Speech Recognition Engines

Nested
How create xxx_test.fileids and xxx_test.transcription
User: psoto
Date: 7/27/2009 1:01 am
Views: 4029
Rating: 1

 

Hi, I'm new to sphinx3, and I want to create a recognizer for words in Spanish.
I followed the tutorial CMU, using a similar case to use an4, supplemented with information from a blog bakuzen (Acoustic model creation using SphinxTrain, www.bakuzen.com/?p16) 
I've done so far is the stage of training, for it generates the following files:

audio.dic======an4.dic

audio.filler=====an4.filler

audio.phone====an4.phone

audio_train.fileids====an4_train.fileids

audio_train.transcription======an4_train.transcription

saved in the folder /etc 

in addition to audio files in the folder to save /wav

but following the tutorial of CMU, then run 

perl scripts_pl /Runall.pl 

So far, no problems

I run

perl scripts_pl/make_feats.pl -ctl etc/an4_test.fileids

but, I don`t have the file an4_test.fileids, whish brings you to the tutorial along with an4_test.transcription

Any suggestions as to create it?

 

--- (Edited on 7/27/2009 1:01 am [GMT-0500] by psoto) ---

Re: How create xxx_test.fileids and xxx_test.transcription
User: nsh
Date: 7/27/2009 7:38 am
Views: 1704
Rating: 1

The same way you created audio_train.fileids and audio_train.transcriptoin, don't you think so?

 

--- (Edited on 7/27/2009 7:38 am [GMT-0500] by nsh) ---

PreviousNext