I am trying to modify the transcriber demo to recognize a list of 5000 words.
I pipe delimited all my words so the format matches the transcriber demo, but
the results returned are never even close to the spoken word. What does
someone need to help me get this setup and working? I can get copies of the
wave files I am using and my word list if needed. Please help.
Thanks in advance.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
I then just run the transcriber demo and it doesn't even come close with my
wav files. Also if I remove all the words down to say Paint | Carpet |
Drywall. I still can't get it to recognize the word "Paint" pretty much I can
only get paint to be returned if it is the only word in the list.
What is the next step? Am I using the correct type of file format with
digits.gram? I haven't changed any other code other than the word list. Do I
need to change parameters in the config.xml to recognize the other words?
Thanks again.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
I am trying to modify the transcriber demo to recognize a list of 5000 words.
I pipe delimited all my words so the format matches the transcriber demo, but
the results returned are never even close to the spoken word. What does
someone need to help me get this setup and working? I can get copies of the
wave files I am using and my word list if needed. Please help.
Thanks in advance.
hmm, money?
You must have shared them without asking.
I posted the wav files here that I am using for Transcriber.java that I am
using.
http://www.gigasize.com/get.php?d=8kps225xjtb"> [http://www.gigasize.com/get.php?d=8kps225xjtb</a]
(http://www.gigasize.com/get.php?d=8kps225xjtb%3C/a)>
http://www.gigasize.com/get.php?d=8kps225xjtb
http://www.gigasize.com/get.php?d=9hy9bq5p2td"> [http://www.gigasize.com/get.php?d=9hy9bq5p2td</a]
(http://www.gigasize.com/get.php?d=9hy9bq5p2td%3C/a)>
http://www.gigasize.com/get.php?d=9hy9bq5p2td
This is my digits.gram file I used to try and recognize words
http://www.gigasize.com/get.php?d=jxg6vx2q44d"> [http://www.gigasize.com/get.php?d=jxg6vx2q44d</a]
(http://www.gigasize.com/get.php?d=jxg6vx2q44d%3C/a)>
http://www.gigasize.com/get.php?d=jxg6vx2q44d
I then just run the transcriber demo and it doesn't even come close with my
wav files. Also if I remove all the words down to say Paint | Carpet |
Drywall. I still can't get it to recognize the word "Paint" pretty much I can
only get paint to be returned if it is the only word in the list.
What is the next step? Am I using the correct type of file format with
digits.gram? I haven't changed any other code other than the word list. Do I
need to change parameters in the config.xml to recognize the other words?
Thanks again.
Sample rate is wrong.
6d845fea-0a1d-4446-ab1b-01897da35709.wav: RIFF (little-endian) data, WAVE
audio, Microsoft PCM, 16 bit, stereo 44100 Hz
Has to be 16k or 8k or whatever your model is.