2010-04-13, 08:21 | Link #1 |
AS Oji-kun
Join Date: Nov 2006
Age: 74
|
Will VP8 become the new standard for fansubs?
It's beginning to look like Google will release the VP8 codec it bought from On2 in the next few weeks. If, as expected, VP8 will be released as an open standard, free of licensing issues, might we expect to see it displace the encumbered H.264 codec for fansubs?
__________________
|
2010-04-13, 11:23 | Link #4 |
x264 Developer
Join Date: Feb 2008
|
It's rather hard to use vaporware for fansubs.
Call me back when it exists, and even then, call me back when On2 releases something that isn't garbage. They never have and never will, probably because they spend all of their time bragging about how their next abortion will be X times better than H.264 instead of actually making it not suck. |
2010-04-15, 12:52 | Link #9 |
Spoilaphobic
Join Date: Jan 2009
Location: USA
Age: 37
|
Should it become a new standard, it wouldn't happen for a long time. One reason why h.264 is used so widely is because of its compatibility (PC, Mac, PS3, etc) and versatility (low res to high qual 1080p).
__________________
|
2010-05-19, 12:15 | Link #10 | |
Pioneer in Fansub 2.0
Join Date: Aug 2007
|
http://x264dev.multimedia.cx/?p=377
First indepth analysis of VP8. Summary: Quote:
__________________
|
|
2010-05-19, 12:53 | Link #11 |
King of Braves
Fansubber
Join Date: Dec 2003
Location: Toronto, ON
Age: 45
|
Who honestly cares if fansubs are 'patent-encumbered' ?? It's not like fansubs aren't skirting the law already. Fansubs will always go with the easiest-to-use/best-performing/l33test codec/container there is.
Semi-agree with Daiz I guess. Or will this become the new way to act smug? "MY fansubs are not encumbered by video encoding patents!" :\ |
2010-05-19, 19:46 | Link #14 |
ひきこもりアイドル
IT Support
Join Date: Feb 2009
Location: Pennsylvania , United States
Age: 34
|
Doubt it. Although VP8 is a open codec, it's a bit too young to be used for something productive. I would give it a few years before considering it as a viable alternative.
__________________
|
2010-05-20, 10:10 | Link #16 |
Translator, Producer
Join Date: Nov 2003
Location: Tokyo, Japan
Age: 44
|
VP8's lack of more than 3 reference frames would hurt its use for anime even more than for usual video, I think.
Plus it's not just 3 reference frames, it's more restrictive than that (from what I read in the spec anyway), where 1 of those 3 is ALWAYS the previous frame, and 1 is always the most recent I frame... The spec did not impress me at all. I wonder if it's not time for some enterprising genius to start developing a non-macroblock based video compression algorithm and really get away from mpeg once and for all. Shouldn't there be some reasonable way of describing a grid using n-gons and tessellations? combined with a final ray-trace to recover pixels? The you use some kind of tree structure as a scan algorithm to go through touching triangles. Triangle prediction modes could be similar to mpeg for intra (based on extending the pixels along the matching edge, etc), and motion prediction could be based off motion of the triangle boundary vertices through the time axis instead of motion vectors of the blocks themselves... The trick would be to construct it in a way that uses mainly integer arithmetic...
__________________
|
2010-05-20, 11:05 | Link #17 | ||
Far out, man!
Fansubber
Join Date: Jul 2008
Age: 40
|
Quote:
Quote:
|
||
2010-05-20, 11:52 | Link #18 | |
Translator, Producer
Join Date: Nov 2003
Location: Tokyo, Japan
Age: 44
|
Quote:
n-gons (dealt internally triangulated, I think) could theoretically give much better intra compression because they aren't limited by vertical and horizontal boundaries... Furthermore, with triangles you could have 2 out of the 3 side's worth of data to interpolate from compared to square block's 2/4 max. But encoding would be incredibly intensive. Rather than just a mode decision, you'd have to somehow find optimal tesselations...
__________________
|
|
2010-05-20, 12:33 | Link #19 | |
Far out, man!
Fansubber
Join Date: Jul 2008
Age: 40
|
Quote:
Additional note, you could compute the entropy of each frame, and based on that assign a number of prototypes to it, then run stuff like RLVQ on it to come up with midpoints for your n-gons. You would expend ridiculous amounts of computation time per frame though when encoding, but decoding should be pretty straightforward. |
|
2010-05-20, 13:18 | Link #20 | |
x264 Developer
Join Date: Feb 2008
|
Quote:
|
|
Tags |
codecs, google, h.264, mpegla, patents, vp8 |
|
|