SPTK 4.0 (Under Construction)

The Speech Signal Processing Toolkit (SPTK) is a software for speech signal processing tools for UNIX environments.

Documentation

See this page for a reference manual.

Requirements

• GCC 4.8+

Installation

The latest release can be installed through Git:

git clone https://github.com/sp-nitech/SPTK.git
cd SPTK
make

Then the SPTK commands can be used by adding SPTK/bin/ directory to the PATH environment variable. If you would like to use a part of the SPTK library, please link the static library SPTK/lib/libsptk.a.

Examples

The SPTK provides some examples. Go to an example directory and execute run.sh, e.g.,

cd egs/analysis_synthesis/mgc
./run.sh

The below is a simple example that decreases the volume of input audio. You may need to install sox command on your system.

sox -t wav input.wav -c 1 -t s16 -r 16000 - | \
   x2x +sd | sopr -m 0.5 | x2x +ds -r | \
   sox -c 1 -t s16 -r 16000 - -t wav output.wav

Changes from SPTK3

• Input and output types are changed to double from float
• New features:
  • Provide signal processing classes written in C++
  • Support conversion from/to log area ratio (lar2par and par2lar)
  • Support nonrecursive MLPG (mlpg)
  • Support pitch extraction by DIO used in WORLD (pitch)
  • Entropy calculation (entropy)
  • Huffman coding (huffman, huffman_encode, and huffman_decode)
  • Subband decomposition (pqmf and ipqmf)
  • Mel-filter-bank extraction (fbank)
  • Mel-cepstrum postfilter (mcpf)
• Obsoluted commands:
  • acep, agcep, and amcep -> amgcep
  • bell ->
  • c2sp -> mgc2sp
  • cat2 and echo2 ->
  • da ->
  • ds, us, us16, and uscd -> sox
  • gc2gc -> mgc2mgc
  • gcep, mcep, and uels -> mgcep
  • glsadf, lmadf, and mlsadf -> mglsadf
  • ivq -> imsvq
  • lsp2sp -> mglsp2sp
  • mgc2mgclsp and mgclsp2mgc -> mgc2mgc
  • raw2wav, wav2raw, wavjoin, and wavsplit -> sox
  • smcep ->
  • vq -> msvq
• Separated commands:
  • dtw -> dtw and dtw_merge
  • mglsadf -> mglsadf and imglsadf
  • train -> train and mseq
  • ulaw -> ulaw and iulaw
  • vstat -> vstat and median
• Renamed commands:
  • c2ir -> c2mpir
  • mgclsp2sp -> mglsp2sp

Relationship at a glance

Authors

• Keiichi Tokuda - Produce and Design - Nagoya Institute of Technology
• Keiichiro Oura - Nagoya Institute of Technology
• Takenori Yoshimura - Main Maintainer - Nagoya Institute of Technology
• Takato Fujimoto - Nagoya Institute of Technology
• Yoshihiko Nankaku - Nagoya Institute of Technology

Contributors

• Cassia Valentini - The University of Edinburgh
  • Calculated the coefficients of the 6th- and 7th-order modified Pade approximation.

License

This software is released under the Modified BSD license - see LICENSE file for details.