/DTW

DTW

Primary LanguageJupyter Notebook

DTW

First we record 10 voices from ourselves and 10 voices from others and we want to use the threshold level to get the differences.

  • steps:
Screen Shot 2022-06-18 at 12 35 44 PM
  • Remove silence :
Screen Shot 2022-06-18 at 12 37 37 PM

Feature extraction:

    1. Zero crossing rate:
Screen Shot 2022-06-18 at 12 47 48 PM
    1. Spectrogram:
Screen Shot 2022-06-18 at 12 48 57 PM
    1. Mel spectrogram(mfcc):
Screen Shot 2022-06-18 at 12 49 31 PM
Screen Shot 2022-06-18 at 12 49 44 PM

DTW:

Screen Shot 2022-06-18 at 12 50 50 PM

  • Threshold level:
Screen Shot 2022-06-18 at 12 54 37 PM