Currently I'm working on a single word recognition like digitis from 1 till 10 and and always get a hypothosis even if you say a none different word that doesn't even sound the same.
The example with 'oh mighty computer' works more what I'm looking for, it doesn't respond at al when talking gibberish.
Can this behaviour been achieved when recognizing single words with ps_process_raw() ?
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
I'm not sure what you define by "gibberish". Our keyword spotting recognizes words which are properly pronounced and rejects everything else. If yo want to recognize some other sounds, you can define them as words with their own pronunciation.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Maybe the keyword spotting is something for me, I havent thought about that.
I always thought this was some feature for only 1 word, to activate a sleeping sphinx system.
Just to be clear, you mean using ps_set_keyphrase ?
And can this be used with ps_process_raw ?
PS: Currently I'm using the Grammar approuch for single words pre-defined as .raw audio samples.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
You can recognize several keywords at once, you can write the list file as described in tutorial http://cmusphinx.sourceforge.net/wiki/tutoriallm and then enable it with ps_set_kws which accepts the list file. Later you can process audio with ps_process_raw
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Currently I'm working on a single word recognition like digitis from 1 till 10 and and always get a hypothosis even if you say a none different word that doesn't even sound the same.
The example with 'oh mighty computer' works more what I'm looking for, it doesn't respond at al when talking gibberish.
Can this behaviour been achieved when recognizing single words with ps_process_raw() ?
I'm not sure what you define by "gibberish". Our keyword spotting recognizes words which are properly pronounced and rejects everything else. If yo want to recognize some other sounds, you can define them as words with their own pronunciation.
Nickolay thanks for the info.
Maybe the keyword spotting is something for me, I havent thought about that.
I always thought this was some feature for only 1 word, to activate a sleeping sphinx system.
Just to be clear, you mean using ps_set_keyphrase ?
And can this be used with ps_process_raw ?
PS: Currently I'm using the Grammar approuch for single words pre-defined as .raw audio samples.
You can recognize several keywords at once, you can write the list file as described in tutorial http://cmusphinx.sourceforge.net/wiki/tutoriallm and then enable it with
ps_set_kwswhich accepts the list file. Later you can process audio withps_process_rawWouw, thats great.
Im gona try it as soon as possible!
Thanks