Hi Guys,
I just wanted to inquire on the feasability of implementing a voice recognition system ( to be able to detect basic speech commands, speaker independant) on an FPGA or DSP ( basically without the use of a computer) . The idea is to be able to use such speech to control motors for various applications.

If anyone has experience with such a field, I would appreciate your comments.

Thanks

P.s: I also wanted to know whether or not it would require the user to "train" the DSP or FPGA.

 

Might be better off with a PIC than an FPGA. We did it with a TRS-80 back in the '70s. I suspect any old modern PIC would run circles around the TRS-80.

Training was required back then. We had to train for each and every different human, and had problems if someone came down with a cold.

 

There are simple and easy to use speech recognition ICs available. These are quite good with excellent rate of detection but usually are limited to ~ 20 to 40 words (commands) or so. They are not that expensive, less than $20 each depending on model, and not difficult to link to a microcontroller system. Search for HM2007, Sensory, OKI or Philips.

 

There are simple and easy to use speech recognition ICs available. These are quite good with excellent rate of detection but usually are limited to ~ 20 to 40 words (commands) or so. They are not that expensive, less than $20 each depending on model, and not difficult to link to a microcontroller system. Search for HM2007, Sensory, OKI or Philips.

is hm2007 still in store?