cc-cherie / kaldi-enhan

Tools for speech enhancement based on kaldi

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

SETK: Speech Enhancement Tools based on Kaldi

This is some speech enhancement tools implemented based on kaldi. I use them for quick experiment.

Updating...

Finished

  • Compute kinds of masks(ibm, irm etc)
  • Compute (phase angle/power&magnitude spectrogram/complex STFT results) of input wave
  • Seperate target component from input wave according to input masks
  • Estimate wave from enhanced spectrogram and reference wave
  • Complex matrix/vector class
  • MVDR/max-SNR beamformer(depend on T-F mask)
  • Fixed beamformer
  • Compute angular spectrogram based on SRP-PHAT
  • RIR generator(reference from RIR-Generator)

Install

Compile pass on Mac OSX, Ubuntu and RedHat.

Patch matrix/matrix-common.h in Kaldi

typedef enum {
    kTrans          = 112,  // CblasTrans
    kNoTrans        = 111,  // CblasNoTrans
    kConjTrans      = 113,  // CblasConjTrans
    kConjNoTrans    = 114   // CblasConjNoTrans
} MatrixTransposeType;

Then run

mkdir build
cd build
export KALDI_ROOT=/kaldi/root/dir
# if on UNIX, need compile kaldi with openblas
export OPENBLAS_ROOT=/openblas/root/dir
cmake ..
make

About

Tools for speech enhancement based on kaldi


Languages

Language:C++ 70.3%Language:Python 19.9%Language:Shell 8.5%Language:CMake 1.3%