Ticket #445 (new enhancement)

Opened 12 years ago

Noise modelling

Reported by: kmaclean Owned by: kmaclean
Priority: major Milestone: SpeechSubmission 0.1.7
Component: Acoustic Model Version: Acoustic Model 0.1.1
Keywords: Cc:

Description

See nsh's post:

Actually it's a bad thing, since real data for recognition will always have noise. In proper databases noise is just marked as such in labels and it's possible to train noise model. Of course there are different types of noise marked as fillers:

++AH++ +AH+ ++BEEP++ +TONE+ ++NOISE++ +NOISE+ ++DOOR_SLAM++ +SLAM+ ++BREATH++ +BREATH+ ++GRUNT++ +GRUNT+ ++LIP_SMACK++ +SMACK+ ++PHONE_RING++ +RING+ ++CLICK++ +CLICK+

Note: See TracTickets for help on using tickets.