🎯 Goal

The goal of this PR is to improve how PiP participant is determined and also improve the video rotation for me participant.

🛠 Implementation details

1. PiP participant rules

• If we have two participants only, always take the not-me one.
• If we have a list of activeSpeakers find the first one who is not me
• If we have a dominant speaker who is not me show it
• If we currently have no other option, show me

2. VideoConfig changes

• The VideoRenderer now accepts a VideoRendererConfig which apart from the scale type and the fallback content also defines a parameter called mirrorStream if yes, the stream will be mirrored for a me participant in case the front camera is selected.