ysdede commited on
Commit
9d8a025
·
verified ·
1 Parent(s): 1067bdf

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +17 -0
README.md CHANGED
@@ -86,6 +86,12 @@ Both approaches are valid, but mixing them could lead to evaluation issues.
86
 
87
  ## Audio Processing and Quality Improvements
88
 
 
 
 
 
 
 
89
  ### Silence Trimming
90
  I processed all audio files to remove unnecessary silence and noise:
91
  - Used Silero VAD with a threshold of 0.6 to detect speech segments
@@ -201,3 +207,14 @@ The final dataset shows significant improvements:
201
  - Maintained original metadata (age, upvotes, etc.)
202
 
203
  These improvements make the dataset more suitable for training speech recognition models while maintaining the diversity and richness of the original CommonVoice collection.
 
 
 
 
 
 
 
 
 
 
 
 
86
 
87
  ## Audio Processing and Quality Improvements
88
 
89
+ ### Audio Resampling
90
+ All audio files were resampled to 16 kHz to:
91
+ - Make the dataset directly compatible with Whisper and similar models
92
+ - Eliminate the need for runtime resampling during training
93
+ - Ensure consistent audio quality across the dataset
94
+
95
  ### Silence Trimming
96
  I processed all audio files to remove unnecessary silence and noise:
97
  - Used Silero VAD with a threshold of 0.6 to detect speech segments
 
207
  - Maintained original metadata (age, upvotes, etc.)
208
 
209
  These improvements make the dataset more suitable for training speech recognition models while maintaining the diversity and richness of the original CommonVoice collection.
210
+
211
+ ## Tools Used
212
+
213
+ This dataset processing work was completed using [ASRTK (Automatic Speech Recognition Toolkit)](https://github.com/ysdede/asrtk), an open-source Python toolkit designed to streamline the development and enhancement of ASR systems. ASRTK provides utilities for:
214
+
215
+ - Audio processing with advanced splitting and resampling capabilities
216
+ - Text normalization and cleaning
217
+ - Forced alignment using Silero VAD models
218
+ - Efficient batch processing with multi-threading support
219
+
220
+ The toolkit is available under the MIT license and welcomes contributions from the community.