Speech data preprocessing toolkit
This toolkit implement variational method for process speech data.
- Speech balancing: Use three method (Box-cox, Yeo-John, Quantitle transform) for transform max speech amlitudes of each audio data into normal space.
Requirements: See requirements.txt