Feature #883
audio normalisation
| Status: | New | Start date: | 2014-03-04 | |
|---|---|---|---|---|
| Priority: | Normal | Due date: | ||
| Assignee: | - | % Done: | 0% | |
| Category: | - | |||
| Target version: | - |
Description
Before analysis starts, it would be great if the audio could be normalised to either
- max(abs(audio)) = 1, or
- percentile(abs(audio), 0.95) = 0.5 (or related percentile normalisation)
This would be good for analysis and display purposes.
History
#1
Updated by Matthias Mauch almost 11 years ago
This would be good for analysis and display purposes.
Just to flesh this out briefly:
- for visualisation it would obviously be helpful if the waveform and spectrogram representations made good use of the range they're given
- for analysis one could assign lower probability to low-amplitude signals, thus reducing the number of false positive pitch detections (and hence faster annotation)