So do you expect your implementations on devices with hardware acceleration to
have any limits on resolution of images they can decode? I can't imagine how I
could implement the frame buffers in VP8 in VLSI without having an upper limit
on both the width and height of the image. How do you deal with that?
I don't know if you ever got the Google VHDL code for VP8 but I have never got
it so I don't know what it does but if you do, that would be great.
On Jul 24, 2013, at 12:57 PM, Timothy B. Terriberry
<tterribe(_at_)xiph(_dot_)org> wrote:
Cullen Jennings (fluffy) wrote:
There is one thing that as far as I can tell that all the implementors agree
on. None of the implications control the resolution using
m=video 62537 RTP/SAVPF 96
a=rtpmap:96 VP8/90000
a=fmtp:96 max-fr=30;max-fs=3600
What resolution do you think "max-fs=3600" is? I have no idea. It is not
possible to implement so it is not surprising no one is doing it. However,
this draft-ietf-payload-vp8 draft says the resolution is specified using the
max-fs and max-fr.
I can't speak for other implementations, but here at Mozilla, our
interpretation of RFC 6236 was that the values specified in imageattr can be
completely ignored by either side, if they so choose. That leaves max-fs and
max-fr as the only mechanism to indicate a resource constraint that cannot be
ignored, and we plan to use it as such.
See <https://bugzilla.mozilla.org/show_bug.cgi?id=881935> for details.
_______________________________________________
payload mailing list
payload(_at_)ietf(_dot_)org
https://www.ietf.org/mailman/listinfo/payload