class ShortTimeFourierTransform(BaseModel):
meta = JSONFeature(MetaData, store=True, encoder=AudioMetaDataEncoder)
raw = ByteStreamFeature(ByteStream,
chunksize=chunksize_bytes,
needs=meta,
store=False)
ogg = OggVorbisFeature(OggVorbis, needs=raw, store=True)
pcm = AudioSamplesFeature(AudioStream, needs=raw, store=False)
resampled = AudioSamplesFeature(Resampler,
needs=pcm,
samplerate=resample_to,
store=store_resampled)
windowed = ArrayWithUnitsFeature(SlidingWindow,
needs=resampled,
wscheme=wscheme,
wfunc=OggVorbisWindowingFunc(),
store=store_windowed)
fft = ArrayWithUnitsFeature(FFT,
padding_samples=fft_padding_samples,
needs=windowed,
store=store_fft)
class Resampled(BaseModel):
meta = JSONFeature(MetaData, store=True, encoder=AudioMetaDataEncoder)
raw = ByteStreamFeature(ByteStream,
chunksize=chunksize_bytes,
needs=meta,
store=False)
ogg = OggVorbisFeature(OggVorbis, needs=raw, store=True)
pcm = AudioSamplesFeature(AudioStream, needs=raw, store=False)
resampled = AudioSamplesFeature(Resampler,
needs=pcm,
samplerate=resample_to,
store=store_resampled)
class Document(BaseModel):
raw = ByteStreamFeature(ByteStream,
chunksize=2 * 44100 * 30 * 2,
store=True)
ogg = OggVorbisFeature(OggVorbis, needs=raw, store=True)
pcm = AudioSamplesFeature(AudioStream, needs=raw, store=True)
resampled = AudioSamplesFeature(Resampler,
needs=pcm,
samplerate=samplerate,
store=True)
windowed = ArrayWithUnitsFeature(SlidingWindow,
needs=resampled,
wscheme=windowing_scheme,
wfunc=OggVorbisWindowingFunc(),
store=False)
fft = ArrayWithUnitsFeature(FFT, needs=windowed, store=True)
chroma = ArrayWithUnitsFeature(Chroma,
needs=fft,
frequency_band=band,
store=True)
bark = ArrayWithUnitsFeature(BarkBands,
needs=fft,
frequency_band=band,
store=True)
bfcc = ArrayWithUnitsFeature(BFCC, needs=bark, store=True)
bfcc_sliding_window = ArrayWithUnitsFeature(
SlidingWindow,
needs=bfcc,
wscheme=windowing_scheme * Stride(frequency=2, duration=4),
store=True)
bfcc_pooled = ArrayWithUnitsFeature(Max,
needs=bfcc_sliding_window,
axis=1,
store=True)
class AudioGraph(BaseModel):
meta = JSONFeature(MetaData, store=True, encoder=AudioMetaDataEncoder)
raw = ByteStreamFeature(ByteStream,
chunksize=chunksize_bytes,
needs=meta,
store=False)
ogg = OggVorbisFeature(OggVorbis, needs=raw, store=True)
pcm = AudioSamplesFeature(AudioStream, needs=raw, store=False)
resampled = AudioSamplesFeature(Resampler,
needs=pcm,
samplerate=resample_to,
store=False)
windowed = ArrayWithUnitsFeature(SlidingWindow,
needs=resampled,
wscheme=HalfLapped(),
wfunc=OggVorbisWindowingFunc(),
store=False)
dct = ArrayWithUnitsFeature(DCT, needs=windowed, store=True)
fft = ArrayWithUnitsFeature(FFT, needs=windowed, store=store_fft)
bark = ArrayWithUnitsFeature(BarkBands,
needs=fft,
frequency_band=band,
store=True)
centroid = ArrayWithUnitsFeature(SpectralCentroid,
needs=bark,
store=True)
chroma = ArrayWithUnitsFeature(Chroma,
needs=fft,
frequency_band=band,
store=True)
bfcc = ArrayWithUnitsFeature(BFCC, needs=fft, store=True)