Audiomixer with CSRCs

Hi experts :slight_smile: !

I’ve implemented the following pipeline:

udpsrc → rtpbin → rtpg722depay → avdec_g722 → volume → avenc_g722 → rtpg722pay → udpsink

payloader and depayloader have both the property “source-info” set to true.

In this scenario the pipeline works as expected. I send a test RTP stream to the udpsrc and it is sent out by udpsink. In the wireshark trace I can see that the outbound RTP packets have the CSRC header set to the SSRC value of the inbound stgream.

Now if I extend the pipeline by an audiomixer as follows the stream is still forwarded but the CSRC headers are not present anymore in the outgoing packets:

udpsrc → rtpbin → rtpg722depay → avdec_g722 → volume → audiomixer → avenc_g722 → rtpg722pay → udpsink

So obviously the source-info gets lost in the audiomixer(?).

Is there a way to fix it?

Every hint would be greatly appreciated!