Hello!
I'm attempting to implement Speex AEC support in GStreamer, using work
started
by Olivier Cr?te. In the Speex manual, I see this text:
"Using a different soundcard to do the capture and plaback will *not* work,
regardless of what you may think. The only exception to that is if the two
cards can be made to have their sampling clock ``locked'' on the same
clock
source."
This is of course a major issue, because people may have webcams with embedded
mics, and cancellation really ought to work with their separate speakers. Is
this "same card" limitation a property of the Speex AEC only or AECs
in
general? If the latter, then I'd be curious how Skype does it, as I've
used
a USB mic and onboard output together with Skype successfully.
In any case, what about using separate cards is a problem? Is it that the two
cards might process the same sample rate at slightly different speeds? For
example, 16000 Hz on one card ends up being 16001 Hz on another card?
I'm using GStreamer, and it forces the same clock to be used for all
simultaneous operations. This means that if you capture and playback
simultaneously, the whole process will be synced to the clock of one of those
devices, and clock skew compensation will be applied to the other. So if you
have a 16000 Hz input and a 16001 Hz output, and the input clock is the
master, then 1 Hz of output will be thrown away periodically. You can also
use a software clock, in which case compensation may need to be applied to
the handling of both devices. To be clear, all of the compensation occurs in
software; there is no literal tuning of a hardware clock.
Is the above GStreamer behavior enough to satisfy what the manual states as
having the sampling clocks of the cards locked to the same clock source?
Thanks,
-Justin
