def chordBeats(infile, outfile):
mywindow = np.array([
0.001769, 0.015848, 0.043608, 0.084265, 0.136670, 0.199341, 0.270509,
0.348162, 0.430105, 0.514023, 0.597545, 0.678311, 0.754038, 0.822586,
0.882019, 0.930656, 0.967124, 0.990393, 0.999803, 0.999803, 0.999803,
0.999803, 0.999803, 0.999803, 0.999803, 0.999803, 0.999803, 0.999803,
0.999803, 0.999803, 0.999803, 0.999803, 0.999803, 0.999803, 0.999803,
0.999803, 0.999803, 0.999803, 0.999803, 0.999803, 0.999803, 0.999803,
0.999650, 0.996856, 0.991283, 0.982963, 0.971942, 0.958281, 0.942058,
0.923362, 0.902299, 0.878986, 0.853553, 0.826144, 0.796910, 0.766016,
0.733634, 0.699946, 0.665140, 0.629410, 0.592956, 0.555982, 0.518696,
0.481304, 0.444018, 0.407044, 0.370590, 0.334860, 0.300054, 0.266366,
0.233984, 0.203090, 0.173856, 0.146447, 0.121014, 0.097701, 0.076638,
0.057942, 0.041719, 0.028058, 0.017037, 0.008717, 0.003144, 0.000350
])
l, beats, semitones = essentia_chord_utils.loadBeatsAndChroma(infile)
chroma = np.zeros((semitones.shape[0], 12))
for i in range(semitones.shape[0]):
tones = semitones[i] * mywindow
cc = chroma[i]
for j in range(tones.size):
cc[j % 12] = cc[j % 12] + tones[j]
chroma = essentia_chord_utils.smooth(chroma,
window_len=int(1.5 * 44100 / 2048),
window='hanning').astype('float32')
chords = ChordsDetectionBeats(hopSize=2048)
syms, strengths = chords(chroma, beats)
segments = essentia_chord_utils.toMirexLab(0.0, l / 44100.0, beats, syms,
strengths)
with open(outfile, 'w') as content_file:
for s in segments:
content_file.write(str(s) + '\n')
def chordBeats(infile, outfile):
print 'Loading audio file...', infile
audio = essentia.standard.MonoLoader(filename=infile)()
bt = BeatTrackerMultiFeature()
beats, _ = bt(audio)
#beats = beats[::4]
parameters = {}
stepsize, chroma = vamp.collect(audio,
44100,
"nnls-chroma:nnls-chroma",
output="chroma",
step_size=2048)["matrix"]
# to essentia convention
#chroma = np.roll(chroma, 3, 1)
chords = ChordsDetectionBeats(hopSize=2048)
syms, strengths = chords(chroma, beats)
segments = essentia_chord_utils.toMirexLab(0.0,
len(audio) / 44100.0, beats,
syms, strengths)
with open(outfile, 'w') as content_file:
for s in segments:
content_file.write(str(s) + '\n')
def chordBeats(infile, outfile):
# initialize algorithms we will use
tuningFreq = essentia_chord_utils.tuning(infile)
chordHopSize = 2048
frameSize = 8192
loader = MonoLoader(filename=infile)
framecutter = FrameCutter(hopSize=chordHopSize, frameSize=frameSize)
windowing = Windowing(type="blackmanharris62")
spectrum = Spectrum()
spectralpeaks = SpectralPeaks(orderBy="magnitude",
magnitudeThreshold=1e-05,
minFrequency=40,
maxFrequency=5000,
maxPeaks=10000)
hpcp = HPCP(size=12,
referenceFrequency=tuningFreq,
harmonics=8,
bandPreset=True,
minFrequency=40.0,
maxFrequency=5000.0,
bandSplitFrequency=500.0,
weightType="cosine",
nonLinear=True,
windowSize=1.0)
# use pool to store data
pool = essentia.Pool()
# connect algorithms together
loader.audio >> framecutter.signal
framecutter.frame >> windowing.frame >> spectrum.frame
spectrum.spectrum >> spectralpeaks.spectrum
spectralpeaks.magnitudes >> hpcp.magnitudes
spectralpeaks.frequencies >> hpcp.frequencies
hpcp.hpcp >> (pool, 'chroma.hpcp')
# network is ready, run it
essentia.run(loader)
# don't forget, we can actually instantiate and call an algorithm on the same line!
print 'Loading audio file...', infile
audio = essentia.standard.MonoLoader(filename=infile)()
bt = BeatTrackerMultiFeature()
beats, _ = bt(audio)
beats = beats[::4]
chords = ChordsDetectionBeats()
syms, strengths = chords(pool['chroma.hpcp'], beats)
segments = essentia_chord_utils.toMirexLab(0.0,
len(audio) / 44100.0, beats,
syms, strengths)
with open(outfile, 'w') as content_file:
for s in segments:
content_file.write(str(s) + '\n')
def chordBeats(infile, chromafile, outfile):
print 'Loading audio file...', infile
audio = essentia.standard.MonoLoader(filename=infile)()
bt = BeatTrackerMultiFeature()
beats, _ = bt(audio)
beats = beats[::1]
chroma = np.loadtxt(chromafile, dtype='float32')
# to essentia convention
chroma = np.roll(chroma, 3, 1)
chords = ChordsDetectionBeats(hopSize=2048)
syms, strengths = chords(chroma, beats)
segments = essentia_chord_utils.toMirexLab(0.0,
len(audio) / 44100.0, beats,
syms, strengths)
with open(outfile, 'w') as content_file:
for s in segments:
content_file.write(str(s) + '\n')
def chordBeats(infile, outfile):
# to essentia convention
#chroma = np.roll(chroma, 3, 1)
l, beats, semitones = essentia_chord_utils.loadBeatsAndChroma(infile)
treblewindow = np.array([
0.000350, 0.003144, 0.008717, 0.017037, 0.028058, 0.041719, 0.057942,
0.076638, 0.097701, 0.121014, 0.146447, 0.173856, 0.203090, 0.233984,
0.266366, 0.300054, 0.334860, 0.370590, 0.407044, 0.444018, 0.481304,
0.518696, 0.555982, 0.592956, 0.629410, 0.665140, 0.699946, 0.733634,
0.766016, 0.796910, 0.826144, 0.853553, 0.878986, 0.902299, 0.923362,
0.942058, 0.958281, 0.971942, 0.982963, 0.991283, 0.996856, 0.999650,
0.999650, 0.996856, 0.991283, 0.982963, 0.971942, 0.958281, 0.942058,
0.923362, 0.902299, 0.878986, 0.853553, 0.826144, 0.796910, 0.766016,
0.733634, 0.699946, 0.665140, 0.629410, 0.592956, 0.555982, 0.518696,
0.481304, 0.444018, 0.407044, 0.370590, 0.334860, 0.300054, 0.266366,
0.233984, 0.203090, 0.173856, 0.146447, 0.121014, 0.097701, 0.076638,
0.057942, 0.041719, 0.028058, 0.017037, 0.008717, 0.003144, 0.000350
])
basswindow = np.array([
0.001769, 0.015848, 0.043608, 0.084265, 0.136670, 0.199341, 0.270509,
0.348162, 0.430105, 0.514023, 0.597545, 0.678311, 0.754038, 0.822586,
0.882019, 0.930656, 0.967124, 0.990393, 0.999803, 0.995091, 0.976388,
0.944223, 0.899505, 0.843498, 0.777785, 0.704222, 0.624888, 0.542025,
0.457975, 0.375112, 0.295778, 0.222215, 0.156502, 0.100495, 0.055777,
0.023612, 0.004909, 0.000000, 0.000000, 0.000000, 0.000000, 0.000000,
0.000000, 0.000000, 0.000000, 0.000000, 0.000000, 0.000000, 0.000000,
0.000000, 0.000000, 0.000000, 0.000000, 0.000000, 0.000000, 0.000000,
0.000000, 0.000000, 0.000000, 0.000000, 0.000000, 0.000000, 0.000000,
0.000000, 0.000000, 0.000000, 0.000000, 0.000000, 0.000000, 0.000000,
0.000000, 0.000000, 0.000000, 0.000000, 0.000000, 0.000000, 0.000000,
0.000000, 0.000000, 0.000000, 0.000000, 0.000000, 0.000000, 0.000000
])
mywindow = np.array([
0.001769, 0.015848, 0.043608, 0.084265, 0.136670, 0.199341, 0.270509,
0.348162, 0.430105, 0.514023, 0.597545, 0.678311, 0.754038, 0.822586,
0.882019, 0.930656, 0.967124, 0.990393, 0.999803, 0.999803, 0.999803,
0.999803, 0.999803, 0.999803, 0.999803, 0.999803, 0.999803, 0.999803,
0.999803, 0.999803, 0.999803, 0.999803, 0.999803, 0.999803, 0.999803,
0.999803, 0.999803, 0.999803, 0.999803, 0.999803, 0.999803, 0.999803,
0.999650, 0.996856, 0.991283, 0.982963, 0.971942, 0.958281, 0.942058,
0.923362, 0.902299, 0.878986, 0.853553, 0.826144, 0.796910, 0.766016,
0.733634, 0.699946, 0.665140, 0.629410, 0.592956, 0.555982, 0.518696,
0.481304, 0.444018, 0.407044, 0.370590, 0.334860, 0.300054, 0.266366,
0.233984, 0.203090, 0.173856, 0.146447, 0.121014, 0.097701, 0.076638,
0.057942, 0.041719, 0.028058, 0.017037, 0.008717, 0.003144, 0.000350
])
chroma = np.zeros((semitones.shape[0], 12))
for i in range(semitones.shape[0]):
tones = semitones[i] * mywindow
cc = chroma[i]
for j in range(tones.size):
cc[j % 12] = cc[j % 12] + tones[j]
chroma = essentia_chord_utils.smooth(chroma,
window_len=int(1.3 * 44100 / 2048),
window='hanning').astype('float32')
#chroma = preprocessing.normalize(chroma, norm='max')
#beats = beats[::3]
#chords = ChordsDetectionBeats(hopSize=2048)
#syms, strengths = chords(chroma, beats)
noteNames = [
"A", "A#", "B", "C", "C#", "D", "D#", "E", "F", "F#", "G", "G#"
]
chordNames = np.empty(24, dtype='object')
for i in xrange(12):
p = i * 2
chordNames[p] = noteNames[i]
chordNames[p + 1] = noteNames[i] + ":min"
probs = chordPseudoProbabilities(2048, 44100, chroma, beats)
#indices = np.argmax(probs, axis=1)
#
#transitionMatrix = np.ones((24, 24)) * 1.0/24
#logMatrix = np.log(transitionMatrix)
#penalty = np.ones((24, 24)) * 0.15
#penalty[xrange(24), xrange(24)] = 0
#logMatrix = logMatrix - penalty
#
# temperley
pOthers = np.array([
113, 113, 1384, 1384, 410, 410, 326, 326, 2266, 2266, 20, 20, 2220,
2220, 412, 412, 454, 454, 1386, 1386, 162, 162
],
dtype='float')
pOthers = pOthers / sum(pOthers)
pSelftransition = 5.0 / 8.0
pModeChange = 1.0 / 24.0
pMaj = np.concatenate((np.array([pSelftransition, pModeChange]),
pOthers * (1.0 - pSelftransition - pModeChange)))
pMin = np.concatenate((np.array([pModeChange, pSelftransition]),
pOthers * (1.0 - pSelftransition - pModeChange)))
pMaj = smoothProbabilities(pMaj, 1.055)
pMin = smoothProbabilities(pMin, 1.055)
transitionMatrix = np.zeros((24, 24))
for i in xrange(12):
transitionMatrix[2 * i, :] = np.roll(pMaj, 2 * i)
transitionMatrix[2 * i + 1, :] = np.roll(pMin, 2 * i)
#
#
#transitionMatrix = np.ones((24, 24), dtype='float') * 1.0/24
# simplest one.
#p = 0.046
#transitionMatrix = np.ones((24, 24)) * (1 - p) / 24
#transitionMatrix[xrange(24), xrange(24)]=p
logMatrix = np.log(transitionMatrix)
indices = viterbi(probs, logMatrix)
syms = map(lambda x: chordNames[x], indices)
strengths = probs[xrange(len(indices)), indices]
segments = essentia_chord_utils.toMirexLab(0.0, l / 44100.0, beats, syms,
strengths)
#chords = ChordsDetection(windowSize=0.2)
#syms, strengths = chords(chroma)
#endTime = l / 44100.0
#stamps = np.arange(0, endTime, float(2048/44100.0))
#segments = essentia_chord_utils.toMirexLab(0.0, endTime, stamps, syms, strengths)
with open(outfile, 'w') as content_file:
for s in segments:
content_file.write(str(s) + '\n')
def chordBeats(infile, outfile):
# chroma_cq = librosa.feature.chroma_cqt(y=y, sr=44100)
chordHopSize = 4096
frameSize = 8192
tuningFreq = essentia_chord_utils.tuning(infile)
# initialize algorithms we will use
loader = MonoLoader(filename=infile)
framecutter = FrameCutter(hopSize=chordHopSize, frameSize=frameSize)
windowing = Windowing(type="blackmanharris62")
spectrum = Spectrum()
spectralpeaks = SpectralPeaks(orderBy="magnitude",
magnitudeThreshold=1e-05,
minFrequency=40,
maxFrequency=5000,
maxPeaks=10000)
hpcp = HPCP(size=12,
referenceFrequency=tuningFreq,
harmonics=8,
bandPreset=True,
minFrequency=40.0,
maxFrequency=5000.0,
bandSplitFrequency=500.0,
weightType="cosine",
nonLinear=True,
windowSize=1.0)
# use pool to store data
pool = essentia.Pool()
# connect algorithms together
loader.audio >> framecutter.signal
framecutter.frame >> windowing.frame >> spectrum.frame
spectrum.spectrum >> spectralpeaks.spectrum
spectralpeaks.magnitudes >> hpcp.magnitudes
spectralpeaks.frequencies >> hpcp.frequencies
hpcp.hpcp >> (pool, 'chroma.hpcp')
## network is ready, run it
#essentia.run(loader)
## don't forget, we can actually instantiate and call an algorithm on the same line!
print 'Loading audio file...', infile
audio = essentia.standard.MonoLoader(filename=infile)()
bt = BeatTrackerMultiFeature()
beats, _ = bt(audio)
beats = beats[::4]
#y, sr = librosa.load(infile, sr=44100)
#chroma_stft = librosa.feature.chroma_stft(y=y, sr=44100, n_chroma = 12, n_fft = 8192, hop_length=2048)
#chroma_stft = chroma_stft.transpose()
#chroma_stft = chroma_stft.astype('float32')
## to essentia convention
#chroma_stft = np.roll(chroma_stft, 3, 1)
#
#chords = ChordsDetectionBeats()
#syms, strengths = chords(chroma_stft, beats)
#y, sr = librosa.load(infile, sr=44100)
#C = np.log(np.abs(librosa.cqt(y, sr=sr,
# hop_length=2048,
# fmin=None,
# n_bins=12 * 7,
# bins_per_octave=12,
# tuning=None,
# norm=None,
# scale=False)) * 100 + 1)
#chroma_cqt = librosa.feature.chroma_cqt(y, sr, norm=2, C=C)
#chroma_cqt = librosa.feature.chroma_cqt(y, sr, hop_length=2048, norm=2, threshold=0.0)
#chroma_cqt = chroma_cqt.transpose()
#chroma_cqt = chroma_cqt.astype('float32')
## to essentia convention
#chroma_cqt = np.roll(chroma_cqt, 3, 1)
#
#chords = ChordsDetectionBeats()
#syms, strengths = chords(chroma_cqt, beats)
chords = ChordsDetectionBeats()
syms, strengths = chords(pool['chroma.hpcp'], beats)
print 'size: ', len(pool['chroma.hpcp'])
#name, ext = os.path.splitext(outfile)
#YamlOutput(filename=name+'.chroma.json')(pool)
segments = essentia_chord_utils.toMirexLab(0.0,
len(audio) / 44100.0, beats,
syms, strengths)
with open(outfile, 'w') as content_file:
for s in segments:
content_file.write(str(s) + '\n')
def chords(infile, outfile):
"""
Algorithm* peaks = factory.create("SpectralPeaks",
"maxPeaks", 10000,
"magnitudeThreshold", 0.00001,
"minFrequency", 40,
"maxFrequency", 5000,
"orderBy", "magnitude");
Algorithm* skey = factory.create("Key",
"numHarmonics", 4,
"pcpSize", 36,
"profileType", "temperley",
"slope", 0.6,
"usePolyphony", true,
"useThreeChords", true);
Algorithm* hpcp_chord = factory.create("HPCP",
"size", 36,
"referenceFrequency", tuningFreq,
"harmonics", 8,
"bandPreset", true,
"minFrequency", 40.0,
"maxFrequency", 5000.0,
"bandSplitFrequency", 500.0,
"weightType", "cosine",
"nonLinear", true,
"windowSize", 0.5);
"""
# initialize algorithms we will use
tuningFreq = essentia_chord_utils.tuning(infile)
print "tg = ", tuningFreq
chordHopSize = 2048
frameSize = 8192
loader = MonoLoader(filename=infile)
framecutter = FrameCutter(hopSize=chordHopSize, frameSize=frameSize)
windowing = Windowing(type="blackmanharris62")
spectrum = Spectrum()
spectralpeaks = SpectralPeaks(orderBy="magnitude",
magnitudeThreshold=1e-05,
minFrequency=40,
maxFrequency=5000,
maxPeaks=10000)
hpcp = HPCP(size=12,
referenceFrequency=tuningFreq,
harmonics=8,
bandPreset=True,
minFrequency=40.0,
maxFrequency=5000.0,
bandSplitFrequency=500.0,
weightType="cosine",
nonLinear=True,
windowSize=1.0)
chords = ChordsDetection()
# use pool to store data
pool = essentia.Pool()
# connect algorithms together
loader.audio >> framecutter.signal
framecutter.frame >> windowing.frame >> spectrum.frame
spectrum.spectrum >> spectralpeaks.spectrum
spectralpeaks.magnitudes >> hpcp.magnitudes
spectralpeaks.frequencies >> hpcp.frequencies
hpcp.hpcp >> chords.pcp
hpcp.hpcp >> (pool, 'chroma.hpcp')
chords.chords >> (pool, 'tonal.key_chords')
chords.strength >> (pool, 'tonal.key_strength')
# network is ready, run it
print 'Processing audio file...', infile
essentia.run(loader)
audio = essentia.standard.MonoLoader(filename=infile)()
endTime = len(audio) / 44100.0
stamps = np.arange(0, endTime, float(chordHopSize / 44100.0))
# workaround for Essentia behaviour I don't quite undestand
syms = list(pool['tonal.key_chords'][:-1])
strengths = list(pool['tonal.key_strength'][:-1])
segments = essentia_chord_utils.toMirexLab(0.0, endTime, stamps, syms,
strengths)
with open(outfile, 'w') as content_file:
for s in segments:
content_file.write(str(s) + '\n')