Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Chord progressions of 25,000 songs

amitkohli.com

Post reply on HN
Print view View original

91–100 of 113 posts

Re: Chord progressions of 25,000 songs

#91

Post by elihu »

gnaritas wrote: post #87
elihu wrote: post #59

Earlier quoted context omitted.

Incidentally, a just (non-tempered) guitar fingerboard might look something like this: http://jsnow.bootlegether.net/cbg/fingerboard_at_a.svg I swapped fingerboards on an old Harmony Archtone with one based on the above design. Admittedly, 20 notes per octave may have been a bit much; it's kind of awkward to play. The next one I do will be simpler. Sounds nice, though.

Just remove the frets, problem solved.

Fretless is difficult to play accurately, especially for chords. Also, you get a louder sound from frets because your fingers aren't muting the strings quite so much.

Re: Chord progressions of 25,000 songs

#92

Post by elihu »

codeulike wrote: post #84
yodsanklai wrote: post #75

Earlier quoted context omitted.

Good question. I don't know the exact answer, but yes I think there's nothing special that distinguish naturals and accidentals from a "listening" point of view. However, this distinction is handy for written music. For example, when you play a piece of music in a given key, let say G major, the notes are G A B C D E F#, there's only one accidental, so you can put this information in the key signature at the beginnin…

But you're just describing a notation, and the notation works OK because of the convention that in any given key you only use 7 of the available 12 tones. But why 7, and why those 7?

Sounds that sound good together have frequencies that are whole number ratios of each other. (This means they have harmonics that align. If they don't align, you get a beat frequency which sounds objectionable.)

A 3:4:5 ratio is a major triad. A good scale should have as many major triads as possible with the least number of notes.

So, let's call the root of our scale 1/1. You might want to build a major triad there, so you have 1/1, 5/4, and 3/2. If you build a triad off of 3/2, you'll end up adding a couple more notes: 15/8 and 9/4. If you add 4/3 as a note and build a triad there, you'll end up adding 5/3 and 2/1. 2/1 is the octave, so we can pretend it's equivalent to 1/1, and 9/4 is an octave above 9/8, so we can substitute 9/8.

So, our scale is 1/1, 9/8, 5/4, 4/3, 3/2, 5/3, and 15/8, and we have the major scale. It is possible to construct 3 major chords and 2 minor chords (10:12:15 ratio) from those seven notes. If you add 10/9 (slightly flat of 9/8), you can add one more minor chord.

By a mathematical coincidence, we can approximate these values pretty closely with various powers of the twelfth root of 2:

x = 2^(1/12) 1, x^2, x^4, x^5, x^7, x^9, x^11

The latter scale is equal tempered, whereas the former is "just" (untempered). The ET scale has very accurate fourths and fifths, but thirds and sixths are far enough off that they don't sound quite as nice in chords. We tend to put up with that in modern music, though.

Re: Chord progressions of 25,000 songs

#93

Post by twerquie »

ssharp wrote: post #63
cjslep wrote: post #4

> Even though there is a great variety of chords and chord progressions, progressions involving 1,4,5, and 6 are favoured, probably because they ‘sound good’ to our brain. I drive my girlfriend nuts because of this. I don't typically listen to the pop/country songs that are popular with our local radio but when she has it on in the car, every once in a while a song will come on and I'll think "I know this" and start…

Do you really listen to the chords that closely? For a lot of music, the chords are either largely hinted at (bass notes, the lead guitar or rhythm guitar lick, and how the melody move) and if there is an instrument banging out the exact chord progression, it's usually buried under the stuff or at least the vocal. If I'm writing a song, I'm spending almost all of my time doing things tangential to the chord progressi…

OP mentioned the pop/country genre, which is typically very heavy-handed with the chord progressions.

Re: Chord progressions of 25,000 songs

#94

Post by joeyspn »

Just learnt about http://www.hooktheory.com/ thanks to this article, and I must say that I really love the concept of the website. I've always had music production as a hobby but w/o proper music theory background I struggled to analyse popular songs. This makes the process really easy and fun!

I feel this site could be huge with a couple of tweaks: real-time and proper midi interfacing. I hope the site will evolve towards more complex collaborative songwriting...

Re: Chord progressions of 25,000 songs

#95

Post by ta83838 »

codeulike wrote: post #71
snorkel wrote: post #53

For non-musicians, here's some terminology behind this: There are 7 natural notes in (western) music: A, B, C, D, E, F, G (the white keys on a piano keyboard) ... plus 5 accidental notes that half-way between: A#, C#, D#, F#, G# also known as Bb, Db, Eb, Gb, Ab ... the black keys on a piano keyboard. # means sharp (go up in pitch one-half step) b means flat (go down in pitch one-half step) 7 naturals + 5 accidentals…

But the question is, what's natural about the Naturals, and whats accidental about the Accidentals? Because taken together they are just 12 equally spaced notes.

The notes of the scale evolved gradually. Originally only the 7 notes of the diatonic (major) scale were used, sometimes fewer than seven. But if you take one of the notes in a diatonic scale other than the first and try to start a new scale starting with this note, you will find that you need new notes, the existing 7 don't allow you to create a new scale that "sounds the same" (has the same ratios between the frequencies of each note). The new notes that need to be added to allow these new scales became what are today the accidentals.

This link has some information about this:

http://en.wikipedia.org/wiki/Accidental_%28music%29#History_...

Re: Chord progressions of 25,000 songs

#96

Post by darkmighty »

tajen wrote: post #86
codeulike wrote: post #71

Earlier quoted context omitted.

But the question is, what's natural about the Naturals, and whats accidental about the Accidentals? Because taken together they are just 12 equally spaced notes.

Harmonics. They resonate well together because they have some common frequency. 13 equally spaced notes wouldn't match on multiples of 2, 3, 4 and 6. Example: The major scale (C, D, E, F etc) are the notes 0/12, 2/12, 4, 5, 7, 9 and 11. You'd expect the 6th note to sound nice, wouldn't you? Wrong: The scale is logarithmic. A nice C chord is made of 0th, 4th, 7th/12, which are 523Hz, 659Hz, 783Hz. Notice how the secon…

Yes (from the little I know) a major aspect in our hearing is that nature doesn't produce pure sinusoids. Mechanical vibration materials is bound to have a series of resonances multiples of some basic frequency. So if you have a single person, or single object, emitting sound, you expect to see this nice harmonic lineup. When the fundamental frequencies don't have a pair (a,b) s.t. a * f1=b * f2, their harmonic multiples don't match and you can tell something is odd.

For f2=125% f1, for example, every 4th harmonic from f2 lines up perfectly; for f2=150% f1, every 2nd harmonic lines up; for f2= 200% f1, every harmonic actually lines up. I guess it's like looking at combs with different spacings and finding ones that superpose neatly.

Re: Chord progressions of 25,000 songs

#97

Post by markbnj »

azakai wrote: post #64
markbnj wrote: post #28

To see how popular the 1-5-6-4 progression is you need only watch Axis of Awesome ... https://www.youtube.com/watch?v=5pidokakU4I

I was surprised though to see that in the data the 1-5-6-4 progression is not the most common. Maybe I'm reading it wrong? Looks like 1-4-* is more popular.

I think what the amusing videos linked above show is that 1-5-6-4 is very common in certain styles of music, especially modern pop. But when you consider music in its entirety then I'm not surprised to see it recede. So much folk and blues, for example, are built around completely different progressions.

Re: Chord progressions of 25,000 songs

#98

Post by a1k0n »

protonfish wrote: post #40
stephen_g wrote: post #5

Earlier quoted context omitted.

Nope. 1 is the root (or tonic) of the key, not necessarily the first chord.

Then what is the first chord? The beginning is labelled "Start" What does this mean? Either less than 25% of songs start on the 1 (no way this is possible) or the second chord is mislabeled as a chord change when it stays the same or those with a 1 in the second column started on a chord other than the root and instead of listing that chord, the data is thrown away. Something is messed up here.

> Very lazily, I just normalized all probabilities across each transition so that each transition “mega bar” is kind-of the same height. I’m sure there’s a better way to do it, the community is invited to improve!

The size of the bars is normalized, not necessarily indicative of probabilities. That's why I isn't any bigger than the rest.

Re: Chord progressions of 25,000 songs

#100

Post by BurningFrog »

haberman wrote: post #29
luch wrote: post #19

Earlier quoted context omitted.

The comma is not a big deal : it's a problem caused by tone equalizing (aka temperament). Think of instruments as ladders : they have all the same steps, but start at different heights. When combining several types of instruments, the resulting steps are not aligned. Equal temperament solve that problem by normalizing the steps' absolute height. It's a trade-off between the instrument's perfect harmony, and the orche…

I don't mean to be rude, but I'm not sure you understand the comma. It has nothing to do with combining types of instruments. It's a problem that exists even with a single instrument. The problem is that, even on a single instrument (say a piano), if you start on some note (say C0), going up seven octaves should land you on exactly the same note as going up twelve perfect fifths (C7). The problem is that the precise…

I understand that pianos have to pick a frequency for each key.

I wonder if there is any effort in electronic instruments to produce "better" frequencies on the fly? I'll explain:

On a piano hitting C+G plays something close to a perfect fifth.

An electronic instrument could move the G slightly to produce a perfect fifth when that key combination is played.

Post reply on HN