Skip to content

octaviotron/wave-peaks

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

7 Commits
 
 
 
 
 
 
 
 

Repository files navigation

wave-peaks

Wave data analysis and processing for finding chunks and segments of audio.

This script takes a WAV file as input and detects segments signal data over a thresshold and when a quorum is reached, the lapse is stored in a separated WAV file. This is mainly used in audio files with a long silences or long noise parts where you need to have separated the sound over this thresshold.

Lets take the following image as example of the waveform of a recorded audio with a lot of bakcground noise. In the white areas are sound wich we want to export and in the gray areas are the noisy parts we need to ignore:

wave

The first thing this script does is read wave data and convert all values into a positive number between 0 and 32768 (16 bits), and using this range the script algorithm takes a double thresshold value:

  • MAX: Defines the level above all wave data values are selected as "posibly part of the white zone". In the image is the horizontal dashed line of the red area
  • TOP: Defines the value beyond all wave data values are selected as "posibly not noise"
  • TASA (0.1-0.99): defines the amount of concurrent peaks ina segment to be considered "not-a-single-peak"

Finally the selected secctions are exported as separated audio files.

About

Wave data analysis and processing for finding chunks and segments of audio.

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages