GitLab

* Add basic filtfilt implementation

* Add filtfilt to functional package; add tests
Co-authored-by: V G <vladislav.goncharenko@phystech.edu>

`torchaudio` treats `soundfile` as an optional dependency, and it assumes either `soundfile` is properly installed or it is not installed at all. However, there is a third state where `soundfile` is installed but the backing library `libsndfile` is not installed, and in this case, `import torchaudio` fails.

This commit resolves it by further checking if `soundfile` is importable. 

See also: https://github.com/pytorch/audio/issues/1687

When building torchaudio from source, `get_version.sh` from kaldi is executed everytime,
which results in kaldi-bindings to be always rebuilt.

This commit add "if" guard to the part so that they are not always executed.

* Reduce length of waveform in pitch_shift batch_consistency test

Summary: To address the test failure in T96406395

Reviewed By: carolineechen

Differential Revision: D30163741

fbshipit-source-id: f88d86b3da7b1ee52518934567b0b0a62700ee58

* Fix batch consistency test in transforms

Summary: The stress test still fails. Add n_fft to address it.

Reviewed By: mthrok

Differential Revision: D30218279

fbshipit-source-id: 7858efd3e5ac0073193a7883fd314486efc73814
Co-authored-by: Zhaoheng Ni <zni@fb.com>

BUILD_VERSION gets set by upstream scripts and should be available
whenever we're attempting to build audio in conda environments.

This solves the issue where conda packages appeared to be building a
development version instead of the actual version we were targeting
Signed-off-by: Eli Uriegas <eliuriegas@fb.com>

- Provide InverseSpectrogram module that corresponds to Spectrogram module
- Add length parameter to the forward method in transforms
Co-authored-by: dgenzel <dgenzel@fb.com>
Co-authored-by: Zhaoheng Ni <zni@fb.com>

ref: D29989551

D30080845