Difference in client CPU usage between meet.jitsi and self-host

Difference in client CPU usage between meet.jitsi and self-host

Hello. During the jitsi meet test, I have a lot of questions.
What many people say is that “when using Jitsi meet, the CPU usage of the client is high”.
I tried various methods such as interface to reduce CPU usage.
However, when I tried using meet.jitsi, not the self-hosting method, a fundamental question arose.

The CPU usage is different when using the self-hosting method and meet.jitsi.
Based on camera off status in 1 to 1 conference (PC-embedded board)
embedded board cpu usage
self-hosting: about 76%
meet.jitsi: about 35%

Looking at the above results, I think that self-hosting can be sufficiently reduced by 35%.
I wonder.
Is there a difference between the version provided by jitsi, or is it necessary to optimize the build file separately when applying self-hosting? (jitsi-videobridge or libjitsi etc…)

The difference is the versions, the version on meet.jit.si is almost the latest from unstable while stable is a little bit older.
The other thing is settings, sometimes there are settings not turned on by default but enabled on meet.jit.si.

Thank you~ Damencho
Can I just ask one more question?
Is it possible to self-host with almost the latest version of meet.jit.si?

Yes. As easy as follow https://jitsi.org/qi replacing https://download.jitsi.org stable/ with https://download.jitsi.org unstable/.
For example the current meet.jit.si versions are now in testing repo: https://download.jitsi.org testing/ so you can try those.