Re: [rtcweb] Proposed Video Selection Process

[Justin Uberti <[e-mail]>]

iOS doesn't expose APIs for said hardware decoding, so you're doing
software no matter what, and VP8 and H.264 are therefore on level ground.

That said, I think the general understanding here is that this is no longer
a technical decision.


On Thu, Nov 21, 2013 at 12:28 PM, Stefan Slivinski
<[e-mail]>wrote:

> So this is my concern.  On iOS for example it's pretty easy do decode
> large resolutions in software only using either h.264 or vp8 (and for that
> matter vp9 and h.265).  So making both h264 and vp8 decoders mandatory
> isn't a big technical problem.  The encoder however is a bigger issue.  A
> reasonably good implementation requires roughly 3x the compute to do this
> in software.  iOS however supports hardware encoding using H264 so there is
> an obvious advantage to using that specific codec on that device.  The same
> might be true for VP8 on other architectures.
>
> So requiring all webrtc implementations to support H264 and VP8 gives the
> flexibility to pick the most appropriate encoder for a given architecture.
>
>
> ----- Original Message -----
> From: Basil Mohamed Gohar [mailto:[e-mail]]
> Sent: Thursday, November 21, 2013 02:15 PM
> To: [e-mail] <[e-mail]>
> Subject: Re: [rtcweb] Proposed Video Selection Process
>
> On 11/21/2013 03:02 PM, Stefan Slivinski wrote:
> > I in no way intended to suggest  a specific implementation of a video
> codec.  My question was around whether we are voting on requiring decoders
> (my assumption) or both encoders and decoders
>
> The discussions, up to this point, have been about choosing an MTI video
> format so that rtcweb endpoints can communicate with one-another.  I am
> not sure how this goal can be achieved without having an encoder and a
> decoder on both ends, sharing at least one common format, whether it is
> H.261, Theora, H.264, VP8, or something else.
>
> --
> Libre Video
> http://librevideo.org
> _______________________________________________
> rtcweb mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/rtcweb
> _______________________________________________
> rtcweb mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/rtcweb
>