bluray and hddvd playback

[chrisbirkinshaw]

by the MD's CPU. The core merely serves the files.

A gigabit LAN could probably let the Core serve cheap MDs with decoded streams, maybe encoded with something really lightweight like HuffYUV (~60% as big). ffmpeg and MPlayer support HuffYUV, even a 2:1 YV12 version. The Core would have to be faster to accommodate. There's got to be a way to harness multiple videocards on the core without just using VNC, because that has no sound and has extra overhead for all the other VNC features - and the VNC client isn't really that lightweight.

This doesn't make a lot of sense to me. Why would you want to put all the graphics hardware in the core? You can get a good selection of fanless graphics cards so I have no problem having one in my living room.

BTW there is no way to decode H.264 in linux using current display hardware.

Even a massively beefy MD cannot decode H.264 1080 content in realtime. Asking the core to do this for multiple MDs is not really going to be possible.

Chris

[tschak909]

actually, if you follow the ffmpeg lists, the issue deals a lot with the fact that x264 isn't all that optimized at all. None of it is multithreaded. This becomes especially apparent if you use CoreAVC on windows (which in its standard form does not rely on any GPU accelleration etc) which can decode a 1080p H.264 stream on a 2Ghz Core 2 Duo comfortably.

There have been a lot of commits funneling into the ffmpeg stack on this issue starting in late 2006, and there still is a lot of work left to be done.

-Thom

[1audio]

H.264 playback in windows requires some serious hardware as well (or an Nvidia card and matching drivers). On my T7100 w/ Intel G965 I see 80m- 90% CPU on an H.264 HD file.

Apparently the H.264 HD files can be played with a BE2400 processor on an Nvidia card.

I had a long talk with an H.264 specialist about video processing. He pointed out that its all about multiplying bits. Each stage multiplys the bits a lot. and once its handed off to the video card its gone in terms of control. We were discussing H.264 HD playback on the Via platform. They had done a lot of work on it but confirmed that the existing hardware cannot do it. The platform needs hardware acceleration to do it.

[chrisbirkinshaw]

Does anyone know the status of the work to get multithreaded h.264 decoding? I can find posts from 2006 which state it was added to ffmpeg but only to decode slices in parallel. This means it will not work for a recording which was not encoded with slices. The more useful way (and difficult way) is to decode frames in parallel, rather than decoding slices of one frame in parallel, however I can't see any indication of when this might be done. Apparently the x264 encoder has been changed from encoding slices in parallel to encoding frames in parallel and not using slices at all. It would seem this is the way the ffmpeg h.264 decoder needs to go. This also of course means that streams encoded with x264 will not make use of the current ffmpeg multithreaded decoding.

Until this is sorted out HDTV h.264 is really not feasible on most computers available at the moment. The drive towards lower processor speeds and quad cores is not going to help this situation in the future. And we shouldn't hold out for any help from nvidia or ati in getting GPU acceleration.

Presumably some of the integrators (Andrew?!) are following this sort of development carefully. Any ideas?

Chris

[chrisbirkinshaw]

I suppose there's always coreavc...

http://code.google.com/p/coreavc-for-linux/