#466 startRecognition loads dictionaries every time

[Stefan Reich]

Hi!

I would like to make the recognition faster. When I call startRecognition()/stopRecognition() multiple times, each time it appears to load certain dictionaries:

14:19:39.612 INFO dictionary Loading dictionary from: jar:file:/home/stefan/.tinybrain/snippet-cache/data_1004793.jar!/edu/cmu/sphinx/models/en-us/cmudict-en-us.dict
14:19:39.948 INFO dictionary Loading filler dictionary from: jar:file:/home/stefan/.tinybrain/snippet-cache/data_1004793.jar!/edu/cmu/sphinx/models/en-us/en-us/noisedict
14:19:39.949 INFO trieNgramModel Loading n-gram language model from: jar:file:/home/stefan/.tinybrain/snippet-cache/data_1004793.jar!/edu/cmu/sphinx/models/en-us/en-us.lm.bin

Could this be optimized (keep stuff in memory) so each recognition runs faster?

Example program

[Stefan Reich]

Ah. Fixed it!

http://tinybrain.de/1006994

I am now down to <5 seconds after warmup for recognizing a short word ("hello"). Not realtime, but getting there