And those modules have to be trained.
Not all of them.
For example, the Tigal board features, in addition to user trained recognition,
Speaker Independent Recognition (do not require user training) available in US English, UK English, German, French, Italian, LA Spanish, Japanese and Korean
.
I think i'm gonna try without filters and see how it works, may be the noise is not enought to be a problem. Otherwise i could try with a high pass and low pass filter to avoid at least some noises?
Excellent! It will be a great learning experience.