For example, while running the digits demo (which will only recognize spoken numbers) I speak "sunny spells on wednesday" and PocketSphinx recognized "two six oh six oh one one zero". Not even close. I would have expected PocketSphinx to recognize nothing until it hears a word that's included in the grammar.
I'd imagine that there'd be a way to configure PocketSphinx to ignore words that aren't in the grammar, or meet some minimum threshold. Can this be done? How? Thanks!
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Pocketsphinx supports keyword spotting mode where you can specify the keyword list to look for. The advantage of this mode is that you can specify a threshold for each keyword so that keyword can be detected in continuous speech. All other modes will try to detect the words from grammar even if you used words which are not in grammar.
hmmm, this seems to be giving me some better results. The thresholds seem to really affect the detected words, so I'll see if adjusting them helps. Thanks!
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
I'm running the Android PocketSphinx demo and when I run the digits recognition part using the limited grammar it "recognizes" words that are part of the grammar but nowhere close to what I spoke.
For example, while running the digits demo (which will only recognize spoken numbers) I speak "sunny spells on wednesday" and PocketSphinx recognized "two six oh six oh one one zero". Not even close. I would have expected PocketSphinx to recognize nothing until it hears a word that's included in the grammar.
I'd imagine that there'd be a way to configure PocketSphinx to ignore words that aren't in the grammar, or meet some minimum threshold. Can this be done? How? Thanks!
From tutorial:
Pocketsphinx supports keyword spotting mode where you can specify the keyword list to look for. The advantage of this mode is that you can specify a threshold for each keyword so that keyword can be detected in continuous speech. All other modes will try to detect the words from grammar even if you used words which are not in grammar.
http://cmusphinx.sourceforge.net/wiki/tutoriallm
what can be done in continuous mode?
what can be done in continuous mode?
what can be done in continuous mode?
what can be done in continuous mode?
hmmm, this seems to be giving me some better results. The thresholds seem to really affect the detected words, so I'll see if adjusting them helps. Thanks!
@Micheal Nistler, How you adjusted threshod for key phrases?