Ticket #12 (closed task: wontfix)
Get Users to Add a Blank Track to their Speech Audio Recordings
| Reported by: | kmaclean | Owned by: | kmaclean |
|---|---|---|---|
| Priority: | minor | Milestone: | Unassigned |
| Component: | Audio | Version: | 0.1-alpha |
| Keywords: | blank track noise levels spectral subtraction | Cc: |
Description
Do we need a blank track to determine noise in a user submission ... so we can use audacity to remove noise over a certain threshhold.
Julius uses Spectral Substraction to remove noise!!!! how can this be used to acoustic model creation
Change History
comment:1 Changed 16 years ago by kmaclean
- Keywords track added; tracl removed
- Summary changed from Add Blank Track to User Recordings to Get Users to Add a Blank Track to their Speech Audio Recordings
comment:3 Changed 16 years ago by kmaclean
- Status changed from new to closed
- Resolution set to fixed
directing users to add one second before and after every file they record - this should be enough to determine noise levels of the recording.
comment:4 Changed 16 years ago by anonymous
- Status changed from closed to reopened
- Resolution fixed deleted
- Milestone changed from 0.1-beta to 0.1-gamma
See if Julius' mkss command (to compute average spectrum from mic input for Spectral substraction - to improve speech recognition) can be used to help automatically clean audio that a user submitts. Likely have to stick with audacity's utilities to clean the audio ...
Note: See
TracTickets for help on using
tickets.
