all obvious, I would say: isn’t it nice these facts are independently discoverable?
Anyway, if you print and play guitar chords & lyrics you might like this tool: http://joewlarson.com/onePageChords/
So, the analysis showed that when analyzing tonal music, it’s tonal?
Hey Dave, really interesting stuff, looking forward to more. I was surprised at the graph showing the root being used less than V and IV in most songs, until I noticed you’ve got I and VI in there separately regardless of whether the tune’s in a major or minor key. I understand that theoretically, Am is the sixth part of the key, but if you’re comparing the relative use of chords wouldn’t it be more useful to have minor key tunes and major key tunes displayed separately? I imagine then you’d see the ‘root’ – C or Am – being used comfortably more than any other chord?
Also, I’d guess the ‘number of sharp/flats but this isn’t universal’ is a secondary factor in the distribution of keys, and that the more common keys you see just come easier to guitar/keyboard players. Hence the relative popularity of, say, Eb and F (keyboard) and A (guitar). I know Bob Dylan wrote a lot of songs in F and Bb simply because he found the black keys easier to play. The two factors are linked of course, but the ‘isn’t universal’ element is probably an instrument thing.
Anyway, keep up the good stuff.
Ben
Ben,
definitely agree about the ordering of the keys likely being at least somewhat related to ease of playing on the piano/guitar.
I’ll definitely be talking more about minor/major modes and how the data is organized. The short answer: popular music tends to flirt between the major and minor mode so frequently that often times it’s subjective as to whether a song should be analyzed in the minor or major mode. This is itself very different from most classical music, and definitely worthy of further discussion.
Addendum – I’d love to see a post on tunes that don’t resolve back to either I or VI at any point, always enjoy those.
Fascinating! I’ve just been thinking about an analysis like this the other day. Some of the questions I’ve had were:
* is there a pattern of repetition of chords that is universal throughout most pop songs?
* what repetition patterns do more “risky” songs follow?
* is there a connection between genres and chord patterns?
[…] folks behind Hook Theory, a music theory app for the iPad, analyzed the chords of 1300 popular songs for patterns. The results are of particular interest to songwriters but any musician will find them […]
You’ve got a nice dataset here to do some higher-order analyses, along the lines of what David Cope has done to generate Bach-ish music, etc: http://artsites.ucsc.edu/faculty/cope/experiments.htm
In fact, there’s lots of work along these lines in the field of Music Information Retrieval: http://www.ismir.net/
e.g., http://www.iro.umontreal.ca/~paiemeje/articles/ismir.pdf
[…] I analyzed the chords of 1300 popular songs for patterns. This is what I found. | Blog – Hooktheor….Related posts: […]
There is work being done in this vein right now in musicology. This sort of analysis might be of interest to the likes of Allan Moore (the musicologist, not the author), Shaugn O’Donnell or Robert Walser. This large scale statistical analysis is not the sort of thing that captures the heart of many musicians, but is exactly the kind of groundwork musicology of popular music needs.
In popular music theory right now, the prevailing question is, “What is the ‘tonality’ of popular music?” Tonality, capitol T, is a very specific word in music theory, applying only to a specific use of specific chords (specifically, that the V is the chord that points back to tonic). But popular music doesn’t always do that. And yet, the chord progressions are not the chaos of avant-garde or 20th century music – so if they are organized in some fashion, what is it? What rules govern popular music chord progressions? These kinds of statistics – perhaps on a yet grander scale – could begin to plumb the depths of these questions.
[…] One Minute Vacation – BoingBoing.net Escape Artist – Zoe Keating Nick, Software Engineer Pattern Analysis for the Chords of 1300 Popular Songs – HookTheory.com Voyager – Daft Punk Julian, Development Intern Heavy Rain Creator […]
I remember reading something that Paul Simon wrote about a million years ago – that most of the songs he writes are in ‘sharp’ key signatures because the guitar is more suited to those keys, whereas songs written on keyboard tend to be in ‘flat’ key signatures. I found that interesting and pretty much true.
Horn players also prefer flat keys.
writing is not a easy thing to do, it takes creativity and good sense, and you’ve got it!http://www.rebite.net
I just want to point out that you don’t really have enough evidence in your study to support your conclusion. What you need to do is analyze 1300 unsuccessful songs and compare. Right now you can say that you looked at 1300 popular songs and found certain trends. But you cannot yet conclude that these harmonic trends are the reason those songs were successful. The songs you looked at may have been popular for any number of other reasons unrelated to the chords used. I’m guessing that if you analyze another 1300 songs, same genres but unsuccessful, you would find the same chord trends in those songs as well. I’m guessing that what your study has found is what chords are more popular with songwriters, rather than with listeners.
Was it stated that the chord progressions were the reason for the success of a song? If it was, I missed it.
It would be an interesting study to compare the trends in the Billboard top-100 songs vs, say songs in the 900-1000 range (if there is such a thing). I have a feeling it would be fairly similar to the top 100s: mostly familiar and repetitive with a smaller number of odder ones. This is pure conjecture, or course.
While very interesting, what part does the beat play in a successful song? Would a comparison of your database with another which analyzes the beat, as well as the chord progressions in general give you a more detailed formulaic answer? What about older songs? What about pop instrumentals such as those made popular by the Tijuana Brass? So many questions lol. I have listened to remakes of original songs by the original artist, such as layla and the ones Carly Simon and Joni Mitchell did of their hits, same song done so very differently, therefore I would say it is not just the song chord progression itself but a very complex combination of tonality, beat, style, chord, and subject matter. keep up the good work!!
Interesting article. This might be relevant to it: http://www.ted.com/talks/bobby_mcferrin_hacks_your_brain_with_music.html
The gist: pentatonic scale is burned into our MINDS
But for a less serious take on the same subject, british comedy rockers certainly figured it out: http://www.youtube.com/watch?v=5pidokakU4I
Like others have noted, if you have any knowledge of music theory the findings here are not all that surprising. It’s all very interesting and I appreciate the work you’ve put into this. Still, I’m sort of left with the question, “So what?” I can’t help but feel quantifying music detracts from the non-quantitative aspects of songwriting. You state in your discussion that “If you write a song in C with an E minor in it, you should probably think very hard if you want to put a chord that is anything other than an A minor chord or an F major chord.” To me this sort of statement is a call to stifle creativity. Writing is a journey filled with discovery, wild tangents, and flashes of serendipity. A songwriter should not have to “think very hard” about going from Em to A instead of Am just because statistics show that’s a much more common move. If it sounds right and helps convey the emotion and meaning you’re trying to express, then go for it.
Note: I’m not a professional songwriter and so perhaps my relationship to songwriting is more idealistic than those who do so to put food on the table. Maybe not!
First of all, this is amazing!
Secondly, here’s something I might be interested in. It looks like you’ve built a Markov Model of popular music. I am now wondering: can you use this model to generate melodies, automatically? I’m talking like what Claude Shannon did in his 1948 article, “A Mathematical Theory of Communication”.
Similar to this: http://www.cs.princeton.edu/courses/archive/spring12/cos126/assignments/markov.html
[…] I analyzed the chords of 1300 popular songs for patterns. This is what I found. | Blog – Hooktheor…: “” […]
There is a perfectly good reason why the Am chord is more common than AMaj. And why Gmaj is more common than Cmaj whenever Cmaj is the most frequent key. Its pretty simply really. C major and G major share common chords. A minor is in C major, G major, F major.
Cmajor is the most common key because it is all the white keys on a piano/synth and therefore the easiest to write in. F# major on the other hand is hard to play on the piano, that is why its the least common. G major (e minor LOL) is very common on guitar because its very easy to play in (and tuned for it), and it is because of this reason that it is a very popular key in guitar-centered music. Not because it has some special quality that makes it sound better.
I hate articles like this that stand to mystify music more than educate on it. Your “analysis” is completely worthless.
It doesn’t make sense, I think, to group the major with the relative minor together when analyzing the most common keys — even though they share the same notes, the major and its relative minor keys sound drastically different, just like every other mode does — they should be analyzed against each other and not grouped together when asking what the most common keys are. Usually when a song is in a minor key, while it may borrow from its relative major, the feel remains minor throughout the song (and vice versa if the song were in a major key) that’s why it doesn’t make sense to me to have those grouped together!
Please transcribe all the songs in a major cord into C and redo the analysis. This should be easy mathematically.
Also please analyze all A minor songs differently from C major. There’s really no comparison. Songs in a minor key should be seperated out. However songs that mix major and minor should probably be in a third category. Yet songs with a key shift should be transcribed so that the part before and after the key change both transcribe back to C. At least for the kind of analysis you have done so far.
I think it is interesting to note that the IV and V are used more than I. I wonder if this is true for current music as well as music from previous decades. Has the balance changed between the relative frequency of use of these three chords, or has it remained the same? I wonder this because there is a tendency in contemporary music to distill the tonal center by ending on the subdominant. That would certainly contribute to the less frequent use of tonic. However, it seems this may not have been true in music of past decades.
It would also be interesting to note the order of subdominant and dominant in progressions. Again, common practice theory (through JS Bach) would indicate that retrograde progression is frowned upon. However, progressions in the modern era reflect different ears and the “retrograde progression” (dominant to subdominant) is much more common – to the point that it no longer sounds “wrong” to our ears. I like how you have indicated that IV almost always follows iii, but I would be interested to see your stats on IV and V in relation to each other.
Interesting. Such as the usual V-IV in rock-based 12-bar blues in measures 9 & 10? This is one of the most familiar V-IV contexts, methinks.
[…] […]
The best music I ever heard was never produced up to broadcast level and
never broadcast but of that which was I will name as examples two bands
which can be seen on youtube etc
1/ the American band `Rain Parade` ( mid eighty`s )
2/ the British band `Love ` again from the mid eighty`s
This research confirms that most people that write popular music don’t have much or any training in theory. Popular music sounds monotonous a lot of the time because it is. It’s all the same, easy chords options put together in very similar ways. It’s amazing what people get famous for these days. Actually, it’s not amazing. That’s the point.
most of your comments would holod true for both classical and jazz when painting with a broad brush. They are not unique to pop music
I was slightly surprised at how many songs in C (sure that wasn’t the default for songs of indeterminate key?). But both of its neighbours are more difficult, while ones like F & G have other alternatives nearby. It would be interesting to see the breakdown by dominant instrument. When I look at church song books, the shift from keyboard instruments to guitar can be seen in the disappearance of key sigs with a few flats – some piano/organ players love to play in Ab or Db, but guitarists prefer A & D.
The perfect cadence is a well-established thing – you can’t go wrong moving that way in the circle of fifths if you don’t know what chord to write next.
I’m surprised too that IV & V came up more than I.
An analysis of key changes would be interesting too. I think +1, +2 and +3 would be common – in my compositions I have included these, as well as -1, -2 and -7.
[…] podesłał mi benedyktyńską (dwa lata pracy) analizę harmoniczną 1300 przebojów zaczerpniętych głównie z listy Billboardu. Autorzy sprawdzili, z jaką częstotliwością […]
I’d like to hear what song you come up with if you used all that info.
you realize that music theory basically covers all this already? No statistical analysis necessary! It’s a little bit like examining 1300 novels to find out what sentence structure and words they have in common. Things that can basically be extrapolated from the rules of grammar and vocabulary of a specific language.
[…] Chords To 1300 Songs Analyzed Statistically For Patterns Lies, damn lies and statistics… I analyzed the chords of 1300 popular songs for patterns. This is what I found. | Blog – Hookt… […]
“We transposed every song in the database to be in the key of C to make them directly comparable.”
Reduce all songs to the same major key…and the most popular chords found are the I IV V, closely followed by the relative minor. Stunning finding.
superb! i’ve gotten through the 5 newest posts, and i’m lovin’ it! i’ve really enjoyed your blog.http://www.luftec.net
Also check out Axis of Awesome’s 4 Chord Songs 🙂
http://www.youtube.com/watch?v=oOlDewpCfZQ
Wow! cool stuff.
What I would love to see extracted from your data base if possible, is the progression of chords that go down or up in scale, with minor variation, then repeat. Like for instance a song might have a melody that generally goes up the scale, A B C D E F G, then repeats. So the listener has an “up lifted” feeling.
In short, I would be interested in categorizing what songs have a more up-ward-bound scale theme, and which have a down-ward scale theme.
It’s my theory that this effects mood greatly in the listener.
I’d love to see an analysis of the cadences used by the songs in your database. In other words, how are they getting back to tonic? Many of us learned the “Circle of Fifths”, and we’re used to hearing cadences like I-IV-V7-I (e.g. C-F-G7-C), but my ear tells me that modern music is often employing more unusual cadences – and, on occasion, some rather abrupt, almost “ear-jarring” resolutions to tonic…
I’d also like to see a more “architectural” chord analysis. How often do suspended chords (heavily used in “Pinball Wizard”, for example) occur? Diminished 7ths?
“Next chord” is interesting, but you also need to ask “does the next chord differ, based on the key of the song”? In classical music at least, there are theories that the chord progression depends on the key, as does the emotion of the song. Does any of that hold in pop music?
I’d be curious to know about which chords are actually remembered. That is to say, given a fragment of a song, can you identify that song, and if so, the notes/chords/whatever in that fragment were actually remembered and linked to the song in the listener’s mind. Those chords should have higher weight in your analysis, I think!
Good luck surviving the slashdot effect! 🙂
[…] analyse van pop muziek Zie: I analyzed the chords of 1300 popular songs for patterns. This is what I found. | Blog – Hookt… […]
Strongly suggest using the Nashville system instead of supposing the key of C. That is my sole criticism of this great information, though I do have a question.
Have you adjusted for when songs change keys during the middle 8 or chorus? This is a fairly common practice, and could explain the prevelance of the IV and V chords over the I, since the key often shifts to either the subdominant or dominant in these instances. (Can’t think of a an example off the top of my head. Anybody?)
Great stuff. Many thanks. Looking forward to more.
[…] I was busy on packing boxes and criticizing another study highliting recurrents chords in popular music that Axel Helios sent me, some tracks from Alva Noto were in charge of the sound design of my […]
A simple fact is that a great deal of popular chord progressions are dictated by ease of play – that C major and A minor are simply the two easiest major and relative minor chords to play…etc…
I think this article I interesting but it also creates an illusion that most pop artists focus on chord progressions alone. It seems to me that there are three or four major elements of pop music, chord pogression, rythym, and melody/harmony. If you can ‘define’ chords and how they relate to eachother then the same could be done for the other elements. It seems to me that many of the most pop artists focus on how these elements work in sync to create one sound, giving us that emotion mentioned above. Chord progressions are merely one element of the 1300 songs mentioned, and anybody can play those chords in any order, but that doesn’t mean they can create the other elements that will complement one another in such a way to create one specific sound.
Interesting, but in the end useless, I believe. Inspiration for all art is intuitive. Analyzing the stuffing out of anything is fun, but can not be used as a substitute for creation. The current “machinery” of the popular music “industry” is using formulas derived from statistics and is destroying the airwaves as a result. What’s next, computers composing? Truly gifted musicians, from Beethoven to Basie to Beatles, never worked “reverse-engineering” like this. They often broke the “rules” that so many before them have derived from the music that came before. Can you paint a Rembrandt or Picasso by analyzing their color choices? their brushstrokes or the size and shape of each form?
Hankus, are you not aware that most musicians learn music theory by “reverse engineering” Bach…? That’s what music theory classes are. And also, knowing the rules doesn’t prevent a composer from breaking them…
Have a look at iphone app jazz practice as it explains the theory in a good way and how to play chords on piano and guitar
I’d be more interested in the emotional context of the chord progression. What is the most popular chord progression for a slow sad song, compared to a fast happy song, or a sober serious song. A song of triumph compared to a song of loss. Even down to the individual chord changes would be very interesting, though I suppose ALOT more work.
Take your example of the most popular chord after Em. I’d be willing to bet that F is most chosen in a happy song, or at least a point of resolution within the emotional context of the song, where as Am would be used more in a sadder, tragic sort of context.
I’ve seen a charts that listed the relative chords in a key, and basic emotional “purpose” of each, such as tension, release, rise of emotion, fall, and resolution ( most generally by going back to the root ) along with a description of the emotion conveyed by each change from one to another. It would be really interesting to see what the actual statistical use of those chords and changes show. That is to say which actually “work best”, to convey the emotional context of the lyric. That is of course making the basic ( though admittedly debatable ) assumption that the most popular song, is the one which best conveyed it’s intended emotion.
[…] Analysis of over 1300 popular songs They analyzed over 1300 popular songs, and the results are not surprising at all: http://blog.hooktheory.com/2012/06/0…-what-i-found/ […]
Do you have any plans to do relative analysis on the chords as well? That is, publishing which songs consist of I-vi-IV-IV-repeat and also what modulations do (or don’t) take place when moving to different sections (chorus, bridge) of the song? Finally, is there any way to access your raw data?
Let me end by saying, freaking awesome.
Isn’t all of this obvious to anyone who ever plays music? In the “Chord Use” chart, doesn’t it go without saying that the chords IN THE KEY would be used more than the chords OUTSIDE the key…? Duh.
And by the way, C major and A minor are not the same key, so why are they conflated in the “Most Popular Keys” chart…? Someone doesn’t know much about music…
[…] half) augmented and whole tone and all the modes to these scales (not fun)Powered by Yahoo! AnswersMichael asks…All blues for jazz guitar. Are these good books?Someone suggested this book to me, bu…dtm-content">All blues for jazz guitar. Are these good books?Someone suggested this book to me, but […]
Hi folks,
I would be interested in working on this project. I am a composer and a psychologist who is actively teaching Psych of Music. There are vague answers to a lot of what you are doing, but music is way under-studied. The choice of chords usually comes down to 1) frequency relations being relatively low in “dissonance” in the sense of conflicting frequencies, but 2) having enough dissonance to stimulate and 3) enough novelty to be interesting. Some evolutionary theories tie consonance/dissonance preference back to jungles and animal sounds. The unexpected avoidance of the I chord for much of some songs is part of a learned emotional activation that keeps us feeling a little anxious or expectant. There is still plenty of room for evidence and better resulting explanations.
Best wishes, Dr. Fox
It’s definitely interesting to hear about the psychology behind all of this. If there are any online course materials, I’m sure the community would be interested in learning more.
And as Darren Bean said, freakin awesome.
It’d be very interesting to see this analysis applied to modes as well – how many songs are in strictly major diatonic, compared with major blues-keys with flattened sevenths, how many in “classical” minors, how many in Dorian / blues minors. That would probably be a lot more work though, I don’t know how sophisticated automated analysis techniques are for reading this off.
Anyone saying “musical theory predicts all this” needs to better understand the value of empirical observation and measurement in science. Yes, people try to write down “rules of grammar” in linguistics, but anyone saying that this makes statistical linguistics a waste of time needs to catch up on the last 20 years of research and development.
Sorry but some of your conclusions are daft. With so much music out there these day, anyone wishing to be successful should find ways to differentiate themselves from everyone else. One way is to be more adventurous and inventive – rather than simply grabbing a guitar and strumming the same sequences as everyone else.
http://www.youtube.com/watch?v=Lsi9ljzclc8
One of many life-changing 4-chord songs. Just the tonic, 5th, 4th, and relative minor (for spice).
Structure’s only half of it. Just a bed for the message to rest on. Oftentimes still the recipe for “success” 4 chords and the truth.
Truly though, to be in it for the success is to be in it for all the wrong reasons.
[…] 26% das canções populares são escritas em dó/lá menor (C/am) (Fonte da imagem: Hooktheory) […]
i was looking for such easy task from a long time it will help me in increasing my knowledge. thanks a lot.http://www.celularcomchip.com
i love this.http://www.hpbrasil.org
If you come up with any reason for the (relative) popularity of C/Am other than the fact that they incorporate all the white notes of a piano and are thus incredibly piano and keyboard friendly, I would be shocked and very interested.
In fact, they are also comparatively easy chords to play on a guitar as well, and most beginner musicians delight at the simple change in finger positions required to go from Am to C.
Fantastic article, I would love to get my hand on more of your analytics! Just for example, more data on “the most likely chord to follow”, but for different keys and chords of course. Where could I read more of your findings?
[…] we say about the probability for what the next chord will be that comes after it?More information: Hooktheory No CommentsShare/Bookmark wpa2a.script_load(); googletag.display('div-gpt-ad-1339048641543-3'); […]
What would be also interesting is to identify the 7ths — since pop music is derived from the blues, and the blues use 7ths all the time, it would be interesting to see how much 7th and 7thless music there is (that is, 7ths not being used simply as the V7 chord at the resolution).
Another interesting question would be descending bass runs, or movements descending through the circle of fifths — virtually all of Dylan is marked by descents of either kind (e.g. Sad Eyed Lady of the Lowlands).
I find it incredibly unlikely that, with all songs transposed to C, C is not the most frequently-occurring chord.
So either I’ve misunderstood something fundamental or I think there’s something wrong with your methodology. What are the frequencies in “Chord Use” based on, exactly? What does “chord progressions containing a given chord” mean?
Say I have a song which repeats C, F, G a hundred times, then has a break verse which features Am, Dm, G7 once. What’s the “Chord Use” derived from that? Does C score one hundred, or one?
I was assuming that it was based on a measure count. If is just frequency of selection, that would explain IV and V occurring more frequently than one. What do you say, esteemed blog author?
Interesting analysis. But there are others who have tried to do something similar as one commenter states. Anyway I don’t think this will turn out very useful for songwriters at least. You can analyze popular words in famous literature and list them up. That does not mean using those popular words in your writing will guarantee a good literature at all. I think same thing goes for music as well.
This article is just amazing. But I think future researches in this area would achieve better results if you segment by music style. As a musician, music style for me plays a strong role in determining which chords I will use for a song in a key. Of course I’m saying that based on my experience, and not scientific methods.
By doing that you will probably find which chords are most commonly used for a specific music style. Also, you transposed all the songs to the same key in order to compare, and it was C. Instead of doing that I would also suggest to use the chord scale, like F is the 4th Major in the C key. And A is also the 4th Major in the E key. I’m a Brazilian, so I don’t know if there is a specific name for this in English, but I hope you got the idea.
Anyway, IT WAS AN AMAZING STUDY! Great Work.
[…] статистический анализ 1300 песен и опубликовали результаты: какие созвучия чаще всего встречаются в поп-музыке и […]
[…] статистический анализ 1300 песен и опубликовали результаты: какие созвучия чаще всего встречаются в поп-музыке и […]
Excellent analysis, thank you for making it available. Looking forward to more. Some arrogance and incredulity in some of the preceding comments. And I see little point in creating separate realities for major and minor keys or parallel major/minor. The chart focusing on keys makes it clear some distinction is being made, but a tonal centre is a tonal centre, if even ambiguous in relation to a secondary tonal centre of a relative major or minor.
Count me among the folks who are fascinated by the prevalence of the dominant and subdominant over the tonic, though not surprised. This goes back at least to the 1970s in pop with a song like Fleetwood Mac’s “Dreams” (not in this study), where we never hear C major, only F and G with a short reveal of A minor. A commercially popular example with cross-generational influence of C/the tonic being de-privileged to the point of exclusion.
Given this sampling of songs it points to a small twist on classical harmony: we like to avoid the cadence and sustain tension. An examination of tonal tendencies across a larger spread of style periods/decades would be interesting. When did we become less interested in C/the tonic? My guess would be the 1960s/70s with the influence of modal jazz, non-Western music and minimalism and the diminished influence of Tin Pan Alley.
(*Assuming one doesn’t feel it necessary to transpose “Dreams” into C minor, which would be valid.)
First of all, this is great! I can’t wait to read more of your analysis. I have actually thought of doing a similar study myself, but limited to rock music and with more of a focus on large-scale trends and how they changed through the years. Secondly, good job on ignoring and/or responding graciously to the wet blanket comments and arm chair music theory experts. A few of my thoughts/questions:
1) Someone else mentioned this, but I’d be interested to see the element of the alternate bass notes included in future posts. I think you said you would be doing this.
2) Regarding alternate bass notes, I’m curious how you named chords when a bass note other than the root or the 3rd was played. For example, a C/G almost always resolves to a G (at least in classical music), and thus function more like a doubly suspended G chord; so would you call it a C or a G chord? That raises the question: did you name the chords based on published music or purely from your own listening/analysis? Also related, I’d be interested to see how you analysed songs that were more riff-based or with a riff-based bass line (and drone bass parts).
3) Although this would increase the volume and complexity of your study exponentially, I’d be interested to see analysis with chord extensions (7s, 9s, etc) included. Since you are looking at popular music, I imagine the use of the ‘extra’ notes is more incidental and less structural than in jazz. I’m guessing that a large amount of such notes in guitar-oriented music is due to convenient open strings or voicings that are easier to play with extensions. And I imagine that “sus” chords would have to be ignored in your analysis due to their ubiquity.
4) I assume that songs with the most experimental and complex progressions are rarely in the top-100 (e.g. experimental/prog rock/fusion sounds). However, if any did sneak in, I’d be interested to know if you came across songs in which a chord couldn’t be identified or just had to be noted as “implied” or named merely from the bass note.
5) Since most popular music is major/minor oriented, sticking with that paradigm makes since. But it would be interesting to see analysis of the use of modes given or perhaps included as a footnote when a non-diatonic note implies a mode (e.g. bVII = mixolydian, II = lydian or dorian [when using vi for your minor tonic), III = harmonic minor, bII = phrygian,). Your nomenclature is most logical since your target audience would think of of these non-diatonic notes more as accidentals than as ‘normal,’ albeit modal, notes. Furthermore, just as you noted that in popular music major and minor tonalities are blended and blurred, lots of popular music teases the modes without being committed to a particular mode, so again, your method is probably best.
Keep up the good work! I look forward to more, especially with regard to your analysis of melody.
Thanks for the great comments,
I’ll try to answer some of your questions below.
1. We’ll definitely be talking about inversions in later posts.
2. This is a good point. A lot of inverted chords function very similarly to their root position counterparts, but obviously there are many exceptions to this, the cadential 64 (C/G ) being one such exception. It wouldn’t be too hard to take this into account. Interestingly, the I64 has other (and often creative) uses in popular music that aren’t seen much in classical music, so maybe it’s not so easy. One example is a passing chord between vi and IV. Another is I64 V6/vi vi
3. We definitely are taking note of 7’s (and their inversions) into account. We don’t support 9’s in the editor currently (though we built in support in the playback engine so it’s something that could easily be added in the future)
4. Every now and then a song will use chords that aren’t described easily with conventional harmony. Most of the time this is because guitarists will use a particular finger pattern that adds non-diatonic scale degrees but still preserves the “feel” of a normal chord. Whether or not these alterations are on purpose is another matter, but these chords are certainly interesting and add a different sound to otherwise normal chord progressions. Off the top of my head, John Mayer is an example of a mainstream top 100 artist who probably knows what he’s doing (he studied at Berklee school of music) who sometimes will employ complex harmonies in his music.
5. Dealing with modes is definitely tricky. I think we’re definitely on the same page though. For example, we were playing around with the idea of having a chord called a “Dorian II” in the editor for when it’s clearly not acting as a V/V. Of course, a lot of this stuff has the potential to scare away beginners which is antithetical to a lot of the goals of this project. Lot’s of stuff to think about though!
World class response.
Great response! Please don’t worry about “scaring away beginners.” I’d rather you dive deep on this. Perhaps a section that has a caveat or explanation about needing a better base of music theory to understand would be good. I’d hate to think you’re coming this far only to dumb it down for those of us who find this really interesting!
[…] “I analyzed the chords of 1300 popular songs for patterns. This is what I […]
[…] “I analyzed the chords of 1300 popular songs for patterns. This is what I […]
Personally I was waiting for this article to go into greater detail. Maybe an extended version for musicians. Also, you can change keys on the iPad…. That’s why chords like A, B, and E are not available because they are not in the key of C which is the default key of Garageband.
[…] This is intense – it must taken a lot of work to analyze a gajillion songs (okay, 1300) to extract this data on chords and chord progression. […]
It would be an interesting exercise to do a melodic analysis on the same 1300 songs. Email me if you think this is worth pursuing.
Thanks for the corpus of work. As a software developer/geek this is fascinating stuff and maybe the springboard to a computer writing a ‘statistically popular’ song! BUT as a lyricist and song writer this is cold and doesn’t take into account lyrics and feel of a song into its popularity.
E.g. there are loads of amature songs that could be unusual, technically and brain stimulatingly brilliant, but unheard, and therefore not on the list.
But thanks either way.
craig salon software
http://www.SalonSoftwareSystem.com
[…] Fascinating article—and would be even more fascinating if I knew more about music. It begins: For many people, listening to music elicits such an emotional response that the idea of dredging it for statistics and structure can seem odd or even misguided. But knowing these patterns can give one a deeper more fundamental sense for how music works; for me this makes listening to music a lot more interesting. Of course, if you play an instrument or want to write songs, being aware of these things is obviously of great practical importance. […]
The research seems promising however I think you would be able to find even more out if you looked at each chord relative to the key. For example a C in the key of C is the same as an F in the key of F, the 1. However, an F in the key of C is the same as a C in the key of G, the 4 chord. Further I think if you went to most career songwriters and session musicians they could tell you that most times after a 1 chord you are going to see a 4 chord or a 5 chord and less rarely a 6 chord. Lastly this method seems short sites because it does not take into account things like suspensions, meter, length of chords, melody over the chord, syncopations, and many other tools that are used to build and release tension throughout a song an draw an emotional response from the listener.
I’m going to assume that while 1300 songs sounds a lot there’s likely a strong bias towards certain songwriters eg. lennon/mcartney, dylan, baccarat etc. who would appear multiple times. Not to mention some of the other non-performing prolific songwriters who have had careers spanning several decades or label owned songwriting production lines.
Interesting stuff though, will be waiting to see where you go with this.
I am most intrigued by what happens in the outlier songs that make them “work.” How can there be 1% of pop songs that move from C to A? Are they weird sounding or brilliant?
This is the beginnings of an explanation of basic music theory, approached obliquely via data. I’d be more interested in an acoustics-level data-driven analysis, because a chords-driven analysis has already been exhaustively done within the field of music theory. Music from the early to mid romantic era of classical music follows these same patterns almost exactly (not surprising at all that V is the most common chord, for example). For a far more rigorous, technical, centuries-old explanation of this exact same phenomenon, simply take some time to learn music theory.
… [Trackback]…
[…] Find More Informations here: blog.hooktheory.com/2012/06/06/i-analyzed-the-chords-of-1300-popular-songs-for-patterns-this-is-what-i-found/ […]…
I have read most but not all the comments so forgive me if I am repeating this question. I understand that the songs you analyzed were popular songs. Was there a certain time period in which you focused on or did you categorize the songs by time period and then focus on the popularity of the chords and progressions specific to those time periods? If you did did you see a trend and/or repeating of trends relative to the different time periods.
We didn’t focus on any particular time period. As the dataset gets bigger, I think seeing how the data changes as a function of time would be really neat. At the moment I don’t think their’s enough data to be statistically confident. Great idea though.
What is this an on musical way to approach music. Music one of the human universals. 1-6-2-5 chord progressions? It’s what you create not what you analyze. Ask Mozart 1-5 but at the end 5-1.
this isn’t really news……. nearly every pop song i hear nowadays has about 4 chords. the prevalence of the iv and v over the tonic probably account for the amount of time people spend dicking around on bridges where generally the tonic isn’t introduced
Why did you decide to bundle the relative keys?
“It goes like this
The fourth, the fifth
The minor fall, the major lift
The baffled king composing Hallelujah”
Leonard Cohen.
Yup. There’s a pattern there folks.
Likely
[…] Analyzing the chords of 1300 popular songs for patterns,the first blog post of hooktheory, which “teaches the theory behind popular music for songwriters and musicians.” Via Hacker News. If you know some music theory you won’t be surprised, but it’s interesting. […]
[…] Chords by Marty X on Jun.18, 2012, under Hose Down Daily Blog Band photographer, Mat Ward, sent us this interesting article about the use of specific chords in pop songs. The folks who did the study analysed some 1300 […]
[…] I analyzed the chords of 1300 popular songs for patterns. This is what I found. is a post on the blog for Hooktheory. I found it via Waxy Links. The authors analyzed chord progressions to find which keys and chords are common. While I was slightly surprised that some keys were not represented more often, the basic results are what any musician would predict. It is interesting, though, that the conclusions came from data analysis rather than music pedagogy. […]
Axis of Awesome have already covered this…
http://www.youtube.com/watch?v=5pidokakU4I
Very good information. I do similar research on my blog, http://recordingartsresearch.wordpress.com/
I take Billboard hits and deconstruct them, musically and production-ally (I know that’s not a word, but you get it) and figure out what “makes them tick”. Check it out & follow!
Transpose Pachelbel’s Canon to C (from the original D) and the sequence is:
C – G – Am – Em – F – C – F – G
Just sayin’
While I really really enjoy the key analysis you did, I think its important to understand a chord progression on the relativity to tonic rather than their absolute pitch. When people mention the four chord pop song they are referring to songs that use similar chord progressions, such as I-V-vi-IV, not necessarily songs that use C, G, A minor, and F (or that above progression I listed in the key of C). Also a B diminished chord in the key of C is a very logical way to end a chord progression. Its resolving the diatonic tendencies of the chord built off the leading tone to the tonic. Thanks for this article though!
See the result of an earlier attempt. “Tales from The White Hart”, by Clarke.
What the hell is with the key of the chords?!?!?!?…You should be refering to intervals, instead…(like I, IV, V, etc.)…
This analysis does seem superficial. Let me echo those who said that the focus needs to be on intervals within a key instead of discrete chords like C, F, and G. And individual chords aren’t as interesting as the progressions. Are the hit songs using I-IV-V or I-IV-vi-V or another progression? That’s more useful, and even then, if broken down by styles.
The most common word in English literature is “the”.
Of what value is knowing that?
Should I write a novel that emphasizes “the”?
As it turns out, the value of knowing that is astounding: http://science.discovery.com/tv-shows/through-the-wormhole/videos/through-the-wormhole-information-theory.htm
I don’t know that you’ll cover this, but I’ve no doubt that for songs with primarily classic rock instrumentation such as guitar, bass, keyboards, C major or sharp keys such as G, E and D major are most common. This is so because due to the construction and tuning of guitars, they sound best and feel most comfortable to their players in those keys. Also, while the guitar is one of the most difficult instruments to play with real virtuosity, it is one of the easiest instruments to learn to play without formal instruction, thus most guitarists are guitar players, but not musicians; they don’t know music theory and can’t read music. And lo and behold, among the first chords any self taught guitarist learns are C, G, D, F and E. Non-sightreading guitarists–most of them–usually have great difficulty playing in flat keys (other than F) and sharp keys with more than four sharps.
Songs in flat keys become necessary when brass and woodwinds–particularly the sax–are added. Unlike the guitar, these instruments play most naturally and sound best in flat keys, so when these instruments are scored, guitarists have to adapt, usually through the expedient of barre chords.
I also suspect you’ll find the most common chord/measure structure is the classic strophic form.
I have to join the chorus of people criticizing this analysis. The most common chords are those that beginner musicians learn first. If you analyze the Beatles or Jazz music, all sorts of sophisticated chords would be equally dispersed throughout.
For high-level professionals, form follows content. For amateurs, it’s the other way around.
Better than nerdify over ‘Most Popular Keys’, turn some thinking to the lead vocalists of each song analyzed. Most pop hits are set in the key which fits his/her voice best. It’s not the keys that are popular, it’s the singer’s delivery. At best, those ‘Most Popular Keys’ simply express the most common human physiologies.
I would agree if by far the three most popular keys weren’t also by far the three easiest keys to play on the piano. Even the seemingly “out of place” E flat is actually relatively easy to learn and play in.
Except that F, like G, is noteably easier to play in on piano than Eb, while Bb and D are slightly easier, and A (with its 3 sharps) is equally as easy (as Eb with its 3 flats).
That said, the difference in popularity between Eb (9%) and F (8%) is a mere 1 percentage point, and there are only 2 percentage points between Eb and D and A (both at 7%).
Anyway, I agree with your premise that the popularity of a given key seems to largely relate to the ease of playing in it on piano.
Eb definitely sticks out, but perhaps the more surprising one to me is the relative lack of popularity of Bb (5%) — a very popular key in jazz and, with only 2 flats, tied for third easiest (with D) key to play in.
Db, at 7%, also seems unreasonably high (tied with that guitarists’ favorite, E!), somehow vaulting over the apparently lowly Ab (only 4%) and the already mentioned Bb, as well as its opposite mate on the Circle of Fifths, B (3%). Boy, B sure gets little love, tied in popularity with the dreaded F#.
Appropriately though, F# is at least given the place of dishonor, last in line, and C# and especially Cb shall not even be mentioned.
Cb is in the list. It’s called “B”.
Yes, but in some cases in music theory it simply makes more sense to write Cb as opposed to B.
Given that Cb could very easily be transposed up to C — and how ridiculous it is to flat quite literally every note on the staff, I’d rather write in B, to be honest — even as someone who dislikes reading sharps.
I would disagree on that. Most pop songs are not written or arranged by the performer. The are written by songwriters who then ‘shop around’ the song to performers who’s voice would fit the song.
I find it a little surprising that you apparently define “pop tunes” not only as Anglophonic, but (with a few exceptions) as written within very recent memory. There really was music before Buddy Holly was born, kids, and it was pretty popular.
I have to admit, though, that your extremely limited survey confirms my sense that the “pop tunes” in your database are, well, mind-numbingly similar.
Agree. “Popular” songs analysed! It figures. Theres been very little “music” in music for dozens of years because would-be-guitarists have been let loose on the public and this is a huge study of commercialised garbage non-music. Open your ears, go back e.g. to “Mr Sandman” and feel the “music” in the circle of chords, then go back to Beethoven and the masters. They knew about “music”.
reading all of this reminds me of this story:
http://www.thisamericanlife.org/play_full.php?play=88&act=2
[…] guy wasted a lot of time analyzing the chords of popular songs, coming up with something that any songwriter or guitar player could tell him. Still, it’s […]
It’d be cool to do an analysis of song structure, and the chord changes between sections of songs.
For example, the verse-chorus-verse-chorus-bridge-chorus structure, what are the most common changes between verse and chorus, chorus and bridge etc?
And analyse the melodies too! You’ve got time for all this right?
Didn’t Arthur C. Clarke write a story about this? If I remember, the researcher distilled popular music to a series of “perfect” notes.
C/Am/F/G? Of course! That’s an evergreen rock progression that forms the bones of everything from “Where Have All The Flowers Gone” to “D’yer Maker”.
Release the krak…. errrrrrrr… data-set.
[…] I analyzed the chords of 1300 popular songs for patterns. This is what I found. | Blog – Hooktheor… […]
What a great start! (Ignore the pompous fools.)
In the early 80’s, I wrote a program in APL in which I took phrases from a classical music theory book such as “V7 almost always resolves to I” or “ii usually goes to V7, but sometimes to vi,” etc., and applied probabilities to all of these “rules.” Then, using randomization, I had the program spit out progression after progression. Many were really interesting, most were pretty unique, and some were excellent.
I’m sure people have gone far further in this direction than I ever did, but it was a great way to come up with new ideas, and those “rules” stayed with me for decades. (For the hotshot artistes who scoff at such things, try using your innate creativity to play chess against someone who knows opening theory.)
I would be interested in a genre study, to see how the progression probabilities vary between pop, jazz, rock, etc.
Btw, here’s a cool site for guitar players:
http://www.fretmagic.com
Both are surely interesting, you can’t deny the brilliance of some song-writers that do it by ear.
Great site… a few things to keep in mind.
1. Another reason for the “natural” keys to be more common is the occurance of the ‘syntonic comma’, which is a natural phenomenon found in the physics oc acoustics. It occurs when one measures the frequency of “cents’ found within an interval. For instance, the perfect octave , which is the ratio of 2/1 has 1200 cents, but the perfect 5th has 702 cents to sound “pure”. This 2 cent difference is known as the syntonic comma or Pythagoris comma. Traditionally the key of C major was tuned with the perfect 5th between C-annd G to a pure 702 cent 5th. The problem with that is the farther away from tonic one gets the larger this syntonic comma gets making each subsequent 5th in the circle sound more and more out of tune. “Equal temperament” was invented to justify this rift by tuning every 5th to 700 cents thus making every fifth equally out of tune. This is why Bach composed much of his melodic material in C major and his faster more percussive material in C # major.
2. Look into “Alphabeto” notation. This was used in Spain and Italy when guitarists were learning how to teach chards. They called the G chord the A chord because it was the first chord guitarists learned!
While the study documents the (fairly obvious) “what”, it doesn’t tackle the far more interesting question of “why”….why do certain chords sound better together, why do certain combinations of notes forming chords sound “good”, but other combinations never became “chords” because they don’t sound “good”? Why do different cultures prefer different scales?
And perhaps the most important question…how much in Federal grants was spent for this study? [/sarcasm]
[…] that it is popular for a reason. It has to be, right? There has been a lot of talk recently about what makes a hit song. If you look at the number one songs of the last fifty years you’ll see a list that seems […]
As mentioned by comments above, this is immediately evident to anyone with basic music theory knowledge. The most basic of chord progressions, not just in pop music, but in Western music in general, is
iii vi ii V I, which in C Major translates toem, am, dm, GM, CM. This comes from the “natural” tendencies of notes to want to resolve in a particular way, such as the way that a B natural in the key of C feels like it is inexorably pulling toward C, the tonic. IV, or FM, is often used as a substitute for ii, as a pass-through between inversions of I, in a IV-V lift in pop music, and is also found as IV-I in a plagal (or “Amen) cadence.Suggestions for change for GarageBand, at least from a musician’s perspective, would be to include D7, as it is common to temporarily modulate into the dominant and use its dominant (secondary dominant, as it were), and DM, as the dominant of the relative minor.
I absolutely agree about DM and D7. I found that to be missing in this list.
I would love to help you understand more about this topic and assist in you bettering your research, but there are so many holes in this that I don’t even know where to start…
When you study music theory, you’ll understand that, there is a pattern that all chords follow, regardless of key. If you took ALL Western music written since the 17th cen., you would find that the roots generally move in three ways:
up a 4th (same as down a 5th), e.g. iii to vi to ii to V to I,
up a 2nd, e.g. iii to IV, IV to V, etc. and
down a third, I to vi, vi to IV, IV to ii, etc.
In addition, I moves to V and IV to I as they are the pillars of any tonal piece (i.e. the DOMINANTS)
A couple of semesters in music theory will give you a basic understanding about why all of these things happen.
I think you make excellent points but it’s worth noting that chord movement took a huge leap from being functional to ornamental during the romantic period, and a lot of 20th century music strongly reflects that change. What’s more 90% of the music that 90% of the people listen to in the west is mostly post-Beatles and blues-based, which means ornamental harmony is deep in our collective ears and forms our musical ideas on a foundation level as opposed to being an experiment or new trend.
You’re bang-on about how chord theory developed in the west between the 17th century and the romantic era, but I feel the need to add this postscript due to the fact you wrote “ALL” in caps.
Mary Ann has given all the answers you need: root progression, primary chords, functional harmony. You are talking about music from the romantic period, we are talking about POP MUSIC, not Wagner. It’s amazing how people always complicate matters. Ornamental harmony, blah blah….let’s not confuse further people who don’t understand basic harmony in the first place. ‘ALL’ in caps is ok. The basic harmony guidelines still apply, as they have been for hundred of years. Yes, there’s always exceptions, so what, in pop music there’s very few anyways, the biggest one is that a lot of it, just like folk music, is written by using modes. Big deal. It’s all still pretty basic music theory, no need to look into Schoenberg 🙂
she also said generally
yes, i dont understand why you went thru all the trouble, your findings are standard common practice music theory
I’m glad she went to the trouble. It was just what I was looking for as an apprentice.
I read your comments and I thought about a new song from Taylor Swift written with Max Martin and others. It happens to be called Gorgeous. I thought “The chord shifts are magical. I think she/he’s doing something more sophisticated.” I looked it up online. It is literally built around C-Am. I’m stunned. Such simplicity and such beauty. It’s how they joined the melody to that simple pattern.
Max Martin is a HACK. Taylor writes ALL the melodies and Hack Martin
creates a track around it. I’ve seen the ones he sends out to people in
the industry to write over. And he claims half the publishing. He’s a
crook, worthy of zero praise.Not sure why you thought they were ”magical”, I identified them after one listen and switched off right away.
Show me a respectable source that said Taylor writes all the melodies with Max Martin. Taylor is a great songwriter. Max Martin is a great songwriter. He has at least 20 number one hits. How many do you have? Now who’s the hack?
jeez, you actually did this thing?? kudos! Congrats man, that’s a great start. I have always felt that most of the western songs share some pattern of chords, but it is kinda different with Indian music, songs set on Indian Raagas. Some of them have unusual combinations. Random chords.
[…] a system for learning to write music, analyzed 1,300 popular songs for how chords were used. The above shows chords that followed an E minor chord. This result is striking. If you write a […]
I did a smaller and less rigorous project with classical piano music
I’m having a hard time with this post. Let’s try this again. Here’s the link:
https://docs.google.com/drawings/d/1nBU6PDVK3rv1Q09VrxC3QPZbzpcBQ60E4QqmsqOXGKc/edit?pli=1
I’d be interested to see a further breakdown of the keys used. For example, how many of those CM/am songs are C Major and how many are a minor? Are there a significant number of modal songs (for example, a song with no sharps or flats that’s tonally centered on G, which would be G Mixolydian)? If you’ve already done this in another part, sorry, I just got started with this one. If you haven’t, I understand if you don’t have the time, since it would be a fairly substantial amount of work (which is why I probably won’t be doing much of anything with it). Also, and this has almost definitely been commented on already, there are very logical chord progressions (I-IV-V-I, or C-F-G-C, for example) that have been used for centuries in Classical and various offshoot musical styles, including modern popular music. The primary reason that this became the convention is because it is pleasing to the ear, which in turn is most likely because the ear and brain notice the strong and significant mathematical patterns embedded in this progression and others like it. This is also why the I-IV-V-vi (C-F-G-a, or “deceptive cadence”) progression is so interesting: it sets up the pattern the same way, but finishes with a logical but less-anticipated minor chord. Sorry if I’ve been rambling, but I’m a theory nut. Bravo for going through all this work, though! I am thoroughly impressed.
[…] great post here showing what chords are commonly used in pop songs, and even some details about the sequences that […]
Well we only use A Am B Bm C D Dm E Em F and G in our guitarkaraoke app and we think thats more than enough for beginners to play most of the pop/rock songs!
http://www.fourchords.com
This is great keep up the good work
[…] I analyzed the chords of 1300 popular songs for patterns. This is what I found. | Blog – Hookt…In this article, we’ll look at the statistics gathered from 1300 choruses, verses, etc. of popular songs to discover the answer to a few basic questions. First we’ll look at the relative popularity of different chords based on the frequency that they appear in the chord progressions of popular music. Then we’ll begin to look at the relationship that different chords have with one another. For example, if a chord is found in a song, what can we say about the probability for what the next chord will be that comes after it? […]
Regarding the Apple smart instrument chord selection – they are key-based. Your screen grab shows the chords in C major, you can change these by changing the key in the settings for the app. Your suggestions of D and E don’t belong in the key, but the bVII (By shifting the root of Bdim back a semitone to give you Bb major which is an easier sound on the ears and is a nod to mixolydian) is a common addition. But I agree, it would be great to have some other out-of-key chords available. The problem is that when you are in the key of C and you play a D the only place to really go is G. A bit of knowledge beginners won’t have, and similarly playing an E takes usually you to Am usually.
To get technical here… D & E are secondary dominants to the key of C (V of the V chord & V of the vi chord). So they are in fact found in a lot of music in the key of C (e.g. Elton John comes to mind).
Excellent? More like stupid.
His entire results are pointless without the use of Roman Numeral Analysis.
Point #1 he tackles, most common keys. Duh C major is the most common, b/c it’s the most accessible for non-musicians (no accidentals = easy) In fact, even after equal temperament existed, Classical composers didn’t stray far from C. (No Haydn or Mozart Sonata has more than 4 flats or sharps)
Point #2, of course C, F, and G are the most common chords. Duh Once again, in C major they are I, IV, V respectively.
Point #3, what’s to say that I haven’t said?
What he should have done was look at the actual progression using Roman Numeral Analysis, because there we see the relationship of the chords within the key, and how they are functionally used; the key doesn’t matter so much. For example, How often do we see the progression I, IV, V, IV, I, as opposed to I, vi, IV, V, I? Both are very common progressions in pop music.
How hard would it really be for you to do that in your head? That does seem to be part of the point of moving everything to the same key.
Your reply? More like arrogant.
Unlike you, the OP isn’t thumping his chest; he’s presenting the results of his analysis based on basically one aspect.
But Roman Numeral Analysis…that sounds so impressive.
Nice info collected about songs. Wonder if you could analyze melody lines to determine what/frequency notes make up a melody. i.e. what notes work in a triplet best, what is most popular notes in a bar. Any patterns over longer sequences of bars. Just a crazy thought
This analysis suffers because you do not appear understand music theory, and you’ve conflated ‘more common’ with ‘better’. The advice, for example , that Apple should get rid of the VII chord in their smart guitar (in C, the Bdim) is ridiculous. It may be rarely used, but it’s the right chord in that key if you’re going for that sound
Exactly. I’m no expert but even I thought that was a strange suggestion.
you are stupid, he suggested getting rid of the diminished VII chord which is basically a V with more tension. It doesn’t serve the amateur musician with more creative options as putting in a secondary dominant (like D Major) would.. It is a very prescient observation that Apple should change it, that way their customers will use the program longer and be more excited about the range of sounds/chords they can put together with little/no knowledge.
Necessary to call him stupid? lol
Music theory crowd can be a feisty bunch.
^ dumbass
I was gonna say something similar but you’ve said it already! Bdim is a better choice. Other sounds (7ths. 9ths, etc) can be easily added through the keyboard instrument (not the smart one).
[…] I analyzed the chords of 1300 popular songs for patterns. This is what I found. | Blog – Hooktheor…. […]
Is it possible to access the raw data of your database, in a way that a program can work with? If it’s not, I would love that: i’m sure a lot of people would like to run the tests they want.
Lots of good comments here. It all boils down to one thing I learned in music theory. Most western classical and popular “tonal” music is based on the “tonal axis” which is I, IV, V, I. Any other chord used in the progression is essentially a substitution for one of those chords and must resolve, either directly or through the axis chords back to I (or serve to modulate). Good composers know how to break these rules within a given structure and surprise the ear, making the composition interesting. You can analyze forever music by composers who broke the “rules” (like the Beatles who didn’t know the rules) and attribute their progressions to modal scales or whatever but, in the end, it all comes down to good instincts, imagination and intuitive exploration of sound.
Well said, Steve. If you’re composing a song and following the rules all the time, then it’s not much of an arrangement.
Their are two sides to music, being creative and knowing how to be creative. The Beatles knew what they where playing, their are no rules in music other then the ear but not knowing advanced Chords and how they sound can deprive a music writer or artist from letting loose and adding the creativity.
So I agree, songs that follow rules are silly, but songs that follow theory are usually good. Beauty in music is in the eye of the listener, someone’s rap might by (word removed due to bad nature) to me while someone who listens to rap might laugh at me for listening to Richard Wagner.
Any chance you’ll be releasing the raw data for database/music geeks like me to play around with?
We’re definitely going to start making it easier for the geeks to run their own analyses. We’ll have something more concrete on that to say soon.
[…] in this guy’s pop-song chord analysis does my new favorite catchy pop song “Beez In The Trap” fit […]
This may be a simple point but I would guess that C major, G major and D major are even more common than indicated by virtue of the use of a guitar capo. Dylan, for example, uses a capo extensively to change keys, for whatever reason. So you have a sounding key and a “played in” key. I can’t imagine sitting down with an un-capoed guitar and writing a song in, say, Eb. More likely it was in D and recorded with a capo at the first fret. For popular, guitar-based music there’s just no reason to write outside the common “open” keys. Why write a song in C# major, it’s just an unnecessary pain in the butt. Unless you’re one of those people who claim to be able to tell the difference between keys (I recall Rostropovich describing D major as evoking the color blue in his mind).
Actually it is far more likely to be because of the use of horns/woodwin etc typically they are much nicer to play in keys of Eb and Bb, in the same way that a guitar is suited to e, or a. However far easier for a guitarist to play Eb than a sax player to have to play in say E
About the differences in tonalities, being able to hear the differences has very little to do with musical abilities or even gifts. It’s just we’ve been so accustomed to the equal temperament that’s been in use for so many decades that we may have never heard “real” tonalites, whose specificities were due to the uneven division of the whole-step, and that equal temperament rolled away. That’s why transposing a piece of music in any key with equal-tempered instruments changes virtually nothing in the perception (except for absolute pitches) whereas using pure intervals (like in the 16th century) and transposing music with instruments using this system would lead to different moods for the different keys in which they’d be played. I guess Rachmaninov wasn’t making this up and really heard what he claimed. Or maybe he wasn’t a good musician.
Thanks!
The way you explained this gave me a deeper understanding of this.
Very interesting stuff.
Actually, the Smart chord in Garage band is following the circle of fifth pattern. F C G D A E B. Each of the next chord is the fifth from the previous chord if you count it from left to right.
[…] Post navigation ← Previous […]
this is awesome. I havent seen a lot of this analyzed in the way that you have it here. I might even have to take it into consideration when writing some songs. great post keep up the hard work. check these out IStillGotMyGuitar
This was done a long time ago before you. The analysis consisted of about 42,000 songs. http://forum.ixbt.com/topic.cgi?id=26:41146
What exactly are the four chords in the four chord pop chart, in relation to the one chord?
Very interesting. I’m interested in your comment about 4 chord pop songs. My question is: What percentage of pop songs use 4 chords? What percentage use 3 chords? What percentage use 5 chords? etc.
I think it’s a very basic but important question to ask. I often write songs with just 2 or 3 chords – then feel guilty about it! Should feel guilty or should I feel validated?
That’s an interesting question. It’s definitely true that the majority of the chords in the majority of songs are just the 4 standard chords, but it would be interesting to actually break it down further and look at the total number of chords used per song. My guess would be that even the simpler songs throw in extra chords somewhere like the bridge to add a little variety.
If you had someone who was a mathematician as well as a musician I wonder what interesting facts they would find. Since that say that music is really another form of math. An analysis of music by math would provide some interesting facts. similar to when chemists study all life form molecules and discover that any molecule can be a right handed one or a left handed one, but all life forms use only left-handed molecules. Another theory they say that in heaven there is music that is not possible to humanly reproduce on earth, Why?- are there more notes or a scale with different frequencies that can not be reproduced in the range of human hearing?
I know that if you put three nails into a board at the points of a 3-4-5 triangle, and strung a nylon guitar string around them, they would play a major chord when struck . . . how good would geometry lessons have been with that kind of innovation? :o)
[…] analyzed the chords of 1300 popular songs for patterns. This is what I […]
[…] articles related to this topic, for example the ones of Dave from hooktheory. His work is about musical progressions but the underlying math is more or less a more complex version of what I was doing […]
[…] https://www.hooktheory.com/blog/i-analyzed-the-chords-of-1300-popular-songs-for-patterns-this-is-what… Share this:TwitterFacebookLike this:LikeBe the first to like this. […]
No need to apologise up front for what you’re doing. Maths-phobia doesn’t merit it. What you’ve done is so clearly worthwhile, anyone disputing it on the basis that “but stats! that’s one of those hard classes!” is just being silly.
Wow, you painstakingly did some research here. I do not know much about the principles behind the different keys and chords but I think everything depends on the musician. If one is comfortable with a particular key, then he/she would definitely stick to that. Not all singers can sing so high or low keys and surely the most popular keys or chords are the ones usually at the middle section. This is just my own idea, but I do appreciate how you really put some time and effort to study music.
[…] Source: https://www.hooktheory.com/blog/i-analyzed-the-chords-of-1300-popular-songs-for-patterns-this-is-what… […]
There is mystery in music which no one can analyze – As a musician, I also am trying my best to unravel the mysteries of music as this detective work is very fun ! Though most musicians are not physicists or mathematicians, (like Albert Einstein who really was a violinist than a scientist ) music is all about physics and mathematics. When we play an instrument , we are really following the principles of mathematics – When we play, our feel – good hormones like oxytocin also is released more and more and we also feel on top of the world because of these hormones. – priyantha meemeduma , Sri Lanka
Scales of some songs are difficult to find as these tunes are made just by following the ear of the relevant musician. Even that person is unaware of the scale of that particular tune.
I am loving this new site on how to learn music! I am currently taking a beginners music class as part of my GED and I’ve never been exposed to music before as far as reading it and such, but this site has been helping me break it down to fully comprehend it. Not to mention, the videos are very beneficial as well. Thanks for the support! Keep it up.
Thanks for useful explanation! I learn some new
stuff about music chord and its pattern though I’m just learning.
This my first time studying music and its interesting just knowing how music relate.
[…] tried to google “patterns in popular music” and I came across this interesting post noting the chord patterns of approximately 1300 popular songs. It’s kind of a long post […]
These documents are referring to what is obvious, though it is called analysis it does not go into the deeper questions such as “why” and “so what”…why do notes and chords sound good and combining these form altogether a good combination is not illustrated, but other combinations are not accepted as chords because of a dissonate sound, baffles me.
[…] Fame and Money – In a recent article by Forbes magazine they placed DJ Tiesto at the top of the pile with his yearly earnings at $14 million. Now we all like to be rewarded for our skills and efforts, but does anyone else find this a little excessive? I am sure Tiesto doesn’t mind having all that money, but has his love of the music been tarnished by the necessity of focusing on the next marketing strategy or the bottom line? With money, comes pressures and also a loss of control. Record labels and agencies take control of your creativity, turning your act into a product. There is even a magic formula for chord progressions in modern day pop music, a formula that creates hits. Just go ahead and search the internet for some information. Here is one interesting article I found – Hooktheory. […]
I found out that there’s a lot of patterns involve in music theory. Great job!!!
A isn’t a common chord at all? what is this tom foolery. But in the “common chords coming after em” there are tons of songs that have c coming after them in Em
Just Saying, but your second graph doesn’t make sense. Percentages are meant to add up to 100, not 333. I just think its a shame because it doesn’t do justice to the magnificent analysis you’ve conducted and the results you’ve found.
[…] Fame and Money – In a recent article by Forbes magazine they placed DJ Tiesto at the top of the pile with his yearly earnings at $14 million. Now we all like to be rewarded for our skills and efforts, but does anyone else find this a little excessive? I am sure Tiesto doesn’t mind having all that money, but has his love of the music been tarnished by the necessity of focusing on the next marketing strategy or the bottom line? With money, comes pressures and also a loss of control. Record labels and agencies take control of your creativity, turning your act into a product. There is even a magic formula for chord progressions in modern day pop music, a formula that creates hits. Just go ahead and search the internet for some information. Here is one interesting article I found – Hooktheory. […]
[…] very extensive research can be found over at hooktheory. It’s a very informative read, although they do reach different conclusions than I did […]
This book is very helpful, provind a very affordable alternative, while offering videos and links to help understand music. Thanks!
This seems more like an article on how to write the most cliche predictable music, rather than how to write good music.
Who said it was about how to write good music?
Quiet! Music has spoken. Music can’t be wrong about itself now.
If a song is not somewhat predictable, it does not connect with people..
YouTube the duck song and let me know if your opinion changes. Also look how many have viewed that video/song. ha
Well said, Justin Mitchell. Simplicity sells. It’s been that way forever. The average ear doesn’t care if it’s predictable, all they know is that it sounds familiar, which is key for a songwriter trying to stay relevant in the music business. Some of the simplest tunes on the radio are composed by musical geniuses who understand that if you want to eat, you have to dumb it down. After all, the musically inclined aren’t the ones buying it. Success in the music business comes by way of connecting with the audience, not by confusing them with your musical knowledge.
You’re confusing simplicity with hackneyed. Look back at the 60’s at how different everything was. The music was ”simple” but it was different. You had about three chord structures in the 50’s – and we’ve gone back to that safe shit in this century!! By the way, The Beatles never ”dumbed it down”. Neither did the Bee Gees, The Beach Boys, Bob Dylan, The Rolling Stones. I could go on and on.
Don’t most Beatles songs only use 4 chords?
For example: Here comes the sun literally follows I, IV, V progression and it replaces the vii chord with an A7 instead of the A diminished it should be, which is far from inventive.
Bullshit. Look at thje kind of music that was coming out in the 60’s, 70’s and the 80’s – and then go back to the 50’s and compare them. Everything is night and day away from that boring twelve bar blues, C F G, C Am F G borefest – and we’ve gone back to it in this century.
If you can determine which chord is most likely used following another, is it possible to come up with a “perfect” pop song relating one chord to the last for 3 minutes? I would really like to hear what you would come up with!
[…] songs to discover the answer to some interesting questions about how popular music is structured. Click here to read Part 1. Click here to read Part […]
cool theories dood, just a note on the garage band pic above- reason Bdim is used is bc all those chords are the diatonic harmonization of the Cmaj scale w exception of the Bb which is an oft used bIII of V chord (the b3rd, Bb of G). Bdim is the VII chord of the scale; the leading tone chord. The point which ur making is true tho, those are very popular chords. C, F, G are primary chords which contain all notes of the scale within them… cdefgab. cheers, program looks cool 😉
[…] 2,375 this might help I analyzed the chords of 1300 popular songs for patterns. This is what I found. | Blog – Hookt… […]
this is great exaclty
[…] is called “jazzywalk”; so named for it’s almost frenetic beat, bass lines and chord structures, but simple, smooth melodies. I recorded a previous version of the song with echoed saxophones on […]
Sure it’s been mentioned. But in GarangeBand you can actually modifiy the chords to the ones you want in the song settings.
This is just a confirmation of what all knowledgeable musicians already know, music goes in fourths or fifths. Garage band understands this perfectly well, the Bdim chord is the 7th harmonized tone of the key of C, although its not in the circle it completes the chord in the scale.
This read was very interesting. It analyzed the pattern of many used chords in popular music. It provided many useful constructive videos and links. Looking forward to reading more.
This caught my eye at first then I read about apple not knowing what they’re doing putting a B dim chord in there…. Well B dim is the same thing as G seven flat nine… they are interchangeable…. I stopped reading at that point you don’t know what you’re talking about
Why is there no E sharp?…
When people play music on a certain scale, they don’t play the pitches in that scale sequentially. They just use it as a reference (building block), then build chords, then chord progressions, arpeggios (broken chords), etc. Then the next issue is han…
Please, where can I find your complete chord transition probabilities matrix?
We don’t have a public API but we just released a new feature that lets you interactively explore chord transition probabilities. It is called Hooktheory Trends. http://www.hooktheory.com/trends Enjoy!
[…] Analysis of 1300 popular songs for patterns […]
I smell Markov chains 🙂
Hi, Allan. We just released a new page that is essentially an interactive Markov analysis of every song in our database. I invite you to check it out. http://www.hooktheory.com/trends Cheers.
[…] stumbled across an interesting website with an entry that analyzed patterns in pop songs and realized something: Am is a popular chord. Not only that, […]
[…] have even been studies of how often different chord progressions are used. However, the main ideas I want to get across right now […]
Jodi Parkinson
After Reading Chapters 1 and 2 of the Hooktheory book I was very pleased. I thoroughly enjoyed the teaching style and the clarity music theory is presented with. Being someone
that gets easily frustrated when I don’t get technical aspects of music right away, I really liked this easy to follow guide. It seems to me like a great way for aspiring musicians and songwriters to get a solid foundation in what makes up a song. It is a “Music for Dummies” in a sense where even a non-musician can understand the examples and learn quickly from them. I find it a great alternative for those not enrolled in a music theory class because there’s no teacher to confuse you but instead you get to set your own pace while reading and you can re-read at your discretion.
The text is quite immersive because it has videos to play of audio. You get to visually see what you are hearing, which makes it great for many types of learners. The examples
they provided are songs that are pretty updated and current so it makes you feel as though the material being taught is relevant in today’s music world. It also has a summary called a “Wrap Up” to remind and recap the material you learned in the chapter. This is helpful because it solidifies the concepts that they have presented for you and rephrases and condenses in order for the reader to better comprehend. There is a “Check for Understanding” quiz at the end of Chapter 2, as well. I enjoyed this because you actually get to test yourself to see if you really got the knowledge and understood the information, and if you
didn’t, you can always go back and search for the answer.
One downfall of the chapters I read was that there were some typos like “you you” and also in “Check for Understanding” the last question has the same answer twice both A. and D.
so either one of the choices should be changed or the right answer changed to both A. and D., if I’m not mistaken. Besides that, I say, if just a bit of editing and typos are the only issues that I can find, this method of learning is great. I really enjoy learning Music Theory, and this website and book are great sources for that, because I would like to end up writing some coherent
and technically apt songs.
[…] Originally Posted by Mahlerian There are also the studies saying that pop music has become increasingly homogenized, with fewer types of chord movements used. It's no surprise that if you limit yourself so severely, you'll run into repetition before long. Art music has been renewing itself for centuries, and it doesn't seem to be slowing down now. Mahlerian, that reminds me of this… https://www.hooktheory.com/blog/i-ana…-what-i-found/ […]
This article was very intriguing to me and makes me want to read further. I also think that analyzing the frequency and placement of certain chords could be a huge help when writing our own music. I found it to be a very true statement that many of the most frequently used chords are common becuase they are easier to play on instruments such as the guitar. I know that when i pick up my guitar to play or to try and write a song the first chord that i usually start with is G or C. I strive to break out of the box of the basic chords used in most songs and the patterns that make up many of the popular songs. By using the method that you have presented I can now analyze some of my own favorite songs and fin out if what I love has a certain pattern to it, and perhaps find a way to make that pattern my own in songs that i write.
Hooktheory is a great way to analytically look at music academically and from the beat of the sound. It gives a one on one music set to show you what they are talking about and how it is repetitive etc. It gives lessons on songs people are familiar with so it keeps students interested and focused. I think hooktheory is a wonderful way to learn the rhythm of a piano. It is a easy beginners instrument and has simple key notes to play from to advance in music.
[…] songs for artists you love is one of the most valuable exercises a songwriter can do. Check out this series of blog posts, where the authors analyze over 1300 pop songs and give you data on which […]
Really interesting! As a jazz musician, spending hours a day honing my skills and learning advanced theory, I’m always dismayed by the mediocrity of pop music, and how harmonically bland it is! And here is further confirmation. It’s really a shame the masses can’t seem to appreciate more complex harmony; there are so many more possibilities!
To Mary Ann (though I know it’s months ago): your statement is far from true. If you continue another month or so in classical theory, you’ll find out about secondary dominance, aka tonicization, or implying another key — for anywhere from a beat to a movement. Composers as far back as Bach have done it , but as proven in this article, it is largely ignored in pop music. Bass motion out of your paradigm is required for this very often. Then there’s also the ubiquitous plagal (IV-I) cadence, and when you delve into diminished harmony, you can have a root go virtually anywhere. Not to mention 20th atonalism, or Coltrane matrices (up or down in major thirds, with tonicizing dominant chords)…
There’s such a wider range of emotions that can be conveyed with more harmonic options. For me, an unexpected chord change creates excitement, and I’m more likely to like the tune. The predictability of pop music either makes me tune out totally, or be able to harmonically analyse a tune in my head on the first listen — which usually tends to impede dancing…
What partially frustrates me are all these “arm chair music theory experts,” to quote Joshua Jones, who believe such are the only ways of doing music… Be CREATIVE and EXPRESSIVE, and think outside the box, a little. Theory is just that, a theory: this should work, and it should sound good… but there are other ways of doing things.
My Jazz theory teacher taught me a lot, and then, towards the end of the year said: “Now forget everything I’ve taught you. It’s time to stop thinking functionally, and start thinking about the vibe and the sound you want to create.” But, paradoxically, you need a good grasp of basic theory before you can start doing it, otherwise it’ll probably end up sounding like crap.
Anyway, long story short, I wish pop was more interesting!
(Also, I meant to tag @94508d8109392b81ac110920bc6d9497:disqus in my previous post)
There’s a lot of unfortunate misconceptions and fallacies about music theory.
”Be CREATIVE and EXPRESSIVE, and think outside the box, a little.”
Yes, I have heard that stuff a million times, and a million more times after that. It’s useless. And it’s a common fallacy that music theory and creativity are opposed. They are not, but it is thought they are, by people who learned a little music theory but never mastered anything about it. Music theory gives nothing back to people who don’t go well past the initial stages. You don’t have to master every single thing in it, but if you only learned a couple of scales and some chords and the circle of fifths and then said ‘nah, this is not fun, I’ll just play’, then I am not surprised that you got very, very little out of it.
‘Be creative and EXPRESSIVE’. These are such useless suggestions. This kind of advice is not helpful to the aspiring musician who’s trying to find practical, real answers about how to learn or improve their skills. It’s like telling an aspiring writer who just know how to write and read: ‘Here, this is a pen and this is some paper. Don’t worry. Just be inspired and write me a really interesting and wonderful book.’.
Then there’s the smart aspiring writer who MASTERS grammar from inside out, understand synonyms, understand structure, understand all he can about his language, because he knows that his knowledge about language will be his asset as a writer. Given that both aspiring writers have equally good ideas to write about, who do you think will write a better book ?
And I am no snob, believe me. I love all good music, from medieval popular songs, to rock, to Bartok, and Beethoven, to rock and roll, to jazz and blues and circus music, to horror movie soundtracks and a dozens others styles and genres. I am no ‘arm chair theorist’. This is also a misconception. There’s the kind of musician who starts as a musician and ends up as a professor who teaches music theory in a college and stops making music because of lack of time (hey it’s understandable, you can’t do it all), or because he started a family, or because he needed more financial stability, who knows? All musicians know that it is VERY hard to financially support yourself as a musician, no matter how skillful you are. But there’s no correlation about being an ‘music theorist’ and not being creative. Indeed, it’s exactly the opposite. I have DOZENS of great books written by ‘music theorists’ and ‘boring’ music professors, and in ALL of the books they write their own musical examples. All the musical examples are themes, melodies, pieces, etc of the highest standard. I have personally never known someone who knows a lot of music theory that isn’t able to make, at the very least, good music.
It is true that there isn’t one way to make music. People have written music by looking out of the window, by using random cards, recorded tapes played backward, the fibonacci series, all kind of wacky stuff.
But music theory isn’t there to say how you should write music. It’s there to tell you all it can tell you about music itself, the rest is up to you. It would be logical to believe that if you want to do something, why not knowing as much as you can about how it works? But the reality is, people are lazy. They just want to make music without learning anything. I am not saying that’s you, by the way. Just speaking in general. Cheers
Yeah, damn them kids!
“really a shame the masses can’t seem to appreciate more complex harmony”
So you are shaming people because they do not like your music. People automatically tap their feet and sing along with basic pop songs. Tough luck.
Your tone is antisocial. Your music is a projection of your inner desolation. Your hatred of the “masses” is an echo of your psychic relationship to the world.
Recall “The Fool on the Hill” lyrics:
“Well on the way, head in a cloud, the man of a thousand voices talking perfectly loud. But nobody listens to him…”
That’s why it’s called pop “popular”. Why is it popular? Because it’s simple. People remember it. That’s the goal. Repetition with a simple hook. Every genre can’t follow these rules. If they did, the whole world might stop spinning.
You’re an imbecile. I’ve already explained that you don’t understand the difference between simple – ie: easy to sing along to – and cliche’ ie: riffing over the same chord structure/ same tempo/ same instrumentation, which is what top 40 ”music” mostly does today.
Please don’t use that song as an example moron. That has an incredibly interesting change in the middle, a change that cannot be found in any song in this weeks top 100, or probably this year.
I think song structure specifically is the issue you’re talking about.
You can write an epic in a simple key, you could make it stylistically interesting and keep it simple in other ways.
I’m not arguing with you, just sayin 🙂
[…] year, we discussed the first results of a long term effort to study the patterns found in the chords of popular songs. The reception that we got was incredibly positive, and we received a ton of great […]
[…] Een E mineur akkoord wordt overigens in de meeste gevallen opgevolgd door een F of een A mineur. Hoe ik dat weet? Lees maar eens het artikel ‘I analyzed the chords of 1300 popular songs for patterns. This is what I found.’ […]
[…] a lot of fun to look at, and there’s a blog post you can read about the statistics uncovered from it, but the most interesting part is probably that it can predict what chord will come next in a […]
[…] a lot of fun to look at, and there’s a blog post you can read about the statistics uncovered from it, but the most interesting part is probably that it can predict what chord will come next in a […]
At this point I’ve gone through numerous versions of what I want to comment with, either in my head or in my own temporary document on my computer, and I think I’m finally ready to post it 🙂
First off I just want to say this is totally the kind of thing I would do 🙂 I very much enjoy figuring out how to play songs (piano’s my main instrument), and thinking about what the chords are doing. There are certain characteristics I have identified as very frequently resulting in a sound I really enjoy. For example, there are a number of vi-IV changes that I am a big fan of, as well as iii chords in general are very frequently my favorite chord of a song. And I have my own theory on why that is 🙂
I spent a while being rather confused by some things on this site, such as I7 and V7/IV being treated like two different chords in your Trends section. To me it seemed like they were the same thing, because I was just observing, and for a long time didn’t happen to actually listen to what I7 was sounding like. I play mostly non-classical music for a living, and in any charts I see, C7 will always mean dominant 7, whether the song is in F or C or whatever. When I finally noticed that I7 on your site means major 7, I had to think really hard about: is that really the way I learned it in music theory class, and I’ve just totally forgotten because charts use a different standard? Which does seem to be the case. This put me into research mode, looking up things about roman numeral analysis online, and making a facebook status that actually got a decent discussion going amongst my friends. At this point I have convinced myself that I would Not use an assumed diatonic 7 for any personal analysis I may do, as well as I wouldn’t be so strict to limit analysis in general to diatonic stuff, however that is a discussion for another place (do let me know if you’re interested!). For the purpose of this site, I will accept that 7 is implying diatonic, and in general having to stick to diatonic analysis. But yeah… you could possibly note that somewhere, as I think pretty many people interested in playing pop/rock would tend to think of 7 as implying dominant.
One of the biggest things bugging me, even after accepting that the analysis has to stick to diatonic (and that I can’t put in details like augmented, 9, 5, etc… personally I think that would really add a lot to what all you could analyze), is what to do about a song with what I call “4 over 5”, or really any chord that has a different bass note that is NOT an inversion, 4 over 5 being probably the most common. One such example is the song Easy by the Commodores. Currently it is analyzed as I, iii7, ii7, V. But that is so not an Eb chord. It’s Db over Eb (Db chord with an Eb in the bass). The closest you can get is V7sus. I suppose that many times it is kinda acting as a suspended chord, but I feel like treating it instead as a chord simply with a different bass note would be very beneficial in some circumstances. It’s another reason why I wouldn’t stick to only the classical analysis stuff. Another such example is Come Sail Away (not currently in the database) by Styx… once the rock part starts, it’s C, F/C, G/C, F/C. F/C is IV64, sure… but G/C? Do we really need to downgrade that to calling it Imaj7? The approximations I would have to use in these cases just seem not good enough to me.
At this point this is no longer as relevant, since you can change the mode, but I do like treating songs in minor as the “tonic” being 6 instead of 1, as I think the mode of many songs is often way too fluid to treat them with their own tonics as 1. As I see in one of Dave’s comments, “Pop songs tend to flirt with the parallel minor…” But now that you can analyze them with multiple different modes, problem solved!
One last thing I want to suggest is that you could give two different options of how to analyze the same chord. For example, in Forget You, it’s I, V/V, IV, I. But since V/V doesn’t actually go to V next… maybe we could analyze it as a borrowed lydian II. Also, borrowed bVII vs. IV/IV. I’m sure there are more examples, but those are the first that come to mind.
Well I look forward to future updates, and I most likely will edit and contribute, as I really enjoy this kind of thing 🙂
OH! P.S. I remember seeing a couple songs with a chord analyzed as sus2. That does not appear to be an option for us to analyze a chord as… only sus4… How???
Now, instead of replying to each person’s comment individually, I’m just going to list some of the things I would respond with. First off, I’m amazed at how many commenters didn’t really read your article… it IS relative to the tonic, idiots. Also amazed how many people seem to think you could/would do all this without knowing music theory.
I agree that duration is a good thing to consider when deciding what chords are most common. And yeah, I don’t think Pandora recommending songs based on having the same chord progression would be a good idea haha.
I disagree with the people saying C is not the easiest to play in on piano. Sure, your thumb is lower than your fingers, so a B scale is technically more comfortable, but speaking as a professional keyboardist, I would still say C is easier to play in, because I don’t even have to think about which fingers need to be on a black key. Not to mention, when playing, your hands don’t always play things the way they come up in a scale, so what scale fits the hand best is not the same as what key of a song fits best.
Of course there’s roman numeral notation, which clearly at least at this point is what the site generally uses, but I approve of the idea of transposing everything to C, as that’s frequently what I do in my head when I don’t know what key a song is in… or even sometimes when I do :). It sometimes actually connects faster in my head that way than the roman numeral way of thinking. Especially when you get into alterations. Like I saw an analysis with something like vii°/V… yeah, no, my brain pictures the chord better as the F#° in C, and for me that would even click faster than, say, G#° in D. So both methods are useful I think.
What Philibert mentions concerning how chords can be chained together by easy movements may not apply much to the bass note of the chord, however I would say that definitely has an effect on what kind of that chord they use. For example, I think guitar songs in E are more likely to use Bsus than piano songs in E, simply because that’s easier to play than a normal B on guitar.
A commenter mentioned B diminished not being diatonic. I think this reflects an assumption people make that I HATE, which is that when unspecified, diminished automatically means diminished 7. I think it shouldn’t… to me, diminished only implies the triad unless you actually say that it has a 7. Because otherwise, how would you imply only the triad?
Connected to that thought is another comment about considering scales that a chord/melody are based on, rather than just the chord. There certainly is another level of analysis when considering the scale a part is based on, as opposed to just the chord, but I think frequently considering the whole scale is not necessary. That would be the distinction between whether diminished implies fully diminished 7 or just the triad.
It’s interesting how people comment about analyzing music this much being not a good idea to influence songwriters. The thing is… I love analyzing music like this… but when I write a song, I don’t think about this kind of thing very often – I usually write it the way it seems to go best, however that may be. I guess my point is, which I think you get, is that they aren’t as connected as those people are making it out to be. You can analyze stuff like crazy and still do whatever you want. At the same time, I do occasionally make a point to use something that I know I generally like, which is why all this stuff is not pointless.
Of course “pop tunes” would be ones within very recent memory. If that weren’t the case, you would have to include everything that was at some point popular… I’m sure Beethoven at one point was “popular” music.
Dean Lombard mentions that “more common” doesn’t mean “better”, but it’s definitely worth thinking about it that way, as there’s probably a good reason why it’s more common. And just because it’s in the key doesn’t mean it’s going to sound better than a chord not in the key.
Seeing as at least one commenter was confused, it might be good to mention for the second graph that the percentages are referring to the percentage of sections that include that chord (at least I think that’s what they mean… exactly why you should say it somewhere). Though the percentage of time that is occupied by the chord would be another interesting and possibly different graph.
I agree that analyzing key changes would be an interesting thing to look at, and how it gets to that new key. Also, analyzing what happens in other transitions, which seem to currently not be included, since sections are separated without indiction of which are connected directly to each other, although I see sometimes the first chord of the next section (or last of the previous) is included (that could probably use standardizing). And, in addition to noting how many chords a song has, how many chords each section of the song has.
I really have no clue where to write general bug reports or see what bugs you already know about… so here we go:
The names next to “Analyzed by” does not always update with all of the names, and I can’t figure out any pattern as to why. I added a whole section to American Pie and my name doesn’t show up, I merely edited a few chords on Easy and my name does show up, and I edited/added things on Clocks and philipvanderleest’s name (which wasn’t there before) suddenly showed up along with mine, even though his edit was months ago.
(and when the “Analyzed by” part is not updated, the song is also not bumped up in the New or Recent changes section)
And now after adding the melody to the existing section of American Pie, my name’s appeared, but the others have disappeared!
it definitely looks like something screwy is going on with song edits. Thanks for reporting these issues. It looks like just a display issue so hopefully the fix will be an easy one.
Re. the relative popularity of Eb … that’s probably because Eb is a common “rock tuning” for guitars: drop all the strings by a semi-tone and it’s easier to bend the strings and get all rocky-outy.
[…] I analyzed the chords of 1300 popular songs for patterns. This is what I found. | Blog – Hooktheory オモシロい！ コード進行がこうやって学べるんだ・・・ […]
If you ever find yourself at a loss for something to do, I would be delighted if you would do the same analysis for flamenco. The results will be very different, and very useful, though to a much smaller cadre of musicians.
[…] a lot of fun to look at, and there’s a blog post you can read about the statistics uncovered from it, but the most interesting part is probably that it can predict what chord will come next in a […]
I can’t believe how misguided this all is. You’re actually encouraging people to dumb down pop music further by repeating sequences that you think they should just because of past trends. Nothing evolves if nothing vaires. Key signatures are just a convenience. They don’t have to suck the life out of everything.
There’s no such encouragement. I for one have had the same question as the author. It’s a daunting task and it may not be considered scientifically valid, but it’s a start.
[…] I analyzed the chords of 1300 popular songs for patterns. This is what I found. – Dave Carlton For many people, listening to music elicits such an emotional response that the idea of dredging it for statistics and structure can seem odd or even misguided. But knowing these patterns can give one a deeper more fundamental sense for how music works; for me this makes listening to music a lot more interesting. Of course, if you play an instrument or want to write songs, being aware of these things is obviously of great practical importance. […]
[…] […]
Awesome! Have you considered how many songs will be written on a guitar first? I was changing my guitar strings last night and giving the neck a clean and noticed that I’m starting to wear indentations on particular positions of the fretboard. The most prominent of these was the 2nd fret on the 3rd String as I prefer to write songs with chords that have a lot of open strings in standard tuning – with C and Em chords being the most frequent. I wonder if other songwriters have this tendency?
I really liked the way, Hook Theory made excellent points on iii chords and ii chords. I learned the difference between the two chords. The iii chords has a sound that feels far away from the base. Compared to the other basic chords and it is not used as often. The iii chords have to be used carefully in shorter chord progressions. The ii chords do not have any restrictions as to which chords must follow. II can be used in a number of different ways. Very pleased with the excellent points on the different chords.
[…] | HookTheory, Popular […]
I think your analysis for GarageBand is a little faulty. You may notice that if you go through the options, C is the default key, which you can change anytime. D and E are so common because they’re the V and VI of G, which puts them in a few common chord progressions. It would be wrong to include either of them in the Smart Instruments for the key of C because they simply don’t belong in that key; if you simply changed the key, you’d find them.
[…] Thanks to Grant Potter for pointing #IntroGuitar to this encyclopedic musical resource: Hook Theory. From the Blog: […]
I give you much respect for this endeavor. As a musician with 19 years experience and 6 studio albums to my credit, I have to say, I really don’t give a shit about your conclusions. I don’t consult a book when I write songs, I don’t listen to what is popular to get ideas to express myself, I just write music as I see it. I do have influences that I draw from, they are NOT in the 1300, GUARANTEED!!!! By no means am I trying to insult you or demean your work. Like I said, I respect what you have done here. What I question is the intrinsic value of the opinion of the masses, especially when it comes to creative endeavors.
chill. Your music’s gonna suck either way.
Brian – after 6 albums, you are still utterly unknown and completely oblivious to how much your music sucks. Ironically you have more to gain from this research than most but I’m sure you’ll just keep on truckin’ with your very white, sub-Dave Matthews unusual time signature bass instrumental jams.
This is a database about the most common chords used in pop music. It’s not a personal attack. Really weird reaction.
Maybe you will read this post in a year and look back on how pompous and unnecessarily defensive you sound.
Also, I’d love to see if your music actually follows these chord patterns or not. You can still be using the same chord patterns even if you start in a different key. The common leap from C to Am is no different from the leap from Bb to G.
Two very different songs can be in the same key.
This was just an analysis, he wasn’t suggesting you redesign how you write songs.
Just a note though – mentioning how many albums you’ve made gives you no gold stars, why not link to your work?
[…] I analyzed the chords of 1300 popular songs for patterns. This is what I found. | Blog – Hooktheor… ← The Grocery Store and Wine Debate […]
[…] us how the editor has helped with their songwriting. You created some waves online with your music analysis blog posts last year, using data on real pop songs to explore some facets of chord theory and validate the […]
[…] here. For guitar […]
In reality, it is quite simple, if you really understand harmony, the study of chords. There’s no need to complicate something that, in fact, is NOT complicate at all. In any key, there’s 3 chords, called ‘primary chords’. Taking a key at random, let’s say C as it represents the white keys only, on a piano, the primary chords are C, F, and G. These chords are always used the most because they define the key. The other chords in the key, which are Dm,Em,Am and Bdim, are used with less frequency, and are called secondary chords. If you know that ANY diatonic melody can be harmonized by simply using primary chords, you are already ahead of the pack. Pop music doesn’t uses chords any differently than folklore songs that have been written in the West for hundreds of years. You first use primary chords. The secondary chords are added as variants, if you know how. As for why certain keys are more or less used than others, that too has an easy and logical answer: the keys most used are the ones idiomatic to the instrument that the song has been written in, on guitar it’s the keys of E and A, major or minor, because it uses open strings, and on the piano it’s the key of C, because it uses only the white keys. As for why a chord follows one instead of another, it depends on ROOT PROGRESSIONS and if the two chords have notes in common: if they have notes in common, the ear finds easier to connect them and they sound good because there’s a common thread between them. Guys,
music theory has had all the answers for hundreds of years, yet people
keep avoiding it, thinking it’s difficult, and then go around in circles
for years and look in the wrong places with the wrong tools. There’s no need to fuss over. Also, very important, is to know that a lot of pop and rock songs are written in MODAL keys. Again, music theory has all the answers, there’s no need to chase a rabbit by using a truck.
May I chime in a guess as to why E flat is listed as common. Guitars are often tuned down a half step (Jimi Hemdrix did this). This way, E flat can be played idiomatically, by playing the standard E major chord it sounds the E flat because it is tuned down a half step. One reason for tuning down a half step (so I was told) is that the guitar can be played with wind / brass instruments. Also, it reduces the tension on the strings and changes the sound.
Furthermore, the standard concert pitch was raised after the even tempered scale was adopted to accommodate the piiano and key modulations. The even tempered scale is dissonant vs the Just temperment tuning, therefore, they raised the concert pitch to make the sound brighter and to mask the dissonance.
During the 60’s, eastern music with just intonation (such as the sitar) became popular. Many bands detuned their instruments in an attempt to capture this sound within the western pop context.
Musicologists can explain this better than I can.
Hey all,
Does anybody here actually have and use that iOS garageband app? I am fairly certain you can change the chords available in any key they just give you a template of the usual suspects. It is an amazing app if you are a player or brand new to making music. Really. Try it… I dare ya.
Honestly, as a musician, I could have reiterated this theory by memory without the painstakingly, unnecessary research.
I’m surprised by the outcome of the most-used-key part of the study. E flat major is universally loathed by guitarists, and G, D, A, and E are all easier to play on guitar than C and F. Granted, C is easiest on piano and pianists don’t ever seem to mind E flat; but considering how much pop music is guitar-based, I’m still surprised.
I’ve never understood the oft-repeated argument that E-flat is in the tessitura of most singers. There are high notes and low notes in every key. And if it were true, then why is D major, a key that’s only one semitone away, relatively easy to play on piano, and super easy to play on guitar, so unpopular? A major has the same number of black keys as E flat major, and it’s a guitar-friendly key. Why does it score even worse?
I’m not disputing the evidence, but are there better explanations for the relative popularity of E flat? And why D and A are so unpopular?
I’m guessing that Eb has two things going for it:
1) the graph showing the most commonly used keys shows only major keys on top (i.e. Eb major), but less prominently underneath each major key, it also lists the relative minor key. Each bar in the bar graph thus apparently represents the popularity of a major key AND its relative minor. Accordingly, saying that “C major is the most popular key” isn’t really what the graph shows. The graph rather shows that, leaving modes and other complications aside, *some combination* of C major and A minor are the most popular pair of relative keys. What portion of the given 26% for that pair should be ascribed to C major and which portion to A minor is unknown.
Thus, saying Eb major is the 3rd most popular key is simply not what the graph shows. It shows that some combination of Eb Major and *C minor* is the 3rd most popular pair of relative keys. I’m guessing that in this case, the minor key has a lot to do with this pair’s popularity. A lot of untrained piano composers can only play in C (or, for the more ambitious, plus G and F for variety’s sake). A minor would be the most accessible minor key by far (followed by E minor and D minor — the triumvirate of easy minor keys that also works well for acoustic guitar), but for those used to working in C, flatting the 3rd, 6th and 7th of the C major scale to make C minor would be very natural. To a lesser extent that could apply to guitar too, sticking to at least familiar chord letters (C F G), though in their less familiar minor forms — at least the names are familiar and it gets you out of playing in A minor *again*, lol.
2) Eb is the basic key for alto sax (the most common sax), as well as baritone sax. Also, it’s a very close key for Bb trumpet (just one flat away) and other Bb instruments (clarinet, the popular tenor sax, and of course trumpet). Thus Eb major is a good key for use with horns. I would have thought Bb would have been a more popular key though — it certainly is in jazz.
Oh yeah, and i gotta add that, personally, I hate Eb!
[…] in it – and it’s a key that’s difficult to play even on guitar. According to the Hooktheory blog, about 3 percent of all popular songs analyzed were in the key of B. You may not be able to tell […]
ok I see none of the more chord savvy folks talked about why this was a good study for whom it was done by. The reason I can validate this study for the cat who took the time to do it is for the following reasons. He analyzed the modern “pop” tune. Its important to state, just as I IV V IV I is to blues and II V I is to jazz the I IV chord progression is quite indicative of pop music since the 50’s and 60’s. It is the quintessential singer song writer chord progression. It is played the most because it is easiest to form melodies over, particularly for people who may not take their playing any further than what they need to do in order to write their ideas down. I come from a background of Jazz fusion, bebop, big band, afro cuban, brazilian, african, music from china and other parts of asia, as well as middle eastern music, ragas and the like from india, born and raised in southern cali, grew up in the 80″s so a big hip hop r&b fan, MTV first aired when I was like 10. I’ve travelled the world and played music all over the world. Point is, I have a lithe (not much) but a little background from which to speak. Chiding this crew for their work because their not talking tritone substitution or talking about how the W-H-W-H-W-H-W-H steps relate to the Diminished 11 maj 7 chord which is actually one diminished chord built a half step above another diminished chore i.e. D-F-Ab-C#-E-G etc etc is not cool, because some people actually don’t know that. Me for one, despite all that I did know about chord progression it still took a friend to point out to me as I was embarking to write some pop music , that it too has a structure, much like jazz has it’s b9#11 and classical has its German and French 6th chords, pop too has a method to its madness. Yes it is simple, but remember you don’t have to learn from someone that’s smarter than you, they just need to know something you don’t. Good work guys, I applaud you for taking the time to do the study. Now of course if you take a music theory class, a good instructor would have been able to already tell you that. But still, there’s nothing likeself discovery to prove a question of why you should do something…Thanx for reading midiprinz@gmail.com
You must have been looking at orchestral music. I’m illing to bet that if you’d looked at folk, blues, rock etc, in short any music involving guitars predominantly, you would have found that A and Eb changed places.
I hope apple doesn’t change the diminished chords, for those of us songwriters that like to use more than 4 or 5 chords and also happened to be obsessed with 60’s pop 🙂
Most songs are initially written on the piano, and most people don’t want a song with any black keys in it. So they use the key of C or Am to write their music, and and the chords that go well with each other in that key.
I guess most people don’t know about chromatic mediants, falling or ascending chromaticism, sudden modulation, secondary dominants leading to surprise chords, moments of atonality leading to another key, common tone theory, and other techniques to spice up your chord progressions. There’s nothing wrong with being predictable, but there’s also something nice about being surprised.
Just wondering about how you define the key that a song is in. For example Born This Way by Lady Gaga is in F# Mixolydian (with chords being F#, E, B, F#). So do you consider it to be in F# Major or B Major? When you transpose it to C Major, do you consider the chords to be [C, Bb, F, C] or do you consider them to be [G, F, C, G]?
Or similarly, a song like Lose My Breath by Destiny’s Child which is in G Phrygian… do you consider it to be in G Minor or C Minor?
I heard nietzkov first album and I don’t know which chords he used but sound classy
[…] analyzed the chords of 1300 popular songs for patterns. This is what I found. | Blog – Hooktheory http://www.hooktheory.com yoshi uemura 0 minutes ago music © 2013 Annotary, Inc | Terms […]
[…] this hasn’t stopped people from attempting to come up with the magic formula for creating the perfect pop song (which everyone knows is ABBA’s Dancing […]
It’s interesting that these jazz cats wish pop music was more like jazz, would they really want to go see a jazz band and have to deal with the same crowd that goes to pop concerts? and if pop was more jazz then jazz would be pop therefore I still wouldn’t like it, to quote Les Claypool “I am Antipop; I’ll run against the grain ’till the day I drop”.
Great ideas!
It’s interesting. If I played you 2 seconds of music you probably wouldn’t be able to tell me who the band or what the song was. But you could say whether it was ‘rock’ – probably what type of rock, was it more like zepplin, nirvana etc.
You could probably get a surprisingly large amount information from those two seconds, do you know why?
Because music is derivative. That is how we ‘understand’ music, and how it has ‘meaning’. What chord is ‘expected’ after a preceding chord? Well everybody ‘feels’ what that chord should be, so to be ‘creative’ we choose another one, and go from there. It’s a beautiful thing and pretty difficult to ‘formulate’, I certainly wouldn’t be able to do it but good show old bean. pip pip
Just came across this site… it’s an awesome resource! Starting to play with it a little, I was wondering if there’s a way you could show a list of the songs with the longest sequence of circle of fifths movement., E.g. 2 or 3 chords is interesting, but what about those songs with 5 of 6 steps around the circle in a row? Now that’d be fascinating. I realise it’ll probably be older songs (more Tin Pan Alley type stuff), but it would also be interesting to see how many more modern songs also have longer progressions. How easy would that be to do?
Five Foot Two goes from III7 – VI7 – II7 – V7 – I. And I know you know this song as a fellow barbershopper :). Yes I’m replying six years later.
[…] “The guys over at Apple, Inc. evidently know their music theory. Their latest version of GarageBand lets you play with “Smart Instruments” that “make you sound like an expert musician… even if you’ve never player a note before.” (https://www.hooktheory.com/blog/i-analyzed-the-chords-of-1300-popular-songs-for-patterns-this-is-what…) […]
[…] amazing people at Hooktheory analyzed 1300 songs and found that we do favor certain “sounds” above others. This popular “I V vi […]
[…] I V vi IV, I V vi IV, … The amazing people at Hooktheory analyzed 1300 songs and found that we do favor certain “sounds” above others. This popular “I V […]
[…] https://www.hooktheory.com/blog/i-analyzed-the-chords-of-1300-popular-songs-for-patterns-this-is-what… Share this:TwitterFacebookGoogleLike this:Like Loading… […]
Sweet article, thanks for doing all that work. I like the point about feeling weird about breaking a song down. I ran into that just recently. I want to write songs but this is giving me problems because I’m self taught. God Bless you and God Bless America
All that this study accomplishes is a verification of basic theory. It supports root movement and predominant-dominant passages and sentence structures/periods. So much unnecessary work has been done here.
This thread’s quite old but still I’d like to add sthg: as a musician I am always trying to find new ideas and approaches, and though it’s true the 4ths and 5ths and Is and Vs are what define western harmony I think one analyzes these things in specific moments (mainly when playing with other musicians) while when composing it’s more common to be spontaneous to play what you feel like; and when you play many similar progressions you tend to analyze how they work without getting into deep so you really feel good improvising and then all the harmonies and patterns come out naturally (then you have to sieve and sift, best when recording your jams). Funny coz I thought changing from an Em to a C or a G would be more common than changing to F or Am, but this study adds some light in this sense. I’d rather take a comprehensive look at all the songs from your survey,maybe I’d change quite a few of them for some others but I think your work is worth of respect simply because you made it and, more important, you crossed that hazy territory between musicians and listeners. Thanks and congratulations.
Does anyone know of ANY OTHER song with the same chord structure as “Louie,Louie”? ie: 1,4,minor 5 It seems quite unique and stands alone which is quite a feat as 1,4,5,structure dominates blues,and rock based forms. Thanks,Barry-Los Angeles
[…] and they’ve got a great blog that explains it all. Make sure you read their article: I analyzed the chords of 1300 popular songs for patterns. This is what I found. Explains a […]
[…] I get to my 21 Points, I would like to direct any music-lover’s attention to this article. The article is in three parts, with links to the other two parts found in the article I linked. […]
[…] I analyzed the chords of 1300 popular songs for patterns. This is what I found.. […]
[…] I analyzed the chords of 1300 popular songs for patterns… Leave a reply I analyzed the chords of 1300 popular songs for patterns. This is what I found.. […]
http://www.amazon.com/b?_encoding=UTF8&site-redirect=&node=172282&tag=tabbooingcom-20&linkCode=ur2&camp=1789&creative=9325
yes, but you are only looking at FOLLOWING chords, not necessarily in a suitable order, e.g. Em, B A might be the last note of a chorus with the first 2 notes on a verse, or the last 3 notes of 4 in a hook so they are really connected in such a meaningful way.
[…] As you develop, you start mapping the intervals every time you listen to music. You might learn a song before you’ve even played it! At that point the only thing to figure out is the key – unless you have perfect pitch or are good at guessing (the odds are not that bad: 80 % of songs are in one of the seven most common keys). […]
What an amazing piece of work. Would it be possible to obtain a file with all 1300 songs and their chords so I could do my own analysis? Thanks in advance!
I studied a bit of harmonic theory in the last couple of years with a private teacher. I went in with the idea that chord harmony was the most fundamental aspect of a song, but he argued to me that melody was actually the most important part of a song. He said: “There is no song without melody.” Hard to argue with that. Then I came away with the idea that melody, harmony and progression is actually what, all together, makes a song listenable. We can take basic, well-known melodies like “Happy Birthday to You” or “Twinkle Twinkle Little Star”. They are always great melodies, but it’s hard to repeat them more than a couple of times in one sitting without quickly getting bored with them. But when we vary the harmony and the progressions (and perhaps the rhythms) behind them, we can turn them into songs that can be heard for a half an hour or more. Thus, I get the sense that it does no good to just look for a chord or chord progression that has or hasn’t been used previously. You really do have to consider the melody that is leading the song and what it is doing. You might even vary the melody to introduce something new, but at the end of the day, the melody you choose is going to tell you what direction your harmony can go. The fact that a certain chord is commonly used in pop music doesn’t seem to be the best guide to what will sound good, but more a guide to the conventions that people have most often relied upon to make their songs catchy. Think of past composers, like Debussy, who had an influence on later pop music. Do you think Debussy was looking for an easy out chord to make his song catchy to the listening public or was looking for a sound for his entire progression that would make it stand out as a singular statement? I would argue for the latter and would say that if musicians and composers want to really stand out they should probably go beyond songwriting conventions. BTW, this is just a comment about composing and is in no way meant to disparage the idea of the database, which I think is an interesting and probably very useful tool.
In discussing popular chords, it seems discussing the 7th chords would be absolutely necessary. In the graph the Major and minor chords are relatively apparent. But what about the popular 7 chords?
[…] https://www.hooktheory.com/blog/i-analyzed-the-chords-of-1300-popular-songs-for-patterns-this-is-what… […]
[…] I analyzed the chords of 1300 popular songs for patterns […]
Interesting read. Thank you!
[…] https://www.hooktheory.com/blog/i-analyzed-the-chords-of-1300-popular-songs-for-patterns-this-is-what… […]
The singers are higher pitched today then in the past. More bands do you notice are singing higher pitched
tunes which is bad when the singers get older and can’t hit the notes anymore.
[…] I analyzed the chords of 1300 popular songs for patterns […]
There are major problems with your sample of songs if the chord following Em was F 93% of the time. That is a weird place for Em to go. It is far, far, far, far, far, far more common for the Em to go to C.
I think it is likely that you made a mistake while calculating the data because I can’t imagine what sample of 1300 pop songs could give that sort of result.
yes zach I agree. there is NO way C is the most common key for popular songs. i dont know where he is getting his data.
i would love to see his list of the most popular songs
[…] https://www.hooktheory.com/blog/i-analyzed-the-chords-of-1300-popular-songs-for-patterns-this-is-what… […]
Now your guitar can sound like Van Halen, Joe Satriani and other guitarists studio album guitar sound with BOSS GT-10 Patches by James Limborg… Listen here: http://www.jameslimborg.com/boss-gt-10-patches.html
[…] I analyzed the chords of 1300 popular songs for patterns David Carlton | Hook Theory […]
Only 2% using the VIIb (Bb)? I suspect a much higher percentage of songs are in mixolydian, and that someone (or an algorithm) has improperly transposed a lot of these (to F instead of C). That would inflate your C/G/F counts.
E.g. D-C-G-D is D mixolydian, so a proper transposition to C would make it C-Bb-F-C. But I suspect a lot of such songs in your DB were transposed to G-F-C-G for this analysis.
A someone with a moderately good ability to play chords on the guitar, I was wondering if it is possible to write / work_out the chords for any music at all? Say for Gilbert & Sullivan music. Hopefully where there isn’t a chord-change on every note of the melody. Thanks.
[…] https://www.hooktheory.com/blog/i-analyzed-the-chords-of-1300-popular-songs-for-patterns-this-is-what… […]
[…] conventions established, and the standard refrain of popular music became increasingly concrete. Our ears engage with popular music now with expectations of how and where melody resolves; over 50% of songwriters who produced commercially successful music, when asked to follow a common […]
[…] https://www.hooktheory.com/blog/i-analyzed-the-chords-of-1300-popular-songs-for-patterns-this-is-what… […]