MCU audio mixing

[Bob Fitzgerald]

Hello!

This question is to verify what I think I know. 

For our hypothetical conference, we have 10 participants (a 6000MXP, a SX20, 7 C40s and an audio-only participant)

The 10 participants call in.  Their audio is mixed together and sent out to each participant minus their own audio.

Here is the question;

For each participant the balance of audio levels is exactly the same, with the exception that their audio is removed from the mix, right?

For example, both the 6000MXP and the SX20 receive the same relative volume of the C40s and audio-only participant audio. There's nothing in the MCU audio logic that would increase or decrease the audio-only participant's volume differently for the 6000MXP than the SX20, right?

Thanks!

[Saurabh Gupta]

Thats Correct Bob...

Whenever the Codec Sends out the RTP stream to the MCU - It is combined in one and then sent out

The MCU takes the RTP Stream from each participant and then sends out to all the participants depending upon thier capabilities and limitations.

Thanks,

Saurabh

[sagsheth]

Hi Bob,

When you say increase or decrease the volume of audio only pariticipant, do you mean audio coming from or going out?

I mean you can always apply audio gain.

Regards,

Sagar

[Bob Fitzgerald]

Hi Sagar,

I know the gain can be raised or lowered manually per participant.  What I am referring to is the realtive audio levels when being mixed before being sent back out to the participants.

Using my example from above I'm going to simplify the volume values.  The 7 C40s' audio comes into the MCU at a level of 6.  The audio-only participant's audio comes in at a level of 5.  No AGC or Fixed Gain are applied.  When the audio is mixed together, both the 6000MXP and the SX20 will be sent the same relative volume levels of the C40s and the audio-only participant.  In other words, the following will NOT happen;

The 6000MXP hears the 7 C40s at level 6 and the audio-only participant at level 5

The SX20 hears the 7 C40s at level 6 and the audio-only participant at level 2

I am correct in thinking this, yes?  There isn't anything in the MCU that would change the relative volume levels on a per participant basis, right?

Thanks!

[Bob Fitzgerald]

And a follow up question:  Are the incoming audio packets just being redistributed unaltered to the other participants, or is each outgoing audio stream a collection of completely new audio packets with combined audio from the contributing participants?

Using my example conference: The audio packets from the SX20 go directly to the 6000MXP, C40s and audio-only participant, yes? 

Or, are the audio packets going to each participant newly created by the MCU as a combined audio source?

Thanks!

[sagsheth · Accepted solution]

Ok i got it now...

Yes, the relative audio level remains unchanged.

Now coming to your second question,

MCU may or may not transcode the audio packets. Yes, each outgoing audio stream is collection of completely new audio packets, now the codec may remain same or its gets transcode depending upon SDP/Capsets negotiation.

Each participant (be it audio or video), individual resoureces are assigned, so each participant will have new outgoing audio stream being MCU as source.

Hope this helps

Regards,

Sagar

[Bob Fitzgerald]

Thank you Sagar, that does help.