Slashdot Mirror
VP3.com: Future VP3 Releases To Be LGPL
sudog writes: "According to this vorbis-dev posting and The VP3 Homepage VP3 (QT5-type movie compression scheme) is now under the LGPL! What's not clear is whether they intend to offer it guaranteed royalty and patent free to the community. They're actively looking for help, too. Does this mean that we no longer need the OGG-Tarkin to save us from our movie-less, video-app-less emulating?" Of course, they don't say starting when, exactly.
After all, QT is a metaformat .. which isnt actually doing any compression or anything, but rather specifying which codecs to use. the biggest problem with getting an open source version that handles all QT-files nicely, is the patents and licensing on central codecs (especially the Sorenson Video codec).. more info can be found at http://www.apple.com/quicktime/specifications.html .
mats
One man's ceiling is another man's floor.
___ Dan Miller
(++,) CTO and founder, On2 Technologies
On Wed, 27 Mar 2002 volsung@asu.edu wrote:
> On Wed, 27 Mar 2002, Daniel B. Miller wrote:
>
> > Hi fellow Ogg-oids --
> >
> > I wanted to let everyone know that VP3, our open-source video codec that
> > is commonly used with QT5, is being re-released under the LGPL. We are
>
> Really!?! All I can say is wow. What about the patent issues? Are you
> granting royalty-free license to the required patents along with the license
> to the code?
That's a requirement of the LGPL, so, yes.
...The latest issue of Linux Journal (remember that mag? ) has a good article on the various quicktime players available for Linux, and how to install/use them ...
Check them here : Linux Journal
I lost my concept of community when my community lost all concept of me.
Another video codec sounds good, but after the nasty DivX ;) experience, I don't trust people who make their codecs open source, and then return to being proprietary.
LGPL is a bastard license at best. GPL or nothing is the only way to go.
Future versions of VP3 will be released under the LGPL [...] Stay tuned!
In other news, future versions of Windows will be released under the LGPL; stay tuned and don't forget to tell your children to tell their children to tell...
:)
I have never heard of this codec, but it seems to me that this is more or less what the LGPL is intended for. Take a quick look at the LGPL and note this section:
(Emphasis mine)
Seems to me that the people at VP3 would like as many people as possible to start working with their codec, allowing it to gain ascendancy over other codecs so that someday they will be able to make money selling their own "enhanced" version. Not a bad deal for GNU, because we get something badly needed. I hope that we start to hear more about this codec being used in some interesting projects in the future now that it has become more available.
http://www.xvid.org
XviD is based on the old OpenDivX-Codec but afaik doesn't use any of its code any longer and is completely GPled.
the codec improves at an amazing speed and already beats the shit out of VP3:
http://www.doom9.org/codecs.htm
Don't think of it as a flame---it's more like an argument that does 3d6 fire damage
For those bitter about Divx, have a look at XVID (www.xvid.org). This a GPL video codec based on the source of OpenDivx.
:)
This codec give me excellent results compared to Divx 3.11Alpha and Divx5...
With XVID video and OGG sound all in a OGM file (OGg Media) i get fully legal DVD-Rip !!!
This means that streaming video embedded in webpages for the Linux / Free Software clan is now a reality, but perhaps more importantly, VP3 makes it possible to eliminate adware/spyware like RealVideo and the equally obnoxious and platform-specific Windows Media. Cheers for all the great work, On2 Technologies! This is, in my opinion, one of the most important things to happen on the open source scene for quite some time.
Tarkin is currently working on bringing new technologies such as wavelets and 3-d transforms into video coding. It's not finished yet, but it offers more possibilities for really new technology and further development.
While this is great news, it by no means means that Ogg Tarkin suddenly is obsoleted
--
GCP
> Really!?! All I can say is wow. What about the patent issues? Are you
> granting royalty-free license to the required patents along with the license
> to the code?
That's a requirement of the LGPL, so, yes
i took a look at how they do things and it doesnt look good.
3d wavelets are useless without motion compensation, they have no motion compensation yet.
even with motion compensation they are gonna have a hard time stitching it together so no artifacts are left over.
tarkin is around since 2000, all they have is some obscure 3d wavelet transform and a huffman backend that leads to good quality at 800kbyte/sec but.. no.. something striped moves by one pixel and everything breaks. yes, its a research codec but comeon.. they have been talking on the mailing list of adapting several other codecs (like vp3) and go on and on over different motion compensation aproaches like meshes and stuff without writing any test code.
so.. as long as no one comes by and drop them a pretty done codec tarkin wont get done.
well ok.. its better then indeo 5 in some cases though.
now that vp3 is lgpld theyll probably just write an ogg header and its done.
and here I thought we were talking about virtual pool 3.
don't do that to me so early in the morning, I had sudden visions of making a FFA 8-ball tournament. or 9-ball CTF.
ugh. back to my coffee.
Doom9's site is the premiere site on the web for video encoding. Doom9 actively tries to get his hands on the newest encoding tools, and periodically he tests them to see which codecs give the best results.
It used to be that along with the lastest versions of DivX, he tested WMV and VP3; he doesn't anymore: WMV and VP3 consistently lost and lost badly to div3 sbc, div4, and xvid. You can't say that VP3 is "the next DivX" when it's can't outperform the ancient div3, much less div4 or the newly released div5 / xvid.
[shameless plug]
I really, really, /really/ like xvid. It's an open source reimplementation of Project Mayo, the project that led to the development of div4. Development is fast; I have realized significant gains in quality and usability in even the past two weeks. The codec is fast; on my crappy windows machine it crunches frames faster than div4 and div5 and its playback filter (w/postprocessing!) uses fewer CPU cycles than div4 or div5's.
If you want to play around with xvid, the easiest way to start is to go to the xvid forums at doom9.org and read about what the codec can do for you.
[/shameless plug]
-inq
They're selling the VP4 codec and have a VP5 that's in testing right now. They plan on selling that- if they get an open sourced version of the prior generation out there that can be supported by their newer codecs, that's a win for most people (So long as they provide the decoder for the latest format for free, that is... :-) I'll be posing that question to them on the list. It's great and all, but unless we've got decoders for VP4 and VP5, it's not as good as it could be.
I am not merely a "consumer" or a "taxpayer". I am a Citizen of the State of Texas
Yes, it may not perform as well as xvid, but xvid has a severe drawback, that you should be giving consideration to- patent licenses.
There is not an aspect of MPEG4 (which is what xvid is a codec for...) that is not covered by patents that need to be licensed. Nobody has a royalty free license to any MPEG4 patents for free software implementations so distribution of the codec is violating the patent rights of several companies.
You may not care now, but they're stepping up enforcement efforts of all this stuff and you may well find yourself without a codec or in trouble because you're violating the rights too.
I am not merely a "consumer" or a "taxpayer". I am a Citizen of the State of Texas
To quote the article:
In my experience VP3 only gets noticibly blocky (the tester's major complaint) when it is prevented from creating a keyframe when it wants to. Here, they pretty much prevented vp3 from generating keyframes at all. The keyframe interval should have been left BLANK not set to a stupidly high number.
Additionaly, there is another menu of keyframing options (the one he should have used to set the adaptive keyframe rate rather than locking it) of which he writes nothing. Here, I probably would have set the minimum time to about 1/4-1/2 second, and set the maximum time to the highest supported number.
Furthermore, There is an image quality control which controls the tradeoff between image quality, and the risk of dropping the frame rate. No mention was made of the setting of this control, but the complaints about low detail make me wonder what it was set to.
Finally, turning quick compress on does lower quality. For a test which did not involve encoding speed, I have to wonder why the tester chose to turn that option on, as it trades off quality for faster encoding!
I use vp3 to encode DV streams (in Quicktime) for viewing over the web. Vp3 is a very good quality codec, superior in many cases (unless you are streaming from a QTSS, or the source was shot under unusual light conditions) to the free version of Sorenson. It is excellent under these conditions.