Changes between Version 5 and Version 6 of AcousticModels


Ignore:
Timestamp:
07/28/06 11:08:07 (16 years ago)
Author:
kmaclean
Comment:

--

Legend:

Unmodified
Added
Removed
Modified
  • AcousticModels

    v5 v6  
    66  * [http://www.ldc.upenn.edu/Catalog/CatalogEntry.jsp?catalogId=LDC97S44 1996 English Broadcast News Speech (Hub-4)] - 104 hours of broadcasts 
    77  * [http://www.ldc.upenn.edu/Catalog/CatalogEntry.jsp?catalogId=LDC98S71 1997 English Broadcast News Speech (Hub-4)] - 97 hours of news broadcasts  
     8 
     9== Estimating Storage requirements: == 
     10 
     11  * for 48kHz:16bit audio, 5 seconds of audio takes 500k. 
     12  * therefore about 30 meg per minute! 
     13  * if we want 140 hours of speech, we will need 252000 Meg or around 252Gig (assumes a 1000k per Meg), for Original data. 
     14  * Will likely need at least double that space with the propagation of audio (downsampling, noise reduction, etc.) through version control to create Acoustic Models. 
     15  * !VoxForge server currently holds 200 Gig, but can easily add additional storage. 
     16  * Bandwidth is a greater issue, therefore require Peer-to-Peer sharing of audio files (i.e. Bittorrent) - see ticket #11. 
     17