Ticket #65 (closed enhancement: wontfix)

Opened 15 years ago

Last modified 15 years ago

Adapting Acoustic Model using mfc and own audio

Reported by: kmaclean Owned by: kmaclean
Priority: minor Milestone: Unassigned
Component: Acoustic Model Version: 0.1-alpha
Keywords: Cc:

Description

Update tutorial to include adaptation process to copy VoxForge? Speaker Independent mfc files, user mfc files, update script, and re-run Compile_AcousticModel.sh against everything - Rather than trying to get HTK 3.2.1 HEAdapt to work properly. mfc files are much smaller than wav files.

Change History

comment:1 Changed 15 years ago by kmaclean

.mfc files may be much smaller than wav files, but they are still too big when compared to an Acoustic Model: VoxForge? Speaker Indepdent Acoustic Model - build396 includes the following audio: 

cmu_us_bdl_arctic 
cmu_us_rms_arctic 
cmu_us_slt_arctic 
cmu_us_clb_arctic 
cmu_us_jmk_arctic 
cmu_us_awb_arctic 
cmu_com_kal_ldom 
an4

Original/16kHz:16bit:: 1.2 GB
Main/8kHz:16bit: downsampled to 8kHz:16-bit files: 554MB
MFCC: wav converted to MFCC files:

  • MFCC/8kHz:16bit: 183MB
  • MFCC/16kHz:16bit: ?

comment:2 Changed 15 years ago by kmaclean

  • Status changed from new to closed
  • Resolution set to wontfix
Note: See TracTickets for help on using tickets.