Reputation: 21
I'm currently using dragon to perform phone calls transcriptions on an IVR monitoring system. basically I monitor a folder with dragon, drop wav files in there and wait for the transcription output in another folder. such solution has given quite some headaches lately, due to dragon crashes.
I'd like to have a simple executable performing transcriptions on demand (and thus eliminating the need of a permanently running sw), something like:
cmd -> transcribe in.wav > out.txt
The software should have Italian dictionary and acoustic model available, or at least should be readily trainable to support Italian. I need to run on windowx xp embedded (a quite old version!) with no office suite, so windows speech engine is not an option.
Any lib/dll along with a code snippet of its use in vbscript or python would do, too ?
Upvotes: 0
Views: 1178
Reputation: 25220
You can use pocketsphinx for that:
pocketsphinx_continuous -hmm hmm_dir -lm lm_dir -dict dict_dir -infile file.wav
pocketsphinx is trainable on Italian and can be used on Windows.
Upvotes: 1