Voice recognition on embedded hardware.

unholy9

Just wondering if anyone knows how plausible it would be to implement voice recognition system in hardware (eg a PIC).

My experience so far is purely software, and the applications are quite large. Has anyone heard of such ventures in the past?

lateef

: Just wondering if anyone knows how plausible it would be to implement voice recognition system in hardware (eg a PIC).
:
: My experience so far is purely software, and the applications are quite large. Has anyone heard of such ventures in the past?
:

It's a very possible idea. I've done some speech (not voice )recognition on Windows platform with not to many lines of code. I also intend to implement it in assembly (for an embedded OS). It depends on what u want to lose (accuracy, vocabulary size, etc) and most importantly on how efficiently you can translate your formulas into your target's machine code.

merlin_bar

: : Just wondering if anyone knows how plausible it would be to implement voice recognition system in hardware (eg a PIC).
: :
: : My experience so far is purely software, and the applications are quite large. Has anyone heard of such ventures in the past?
: :
:
: It's a very possible idea. I've done some speech (not voice )recognition on Windows platform with not to many lines of code. I also intend to implement it in assembly (for an embedded OS). It depends on what u want to lose (accuracy, vocabulary size, etc) and most importantly on how efficiently you can translate your formulas into your target's machine code.
:
:
Thanks for your reply. When I made the post I was not aware of the distinction between the two recognition types. I am interested in speech recognition with a vocabulary of less than 100 utterances.

Could you reccomend any reading material or online resources?

Regards,

merlin

vamo

: : : Just wondering if anyone knows how plausible it would be to implement voice recognition system in hardware (eg a PIC).
: : :
: : : My experience so far is purely software, and the applications are quite large. Has anyone heard of such ventures in the past?
: : :
: :
: : It's a very possible idea. I've done some speech (not voice )recognition on Windows platform with not to many lines of code. I also intend to implement it in assembly (for an embedded OS). It depends on what u want to lose (accuracy, vocabulary size, etc) and most importantly on how efficiently you can translate your formulas into your target's machine code.
: :
: :
: Thanks for your reply. When I made the post I was not aware of the distinction between the two recognition types. I am interested in speech recognition with a vocabulary of less than 100 utterances.
:
: Could you reccomend any reading material or online resources?
:
: Regards,
:
: merlin
:
:

Which speech recognition algorithm do you want to use? As I know common reliable methods need a lot of system resources and in this case PIC isn't good choice.

unholy9

: : : : Just wondering if anyone knows how plausible it would be to implement voice recognition system in hardware (eg a PIC).
: : : :
: : : : My experience so far is purely software, and the applications are quite large. Has anyone heard of such ventures in the past?
: : : :
: : :
: : : It's a very possible idea. I've done some speech (not voice )recognition on Windows platform with not to many lines of code. I also intend to implement it in assembly (for an embedded OS). It depends on what u want to lose (accuracy, vocabulary size, etc) and most importantly on how efficiently you can translate your formulas into your target's machine code.
: : :
: : :
: : Thanks for your reply. When I made the post I was not aware of the distinction between the two recognition types. I am interested in speech recognition with a vocabulary of less than 100 utterances.
: :
: : Could you reccomend any reading material or online resources?
: :
: : Regards,
: :
: : merlin
: :
: :
:
: Which speech recognition algorithm do you want to use? As I know common reliable methods need a lot of system resources and in this case PIC isn't good choice.
:


I have not yet chosen an algorithm. My knowledge of DSP is pretty basic. But I think the level of reliablility I want clashes with the low cost that I also want.

grupoid

I was reading about these on last few days, and I find a project based on HC05, with no ADC: it's a project called TinyVoice and I still can't understand how it works.

http://www.circuitcellar.com/pastissues/articles/Stewart91/Stewart-91.pdf

Does any of you have idea of the algo it's using?

Thanks in advance!

Enzo