On Mon, Sep 1, 2008 at 3:36 PM, Victor Librado <vlibrado at bioingenieria.es>wrote:
>
> /
> / Do you not hear any difference at all compared with when the ES is
> disabled
> (with --ec-tail 0)? With the current algorithm, the ES does not
> remove the
> echo completely but rather bring the level down so that it's not too
> annoying. So I could understand that some low audible echo is still
> going
> through.
>
> Indeed, when echo is low (speaker and microphone far one from another)it
> gets lower and suppression seems to work,
> but in a realistic handsfree mode (mic next to speaker, sound in the mic
> coming from the speaker much higher that
> what a person "talks" to mic) echo persists and makes full-duplex an
> annoying experience. Can get better results anyway? ("suppress" more
> the output part from the input changing any parameter?).
>
>
In theory that shouldn't be necessary. The ES determines the suppression
level to be applied to the input during the learning process, so it should
be able to reduce the echo to minimum regardless of the mic level. Of course
this would break if there mic level changes between learning and
conversation, but to fix that, the ES also continuously monitor the residue
level (the leftover signal after ES is applied) and will force soft-reset if
the residue is too high. This happens every 30 seconds or so.
If you want to experiment with suppressing more, you can just change the
factor in echo_supp_cancel_echo() function. I think the code there is self
explanatory.
> I will try to go back to previous echo suppresor and watch results:
>
> pjmedia_silence_det_set_params(ec->sd, 100, 500, 3000);
>
> This parameters should give best results?
>
>
I forgot, and I don't want to support the old ES either. :)
>
> >/ Cpu usage goes to 15% approx (so ok, compared to my test with AEC --of
> />/ course, no way with AEC--).
> /
>
> Yes. 7% usage.
>
>
Is this with a call, or just the ES alone?
Cheers
Benny
>
> Best regards,
>
> Victor Librado Sancho
> Departamento I+D
> ------------------------------------------------------------------------
>
> _______________________________________________
> Visit our blog: http://blog.pjsip.org
>
> pjsip mailing list
> pjsip at lists.pjsip.org
> http://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.pjsip.org/pipermail/pjsip_lists.pjsip.org/attachments/20080902/d1d65f29/attachment.html
