{"id":443,"date":"2012-09-03T02:30:55","date_gmt":"2012-09-03T02:30:55","guid":{"rendered":"http:\/\/garysieling.com\/blog\/?p=443"},"modified":"2012-09-03T02:30:55","modified_gmt":"2012-09-03T02:30:55","slug":"detecting-pitches-in-music-with-r","status":"publish","type":"post","link":"https:\/\/www.garysieling.com\/blog\/detecting-pitches-in-music-with-r\/","title":{"rendered":"Detecting Pitches in music with R"},"content":{"rendered":"

In a previous post, I described a method to detect a chord using a Fourier transform<\/a> in Java\/Scala. I’ve re-implemented the same in R, detailed below.<\/p>\n

This will generate an audio file containing the C-Major chord:<\/p>\n

library(sound)\n\nc<-261.63\ne<-164.81\ng<-196\n\nlen<-1\ncData<-Sine(c,len)\neData<-Sine(e,len)\ngData<-Sine(g,len)\n\naudio<-normalize(cData+eData+gData)\n\nsaveSample(audio, \"out\\\\ceg.wav\", overwrite=TRUE)<\/pre>\n
And a series of helper functions:<\/p>\n
magnitude<-function(x) { sqrt(Re(x) * Re(x) + Im(x) * Im(x)) }\nmaxPitch<-audio$rate\/2-1\n\nfrq<-c(16.35,17.32,18.35,19.45,20.60,21.83,23.12,24.50,25.96,27.50,29.14,30.87)\nnoteNames<-c(\"C\", \"C#0\/Db0\", \"D0\", \"D#0\/Eb0\", \"E0\", \"F0\", \"F#0\/Gb0\", \"G0\", \"G#0\/Ab0\", \"A0\", \"A#0\/Bb0\", \"B0\")\n\nlookup<-data.frame(noteNames,frq)<\/pre>\n
A function to find the middle frequency in each FFT bucket:<\/p>\n
idxFrq<-function(idx) { (idx - .5 ) }<\/pre>\n
And a function to find the closest note in the lowest octave:<\/p>\n
noteDist<-function(note1, note2) { abs(log(note1)\/log(2) - log(note2)\/log(2)) %% 1 }<\/pre>\n
Now to do the actual FFT:<\/p>\n
fftdata<-fft(audio$sound)\nhalf<-fftdata[1:maxPitch]\nindexes<-c(1:maxPitch)\nmagnitudes<-magnitude(half)<\/pre>\n
Now, from each bucket, find the note that is closest to each bucket:<\/p>\n
closest<-function(x){ subset(lookup, select=c(noteNames,frq),subset=(min(noteDist(frq, x)) == noteDist(frq, x)))$noteNames }\n\nnoteList<-mapply(closest, indexes-0.5)\nmag<-data.frame(noteList, magnitudes)\nbarplot(by(mag$magnitudes, mag$noteList, sum))<\/pre>\n
This results in the following image, which has peaks at C, E, and G, as expected.<\/p>\n
\"\"<\/a><\/p>\n
The calculation is very crude – all frequencies are mapped to a note evenly, which isn’t really correct.<\/p>\n
This can be adjusted by de-emphasizing the fft buckets between notes:<\/p>\n
scale<-function(x){ .5 + .5 * sin(pi * x*2 + pi\/2) }\nnote<-function(x) { 12 * log(x\/440)\/log(2) + 49 }\nscaleData<-scale(note(idxFrq(indexes)))\nscaledMagnitudes<-scaleData*magnitudes\nscaledMag<-data.frame(noteList, scaledMagnitudes)\n<\/pre>\n
This looks better – the three highest are the notes in the chord-<\/p>\n
\"\"<\/a><\/p>\n","protected":false},"excerpt":{"rendered":"
In a previous post, I described a method to detect a chord using a Fourier transform in Java\/Scala. I’ve re-implemented the same in R, detailed below. This will generate an audio file containing the C-Major chord: library(sound) c<-261.63 e<-164.81 g<-196 len<-1 cData<-Sine(c,len) eData<-Sine(e,len) gData<-Sine(g,len) audio<-normalize(cData+eData+gData) saveSample(audio, “out\\\\ceg.wav”, overwrite=TRUE) And a series of helper functions: magnitude<-function(x) … <\/p>\n
Continue reading “Detecting Pitches in music with R”<\/span><\/a><\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"om_disable_all_campaigns":false,"_monsterinsights_skip_tracking":false,"footnotes":""},"categories":[6,27],"tags":[109,178,373,450,532],"aioseo_notices":[],"aioseo_head":"\n\t\t\n\t\n\t\n\t\n\t\n\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t