sdqdlgj / kaldi-enhan

Tools for speech enhancement integrated with kaldi

SETK: Speech Enhancement Tools integrated with Kaldi

Here are some speech enhancement tools implemented integrated with kaldi. I use them for quick validation.

Finished based on Kaldi

• 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, may not very stable, need further debug)
• Fixed beamformer
• Compute angular spectrogram based on SRP-PHAT
• RIR generator(reference from RIR-Generator)

Now I mainly work on sptk package, development based on kaldi is stoped.

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 -j