divx and h264

[sjakke]

I don't know if this is a stupid question or not, I'm a newbie. but I was wondering if divx is gonna use h264?

[Doom9]

well.. if you know that h.264 aka mpeg-4 avc is quite different from mpeg-4 asp (divx is an mpeg-4 asp codec), that line of questioning does indeed not make much sense ;) But if you don't, you've just learned that there's a big difference. You can't just "add" AVC features to an ASP codec, it would require an almost complete rewrite and DXN has already stated that they will not create an AVC codec (as reported in my news).

[DigitAl56K]

I think the key point we would argue is that while AVC can offer additional benefits in terms of compression over ASP depending on the implementation, for a number of reasons (including high decoding requirements for HD and a lack of AVC-enabled CE devices) this is the wrong time to introduce an H.264 codec.

That, of course, is not to say that we will never create one, but you have to put any benefits you think you'll gain in the context of the losses you'll suffer (e.g. compatibility) and how a new technology falls into the media ecosystem as a whole (i.e. DivX Certified devices, the new DivX Media Format, etc.).

If you had a very narrow perspective on technology, i.e. looking at AVC solely from the point of compression, you could potentially claim that it was "better" than some other modern codecs. As soon as you widen your view, however, and think about how your content will be viewed by the wider world, including on CE and on less powerful desktop computers, AVC is less attractive.

Technological progress is a wonderful thing, but think MP3 vs. Vorbis: Which is most widely used today and why? Vorbis is free and offers better compression than MP3, but MP3 still retains by far the largest share of the market, with Vorbis generally being relegated to a smaller technically minded group. MP3 has some advantage in being the first technology on the scene, but more than that it plays everywhere from 101 media applications on your PC, to your stereo, to your portable CD player, to little flash sticks you can hang around your neck - heck even cell phones play MP3 these days. Just because one technology slightly improves upon another for a particular task does not make it globally "better". Think about the ecosystem, the way that the technology can be used, and, most importantly: "What can I do with the technology?". Take the DivX Media Format for example - Although we've substantially improved the quality of DivX video recently, the benefit of the DivX Media Format is really that you can now do more with the DivX Format and experience it not only on your desktop, but in CE too.

This story is far too often overlooked by those of us who are strictly focused on the essence of the technology itself...

[bond]

well if you widen your perspective even more you could say that soon all sorts of content creators will ask for encoders able to create HD-DVD/Bluray content and dxn, with its widely known (but of course also pirate related) name could get very important in this upcoming codec manufacturer battle...

of course thats a totally different market than dxn focuses at atm...

[DigitAl56K]

HD-DVD/Bluray are certainly interesting, if not a little chaotic at present :)

[IgorC]

Digital56k. is it more short to say : I´m devop of Divx, that´s why I like Divx.

[DigitAl56K]

Originally posted by IgorC
Digital56k. is it more short to say : I´m devop of Divx, that´s why I like Divx.

It certainly is. However, that's untrue.

Going back to the point I was attempting to convey: A new technology is not necessarily "better" simply because it slightly improves upon a more established existing technology. Using AVC I can't do many of the things I can do with DivX. I can't burn the AVC file to a disk and watch it on my DVD player. I can't create an HD AVC file and play it back without dropping frames everywhere. I can't achieve all of the things that I can with the DivX Media Format with the current generation AVC formats.

Let's talk again about Vorbis vs. MP3. Vorbis can be viewed as an improvement upon MP3. I am very interested in Vorbis, but to this day I have never encoded more than a handful of files into Vorbis format because all my devices work with MP3 and MP3 achieves a quality/compression that I'm comfortable with. I know if I send an mp3 file to a friend and he has a player with "MP3" stamped on it it's going to play the MP3 file just fine.

For similar reasons I encode with DivX 5 rather than XVID. While I appreciate that unbounded by constraints similar to that of DivX, XVID will typically produce slightly better compression, I want to be able to play my files reliably on my DivX DVD player. Sometimes I want to be able to send files to my friends and know that they'll play reliably on their DVD players. In future, I want to be able to add DivX Media Format features to my files and know that they'll still be compatibile anywhere I take them. Sure, if I used a different codec there's every potential for me to get five or ten percent better compression, but the result is that I can't do what I want with my files.

I couldn't possibly argue that DivX is the best solution for every single person on the planet for every use case, but what I will say is that we have put a lot of work into building a strong codec that fits into a much larger ecosystem, and that the ecosystem enables people to exchange content and enjoy a good experience both on and off the desktop, and with some nice interactive features in the pipeline with the new file format. AVC is a cool technology, but I can't apply it to the things I want to achieve today. That is not to say that in future we couldn't develop an AVC-based technology that did, of course.

I think that reply in itself was a little bit cheap, but what I'd really like to see is some further thought into how technologies play into the wider world. I'm actually surprised, given the nature of this forum, that this doesn't happen more often. Always striving for the next codec to give you +.25db PSNR won't necessarily get you anywhere. There will always be something better around the corner. You have to look beyond any one particular piece of the puzzle and look at the entire landscape.

[IgorC]

That´s true. OGG was (and is) a big promise for digital music.But most of people prefer to use standard MP3 (lame).However, for example, now AAC is eveluating quite fast. (at least situation is better than OGG´s). Each time more devices support AAC. I hope it will be the same situation with H.264

[DigitAl56K]

Things will definately progress with AVC, and those here on Doom9 are likely very early adopters. In general though, the world is not ready for it quite yet. If it were, DivX 6 might have been an AVC codec ;) Look at ASP: It's taken DivX 5 years to gain strong industry adoption and compatibility in DVD players. Likewise MP3 existed for years on the desktop PC before it became accessible enough to drive hardware manufacturers to support it. My personal take is that there's a strong value to all of us in contributing to the future, but at the same time recognizing the advantages of what is available today and leveraging it to its full effect is equally beneficial.

And that being said, it's time for me to go home. Until tomorrow.... :)

[CruNcher]

@DigitAl56K
yeah it took you five years right but AVC has cover from Hollywood a pro point that DivX Mpeg-4 Part2 never had look how fast Chips are made now almost every bigo presents a sollution on weekly basis with DivX it took years and not only asian solutions :P

[temporance]

My €0,02:

There are two ways standards are adopted:

1. They are championed by industry heavyweights

The industry (whether Hollywood, broadcasters, hardware manufaturers) needs a new technology. It invests, researches and implements on an industrial scale.

Examples are: audio CD, PAL, NTSC, GSM, UMTS, MPEG-2, DVD


2. They ride the wave of user popularity

Early adopters (like doom9 denizens) find a cool use for a new technology. Word spreads, popularity grows and we have a "de-facto" standard. Manufacturers follow into a market vacuum created by popular demand.

Examples are: audio cassette, mp3, DivX -note that all are associated with piracy!


It looks like AVC will take route 1 into some applications. The jury is out on whether it will also follow route 2 into our PCs and P2P networks.

[DigitAl56K]

Good post :)

Route 1 can take longer than you might think. Certain pieces have to be in place and reasonably proven, e.g. drm, user bases, support in reasonably priced consumer electronics, etc. Think about BluRay and how many years it has been earmarked as the next industry standard, yet we still have nothing. The movie industry appears to be more than content with DVD/MPEG2 right now, their sales are high, and it's really up to someone else to drive a new standard. I think what will happen is that we will see AVC used for things like end to end closed systems, such as digital cable broadcast where media providers have more free reign to adopt new technologies, long before we see compatible implementations in DVD players that would require standardized DRM, menus, audio format, licensing models, etc. to be in place. My belief is still that number 2 will also take a while. I think there will certainly be some AVC content out there but it could be some time before the masses can really make use of it.

It will certainly be interesting to see how it all plays out :)

[LordRPI]

Originally posted by temporance
My €0,02:

2. They ride the wave of user popularity

Early adopters (like doom9 denizens) find a cool use for a new technology. Word spreads, popularity grows and we have a "de-facto" standard. Manufacturers follow into a market vacuum created by popular demand.

Examples are: audio cassette, mp3, DivX -note that all are associated with piracy!



I'd like to think of #2 as the proof of concept method which can open the gates to other forms - think the new Napster and the iTunes Music Store. Notwithstanding, Thomson still collects a stiff royalty for mp3.

[jggimi]

Interesting way this discussion has branched OT.

De Facto standards are choices which are adopted by competitors in a marketplace without any standards body making an agreement or recommendation. They are adopted either due to market dominance, or customer demand. One example of a de facto standard we all know well is a Windows operating system for a PC. It is usually a dominance factor that causes a de facto standard to be adopted by competitors in any market.

There are two reasons why any technology company decides to join a standards body: To push their technical developments for adoption, or... To block competitors from pushing comptitive technical developments for adoption.

[ron spencer]

gonna wade in here.....don't look for HD or Blueray to take off for quite a while..

when my son was born 2 yrs ago blockbuster was still half and half DVDs and VHS....DVD has JUST RECENTLY become mainstream....end users are not going to replace units they just bought for no increase in quality unless you have an HD TV. I remember reading in consumer reports that over 98% of North American TV owners have a TV...just 500 lines of resolution...HD will have no effect on these guys.

I expect that HD and Blueray will start off like DVD-A...a dud...DVD-A sells less than LPs....however marketing will takeover eventually...my forecast is 10 years

[Sharktooth]

If and when DivX network will make an AVC codec every pirate/releaser will start to use it for movies.
It sounds ugly but guess what driven chipmakers to make chipsets for standalone DivX players?
There are plenty of DivX encoded movies out there... and that means BIG MONEY for standalone players manufacturers...

[planet1]

http://images.divx.com/divx/screenshots/vdub4.gif


"DivX® Q H.264 codec" = :confused: DivX 6.x :confused:

I suppose the divx guys want to keep their options open. ;)


maybe related to "project Q":

*March 22nd 2004 - doom9 news*

http://www.doom9.org/index.html?/Old_news/march04.htm

[planet1]

http://images.divx.com/divx/screenshots/audio/vdm15.gif

I wonder which screenshot is older ... :sly:

[DigitAl56K]

Ahh the infamous Q codec! :)

DivX 6 is an ASP codec. That is not to say, of course, that we have never researched AVC. We are, after all, a technology company ;)

I think I've explained above most of the points that have driven our decision to use an ASP-based technology for the video component of DivX 6.

[Elias]

Originally posted by DigitAl56K
Ahh the infamous Q codec! :)

DivX 6 is an ASP codec. That is not to say, of course, that we have never researched AVC. We are, after all, a technology company ;)

I think I've explained above most of the points that have driven our decision to use an ASP-based technology for the video component of DivX 6. ASP rules, keep it up!