We are using OpenVidu Enterprise along with mediasoup media server and we have the enableSimulcastExperimental setting turned on. We are experiencing a problem where the highest quality simulcasted video is not being used by the subscribers.
For example, when we have the resolution set to 640x480, the subscribers are using the 320x240 video from the publisher even though there is plenty of bandwidth. When viewing the WebRTC statistics for the session, there are RTCOutboundRTPVideoStream statistics for three layers. I can view the WebRTC RTCOutboundRTPVideoStream statistics for the 320x240 and 160x120 layers, but the other layer has no frameWidth and frameHeight and all of the other statistics for that layer have a value of zero. As a result, the 320x240 layer is the best available layer and that is what gets used by the subscribers. Obviously there is something wrong with the 640x480 layer.
When setting the resolution to 320x240, two of the layers appear to be invalid and the lowest one (80x60) is the only valid one so this results in horrible video.
When I turn off the enableSimulcastExperimental setting, everything works fine. There is only one layer sent out and that gets used by all of the subscribers without any downgrading of the video.