speech_dtw
speech_dtw copied to clipboard
Dynamic time warping (DTW) functions for specifically speech alignment.
Dynamic Time Warping for Speech Processing
Overview
Tools (functions and scripts) for performing and evaluating speech processing tasks using dynamic time warping (DTW).
Dependencies
- Cython
- python_speech_features: Used in some of the examples.
- joblib: Used to parallelize
query-by-example search in
utils/qbe.py.
Building and testing the code
- Build the Cython components:
make - Run unit tests:
make test - Clean the code:
make clean
Examples
Examples are given in the examples/ directory. The script
examples/rnd_example.py illustrates standard DTW and compares a pure Python
implementation to the Cython implementation.
The IPython notebook speech_dtw_qbe.ipynb illustrates how to do (very basic) query-by-example (QbE) search using standard DTW.