Hi all!
First of all, I m sory of my poor English.
I m using pocketsphinx on embedded system, and I trained an Acoustic model. It
works fine. I read the tutorial, wrote the sample program, I can record audio
with alsa on the embedded device.
But. I want to use the pocketsphinx_continuouse program, but it is not working
on the embedded device, but it works in PC.
So. I made these steps:
1. check recording is fine. (arecord) Check
2. start pocketsphinx_continuous -hmm .. -jsgf ... -dict (it's fine on PC) Check
3.
READY....
Listening...
Stopped listening, please wait...
INFO: cmn_prior.c(121): cmn_prior_update: from < 8.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 >
INFO: cmn_prior.c(139): cmn_prior_update: to < 13.70 -0.97 -0.23 -0.24 -0.20 -0.21 -0.14 -0.16 -0.19 -0.16 -0.12 -0.14 -0.12 >
INFO: fsg_search.c(1030): 623 frames, 61843 HMMs (99/fr), 54554 senones (87/fr), 3927 history entries (6/fr)
ERROR: "fsg_search.c", line 1099: Final state not reached in frame 623
000000000: (null)
READY....
Quit the program.
Check recording again, FAIL
SO after I ran pocketsphinx_continuous the arecord(alsa) broke.
Any Ideas?
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
you can try to modify continuous.c to record audio trough OpenAL. This
framework has a lot of audio interfaces and that work excellent on my Linux
11.04 PC.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Hi! I figured it out. I am use LINE-IN to record audio. After the pocketsphinx
started, It automatically change line-in to microphone.
But how can i use Line-in in pocketsphinx-continuous.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
But how can i use Line-in in pocketsphinx-continuous.
Selection of the Mic element is hard-coded in sphinxbase, there is no way to
switch it to Line-in without the code modification. I'd better change the
input name in the alsa configuration asounrc or alsa.conf so it will be named
"Mic", not line-in.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Hi，xvlabby
I wonder if you have ported the pocketsphinx to embedded systerm success.I'm
trying to compile it to embedded linux,too.But I have some trouble on config
it.Could you give me some suggest?Thank you!
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Hi! What kind of problem have you got?
I could compile it! I knew my models were good, and i can use it on the
embedded system with a diffrent program.
So my only problem was the pocketsphinx_continuous didn't work on PC, only
embedded pc. I figure out there were some problem on my sound card. I use LINE
input for recording instead of MIC. I have to compile sphinxbase end
pocketshinx to my embedded device. So i think the easiest way "rewrite" the
code. I only modified two things in the libspihnxad/ad_oss.c
I modified the INPUT_GAIN (it is not nessesary)
And I modified the sourceMic to SOUND_MASK_LINE.
Recompile and it works. (after compile, install and run do not change anything
mayba you should change ad_alsa.c or anything)
Or Have you got a compile problem?
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Hi,
I can't compile it yesterday,but I had solve it.
But,the problem you faced have come to me,too.I'm sure my alsa is work,but I
don't know how to config it.When I test the pocketsphinx_continuous on my
board,it said"ad_oss.c(103): Failed to open audio device(/dev/dsp): No such
file or directory
FATAL_ERROR: "continuous.c", line 242: Failed top open audio device".And I'm
trying to solve it,now.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Hi all!
First of all, I m sory of my poor English.
I m using pocketsphinx on embedded system, and I trained an Acoustic model. It
works fine. I read the tutorial, wrote the sample program, I can record audio
with alsa on the embedded device.
But. I want to use the pocketsphinx_continuouse program, but it is not working
on the embedded device, but it works in PC.
So. I made these steps:
1. check recording is fine. (arecord) Check
2. start pocketsphinx_continuous -hmm .. -jsgf ... -dict (it's fine on PC) Check
3.
SO after I ran pocketsphinx_continuous the arecord(alsa) broke.
Any Ideas?
The acoustic model is working, I think something wrong with the alsa.
you can try to modify continuous.c to record audio trough OpenAL. This
framework has a lot of audio interfaces and that work excellent on my Linux
11.04 PC.
It works well on my PC too, but i want to use it on my embedded device
(powerpc).
But thanx the reply, i'll check it.
Hi! I figured it out. I am use LINE-IN to record audio. After the pocketsphinx
started, It automatically change line-in to microphone.
But how can i use Line-in in pocketsphinx-continuous.
Selection of the Mic element is hard-coded in sphinxbase, there is no way to
switch it to Line-in without the code modification. I'd better change the
input name in the alsa configuration asounrc or alsa.conf so it will be named
"Mic", not line-in.
I modified the code and it works.
Thanks.
Hi，xvlabby
I wonder if you have ported the pocketsphinx to embedded systerm success.I'm
trying to compile it to embedded linux,too.But I have some trouble on config
it.Could you give me some suggest?Thank you!
Hi! What kind of problem have you got?
I could compile it! I knew my models were good, and i can use it on the
embedded system with a diffrent program.
So my only problem was the pocketsphinx_continuous didn't work on PC, only
embedded pc. I figure out there were some problem on my sound card. I use LINE
input for recording instead of MIC. I have to compile sphinxbase end
pocketshinx to my embedded device. So i think the easiest way "rewrite" the
code. I only modified two things in the libspihnxad/ad_oss.c
I modified the INPUT_GAIN (it is not nessesary)
And I modified the sourceMic to SOUND_MASK_LINE.
Recompile and it works. (after compile, install and run do not change anything
mayba you should change ad_alsa.c or anything)
Or Have you got a compile problem?
Sorry mismatch.
It worked only on Pc, and didn't work on embedded device.
Hi,
I can't compile it yesterday,but I had solve it.
But,the problem you faced have come to me,too.I'm sure my alsa is work,but I
don't know how to config it.When I test the pocketsphinx_continuous on my
board,it said"ad_oss.c(103): Failed to open audio device(/dev/dsp): No such
file or directory
FATAL_ERROR: "continuous.c", line 242: Failed top open audio device".And I'm
trying to solve it,now.
http://cmusphinx.sourceforge.net/wiki/faq#qfailed_to_open_audio_device_dev_ds
p_no_such_file_or_directory
Thank you for your help.
I have solve my problem.With the help of the Wiki and another post,I recompile
the Pocketsphinx.And It works as I hope.