Forums

DSP for simple voice recognition?

Started by Christopher Cole September 19, 2006
What would be a good (&cheap) DSP choice for implementing simple voice 
recognition?  I would like to be able to train about 10-20 words, and hold
the frequency components in on-chip ram or flash.  I would like to develop 
this using a Linux development environment...

I have looked at the dsPIC and the Blackfin, as there exists a GCC toolchain
for each of those.  Any other ideas?

Thanks for any ideas,
-Chris

-- 
| Christopher Cole, Cole Design and Development               cole@coledd.com |
| Embedded Electronics and Software Design                  http://coledd.com |

Christopher Cole wrote:

> What would be a good (&cheap) DSP choice for implementing simple voice > recognition?
There is a number of the low cost single chip solutions already. Look here, for example: http://www.sensoryinc.com/ Vladimir Vassilevsky DSP and Mixed Signal Design Consultant http://www.abvolt.com
We have a C++ code of ISIP Univ that works for Windows CE, porting to
linux won't be much problem, as it basically is Unix

tim
www.research-lab.com


Christopher Cole wrote:
> What would be a good (&cheap) DSP choice for implementing simple voice > recognition? I would like to be able to train about 10-20 words, and hold > the frequency components in on-chip ram or flash. I would like to develop > this using a Linux development environment... > > I have looked at the dsPIC and the Blackfin, as there exists a GCC toolchain > for each of those. Any other ideas? > > Thanks for any ideas, > -Chris > > -- > | Christopher Cole, Cole Design and Development cole@coledd.com | > | Embedded Electronics and Software Design http://coledd.com |