The aubio extension for Python is available for Python 2.7 and Python 3.
From aubio source directory, run the following:
$ ./setup.py clean
$ ./setup.py build
$ sudo ./setup.py install
Once you have python-aubio installed, you should be able to run python -c "import aubio; print(aubio.version)".
Here is a simple script that reads all the samples from a media file:
#! /usr/bin/env python
import sys, aubio
samplerate = 0 # use original source samplerate
hop_size = 256 # number of frames to read in one block
s = aubio.source(sys.argv[1], samplerate, hop_size)
total_frames = 0
while True: # reading loop
samples, read = s()
total_frames += read
if read < hop_size: break # end of file reached
fmt_string = "read {:d} frames at {:d}Hz from {:s}"
print (fmt_string.format(total_frames, s.samplerate, sys.argv[1]))
Here is a more complete example, demo_filter.py. This files executes the following:
#! /usr/bin/env python
def apply_filter(path):
from aubio import source, sink, digital_filter
from os.path import basename, splitext
# open input file, get its samplerate
s = source(path)
samplerate = s.samplerate
# create an A-weighting filter
f = digital_filter(7)
f.set_a_weighting(samplerate)
# alternatively, apply another filter
# create output file
o = sink("filtered_" + splitext(basename(path))[0] + ".wav", samplerate)
total_frames = 0
while True:
samples, read = s()
filtered_samples = f(samples)
o(filtered_samples, read)
total_frames += read
if read < s.hop_size: break
duration = total_frames / float(samplerate)
print ("read {:s}".format(s.uri))
print ("applied A-weighting filtered ({:d} Hz)".format(samplerate))
print ("wrote {:s} ({:.2f} s)".format(o.uri, duration))
if __name__ == '__main__':
import sys
for f in sys.argv[1:]:
apply_filter(f)
Check out the python demos folder for more examples.
A number of python tests are provided. To run them, use python/tests/run_all_tests.