On2 Releases VP6 video codec

A reader writes:"On2 Technologies, the folks who brought you the open-source VP3 video codec (now managed by Theora.org), have released our latest codec, VP6. Highlights include hi-def support with no encoder restrictions, real-time encoding at full D1 resolution, and substantial performance & quality improvements over VP5. Best of all: no "patent pooling" restrictions or external licensing fees, a la MPEG-4."

NOT Free

[Dot.Com.CEO]

People, this codec is not free, it is not open source and I'm not even sure that it will be a free download. Even though VP3 was indeed open-sourced and is as free as can be, VP6 is not. It does not have MPEG4's licencing limitations but it is not open-source. It is, for all intents and purposes, as free as WMV and Real.

Re:Hmm

[PhiberOptix]

if theres demand for the codec, then interest will grow. Many subbed animes are already being released in Xvid format. Also, installing xvid is just as easy as installing divx or any other app in windows, given that you d/l a binary package

http://www.divx-digest.com/software/xvid.html

Doom9 codec tests

[fluor2]

I would really recommend reading this codec comparision by doom9:

http://www.doom9.org/index.html?/codecs-103-1.ht m

Doom9 shows us very nice screenshots from non-keyframes. The conclusion is still: XViD is the best codec around. I hope people understand that whatever other companies claim (e.g. "50% better compression") you should never underestimate what doom9 says: all codecs out there now are VERY similar. None of them is really outstanding.

Re:Doom9 codec tests

[Ishin]

Doom9 is a good site to get news on about what codecs and applications are new, or releasing new versions. It's also a good place for discussion with the developers of many open source applications and codecs/video containers. However, I'd recommend that if you read their video codec reviews, you keep several things in mind. Firstly, the review is very subjective, and though I agree with the conclusion (xvid does the best job in my experience) the reviewer isn't exactly doing a normal comparison. Secondly, they're not very open to criticism, constructive or otherwise, and seem to have the attitude that they are THE EXPERTS on video encoding and codecs, and are thus beyond reproach, although, from what I can tell, the site owners aren't actually actively involved in the development of any codecs or applications. The main two flaws of the review are: A. they use post processing in all of their comparison screenshots and the reviewer used post processed shots to determine which he thought looked best. That in itself pretty much invalidates the results, as the actual output of the codecs isn't the only thing being tested, but also the perceptual quality of their respective post-filtering schemes. B. he was using a special version of xvid that's not available to the public, and that many of the people involved in xvid didn't even realize existed, which, once again completely invalidates his results for the rest of us. So, like I said, good site, great news, even greater discussion, questionable codec review.

and what of H.264

[moogla]

The On guys compared VP6 to something called H.264, which I come to find out later is an official standard (newly released) for MPEG-4 encoding. It looked really damn good, almost as good as VP6 on the frames they chose.

I understand XviD's implementation of MPEG-4 is based on H.263.

So is anybody (including XviD) considering implementing it? I understand it isn't patent-encumbered. I could be wrong...

Install ffdshow and other video tools mentioned

[McQuaid]

Ffdshow [sourceforge.net] is a filter for most mpeg4 codecs. Works with divx 4,5, xvid and other mpeg4 implementations. If all you do is playback, no codec required. Also if use alpha xvid codecs and it doesn't playback properly with ffdshow, you know that your vid isn't mpeg4 compliant. Btw, I capture/encode shows all the time in windows and would like to do this in linux, but it seems really lacking. First off, avisynth [avisynth.org] is an indispensable tool for dealing with video. What first attracted me to it was the best ivtc plugin by Donald Graft [mordor.net]. This processes telecined sources back to their original film frame rate which I use on toons/film sources.

But the versatility goes way beyond that. Here's an animated menu [cultact-server.novi.dk] I made for batman tas for a vcd I was working on, which btw I authored with videopack 5 to include animated menus, galleries with audio and chapter selection (I love pimpin that :) ).

Also worth mentioning is Tmpgenc [tmpgenc.net], probably the best mpeg 1 encoder, which is free. And not to shabby mpeg2 encoding. Also of course is virtualdub [virtualdub.org], which has come in handy on many occasions.

So where are the comparable linux equivalents? I couldn't find them. I'd love to see a write up on video encoding on linux, maybe I'll do one myself.

Re:Doom9 codec tests

[real_smiff]

don't be silly.. neither of those things "completely invalidate the results". postprocessing - it's a real world test of how these codecs are going to be used by most people. "special version" - there are thousands of "special versions" of Xvid, every build by every person who releases builds is a special version... it's in development, that's what CVS is for. You can pick up a build of Xvid now and get results like he got there, there's no fraud like you seem to imply. Having said all that, i think Doom9 is probably biased towards Xvid, but for good reasons. It's forum is really the 2nd home of Xvid.

Re:and what of H.264

[pla]

I understand XviD's implementation of MPEG-4 is based on H.263.

XviD follows the MPEG-4 ASP (advanced simple profile) spec. Virtually all of the current major video codecs out there use some minor variant of this.

H.264 usually refers to the MPEG-4 AVC (advanced video coding) profile. This promises a 2-4x size improvement at similar quality to the ASP. However, it has one major problem...


So is anybody (including XviD) considering implementing it? I understand it isn't patent-encumbered. I could be wrong...

Yes, an AVC implementation exists [hydrogenaudio.org], but it provides its own demonstration of why no one uses it yet despite the improved size and/or quality... Namely, 30-45 seconds per frame at encode time. For a full-length movie, that comes out to two or three days for a single-pass encode.

Additionally, even if you feel inclined to wait that long for the sake of quality (personally, I would), the link I gave above points to more of a proof of concept than a "real" viable codec. It needs quite a lot of tweaking just to make it compare to existing ASP codecs such as XviD.

Re:Linux download? OS X download?

[Anonymous Coward]

Yep! Was very excited by their claims, immediately tried to download the player, which to my horror turns out to be only available for windows. Do we really need another windows only codec? Or is the idea that say, Apple will intedgrate this codec into quicktime?

At the moment though, if you wanna use this codec then you have to dance the dirty with Micro$oft, which is most certainally not free, or without restrictions. We use heaps of video on our sites, streaming, downloadable etc. and our primary concern is cross-compatibility.

So, nil point for ON2.

What about FFMPEG's MPEG4 Codec?

[engine matrix]

I did not see that listed in Doom9's comparison. It seems the MPlayer developers prefer it over Xvid. I've always used FFMPEG MPEG4 with fantastic results.

Re:Doom9 codec tests

[Ishin]

Sorry to misinterpret you, then.

I'm sorry to say that I've yet to find a really satisfactory and impartial comparison, much less, one that is up to date. I've had to do all of my own testing to figure out what settings do, and how codecs compare. Mainly xvid, divx4/5, sbc, and ffvfw. As I stated earlier, I've found xvid in it's latest incarnations(Koepi [goe.net] and Umaniac's [hopto.org] versions are easy to find, and work great, in my experience) are the best, and the doom9 xvid forum is a great place to give feedback to, and get information from, the creators of this robust and customizeable codec.

This [digitalinsanity.ath.cx] is probably the closest thing I could find to an impartial comparison, displaying unpostprocessed, and postprocessed images from many different codecs. Unfortunately, when I tried the link, it didn't respond, hopefully it will be back up.

Hope it helps.

Re:This ad paid for by On2

[Anonymous Coward]

Since there's no download, nor - if history serves as a guide - will there be a download to test, I think this is fairly irrelevant.
Since I've never come across On2-encoded video I suspect that this codec isn't worth the hassle.

This is likely to be VASTLY expensive.

[almaw]

Another poster pointed out that this codec isn't free, and is just as commercial and proprietry as Real and WMV. They're wrong - it's actually much worse than those in terms of cost and lock-in.

Now, I don't know what the terms are for VP6, but our company investigated implementing VP5 a while back. To encode in realtime, you needed dedicated hardware ($15,000 per license). To encode offline, you need software at $10,000 per license. This is licensed on a rolling annual basis - i.e. $10,000 a year. You then additionally need to license the TrueCast on-demand server to distribute content, which is similarly priced.

I'd expect VP6 to be similarly priced to VP5. You'd better be encoding an awful lot of video and saving an awful lot of bandwidth to make it worth your while.

OTOH, the quality of VP5 was extremely good for a given bandwidth (much better than xvid).

Re:Install ffdshow and other video tools mentioned

[Joe Tie.]

Ffdshow is an easy replacement, as it's based on a Linux project, ffmpeg [sourceforge.net].

The extent of my encoding comes from recording TV or VHS tapes and throwing a few filters at them for cleanup and encoding. So I've never had a chance to become familiar with what features avisynth has. Some people over at Doom9's Linux forum have mentioned getting Avisynth running in wine [doom9.org].

Tmpgenc should be a pretty easy replacement. Transcode [uni-goettingen.de] seems to be a favorite for mpeg2 encoding. There was a problem when I first started using Linux, which slowed mpeg encoding to a snails pace for a while, and I've never found a chance to get back to see how it works when fully functional. A lot of peoplel swear by it though, and it's feature list is pretty impressive. Avidemux, as well as quite a few other programs also provide mpeg1 encoding, I think through ffmpeg.

Everything in VirtualDub/VirtualDubMod except video capture should work fine under wine and that's what I used for quite a while before Avidemux [fixounet.free.fr] appeared. The design and functionality are very similar to virtualdub, and it also includes encoding to mpeg1.

One of the two best media players for Linux, Mplayer [mplayerhq.hu], should be noted to have better support for matroska [matroska.org], in cvs, than is available for the format under windows. And these [bunkus.org] tools can be used for creation and editing of matroska files.

For capture from a tv card, I use nuppelvideo [tuwien.ac.at], and then use avidemux to edit and convert to a smaller format.

H.264 not patent free

[benwaggoner]

Actually, H.264 is manifestly NOT patent free. There may be a license-fee free baseline profile for it, but it's certain that the higher profiles will have some kind of license fee ala the current MPEG-4 codec.

Still, that certainly doesn't kill a format in every case. Every DVD player pays $2.50 to MPEG-LA.

Re:and what of H.264

[benwaggoner]

Actually, those encode times aren't that bad at all for a development codec. Back when I started doing professional encoding (1994), we targeted 80 minutes of render time per minute of source files, so a feature film would be more like a week of computer time. Still had a very viable business based around that.

Of course, that was with 80 MHz computers...

LOTS of companies are working on AVC implementations, and they'll certainly compete on speed. There's lots of areas in the standard where speed/quality tradeoffs can be used.

Re:and what of H.264

[benwaggoner]

Xvid uses the Simple and Advanced Simple profiles of MPEG-4. These were extensions of baseline H.263 (aka MPEG-4 short header), but with LOTS of enhancements.

FWIW, H.263 is the standard video codec used in videoconferencing system. Most of the IRAQ video was using it.

All MPEG-4 codecs are patent-encumbered, and will require license fees in some circumstances. However, these tend not to be too onerous. For example, today's MPEG-4 video codecs are free for the first 50,000 units distributed per year.

Re:Hmm

[Bulln-Bulln]

Right up to the point where they download their first XviD video file.

There are no XviD video files. Currently most video files are:
AVI Container
MPEG-4 Video Track
MP3 Audio Track

The difference between DivX files and XviD files is the FourCC ("Four Character Code" or something).
DivX and XviD are just differernt encoders, not different formats. Just like LAME and FhG MP3Enc are differend encoders, but produce the same format.
If you change the FourCC from ''XVID'' to ''DIVX'' the normal DivX codec should be able to play it (I've never heard of any problems).
That's why I think, the people who encoded the video file should change the FourCC to DIVX, try to decode the file using DivX, and (if it works without problems) release the file with the DIVX FourCC.
This way you can reach a wider audiance.

Re:and what of H.264

[Jordy]

Yes, an AVC implementation exists [hydrogenaudio.org], but it provides its own demonstration of why no one uses it yet despite the improved size and/or quality... Namely, 30-45 seconds per frame at encode time. For a full-length movie, that comes out to two or three days for a single-pass encode.

I hardly think this matters for professional encoding. There are real-time H.264 hardware encoders in development by VideoLocus [videolocus.com] and Sand Video [sandvideo.com].

Besides, it took quite a while for the old MPEG-2 and MPEG-4 reference encoders to be anything but dog slow, so I have no doubt that in a years time we'll see at least a 4x speed up with software optimization. Add that to the 2x hardware speed up and you have something that starts to look viable for quite a few applications.