Python port of Lyon's model calculation from Auditory Toolbox. Original version of Auditory Toolbox is written in C and MATLAB by Malcolm Slaney.

• Modified soscascade.c, agc.c and sosfilters.c (removed MEX-related part).
• ctypes wrapper for soscascade(), agc() and sosfilters() calls.
• Translation from MATLAB to Python for files necessary for successfull call to LyonPassiveEar().

If you plan to build manually and run tests, you'll need libcheck installed. On Ubuntu 18.10 run:

sudo apt-get install check

pip install lyon

Build a library and copy to installation location:

make -C c_src lib
cp c_src/liblyon.so lyon/

you can verify that library is properly built by running a test suite:

make -C c_src test

The following code computes cochleogram for a sample sound:

from lyon import LyonCalc

calc = LyonCalc()
waveform, sample_rate = load('audio/file/path.wav')
decimation_factor = 64
coch = calc.lyon_passive_ear(waveform, sample_rate, decimation_factor)
print(coch.shape)

The code above should output shape of resulting auditory nerve response: [<number of samples / decimation_factor>, 86].

See examples/lyon_examples.py for examples on running other functions.

• AuditoryToolbox page
• Report describing function in toolbox