I need make phonetic recognition for russian. I followed tutorial from project page.
In the tutorial I see
You can take a text, convert it to a phonetic strings using the phonetic dictionary for your langauge. Just replace the words with their corresponding transcription. Since number of phones is small, text shouldn't be big either, just a book will do.
But where I can find phonetic dictionary? How I can replace the words with their corresponding transcription in whole book?
Or..where I can find complete phonetic model?
UPDATED
I created file test.txt with line:
SIL i0 lj j a1 i1 lj f j e0 v gj e1 nj i0 j pj e0 t r o1 v z o0 l o0 t o1 j tj e0 lj o1 n o0 k SIL
This is the line form test wav file with replaced russian words to phonetic reprezentation form dictionary (ru.dic).
Then I used SRILM to build mode with command:
ngram-count -interpolate -text test.txt -lm test.phonetic.lm
Next I start it with command:
pocketsphinx\bin\Debug\Win32\pocketsphinx_continuous.exe -samprate 16000 -allphone testModel\test.phonetic.lm -hmm cmusphinx-ru-5.2 -remove_noise no -infile testModel\decoder-test.wav -backtrace yes
But when I use it I see incorrect result with wrong phonems.