[Speex-dev] AEC on a TI C6x - has no effect

[Jean-Marc Valin]

> I have tried running the trivial case where the canceller is given two
> identical buffers, and it works every time.

So it would indicate the code itself performs the same every time.

> The only thing that I know of that can change slightly between runs is
> perhaps a minor change in the delay between the play and record buffers.
> I have verified, however, that the play buffer is always ahead of the
> rec buffer, so the echo canceller is always trying to cancel a signal
> that it has seen before.
> 
> Can slight differences in this delay cause this problem? What could
> cause the PHI and W arrays (sorry I keep calling them that, I don't know
> what their meaning is) to be zeroes only?

I don't know, but the only way to know would be to save the data you're
sending to the AEC. That way you'll be able to see what's going on.
There's no way to do anything when the data is different every time.
Something tells me that the problem is somewhere in your DAC/ADC
interface code.

	Jean-Marc