[jitsi-dev] The Videobridge audio and video


#1

I am pretty sure on how the translator handles video by sending video (or last n) to each participant on the ssrc it arrives at apart from the participant that sends the video.

I remain, however, a bit confused about audio. I thought the audio went through some form of mixing process so that the audio is sent as a single stream to each participant (with one ssrc). Initially I thought the getAudioMixer() method on content did something about this, but clearly is it something related to the rtptranslator.

Can anyone tell me please a bit about how the rtptranslator handles audio and the mixing of audio?


#2

Hi John,

I am pretty sure on how the translator handles video by sending video
(or last n) to each participant on the ssrc it arrives at apart from the
participant that sends the video.

I remain, however, a bit confused about audio. I thought the audio went
through some form of mixing process so that the audio is sent as a
single stream to each participant (with one ssrc). Initially I thought
the getAudioMixer() method on content did something about this, but
clearly is it something related to the rtptranslator.

Can anyone tell me please a bit about how the rtptranslator handles
audio and the mixing of audio?

We support two modes for audio (depending on the "rtp-level-relay-type" of the Content). The default is "translator" mode in which audio is handled just like video (and also uses the RTPTranslator class) -- all the streams are forwarded.

In the "mixer" mode, the audio is mixed. In this case the Content does not use an RTPTranslator, but uses an AudioMixerMediaDevice. This mode has to be enabled specifically (by setting "rtp-level-relay-type='mixer'") and isn't currently supported in jitsi-meet/jicofo.

Hope that makes things clearer.

Regards,
Boris

ยทยทยท

On 23/06/2017 09:58, John Hemming wrote: