I had troubles upon using either ps_alignment or sphinx3_align: for the first I couldn't work out the location or the usage of the binary, and for the second - I couldn't run it on my data, it informed me that "Final state not reached; no alignment for audio.wav".
Running the sphinx3_align I tried putting the <sil> filler at the end and the beggining of the phrase, trying to run it on a single phrase, it didn't work :(
The command is the following:
Understood, thank you!
I proceeded to trying to make feature files from the voxforge audio, but I get a segmentalion error. What could be a possible reason behind it, can you consult me, please?
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Sorry to bother, I defined the arguments as such (for 8000 Hz audio): sphinx_fe -i ru_0022.wav -o 1.mfc -upperf 3500 -samprate 8000
and everything worked.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
As I continued on with the alignment I ran into similar problems, although I have generated the feature files with sphinx_fe tool and using the feat.params of the acoustic model I planned aligning with. This whole situation really confuses me now.
The terminal output is also strange, mentioning various errors:
So the model initialisation seems fine, but it is followed by strange errors, except for the last one - which is a simple file not found error. When I try to align the files with strange "main_align.c" errors, it gives me a simple "final state is not reached" error.
Could you please notify what are the meanings of those strange errors and what could be the problem with the feature files this time?
Thanks in advance,
Olya
Last edit: Dino The Dinosaur 2018-01-04
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
The proper tool for feature extraction is sphinx_fe. It is important to be very accurate, otherwise you'll frequently experience problems like this one.
ERROR: "main_align.c", line 907: Utt ru_0027: Input file read (1-20121125-pgp/wav/ru_0027) with dir (/media/dino/DATA/corpus/voxforge/repository/downloads/Russian/Trunk/Audio/Main/8kHz_16bit) and extension (.mfc) failed
Files are missing because you haven't created them properly, you need to revisit the previous step.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Thank you! :)
I investigated the problem and understood, that this error originally occurs when there is a significant mismatch between the audio and transcript. Besides, I found out that this error might occur when the parameters of the feature files and the parameters of the acoustic model do not match, which seems to be the case. This way I do not really comprehend why I cannot align with this model, since I used its parameters. May there had been something I've missed?
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Hello!
I had troubles upon using either ps_alignment or sphinx3_align: for the first I couldn't work out the location or the usage of the binary, and for the second - I couldn't run it on my data, it informed me that "Final state not reached; no alignment for audio.wav".
Running the sphinx3_align I tried putting the <sil> filler at the end and the beggining of the phrase, trying to run it on a single phrase, it didn't work :(
The command is the following:
sudo sphinx3_align \ -hmm /home/dino/sphinx/AcousticModels/model_parameters/msu_ru_nsh.cd_cont_1000_8gau_8000 \ -dict /home/dino/sphinx/AcousticModels/etc/msu_ru_nsh.dic \ -ctl etc/idi \ -adcin yes \ -senmgau .cont. \ -insent etc/prompt \ -outsent aligned.transcriptions \ -logfn dev/null \ -cepdir /media/dino/DATA/corpus/www.repository.voxforge1.org/downloads/Russian/Trunk/Audio/MFCC/8kHz_16bit/MFCC_0_D/1-20121125-pgp \ -remove_noise no -remove_silence no -upperf 8000 -lowerf 1 -round_filters no -remove_dc yes \ -fdict /home/dino/sphinx/AcousticModels/etc/msu_ru_nsh.fillerMaybe there was something I missed? I cannot determine that mistake myself.
here are also my idi
and prompt file
Last edit: Dino The Dinosaur 2017-12-22
MFC files you download from voxforge are incompatible with cmusphinx, they are for htk. You need to extract features properly first.
Also it is better to use more recent models.
Understood, thank you!
I proceeded to trying to make feature files from the voxforge audio, but I get a segmentalion error. What could be a possible reason behind it, can you consult me, please?
Sorry to bother, I defined the arguments as such (for 8000 Hz audio):
sphinx_fe -i ru_0022.wav -o 1.mfc -upperf 3500 -samprate 8000and everything worked.
As I continued on with the alignment I ran into similar problems, although I have generated the feature files with sphinx_fe tool and using the feat.params of the acoustic model I planned aligning with. This whole situation really confuses me now.
The terminal output is also strange, mentioning various errors:
So the model initialisation seems fine, but it is followed by strange errors, except for the last one - which is a simple file not found error. When I try to align the files with strange "main_align.c" errors, it gives me a simple "final state is not reached" error.
Could you please notify what are the meanings of those strange errors and what could be the problem with the feature files this time?
Thanks in advance,
Olya
Last edit: Dino The Dinosaur 2018-01-04
The proper tool for feature extraction is sphinx_fe. It is important to be very accurate, otherwise you'll frequently experience problems like this one.
Files are missing because you haven't created them properly, you need to revisit the previous step.
Sorry, it was a typo, I generated feats with sphinx_fe.
Happy new year, by the way! :)
Happy New Year! Wish you get through this asap ;)
Thank you! :)
I investigated the problem and understood, that this error originally occurs when there is a significant mismatch between the audio and transcript. Besides, I found out that this error might occur when the parameters of the feature files and the parameters of the acoustic model do not match, which seems to be the case. This way I do not really comprehend why I cannot align with this model, since I used its parameters. May there had been something I've missed?