[Speex-dev] Echo cancellation on Windows -- success :)

[Thorvald Natvig]

I just added ASIO sampling to Mumble (http://mumble.sourceforge.net), so 
that it can sample both loudspeaker mix and microphone at the same time 
(and completely in sync). This gives the input data needed to perform 
successfull echo cancellation. Check out ASIOInput.h/.cpp and 
AudioInput.h/.cpp for how to implement; they are available directly from 
the CVSweb on sourceforge.

As there have been quite a few questions on how to do this, I thought I'd 
announce the availability here ;) Note that while this works, and works 
well, as long as you sample from multiple speakers the result won't be 
perfect, as the ammount of echo from each speaker is different, and I 
just sum them all together. On my 7.1 surround setup, I get about 90% 
reduction, which works well enough that other users don't complain about 
echo.