Hello.
I have the following situation. AEC is used in network chat software
over DirectSound API. Echo and reference signals are almost aligned
(delay is no more than 30ms). When echo is emulated in notebook
(built-in speakers + mic) everything goes fine and echo is cancelled.
But when configuration includes stand-alone speakers and mic no echo is
removed. Audio is in 22050 hz at 16 bit mono format, number of samples
to process were 441 and tail was used 4096.
I've decided to investigate further and took dumps of signals. Buffers
are aligned, as was said previously, but running testecho.exe doesn't
produce any noticable echo cancellation. I want to accent on the fact,
that the same parameters on notebook echo-loop config worked fine. Now
I'm pretty tired with playing with parameters and buffers and I start
thinking that the reason may lay in somewhat audio spectrum or other
physical/math parameters of samples itself.
Please, answer whether it is possible or not. I can provide recorded
dumps so you can test it yourself. This is very crucial situation for
our software and now we're stuck with it.
Paul Novodon
Sr. software multimedia developer
TomskSoft LLC.
