Am new to sphinx3. Have just downloaded it and compiled it a week back. When I run the sphinx3-simple test file I get the following error message :
Hit <cr> BEFORE and AFTER speaking:
INFO: feat.c(943): Feature buffers initialized to 256 vectors
INFO: cmn_prior.c(72): mean[0]= 12.00, mean[1..12]= 0.0
./sphinx3-simple: line 11: 3953 Segmentation fault $S3CONTINUOUS ${ARGS}
Any clues?
I am running RH9 (2.4.20) on a P4 - 2.66, 512 MB RAM, Creative Soundblaster Live 5.1.
The other demo, sphinx3-test worked out okay....although, when I tried to specify a file that I recorded using rawrec (16-bit, 16kHZ raw)....the mic volume is too low....can't seem to get that working properly....I have set it to 100 using aumix....but it's still records at a low volume....don't know why.
Thanks.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Anonymous
-
2004-04-10
Looks like a blatent bug in the demo code. Edit the file src/programs/live.c:
Line 232 reads....
if (live_endutt) /* RAH, It seems that we shouldn't throw out this data */
fe_end_utt(fe,dummyframe); /* Flush out the fe */
After this add...
if (!mfcbuf) return 0;
Then redo your make.
Regards,
CFB
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Hello:
Am new to sphinx3. Have just downloaded it and compiled it a week back. When I run the sphinx3-simple test file I get the following error message :
Hit <cr> BEFORE and AFTER speaking:
INFO: feat.c(943): Feature buffers initialized to 256 vectors
INFO: cmn_prior.c(72): mean[0]= 12.00, mean[1..12]= 0.0
./sphinx3-simple: line 11: 3953 Segmentation fault $S3CONTINUOUS ${ARGS}
Any clues?
I am running RH9 (2.4.20) on a P4 - 2.66, 512 MB RAM, Creative Soundblaster Live 5.1.
The other demo, sphinx3-test worked out okay....although, when I tried to specify a file that I recorded using rawrec (16-bit, 16kHZ raw)....the mic volume is too low....can't seem to get that working properly....I have set it to 100 using aumix....but it's still records at a low volume....don't know why.
Thanks.
Looks like a blatent bug in the demo code. Edit the file src/programs/live.c:
Line 232 reads....
if (live_endutt) /* RAH, It seems that we shouldn't throw out this data */
fe_end_utt(fe,dummyframe); /* Flush out the fe */
After this add...
if (!mfcbuf) return 0;
Then redo your make.
Regards,
CFB