On Fri, 7 Mar 2014, Vittorio Giovara wrote:
From: Andrey Utkin <[email protected]>
---
Maybe worth a changelog entry?
Vittorio
Yes, but before it goes in a bunch of questions do need to be answered:
- Was this tested against some actual separate existing implementation -
if yes, which one?
- Which RFC/spec does this conform to? (Namely, I see one spec in RFC
3551, which declares the static payload type 3, which is mapped in
libavformat/rtp.c, currently to AV_CODEC_NONE, and another spec in RFC
5993 which seems to be slightly different. If this is the RFC 3551 thing,
I would guess we would achieve the same by just setting the right codec id
in the rtp_payload_types table in rtp.c.)
- Are we sure that the bitstream formatting actually is the same as used
in our GSM decoder? This would require testing against one known-compliant
peer. GSM is IIRC one of the formats with a quite loose bitstream format
spec leading to multiple over-the-wire formats (e.g. the libgsm "toast"
format vs MS GSM)
Until these are answered (and at the very least an example is shown where
it does help interoperability with some other implementation) this
otherwise just risks adding incompatibility by pretending to handle a
format.
// Martin
_______________________________________________
libav-devel mailing list
[email protected]
https://lists.libav.org/mailman/listinfo/libav-devel
