I analyzed chord progressions in 680k songs • Hacker News Board

Re: I analyzed chord progressions in 680k songs

#91

Post by kube-system » Fri, Apr 18, 2025, 1:34 PM UTC

I was disappointed to see they use Ultimate Guitar data for this. Back when I played regularly under the study of a professional, I quickly found that the majority of tabs on the site are just plain wrong.

Likely some of the “simplicity” observed in the chords here aren’t actually due to the music being simple, but that the user who contributed the tab made their best simplified guess as to what it was.

Unfortunately the analysis is probably flawed because of the poor quality of data.

Re: I analyzed chord progressions in 680k songs

#92

Post by kube-system » Fri, Apr 18, 2025, 1:45 PM UTC

otabdeveloper4 wrote: post #68

Slow_Hand wrote: post #67

Earlier quoted context omitted.

Literally everyone? Have you got a source for that claim? I don’t disagree that music performance was a pastime for many people before recorded music, but let’s be real here.

There was no recorded or productionized music back then. And yet people liked music as much as we do now. So the only way to enjoy music was to do it yourself. Singing and playing an instrument was just a basic life skill that everyone had back then. (Say, like driving a car or using a computer is today. Not everyone is a professional driver or computer programmer, but not being able to use a computer at all today wo…

> There was no recorded or productionized music back then. And yet people liked music as much as we do now. So the only way to enjoy music was to do it yourself.

Or listen to live music in your community

Re: I analyzed chord progressions in 680k songs

#93

Post by moefh » Fri, Apr 18, 2025, 1:52 PM UTC

zenogantner wrote: post #56

memset wrote: post #36

The way this analysis, and the original dataset were created, makes no sense. This is, in part, not the author's fault, since the original data [1, 2] is flawed. First, the original data was constructed like this: "...The next step was to format the raw HTML files into the full chord progression of each song, collapsing repeating identical chords into a single chord (’A G G A’ became ’A G A’)..." Already this makes n…

The problem with collapsed repeated chords comes not only from the data processing -- most Ultimate Guitar songs are written down entirely ignoring how often a chord is repeated -- the classic "lyrics plus chords" format is incomplete and requires the player to somewhat know the structure of the song anyway. The write-up usually just gives hints where, relative to the lyrics, the chord changes.

Exactly. In my experience, it's not just Ultimate Guitar, all of these sites with chord progressions assume you already know how the music sounds. They're not enough for someone to lean a song having never heard it, so they're almost certainly not enough to automate analysis of the chord progressions.

Re: I analyzed chord progressions in 680k songs

#94

Post by ben7799 » Fri, Apr 18, 2025, 1:53 PM UTC

As others have said this is interesting but use of Ultimate Guitar is flawed as the tabs/scores are so bad on that site, very often not even being close to the real chords.

On top of being simplified tons and tons of songs get rewritten with a Capo so people can just play G-C-D shapes, if your analysis doesn't look for "Capo" and then transpose all the chords then you end up overrepresenting the key of G and it's chords. Then very often 7th chords, Sus chords, etc.. all get transcribed down to major chords & minor chords due to the beginner focus. Interestingly he doesn't include 6th chords as their own thing.

To be fair there are tons of songs that do actually use those chords, so they may still end up coming out as the most popular.

I have a grandfathered in lifetime membership to UG that I only had to pay once for. It was cheap so worth it, but I really find the site kind of icky as they are mostly monetizing crowd sourced low quality work and it's very often wrong. And they nerfed their iPad app recently which is really annoying.

Re: I analyzed chord progressions in 680k songs

#95

Post by calibas » Fri, Apr 18, 2025, 1:55 PM UTC

From my experience playing guitar, the average punk or metal song is almost entirely power chords, while this data says power chords are only 5% of chords in both genres...

I thought maybe there's types of metal and punk that I don't know about, but Wikipedia, LLMs and guitar tab sites all agree with me. Punk and metal is overwhelmingly power chords, so I don't see how the data comparing chord types can be correct.

Re: I analyzed chord progressions in 680k songs

#96

Post by xfalcox » Fri, Apr 18, 2025, 1:59 PM UTC

calibas wrote: post #95

From my experience playing guitar, the average punk or metal song is almost entirely power chords, while this data says power chords are only 5% of chords in both genres... I thought maybe there's types of metal and punk that I don't know about, but Wikipedia, LLMs and guitar tab sites all agree with me. Punk and metal is overwhelmingly power chords, so I don't see how the data comparing chord types can be correct.

I guess one aspect missing here is weighting more popular songs on that analysis.

I assume that the analysis is simply counting every song chords, so a unknown band you've never heard about has the same impact as The Ramones.

I'd like to see the same graph weighted by band popularity using either YouTube or Spotify data.

Re: I analyzed chord progressions in 680k songs

#97

Post by sysrestartusr » Fri, Apr 18, 2025, 2:08 PM UTC

huimang wrote: post #64

Using absolute chord analysis instead of relative chords (i.e. roman numeral analysis) doesn't make sense. As others have noted, the original dataset is flawed because the structure of a song is critical, you cannot omit repeating chords. Programmers/analysts should take more care to understand music theory or the underlying field at hand, before compiling datasets or doing analysis. "Most common chord" is mildly int…

> and the average songwriter isn't making data-driven decisions on how to settle on the chord structure for their song

aren't decisions like that implicit to the source of learning/inspiration? it's not data-driven on the surface of the writers awareness, and maybe not data-driven in the statistical sense, but "intuitively", "that which sounds good successively", is based on what one heard so far within the context of the song ... so it's one hundred percent data-driven, just not data that one has consciously quantified.

IMO: average songwriters and musicians and producers are the top exactly because they hit exactly that big fat belly of the bell curve/ G distribution ... I'd say you have it backwards... there's much more experimentation and less data-stuff going on left and right of the average

Re: I analyzed chord progressions in 680k songs

#98

Post by clonedhuman » Fri, Apr 18, 2025, 2:21 PM UTC

navane wrote: post #75

peanut-walrus wrote: post #73

Earlier quoted context omitted.

Wouldn't using relative chords simply show that 99% of songs use the I chord? :)

It's like analyzing music by looking at the amplitude of the sound wave instead of the frequency. Music is all about the changes.

Yeah. It's all about what changes, what doesn't, and when and where those changes occur. Stability and novelty.

Re: I analyzed chord progressions in 680k songs

#99

Post by uoaei » Fri, Apr 18, 2025, 2:28 PM UTC

huimang wrote: post #64

Using absolute chord analysis instead of relative chords (i.e. roman numeral analysis) doesn't make sense. As others have noted, the original dataset is flawed because the structure of a song is critical, you cannot omit repeating chords. Programmers/analysts should take more care to understand music theory or the underlying field at hand, before compiling datasets or doing analysis. "Most common chord" is mildly int…

The parallels between your critique of music analysis, and linguists' critique of LLMs, bear remarkable similarities. "Language/thought is more than sequences of tokens" will still be true no matter how much data we throw at the problem to smooth the rough edges.

Re: I analyzed chord progressions in 680k songs

#100

Post by epiccoleman » Fri, Apr 18, 2025, 2:36 PM UTC

toolslive wrote: post #76

CuriouslyC wrote: post #70

Earlier quoted context omitted.

I think most musicians know that I-IV-V-I is the zero thought default for in key chord progression, it's so overused you don't need fancy analysis to figure it out. For me, I'm more interested in the intervals and voicing pairs, because those tell you something deeper about the music that you don't get from the chord progression.

I-IV-V-I, II-V-I and maybe I-VII-VI-V and you can consider yourself "advanced" ;)

I have an almost irrational love for I-IV-VII-V. It's got a sort of happy, laid-back nostalgic vibe - sort of the best way I know to smuggle an extra major chord into a key. It can be approached in some fun different ways - can be thought of a "mixolydian" progression off the tonic, but it's also two I-IVs stuck together - almost a little mini-modulation if you wanna think of that way.

Sunrain[1] by Lotus is probably my favorite example (listen for the chords that come in under the main riff). But it's a staple in tons of rock music, and once you get it into your ears you'll hear it all over.

[1]: https://www.youtube.com/watch?v=vAc-B5eDKmI