Re: [Ccrtp-devel] ccRTP audio mixer

[David Sugar]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

The mixing algorithm is much simpler if you have or convert the packets
to linear samples; you can then directly sum the samples of your
incoming packets from all the sessions.  You should also mix each
session separately so that it includes only the sum of packets received
from the other sessions it is joined to and not it's own, so you do not
carry back echo.  Of course if your doing multicast, then usually the
mixing is distributed to the endpoints themselves.

Other issues that make conferencing more interesting and complex include
detecting packet energy level to determine who are the "loudest" or
"active" speakers at the moment, so that you only sum the first (few)
speakers rather than all sessions.  GNU ccAudio2 includes both codec
functions to convert packet frames to/from linear, and functions to
measure energy level.  Overall it would be very nice to see a generic
conference server written with ccrtp.

Dan Weber wrote:
|
| Simple.  Since ccrtp uses packet queues, appends all the rtp data to the
| same queue.
|
| Dan
|
|
| On Wed, 2 Mar 2005 address@hidden wrote:
|
|> I'm writing on my master thesis and looking for a rtp mixer. I found some
|> hints that it may easily possible to implement with ccRTP.
|> But I'm not sure! So has anybody experience how to implement an audio rtp
|> mixer?
|>
|> maik
|>
|>
|>
|>
|> _______________________________________________
|> Ccrtp-devel mailing list
|> address@hidden
|> http://lists.gnu.org/mailman/listinfo/ccrtp-devel
|>
|
|
| _______________________________________________
| Ccrtp-devel mailing list
| address@hidden
| http://lists.gnu.org/mailman/listinfo/ccrtp-devel
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.6 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org

iD8DBQFCJdJcJlWtSTiTVsYRAkfwAJ9eJ6m0sv7KqPUaRjwJhzX53mijCgCePeZl
QYRgFIlSrfAz/N78u398NEE=
=v2nW
-----END PGP SIGNATURE-----

dyfet.vcf
Description: Vcard