Performance information when using wav file – Politepix /forums/topic/performance-information-when-using-wav-file/feed/ Tue, 23 Apr 2024 14:59:48 +0000 https://bbpress.org/?v=2.6.9 en-US /forums/topic/performance-information-when-using-wav-file/#post-1017429 <![CDATA[Performance information when using wav file]]> /forums/topic/performance-information-when-using-wav-file/#post-1017429 Fri, 07 Jun 2013 12:27:09 +0000 nh3500 Hi all,

I have a question about retrieving log of pocketsphinx,

I know we can simply set pocketsphinxController.verbosePocketSphinx = TRUE to get the log of pocketsphinx, and I did. But there is no log showed in the debugging window when I choose wav file as the source. Although it worked fine when it is in normal mode(continuous recognition).

The information I need is processing time. I can find it in normal mode.
But how can I find it when using wav as the source?
Should I turn on some parameter setting to allow pocketsphinx output the log?
If so, where can I find it?

I would appreciate any suggestions.

]]> /forums/topic/performance-information-when-using-wav-file/#post-1017434 <![CDATA[Reply To: Performance information when using wav file]]> /forums/topic/performance-information-when-using-wav-file/#post-1017434 Fri, 07 Jun 2013 12:38:36 +0000 Halle Winkler Welcome,

Logging should behave identically for runRecognitionOnWavFileAtPath:. If it isn’t I would take normal troubleshooting steps such as making sure that you have all logging turned on in Xcode’s console, etc. But I don’t recall there being anything like overall processing time in either log — what does it look like when you see it with continuous?

]]> /forums/topic/performance-information-when-using-wav-file/#post-1017435 <![CDATA[Reply To: Performance information when using wav file]]> /forums/topic/performance-information-when-using-wav-file/#post-1017435 Fri, 07 Jun 2013 12:44:42 +0000 nh3500 It looked as follow:

INFO: ngram_search.c(888): bestpath 0.00 CPU 0.001 xRT
INFO: ngram_search.c(891): bestpath 0.00 wall 0.002 xRT

Do I misunderstand the meaning?

]]> /forums/topic/performance-information-when-using-wav-file/#post-1017436 <![CDATA[Reply To: Performance information when using wav file]]> /forums/topic/performance-information-when-using-wav-file/#post-1017436 Fri, 07 Jun 2013 12:51:25 +0000 Halle Winkler Hi,

That tells you how much the speed was compared to the real time of the utterance, but not the overall processing time.

]]> /forums/topic/performance-information-when-using-wav-file/#post-1017437 <![CDATA[Reply To: Performance information when using wav file]]> /forums/topic/performance-information-when-using-wav-file/#post-1017437 Fri, 07 Jun 2013 12:55:25 +0000 nh3500 Hi,

Yes, I understand the meaning. For my research, that is enough.

Thanks for your reply. I will keep working on finding a solution.

]]> /forums/topic/performance-information-when-using-wav-file/#post-1017438 <![CDATA[Reply To: Performance information when using wav file]]> /forums/topic/performance-information-when-using-wav-file/#post-1017438 Fri, 07 Jun 2013 13:00:54 +0000 Halle Winkler What do you see in the logging output rather than the pocketsphinx run log?

]]> /forums/topic/performance-information-when-using-wav-file/#post-1017439 <![CDATA[Reply To: Performance information when using wav file]]> /forums/topic/performance-information-when-using-wav-file/#post-1017439 Fri, 07 Jun 2013 13:14:01 +0000 nh3500 This the log showed with wav file:

2013-06-07 21:03:47.928 OpenEarsSampleApp[1758:907] The received hypothesis is 床 with a score of -305 and an ID of 000000000
2013-06-07 21:03:49.395 OpenEarsSampleApp[1758:907] Flite has started speaking
2013-06-07 21:03:50.732 OpenEarsSampleApp[1758:907] Flite has finished speaking

And this is the log with continuous recognition:

2013-06-07 21:05:03.054 OpenEarsSampleApp[1770:907] Pocketsphinx has detected speech.
2013-06-07 21:05:04.330 OpenEarsSampleApp[1770:907] Pocketsphinx has detected a second of silence, concluding an utterance.
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 < 11.95 -0.01 -0.16 0.08 -0.27 -0.20 -0.07 -0.31 -0.24 -0.19 -0.23 -0.10 -0.24 >
INFO: ngram_search_fwdtree.c(1549): 468 words recognized (5/fr)
INFO: ngram_search_fwdtree.c(1551): 9619 senones evaluated (98/fr)
INFO: ngram_search_fwdtree.c(1553): 3634 channels searched (37/fr), 188 1st, 3446 last
INFO: ngram_search_fwdtree.c(1557): 849 words for which last channels evaluated (8/fr)
INFO: ngram_search_fwdtree.c(1560): 175 candidate words for entering last phone (1/fr)
INFO: ngram_search_fwdtree.c(1562): fwdtree 0.88 CPU 0.896 xRT
INFO: ngram_search_fwdtree.c(1565): fwdtree 1.78 wall 1.821 xRT
INFO: ngram_search_fwdflat.c(302): Utterance vocabulary contains 4 words
INFO: ngram_search_fwdflat.c(937): 410 words recognized (4/fr)
INFO: ngram_search_fwdflat.c(939): 11448 senones evaluated (117/fr)
INFO: ngram_search_fwdflat.c(941): 4525 channels searched (46/fr)
INFO: ngram_search_fwdflat.c(943): 968 words searched (9/fr)
INFO: ngram_search_fwdflat.c(945): 237 word transitions (2/fr)
INFO: ngram_search_fwdflat.c(948): fwdflat 0.75 CPU 0.763 xRT
INFO: ngram_search_fwdflat.c(951): fwdflat 0.75 wall 0.765 xRT
INFO: ngram_search.c(1214): </s> not found in last frame, using <sil>.96 instead
INFO: ngram_search.c(1266): lattice start node <s>.0 end node <sil>.46
INFO: ngram_search.c(1294): Eliminated 23 nodes before end node
INFO: ngram_search.c(1399): Lattice has 62 nodes, 35 links
INFO: ps_lattice.c(1367): Normalizer P(O) = alpha(<sil>:46:96) = -252498
INFO: ps_lattice.c(1405): Joint P(O,S) = -252548 P(S|O) = -50
INFO: ngram_search.c(888): bestpath 0.01 CPU 0.007 xRT
INFO: ngram_search.c(891): bestpath 0.00 wall 0.002 xRT
2013-06-07 21:05:05.254 OpenEarsSampleApp[1770:907] The received hypothesis is 起 with a score of -50 and an ID of 000000000
2013-06-07 21:05:05.620 OpenEarsSampleApp[1770:907] Pocketsphinx is now listening.
2013-06-07 21:05:05.625 OpenEarsSampleApp[1770:907] Flite has started speaking
2013-06-07 21:05:07.418 OpenEarsSampleApp[1770:907] Flite has finished speaking

And I checked the code of PocketsphinxController.mm. I found in the method runRecognitionOnWavFileAtPath, there is no following line of code:
if(self.verbosePocketSphinx == 1) verbose_pocketsphinx = 1;

May this be the reason why there is no log from pocketsphinx? However, I have no idea how to change and test it directly in OpenEarsSampleApp project…I found the code from OpenEars project, but in OpenEarsSampleApp project, it seems working with one complicated object named OpenEars in following path OpenEarsDistributionPreview/Framework/OpenEars.framework/Versions/A/OpenEars

]]> /forums/topic/performance-information-when-using-wav-file/#post-1017440 <![CDATA[Reply To: Performance information when using wav file]]> /forums/topic/performance-information-when-using-wav-file/#post-1017440 Fri, 07 Jun 2013 13:22:06 +0000 Halle Winkler Aha, thank you for the info. All you should have to do to make the change in your local version is open OpenEars.xcodeproj, make the change, and build it. A product will be created in Xcode that is another OpenEars.framework which you can replace the old one with in your sample app.

]]> /forums/topic/performance-information-when-using-wav-file/#post-1017441 <![CDATA[Reply To: Performance information when using wav file]]> /forums/topic/performance-information-when-using-wav-file/#post-1017441 Fri, 07 Jun 2013 14:00:42 +0000 nh3500 I add the code and rebuild the project. Then delete the framework from Build phase tab then add new OpenEars.framework. But it failed with following errors when I tried to run the project. In fact, when I built the OpenEars.xcodeproj, there were some errors also, but it is just about the location of some files and it built successfully. So I don’t think it is relevant.

Undefined symbols for architecture armv7:
“_fast_logmath_add”, referenced from:
_ptm_mgau_senone_eval in OpenEars(OpenEars)
_get_scores_4b_feat_all in OpenEars(OpenEars)
_get_scores_8b_feat_all in OpenEars(OpenEars)
_get_scores_8b_feat_6 in OpenEars(OpenEars)
_get_scores_8b_feat_5 in OpenEars(OpenEars)
_get_scores_8b_feat_4 in OpenEars(OpenEars)
_get_scores_8b_feat_3 in OpenEars(OpenEars)
…
ld: symbol(s) not found for architecture armv7
clang: error: linker command failed with exit code 1 (use -v to see invocation)

It might be caused by some problem related to setting of libraries, but I’ve added the new generated framework to the project. Any suggestion?

]]> /forums/topic/performance-information-when-using-wav-file/#post-1017442 <![CDATA[Reply To: Performance information when using wav file]]> /forums/topic/performance-information-when-using-wav-file/#post-1017442 Fri, 07 Jun 2013 14:10:32 +0000 Halle Winkler What were the framework errors?

]]>