https://w3c.github.io/webrtc-encoded-transform/#sframe-transform-algorithm may trigger lots of error events

[youennf]

https://w3c.github.io/webrtc-encoded-transform/#sframe-transform-algorithm states that, for each packet, an error event might be fired, say in case of missing key or failed decryption.

This might be way too many events.
We might want to limit the number of events somehow.

[k-wasniowski]

Since as stated with avtcore, we want at first to pre-buffer all packets which will be used to create a frame, maybe instead of triggering an event for each packet, we could trigger an event once for all of the packets correlated to single frame.
Since if we will fail to decrypt one of packets, there is no point in running decryption of other packets of this frame, and we should drop all the packets after first failure.
So in fact, even though we will fail to decrypt single packet, we will drop all packets related to specific frame. In such casem we will end up in the same situation as for per-frame, which also can trigger an event once-per-frame.
The question is, if we should expose packet which failed to decode in 'frame' field, or simply keep it null for per-packet scenario.