Speech Signal Processing Toolkit (SPTK) Tutorials

Install the SPTK

Install the SoX

1
apt-get install sox

Clone the Git Repo

1
git clone https://github.com/sp-nitech/SPTK.git

Compile the C++

1
cd SPTK; make

Export the Environmental Variable

1
vim ~/.bashrc

image-20220825050436624

Just export the bin path to your environmental variable.

1
export PATH=$PATH:/mnt/c/Users/ABC/SPTK/bin

Analysis Synthesis

Pitch Extraction

1
pitch -s 16 -p 80 -L 80 -H 200 -o 1 < test.wav > data.f0

LPC

When we reviewed the lpc directory, which is under /SPTK/egs/analysis_synthesis/lpc.

image-20220912232026981

When we finally compare the output file:

data.syn.raw:

image-20220912233129993

image-20220912232904531

image-20220912232118579

And compared with the original wav file, data.org.wav.

image-20220912233057842

image-20220912232740976

image-20220912232138056

We can found out that the sound pressure is became more linear.

Python Version SPTK

we can have a python version SPTK and it combines the same functions:

https://nbviewer.org/github/r9y9/pysptk/blob/master/examples/pysptk%20introduction.ipynb

Speech Signal Processing Toolkit (SPTK) Tutorials
http://xiaos.site/2022/08/25/Speech-Signal-Processing-Toolkit-SPTK-Tutorials/
Author
Xiao Zhang
Posted on
August 25, 2022
Licensed under