Hello i am doing Confidence measure in speech recognition. I read a paper in
which i came up with a method in which phone durations are compared..
Basically in each utterance i want to extract the phone boundaries.
I ran sphinx3_align as
sphinx3_align
-hmm model_1/ (dir in which the model is present)
-dict marathiAgmark1500.dic (dictionary)
-fdict 850spkr.filler (filler dict)
-ctl docs/fileid.txt (contains the fileIds)
-insent phone.insent (input transcription file)
-cepdir features/ (mfcc features)
-phsegdir phonesegdir/
-phlabdir phonelabdir/
-stsegdir statesegdir/
-wdsegdir aligndir/
-outsent phone.outsent
I ran this and got a error
INFO: main_align.c(1009): M13MH01A0001I501: 96 input frames
not in dictionary", line 889: (M13MH01A0001I501)
ERROR: "main_align.c", line 826: No sentence HMM; no alignment for
M13MH01A0001I501
Please help me out.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
and this is the corresponding word in the dict "SOMETHING HERE"kaapuusa k aa
p u s
It doesn't mean that the word is in dictionary. The word transcription might
have incorrect phones or might be stripped for some other reason for example
because file format is wrong. If the word is in the last line you might miss a
proper newline after it. You should have earlier message about it in the log.
You need to read the whole log, not just the last line of it. To get faster
answer share the files you are using. You can do it through dropbox.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Hey thanks a lot.. i got the error.
Actually My transcription file was from Windows and i read it in linux so an
extra carriage return was added at the end..
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Hello i am doing Confidence measure in speech recognition. I read a paper in
which i came up with a method in which phone durations are compared..
I ran sphinx3_align as
I ran this and got a error
INFO: main_align.c(1009): M13MH01A0001I501: 96 input frames
not in dictionary", line 889: (M13MH01A0001I501)
ERROR: "main_align.c", line 826: No sentence HMM; no alignment for
M13MH01A0001I501
Please help me out.
with reference to
https://sourceforge.net/projects/cmusphinx/forums/forum/5471/topic/4687509
In those three words:
which word do you have trouble to understand?
But the word in the transcription is already there in the dictionary..
This is my phone.insent file
kaapuusa (M13MH01A0001I501)
and this is the corresponding word in the dict
kaapuusa k aa p u s
I am not getting this error..
It doesn't mean that the word is in dictionary. The word transcription might
have incorrect phones or might be stripped for some other reason for example
because file format is wrong. If the word is in the last line you might miss a
proper newline after it. You should have earlier message about it in the log.
You need to read the whole log, not just the last line of it. To get faster
answer share the files you are using. You can do it through dropbox.
Hey thanks a lot.. i got the error.
Actually My transcription file was from Windows and i read it in linux so an
extra carriage return was added at the end..