Advanced I/O Use Cases¶
This section covers advanced use cases for input and output which go beyond the I/O functionality currently provided by librosa.
Read specific formats¶
librosa uses audioread for reading audio. While we chose this library for best flexibility and support of various compressed formats like MP3: some specific formats might not be supported. Especially specific WAV subformats like 24bit PCM or 32bit float might cause problems depending on your installed audioread codecs. libsndfile covers a bunch of these formats. There is a neat wrapper for libsndfile called PySoundFile which makes it easy to use the library from python.
See installation instruction for PySoundFile here.
Reading audio files using PySoundFile is similmar to the method in librosa. One important difference is that the read data is of shape
(nb_samples, nb_channels) compared to
(nb_channels, nb_samples) in
<librosa.core.load>. Also the signal is not resampled to 22050 Hz by default, hence it would need be transposed and resampled for further processing in librosa. The following example is equivalent to
1 2 3 4 5 6 7 8 9
import librosa import soundfile as sf # Get example audio file filename = librosa.util.example_audio_file() data, samplerate = sf.read(filename, dtype='float32') data = data.T data_22k = librosa.resample(data, samplerate, 22050)
For large audio signals it could be benficial to not load the whole audio file into memory. PySoundFile supports blockwise reading. In the following example a block of 1024 samples of audio are read and directly fed into the chroma feature extractor.
1 2 3 4 5 6 7 8 9 10 11 12 13
import numpy as np import soundfile as sf from librosa.feature import chroma_stft block_gen = sf.blocks('stereo_file.wav', blocksize=1024) rate = sf.info('stereo_file.wav').samplerate chromas =  for bl in block_gen: # downmix frame to mono (averaging out the channel dimension) y=np.mean(bl, axis=1) # compute chroma feature chromas.append(chroma_stft(y, sr=rate))
Read file-like objects¶
If you want to read audio from file-like objects (also called virtual files) you can use PySoundFile, as well.
E.g.: read files from zip compressed archives:
1 2 3 4 5 6 7 8
import zipfile as zf import soundfile as sf import io with zf.ZipFile('test.zip') as myzip: with myzip.open('stereo_file.wav') as myfile: tmp = io.BytesIO(myfile.read()) data, samplerate = sf.read(tmp)
This is a example does only work in python 3. For python 2 please use
from urllib2 import urlopen.
Download and read from URL:
1 2 3 4 5 6 7 8
import soundfile as sf import io from six.moves.urllib.request import urlopen url = "https://raw.githubusercontent.com/librosa/librosa/master/tests/data/test1_44100.wav" data, samplerate = sf.read(io.BytesIO(urlopen(url).read()))
Write out audio files¶
librosa uses scipy.io.wavfile to write out wav files. Please be aware, that this function uses the numpy dtype to determine the PCM subtype. For example if your processed audio array is of dtype
np.float64 (which is the default on most machines), your resulting wav file would be of type 64bit float as well. This is not considered to be a standard PCM wavfile. If you would like to write 16bit PCM you could convert your array before hand:
1 2 3 4 5 6 7 8 9 10
import numpy as np import librosa rate = 44100 data = np.random.randn(rate * 10, 2) maxv = np.iinfo(np.int16).max librosa.output.write_wav( "out_int16.wav", (audio * maxv).astype(np.int16), rate )
Writing audio files using pysoundfile is similar to the method in librosa, however it can automatically convert to a given PCM subtype and additionally support several compressed formats like FLAC or OGG:
1 2 3 4 5 6 7 8 9 10 11 12 13 14
import numpy as np import soundfile as sf rate = 44100 data = np.random.randn(rate * 10, 2) # Write out audio as 24bit PCM WAV sf.write('stereo_file.wav', data, samplerate, subtype='PCM_24') # Write out audio as 24bit Flac sf.write('stereo_file.flac', data, samplerate, format='flac', subtype='PCM_24') # Write out audio as 16bit OGG sf.write('stereo_file.ogg', data, samplerate, format='ogg', subtype='vorbis')