Slashdot Mirror

Of course, you could always use a better MPEG-4 codec that happens to be open source, like XviD, or ffmpeg. When I started ripping DVDs, I compared DivX 5.0.5 to XviD to ffmpeg, and IMO XviD provided the best quality (although all three are very close together ... ffmpeg was the fastest but lost some quality compared to the other two).

When you consider that XviD and ffmpeg are open sourced, why would anyone use DivX 5?? (of course, this is using *nix as your OS, but according to the ffmpeg homepage you can use ffmpeg under Win32 as well ...)

I know this a dupe, but it still raises some interesting points.

Like GIF and MP3 before it, it's a widely used proprietory technology. Where possible we should be making the effort to move over to a GPL'd alternative such as XVID or OGG

It's not beyond the realms of possibility that this change will take place in the future, and there will be no free/clean version. If this article has duped article writers and /. readers you can guarantee there will be countless numbers of laypeople (I'm so PC) who install the ad-ware version without realising there is a clean install available.

Ok, bit of history lesson, from the way I heard it. Back in the day, DivX 4's dev team (Project Mayo) split up, half becomming what is now DivX, and the other half becomming what would later be called the XviD team.

Basically, the DivX guys seem more concerned with becomming a closed source but open standard group right now, and making DivX into the next VHS/DVD type standard. XviD seems to be the flip side, free software, version of the coin.

(I should point out that this is just what I heard, and I've also heard things as absurd as that DivX is just a pirated copy of Microsoft's implimentation of MPEG4, and that SCO apparently owns UNIX, Linux, Florida, the plans to the Death Star, and my mother, so, grain of salt, eh? In other words, I might be off in my history lesson. but hey, that doesn't stop Bush, so why should it slow me? :) )

AAAANYWAY. XviD is a MPEG4 codec, much ilke DivX (indeed, DivX and XviD can play each other's files) but XviD is completely open source, liscensed under the GPL.

So, if you are truely fed up with DivX -- and I don't see why you would be, they're just trying to make a living -- you could also go get XviD instead. It's still beta, but it's quite nice. Fast, Slightly smaller files, and has a lot of features DivX doesn't. (Like setting credits to encode in black and white, or in a much lower quality; or intelligently encoding part of the screen at different quality levels depending on how busy that part of the screen is)

Lately, one of the neat tricks I've seen is to use XviD and OGG audio, and combine them in some weird way. Insted of AVI files, you get OGM files, which are *much* neater. OGMs can hold more than one audio stream (English / Japanese, switch at will) and as many subtitles as you want, making them really much closer to the "nifty" factor of DVDs. But I'm no video editor, so someone else would have to chime in on OGMs. :)

Anyway. Go give XviD a try here. http://www.xvid.org/ If you can't compile you're own source, do a search on Yahoo for "Xvid Binaries" for user made binary installers. I like Koepi's or uManiacs for Windows, myself.

What about DivX support in ffmpeg? What about XviD?

Yeah, of course old non-adware versions of the divX codec will still be available for a while, but the point is that there won't be any new non-adware versions except the ones you have to buy.

XviD is a great alternative, which looks just as good as DivX (About 5mb per minute gets you very good quality if encoded properly. 10mb per minute is near DVD quality.)

It's completely free and GPL'd, and it's also already very popular, by my estimates its the second most popular codec, behind DivX, for the (ilegal) online distribution of movies and TV shows, unlike Ogg Theora which is completely unheard of fringe experimental codec that no serious group has ever used for a release.

XviD source code

Nic's XviD binary (best)

A divX digest page with links to several other, older XviD binaries

I have always been a huge fan of ffdshow. Open source, no adware, AND it's way faster than the official DivX codec. My aging P2 400 can watch most movies just fine with ffdshow. I'm really surprised people put up with DivX.

There's also XviD...

I've not seen Ogg Theora in action, but I have seen xvid do wonders. I'm sure there are many other options.

The problem is just getting people encoding the video to realize what a pain in the ass it will be for the end user to view the video they are encoding. If they've been using divX, they're going to continue to do so especially if they've actually purchased the software.

• from Real Networks:
• For Movies: http://www.xvid.org
• for Streaming : http://www.theora.org/

This is probably a good player. But it's just another incarnation of the Sigma Designs DVD kit.
See: Sigma Designs DVD Kits
And remember that Sigma Designs hasn't been too forthcoming with the OSS Community?
DivX Networks Press Release
XVID also has comments (look for "Sigma")
Anyway, there are several players out there already using these kits.
KISS Technology
Revoy -- to name a couple.
I bought one of these players myself (the KISS DP-500) and they are great, but still full of little bugs and the community is just building now to get into the GPL part of the source - just not sure how much of it we will be able to modify and now much will remain closed-source.
Here is a community of KISS owners so you can see what types of issues people are dealing with on these Sigma players.

"VLC works just fine"
Indeed it does.
I also have mplayer, but I find VLC far better.
I have "installed" the Divx-codec for mac, but Quicktime seems to disregard it.
I prefer that the Divx codec is used over all that MS-mediaplayer crap. Although Xvid would even be better.

Huh?

It is an MPEG codec. DivX is an implementation of MPEG-4. If you want source code for a decoder see the ffmpeg (as libavcodec) or xvid codecs. Between then, I've not see an OS with a POSIX layer that's not been able to compile a decoder engine. Granted, there are large bunches of optional parts that the various decoders don't all cover, but I've not yet see any problems with ffmpeg decoder.

If by MPEGS you mean MPEG-1, then yes - that is slightly more portable than MPEG-4 codecs, but not noticably (better support on embeded systems). They do however, have poorer picture quality, and larger bitrates. So, it's not really a good choice for internet distribution. MPEG-2 would also be better than MPEG-1, but it's also not quite as good as MPEG-4, interms of low bitrate quality. And for a web demo, the lower the bitrate, the better.

If you've got a particular platform in mind, then drop a line, and I'll see if I can find a pre-compiled setup for it.

Any reason why people don't use either more often?

I don't now about VP3.2, but MPEG-1 is garbage. Unless you are looking for something in the area of VHS quality, you can toss it. MPEG-1 demands high bitrate, but doesn't give you much in return. Compare that to MPEG-2, which demands high bitrate, but pays you back with beautiful video, the experimental features of mpeg4 codecs such as DivX 5.05 or the open source XviD which allow a low bitrate stream to (nearly?) match the appearance of a high bitrate encode. It is easy to see that mpeg-1 is on its way out. The only use it has is so 1337-k1dz can make (S)VCDs.

The world needs a free video codec.

There already is one. XviD is an open source (gpl) mpeg4 codec. Although there is no 1.0 release yet it is completely useable and can achieve better quality than DivX 5.05 (although encode times are longer). XviD currently supports B frames, chroma searching, VHQ, and host of other compressability improvers and motion tracking aids. I don't develop for it, but I am an avid fan. Check it out if you want to support open source video.

The world needs a free video codec.

Because XviD costs money and isn't Open Source right?

If it's going to be good for Linux users?
Yes

The question is: "Will it be good for the open-source movement and the technologies it develops?"

If you think a little about it, this could be a way for MS to kill a player (Xiph.org) that is doing it's way in the multimedia business.

In the open-source (and at the same time with Linux... cause it's where it gets the most used), we've got Ogg Vorbis, the best patent-free lossy audio format. To get the Unix market, MS has to kill it first. We can see that they've found a good way to do it. They're just forcing the market to adopt their codec as a standart so that unix users will be forced to used it too.

Then, there are theses projects, named Theora and XviD, which could become other good formats for the open-source community, but this time in the domain of video.
We can apply the same theory to this one... MS is just trying to kill them too, to gain the market that will be, again, forced to do a jump its propriatery technology...

So yes, it'll be good for *NIX users, but could kill the projects the open-source community is trying to create...

Actually, although version 4 and higher of DivX are not rips, they are not actually compliant MPEG-4 implementations either. If you want the benefits of compliance like proper synch, responsive seeks and all that, then check out 3ivx or XviD. The latter is open source and fragmented and thus there are often compatibility problems, while the former is very unified but not open source.

I'll download and convert this to XVID AVI, easily playable on Windows with Media Player and Linux with MPlayer. It should be done by tonight, and I'll post it to alt.binaries.multimedia.anime.

It better work with XVID, as Sigma Designs stole XVID's source in violation of the GPL to develop their "REALmagic MPEG-4 Video Codec." I believe most of the GPL problems have been ironed out by now, though.

xVid simply surpasses any DivX codec out there. Small size, superior quality, Open Source, the works.

OT: For those of you unfamiliar with xVid codec, have a look. ;)

But, I do know that any of the MPEG formats are going to be playable for a long, long time to come. (Look how long MPEG-1 has lasted.)

However, if you're after something more like DivX, then XviD is like DivX, but also GPL!

Of course, what is really needed is a third choice, a totally Free Software media codec solution that's competitive with both Windows Media and MPEG 4

There is.... It's called XVID

I *have* seen Divx at high bitrate. Actually, I have the file right here, and it's a trailer for a game. 4.6Mbit/s, 400x300x24, divx5. It's full of rapid high motion sequences; it's in-game video.

If it's a stream of real-time, in-game graphics, we can make three assertions:

1. From frame to frame, there may not necessarily be smooth transitions. For example, a high-action, high-speed scene, done real-time, is likely to be very "choppy" (low FPS). As a result, frames N and N + 1 may have such huge deltas that little common data can be found. This would butcher the quality for most video encoders.

2. A typical game scene typically has very little static imagery. A movie will have actors standing around talking, with the background unchanging, or there may be long sequences where the delta between frames is minimal. Games on the other hand are constantly in motion, often the backgrounds will be moving and special effects, which are more discrete than those seen in movies, will occur.

3. Video games have a higher number of sharp edges than live action film or artificial sequences that are designed to be photorealistic. Real life, when photographed, looks "blurry" (attributed largely to the short-comings of photography equipment). Many codecs use this fact to their advantage of throw out data that does not contribute to the images.

These three things in mind, any MPEG-like codec will probably produce fairly poor quality output or will require high-bitrates.

It's been my personal observation that divx does not give you much improvement in quality past a certain point. The 4.6Mbit video here still isn't anywhere near as good as I'd expect at this bitrate - there are certain imperfections that are just inherent in the format, and I suspect that encoding at 3Mbit would give you almost exactly the same output.

Are you talking about a bit rate or size of the file?

Throwing extra bitrate at divx often yields very little improvement.

I highly disagree. I've found when encoding video that slight increases in bitrate may push the quality of a DiVX ;-) encoding way up. At least, this is with using XviD -- a somewhat superior encoder to commercial DiVX ;-).

The reason I bring this up is when I got LotR:FotR Special Edition, I wanted to encode the MTV parody of the Council of Elrond. I made a number of different versions, and since it encoded so quickly (being short ;) I spent a lot of time tweaking.

MPEG4 in general is optimized for streaming applications and low-bandwidth use. There are certain design descisions in a codec that make it ideal for some applications and suboptimal for others. Try encoding MPEG2 at 500kbit for a dramatic example, the codec is *optimized* for high bitrate applications. Yes, quality is a function of bitrate, but different codecs have different optimum points - the best quality/size compromise. It's really not that difficult a concept.

I know it's not a difficult concept. That's why I pointed it out.

I've seen MPEG-2 streams at low bitrates. Yes, they are inferior to MPEG-4, but in the higher-end, they are easily on par. Another experiment I performed was encoding a dual-layered DVD to ~4.7Gb using MPEG-4 (the standard capasity of DVD+/-R). Aside from the failings of encoding an already encoded stream (video artifacts become more pronounced from the first round) it looked nearly identical.

Now, blizzard is probably settling for an inferior video codec with divx because they got a better licensing deal. Obviously they wouldn't have chosen divx if it truly looked like ass, it doesn't, but I think they made a quality/cost descision more than anything. Slightly less quality, slightly less developed SDK, for a large reduction in cost. It's a good business descision I suppose.

Why does one need a heavy-weight SDK to play a movie in a game in the first place? You're not a general-purpose media player, so you've got limited scope. The process of decoding and blitting MPEG-4 is border-line trivial (relatively). It makes no sense to use something like Bink. Of course it's a good business decision! :)

Now, the fact that you needed to rip on me rather than just set me straight, says something. Maybe my attack on divx offended you at a personal level. My, aren't we protective of our favorite piece of code... unless you wrote part of it yourself, you have no reason to be this argumentative. If you did, then you would consider how such petty bickering reflects on the product you wrote and the company you represent. Maybe you just can't stand to see ignorance in the world and feel it's your duty to arrogantly berate everyone as if you're the final authority on everything.

I prefer Xvid.

If my comment was so full of ignorant rambling, why did you feel the need to correct me in such a vicious manner? Why not simply dismiss my nonsensical rambling? Because I was modded up, because you think I have nothing better to do than validate my existence by trolling or karma-whoring?

Oops, I've just replied to flamebait.

Now really, we have to get that sand out of your vagina. It's making you cranky.

Well, I am sorry for the "reaming". I didn't mean for my response to be so vicious, it just comes out like that sometimes. I'm sure you know how what it's like to start typing feverishly away to defend or attack a viewpoint.

DiVX ;-) is only an implementation of MPEG-4. If you want a DiVX-like encoder that is open source, try here.

MPEG4 is a proprietary format with patent and royalty encumbrance that make it in no way superior to Real for freedom.

No, he means XviD.

Yet another rejected quality story that got rejected by slashdot fuckfaces. Brought to you by Hacker underground.

XBOX Mediaplayer Infringing GPL

Anonymous windows coders have made a mediaplayer for X-box which uses GPL licensed FFMPEG and XVID sources in their closed-source product. Developers can't be contacted because they don't publish their e-mail addresses. As a result XVID development has been stopped. This isn't the first time someone decides to steal open source code - Sigma Designs stole XVID code, but was forced to release the source

Releasing the affected code under the GPL is lip service?
No, it's exactly what is supposed to happen.

They are probably referring to other things than the actual release. They don't give any credit to XVID. From the press release:

"We are pleased to provide the development community with an open source MPEG-4 CODEC, and anticipate that this will accelerate technical improvements and enhance the proliferation of MPEG-4 content," stated Ken Lowe, Sigma Designs' vice president of business development. (emphasis mine)

It makes you think they did all the work. The download page acknowledges that they "had utilized some routines posted by XVID as open source," but the tone is that they did all the hard work. Which isn't what I came to understand when reading the evindence.

Of course, now we have the source, it'll be interesting to make a more detailed study on how much code was copied.

And you, instead, should get "-1, RTFA"

the proof is RIGHT ON THE FRONT PAGE

And while this isn't a court ruling, I think one would have a hard time finding a programmer worth his/her salt that would say that the evidence doesn't look incredibly incriminating.

I know it's posted anonymously, but this news release is important background that is not attainable through the links of the original post.

if by not attainable you mean that clicking on the link to xvid.org then clicking on the press releases in the files section is an impossible feat, then yes the information is not attainable throught the links in the original post. it is indeed a harsh reality.

how about the unattainable information in pdf format.

for more unattainable information i would goto this oracle of truth

I ditched my vcr months ago. Just get a tv capture card with the bttv848 chip for video in [I recommend the winTV-FM, as it also has a stereo decoder and sound capture dsp on the card, leaving your existing sound card free, about $50 street]

Then, all you need is a good audio sync maintaining capture program like NewVideoRecorder and a good MP4 codec, and you're set! Oh, you probably need a least an athlon 1800 or equivilant, to do realtime 640x480 encoding capture with good deinterlacing. Much weaker systems can easily handle 320x240, which isn't much worse than vhs. Add in a few 80gig drives, a fast CDR, and you've got entertainment bliss.
Did I mention that the hauppage card comes with a remote, and it too is supported. So, sit back on the couch, with the computer hooked up to both record and play to your big screen tv, easily controlled by a remote.

It's being done right now, today, on peoples linux boxes. I've been doing it for over 4 months!

The only bad thing is that, currently, I still find the best application for editing commercials out of shows I want to archive, to be virtualdub [a win32 app]. It runs under wine, sure, but it still kind of hurts to have to do it. At least it's GPLd, though.

gprof is maybe not the most impressive tool to use, but it's quite useful. At a IA64 course at university [German, sorry] we used gprof to identify the bottlenecks in the c-code of the xvid-codec. Then we assembler-optimized like mad and got quite a nice speed-up.

Result can be found in our wiki:

I use Xvid. It seems to cope with almost everything and it's open.

Allegedly, transcode has some sort of clustering stuff as well. I can't testify for it, but dvd::rip's cluster mode is great, except that it won't do AC3 audio pass-through.

However, I've found that the XviD codec is fast enough to encode 2:1 on my P3 866. That is, it takes 2x as long to encode as the movie lasts - it gets around 12fps. DiVX 4.02 put out slightly better quality, at 4-6fps on the same system. I haven't done anything with DivX 5.xx yet, so I can't speak for that.

DiVX is a very close variant of MPEG-4 and no longer has its source open. H.26L is open and already provides for 1.5 x better compression than DiVX. XViD is also about 10% faster and is open source and nearly all GPL at this point.

DiVX will just fade away the same as MPEG-4 due to it's too greedy nature.

The problem with Ogg Tarkin it is still pretty much an experiment, using techniques which is way ahead of its time. 3d wavelets haven't been heard of in any other standard which are under development.

XviD, however, exists today. It is a fully GPL'ed MPEG4 codec. However, it cannot exist legally in any form other than an experiment because the MPEG4 license still has to be paid for in order to use XviD. XviD, like LAME, will mostly exist as CVS sourcecodes under guise as an experiment, with many rogue sites around the world providing binaries (usually with automated daily compiling).

Personally, I wouldn't count on the quality of VP4 being anything earthshattering. Tests of VP3 quite clearly shows that it is behind DivX3, DivX4 and XviD in terms of quality, so something has to really shape up. This might probably be due to a lack of 2-pass VBR encoding feature in VP3. Meanwhile I will just continue to encode my rips in XviD, encode the audio with Vorbis and mux them together into an Ogg container. If VP4 or VP5 really turns out to be good enough, I would probably try to find a way to mux that video stream into the Ogg container too.

But some people picked up the opendivx code and kept developping it.

I'd say xvid is about up to par with divx5 now. (Save b-frame support, which is still divx5 only.)

Or maybe he just meant xvid...

Or would you prefer EVERYTHING be rendered on the fly?

In the case of games, in a word, yes. Given the capablilities of the GeForce3 and GeForce4 Ti series and the Radeon 8500 with regards to DirectX 8.1 it shouldn't be that hard to do. Hell even a GeForce2 is overkill for the vast majority of todays games anyway. If you're looking for a good video codec then look no further than XviD or in the future, Tarkin. (Or DivX 3.11 if your into super awsome but illegal codecs.)

MPEG4 is essientally a super snazy version of flash for high quality video. Think about it. It is trying to make many different types of interactive media available on a wide range of platforms simultaneously. It's an attampt to make a single proprietary format that does everything. This is exactly what flash does/is trying to do. True, MPEG4 actually has uses beyond annoying banner ads but it is essentialy the same sort of idea. Just as flash can be surpassed by XML/CSS/DHTML so MPEG4 can be surpassed by XviD/Tarkin and OGG. The reason is because these alternatives fufill the primary purpose of flash and MPEG4, interactive web content and video/audio compression respectively while being free, open, stable, and universal. Yes the flash/MPEG4 paradigm provides cleaner intigration and a nicer package from a development standpoint but when lisenceing costs are factored in the open alternatives win hands down.

It looks like Divx5 is released, and DivXnetworks are resorting to implanting spyware (or charging $30 for the pro version) to offset the licensing costs. This might have some effect on the XviD (open sourced derivative of DivX) though, but looks like it will go the same way as the LAME MP3 encoder, as discussed in this thread

Hardware YUV translation. The only thing that makes video playback bareable in Linux on my box.

Yeah, that probably helps a lot, I use it on my G400. I thought you meant hardware mpeg encoding support.

Can you also watch the item being recorded or does this box act as a standalone video server?

Umm, not sure what you mean exactly. I've patched NVrec:DIVX4rec to use the Matrox G400 back-end scaler (=hardware YUV -> RGB) to display the frames being encoded to disk on the TV as well. If you just wanna watch stuff without recording there's fbtv (WinTV video-in -> Matrox G400 TV-out).

After a recording is finished, I can watch it on the TV using mplayer and the mga framebuffer driver for TV-out.

I have not used the DivX codecs in my testing.

Well, I can really recommend the Xvid MPEG-4 (=DIVX4) encoder. It is open-source, and plug-compatible with the older divx4linux encoder.

I put up some more info on the Freevo project website.

Hmm, kinda. The parts are there, but no-one has put them together in a neat package yet. That's probably since Linux video software has really taken off in just the last 6 months.

There's mplayer which is a great player for any video-format out there. It can even play DVD's, although it doesn't have menu support like Ogle. It can also rip DVD:s to MPEG-4 (a.k.a DivX) using a couple of different encoders. Xvid is my favorite open-source MPEG-4 encoder, it's also got good reviews on Doom9 (good place for DivX info!).

For the TV-in recording part you can use a $50 WinTV card and the Video4Linux drivers. On top of that you need an audio-video capture application that can use encoders such as Xvid and Lame to encode to MPEG-4 (video) and MP3 (audio). I use NVrec. If you try the NVrec suite, use the DIVX4rec app (with the Xvid library instead of divx4linux which isn't maintained anymore). On my P-III 500MHz I can compress 29.97 Hz (NTSC) 320x240 in real-time to 800 kbit/s (video) + 80 kbit/s (audio). It takes about 5 hours to make a one-pass encoding of a DVD, so with a faster CPU it's probably possible to do real-time de-interlacing and encoding of 640x480 video.

A drawback is that NVrec is a command-line app for recording, I'm working on a patch for real-time preview on Matrox G400 TV-out. Or if you have a fast enough computer you might be able to run mplayer on the file as it is being recorded. This would allow for Tivo-like pause and resume. It might be a problem with AVI files from NVrec though since I don't think they're streamable.

Now, to put all of this together you need some kind of control application. That's not really that hard to write compared to all the other pieces (mplayer, xvid, nvrec). I've been working on one for the last couple of months, and have an alpha version that is usuable. It only supports the Matrox G400 for TV-out, and is a little crude, but it works good enough that I have it hooked up to my TV for everyday use. It's controllable by a remote control (see Lirc), using a very simple text-menu system to view tv, play avi/mpeg/mp3/dvd, record tv-in and rip DVD's. I'm getting ready to put it up on Sourceforge as Freevo within the next couple of weeks.

The application is written in Python which is great for stuff like this. Once the basic stuff is done, it might be cool to make a plugin architecture where you could interface to other stuff. For instance, with OSD (on screen display), it is easy to add things like new mail notification while you're watching TV. Or new Slashdot headlines, ICQ chat notif, phone caller ID interface, www control, etc. And, of course, an interface to some kind of tv-guide.

I haven't really found any other complete applications like this. Not that I've looked that hard, I'm always looking for an excuse to write software. mplayer might end up with all these feaures eventually, it is improving at an incredible rate at the moment.

In addition, the adoption of a unique ID written on a Blu-ray Disc realizes high quality copyright protection functions.

I assume this to mean that it'll employ the next generation of CSS encryption. For one thing, MPAA and friends have probably learned their lesson: don't roll your own stream cipher. For another, it's now legal to export products using 128-bit encryption from the United States; the regulations in effect when DVD CSS was standardized permitted only 40-bit.

Or is this just a try to make movies even LARGER so cable and DSL users can't share movies in high quality anymore?

At a point, the detail becomes so fine that the human eye can't distinguish it. XviD (a fork of the last free DivX 4 release) attempts to find that point.

This leaves out the few million people who watch movies on their standalone DVD players.

I can already do that with my laptop that has S-Video out and is actually smaller than most set-top DVD players. But for the rest of consumers, why can't a set-top DVD player maker make its codec system extensible so that users can compile codecs such as xvid (which grew out of the OpenDivX code), put them on CD-ROM, and load them into the player's flash memory?

I don't think DivX supports Dolby Digital 5.1 sound yet

Most casual movie watchers have only stereo anyway.

- Divx;-) (versions 3.xx) are a hacked Microsoft mpeg4 codec. Before Divx 4.xx it was the standard codec for DVD-based online copyright-infringement.

- OpenDivx (aka ProjectMayo) was an open-source project to make a video codec to replace divx3. This was never really able to compete with divx3 in its capabilities.

- Divx 4.xx came from OpenDivx code but is no longer open source. It is written by DivxNetworks and it is this that Fraunhofer is licensing. Despite not being open source it is now widely used; most say it is a better codec than divx 3.11 though it lacks one very slick tool - nandub - which some still claim is the method to use for the best encodes between the two codecs. Also, they don't use the smiley in the name as far as I know.

- XviD deserves a plug here. It's a new open-source mpeg4-compliant (though possibly not completely mpeg4-legal) codec based on opendivx which (developers promise) will be GPL'd as soon as the last few hundred lines of GPL-incompatible code are replaced. Despite being very new and "alpha," it's already competing with divx 4 in terms of quality of encodes. Source can be obtained at the above site; binaries (VfW) are here and elsewhere.

Anyway, so yeah, the submission was wrong on a few points, but we've got everything straight now, don't we?

It's running faster than DivX 3.11 on all CPU's, and only lags DivX 4.x on a P4, though not for long. It also produces smaller bitstreams than either of them.

Anyone interested in MPEG4 video coding should check it out.

BTW, you need to type in the full url http://www.xvid.org to access the XviD site. http://xvid.org does not work.

The guys over at XviD have almost fully rewritten OpenDivX from scratch in GPL'ed code - it's now running faster than both DivX 4.12 and the old DivX 3.11 (MS-MPEG43), as well as creating smaller files with identical quantizers (a standard measure of motion estimation efficiency). MPEG4IP are working on a complete .MP4 file format implementation, also. -h