Draft: rtp: Add MPEG-1 Audio RTP payloader/depayloader

Depayloader

Differences to legacy C depayloader:

• does parsing/checking of content
• outputs parsed data
• sets duration on output frames (important when saving into containers)

Payloader

• Has different aggregate-modes including an auto mode that defaults to zero-latency if upstream is live.
• Also handles the case of multiple frames in the same input buffer correctly (especially if the buffer exceeds the max packet payload size, in which case we want to process audio frames individually and not fragment those N frames over multiple packets).
• Sets marker flag more correctly (should be set at beginning of a talk spurt according to rfc2250 errata, not to signal that the payload ends with the end of an audio frame. (this is being fixed)