So, as I was expirementing with pocketsphinx_continuous with keyword spotting when I came accross this behavior :
I have a keyword "HELLO WORLD" and as soon as the program starts I keep repeating it with the same tonality, distance from mic, pronunciation etc... . The keyword gets successfully recognized 4-6 times in series (of course I am waiting INFO: READY to start speaking). However, after lets say the sixth time it doesn't get recognized. And then again the seventh time it gets recongized again.
So, I exported the raw sound file and I feeded it with -infile again into the program to eliminate false recognition. The result is that it gets successfully recognized when I pass the audio file of my mic recording.
Can you confirm the same behavior?
I am using the default parameters of pocketsphinx_prealpha with -keyphrase and default threshold
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Good morning,
So, as I was expirementing with pocketsphinx_continuous with keyword spotting when I came accross this behavior :
I have a keyword "HELLO WORLD" and as soon as the program starts I keep repeating it with the same tonality, distance from mic, pronunciation etc... . The keyword gets successfully recognized 4-6 times in series (of course I am waiting INFO: READY to start speaking). However, after lets say the sixth time it doesn't get recognized. And then again the seventh time it gets recongized again.
So, I exported the raw sound file and I feeded it with -infile again into the program to eliminate false recognition. The result is that it gets successfully recognized when I pass the audio file of my mic recording.
Can you confirm the same behavior?
You need to provide an audio file