ATI Releases Linux Developers Kit

gravis777 writes, "Well, apparently ATI once again makes noise in the Linux market when they anounced yesterday that they had just released a developer's kit to support video hardware decoding (specifically for MPEG, MPEG2 and DVD) for Linux on their Rage 128 and Rage 128 Pro chips. " The new kit they are touting will "provides a software interface between the iDCT and motion compensation video acceleration of the ATI RAGE 128[tm]-based accelerators, and MPEG-2 and DVD video playback formats."

Comments

So the real question is...

[tallpaul]

Does the software/hardware enforce zones, or is it zone-free? The press release makes no mention of this that I saw other than to say that the driver is released binary only for DVD security (probably required by their DVD licensing agreement) while the playback software will be open source (why couldn't they just make a plugin for xanim?).

And of course the other real question is was it a _Cuban_ cigar?

Wow.. it just keeps getting better

[Dysan2k]

One of the few things I've wanted for Linux has been the ability to make use of hardware MPEG-2 decoders. Most video cards produced in the past couple of years (that I've seen) have had some sort of Mpeg acceleration built in. What I hadn't seen is software that took advantage of it under Linux. Kudos once again to ATI!

One thing that does concern me is the DVD playback features in the devel kit. Does this mean that ATI is standing behind the Linux side of DeCSS? Or do they have a builtin decoding key for playing DVD's? Considering most people these days just want to be able to watch the Matrix and such on their computers like all the WinX users out there, I think either way ATI's definatly hit on something.

They mention in the article that 'well known software developers' will be privy to their SDK, but what about the rest of us? I'd like to see it turn into a set of kernel or X hooks myself to support both ATI and other video card producers (Matrox, CL, etc.)

-Dysan2k

DVD?!!!!!

[kneel]

Let me get this straight... does this mean I'll actually be able to play DVD's in a conventional manner (without DeCSS) in Linux? I sure hope so, because I will soon be buying an ATI All in Wonder 128 and a DVD drive. I assumed I would have to boot into windows to view the DVD's properly, since MPEG playback is very shoddy on my machine.

This makes me very happy... now I can code and watch Anime at the same time.

Whee hee.

-- Neil

Re:DVD?!!!!!

[alhaz]

Be careful, you're treading on marketing waters.

ATi has never claimed their video cards have on-board DVD decoders. What they claim is "DVD Acceleration".

What they mean by that is that there are special features of their video board that make typical mpeg2 display measurably quicker than it would be if you did not use those features.

Specifically, it has exactly what nearly every video card manufactured since 1994 has.

First, it has a bi-linear scaler so it can scale the size of the movie window in a whole two directions. Usually width and height.

Second, it has a pallete-dac that allows your software mpeg2 decoder to write YUV data to the framebuffer and have the video card convert it to RGB on the fly. This is nice, because most compressed video streams are YUV internally, and your monitor is unrelentingly RGB.

Now, these are nice things to have around, I don't deny that at all. But they do not a DVD decoder make.

Think of it like those BASF commercials - "We don't decode your DVD, we just make it better"

Re:DVD?!!!!!

[alhaz]

Oh yeah, these chips also have motion compensation, which is definately cool, but has a lot more to do with video recording than video playback.

Re:DVD?!!!!!

[Wolfier]

NO card manufacturer have claimed their video cards have on-board DVD decoders because nobody actually have them.

BUT, the "special feature" of ATI cards are leaps and bounds better than the "special features" on the other guys like nVidia and S3 (ATI's 1997 cards can accelerate DVD playback better than the newest nVidia cards, for example)

YUV, bi-li scaling and so forth have been available to most video cards since 1994, granted. However what REALLY matter is Motion Compensation and Inverse Discrete Cosine Transform (IDCT), combining to about 60 - 70% of the actual decoding.

ATI cards have had Motion Compensation since 1997 IIRC, and newer cards from S3 are now *STARTING* to implement this feature. IDCT, that provides an even more dramatic acceleration, have been available since the Rage 128. And as of now, the only non-ATI chip that also has it is one from SiS and nobody else.

Well, if you have ~1Ghz processors of course all of them don't matter unless you are running 10's of apps at the same time.

Very cool!

[mTor]

ATi has _the_ best integrated DVD decoding of any multifunction card on the market right now. I own an ATi 128 All-in-Wonder card and its a very good card for my Linux machine. It is lacking an opensource driver but the binary one from SuSE works great. (http://www.suse.de/en/support/xsuse/). The driver is also present in lastest XFree as well (I think it's also opensourced).

The only other thing you need to make your ATi complete is the TV.... get gatos (General ATI TV and Overlay Software ATI-TV) at http://www.core.binghamton.edu/~ins omnia/gatos/

With this SDK we'll be able to combine all these features + DVD into one driver.

Awesome! I only wish other comps. did this... (*cough* nVidia *cough*)

--
GroundAndPound.com News and info for martial artists of all styles.

ATI VHA

[Anonymous Coward]

The VHA is designed for the Rage 128 and 128 Pro adapters. The VHA will consist of a kernel module and a user level library and will give developers access to idct and motion compensation capabilities of these chipsets. This is not a DVD player, or even an MPEG 2 player, but rather it will allow DVD and/or MPEG 2 players take advantage of ATI hardware for accelerated playback. The SDK will include some sample code for MPEG playback though. The reason this has to be binary is so that 3rd party DVD player developers can link in this library into their DVD player and protect CSS encryption. We are going to work with Loki to incorporate VHA into their SMPEG library. This will give other developers a real world example of how to use the library. When this is complete, the patches for SMPEG will be released to the community. David Johnson ATI Developer Relations

Re:ATI VHA

[Sloppy]

The reason this has to be binary is so that 3rd party DVD player developers can link in this library into their DVD player and protect CSS encryption.

I don't understand. If this driver doesn't contain anything CSS-related (the DVD player developer would supply that part), then why can't it be source? They can just as easily compile/link your hardware driver into their app. Just release it under a BSD-style license instead of GPL, so that they don't have to release their player's source.

That would keep CSS encryption just as protected and secret as it is right now. *smirk*

---

Re:ATI VHA

[Ross Finlayson]

Thanks to David Johnson from ATI for this information. However, he notes: "The reason [the VHA library] has to be binary is so that 3rd party DVD player developers can link in this library into their DVD player and protect CSS encryption".

An alternative would be to release the source code to the library under the LGPL, rather than the GPL. This would also allow third-parties to link the library into closed-source applications, while making the library code open-source.

Re:Yay

[adamk]

3d support is being worked on, under contract with ATI, by Precision Insight. If it's not included in XFree86 4.0, it should be out be the end of Q2.

As for your fonts... My suggestion is downloading
XFree86-Rage128-1.3-2mdk, if you can still find it. It works beautifully.

Adam

The real deal?

[captaineo]

"Although the sample code will be released open source, the MPEG-2 driver will be released in binary format in order to maintain DVD security."

Holy sh*t! This could be the real deal... 100% legit DVD decoding on Linux! But...

• The argument that DeCSS is needed to play DVD's on Linux now becomes weaker; the DVDCCA can just say "Look at ATI, they paid for a license blah blah..."
• Presumably the drivers will be 99% open source, with a little blank spot where the DVD crypto stuff goes. So we'll have to rely on ATI to compile the driver... Let's hope they don't just support x86! (DVD playback on PowerPC, Alpha, or Sparc, anyone?)

Re:So what does that really mean ...

[Nodatadj]

There was a patch for GNOME that allowed you to control the menus and stuff like the panel by voice. I'd have tried it out, but I was on a 56k dialup and didn't fancy the 48meg download for the sdk.

This settles it (almost)

[tweek]

Unless nvidia gets off thier collective asses and comes through with better commitment to the linux community, my next card is coming from ATI.
I have been a staunch supporter of nvidia for some time now but while companies like ATI and 3dfx have made firm commitments and provided resource after resource to the opensource community, nvidia hides behind waiting for XFree 4.0. With the news yesterday that 4.0 will be out in March, I will give nvidia until a short time past to follow through on thier commitment. If I can't see any measurable results, I will move over to ATI. Am I being unfair here?

The other side of the coin

[Kagato]

This leads to the question of will this actually hurt the DeCSS case? I can see the lawyers now saying that now that ATI (And Sigma too) have released linux drivers there is no reason for people to use DeCSS.

Of course the reality is that ATI has limited market share at best, and Sigma has even less. But we've gone from 0% market penitration to 5-10%. My fear is that the case will go on so long that the MPAA is going to seriously twist reality on this. (Not that they haven't already) It will no longer be about lawful reverse engineering.

Re:The other side of the coin

[Sloppy]

I can see the lawyers now saying that now that ATI (And Sigma too) have released linux drivers there is no reason for people to use DeCSS.

First of all, this software is tied to a specific hardware product, so it's of no use to people with other video cards. If lawyers can justify restricting CSS by saying, "Just buy an ATI card!" then they could also do the same thing by saying, "Just buy Windoze!" That won't work.

And secondly, there's always *BSD, BeOS, OS/2, Amiga, HURD, QNX, and many others. The world isn't just divided up between 'Doze, Mac, and Linux. (Thank goodness!) Other platforms still need DeCSS-like programs for any chance of watching DVDs.

And thirdly, this dev kit doesn't decode CSS anyway. It's just a tool that would be used by someone who makes a DVD player (which would include CSS decoding).

None of this threatens your right to use or distribute DeCSS and similar tools. None of it ever will.

---

Some thoughts....

[FrankW]

ANY company which suppports Linux should be applauded for it,

BUT

I think it is time to realise that the times when a companies decision to support Linux was driven by some "higher" goal is starting to end.

Linux is becoming a stratetgic platform, companies are realising that getting into the Linux market will be crucial, and it will start returning a profit in the (not to distant) future.

Especially ATI has shown in the past that it does not believe in open support for its product (yes there have been some changes, but only fairly recently).

It seems that there will be a market for a Linux DVD SW player, as countless posts on ./ have shown. The first product to appear is certainly going to get a lot of interest. While it is certainly very important to get DVDs onto the Linux platform, it is worrying that it is only possible with COMMERCIAL SW running on a SPECIFIC graphics card.
We all know the problems most commercial drivers have, when it comes to rapidly changing kernel releases, SMP, etc.

Secondly it would tie anyone who wants to watch DVD on Linux to ATI.

Thirdly, a working closed source solution might kill (or dampen) any effort to get a more open solution, as commercial partners (which seem to be necessary, given the DVD NDA constraints, etc) might fear that the market is already occupied by ATI.

Overall I believe that ATI is mainly doing a very clever strategic move, which might end up doing more harm than good to Linux as a multimedia platform.

I still hope some company will produce a CARD INDEPENDANT DVD player app, probabely with an open interface, to enable graphics card vendors to add card specific support (not absolutely necessary, as with modern processors MPEG-2 decoding can be done entirely in SW).

IMHO this would be more helpful than any graphics card dependant solution.

And lastly let's not forget that ATI only announced the SDK, I haven't heard any product announcements yet....

Just my $0.02

Frank

Re:Some thoughts....

[Anonymous Coward]

Where have you been? Creative has an opensource DVD player that is hardware accelerated (The dxr2 card). I have been using it to watch the Matrix under Linux for some time. There is a small binary, and CSS decoding is done on the card. Everything else is opensource.

Re:Some thoughts....

[GoofyBoy]

>I think it is time to realise that the times when a companies decision to support Linux was driven by some "higher" goal is starting to end.

As long as it provides more tools to expand what I can do on Linux do I really care how greedy a company is? They might not be pure at heart but at least its something.

>Secondly it would tie anyone who wants to watch DVD on Linux to ATI.

I would think it would encourage companies to do the same thing to _compete_ with ATI. "Our product is better and faster than any other" sort of idea. Some company has to be the first.

>Thirdly, a working closed source solution might kill (or dampen) any effort to get a more open solution,

But this solution only works with ATI hardware. There will still be a need for those who don't have an ATI. Thats going to be a huge demand when people see what is possible with an ATI/Linux combo

>lastly let's not forget that ATI only announced the SDK

Again, its better than nothing.

Apps made w/ ViaVocie SDK (Kinda Off-Topic)

[GlitchZ28]

For apps based on ViaVoice SDK try

Xvoice - Dictation in X

http://www.compapp.dcu.ie/~tdoris/Xvoice/

GVoice - Voice Commands for Gnome

http://www.cse.ogi.edu/~omega/gnome/gvoice/

KVoiceControl - Voice Comands for KDE

http://www.kiecza.de/daniel/kde/index.html

Re:Unfair?

[Twisted Mind]

ATI is allowed to have/use the specifications to develop DVD-playback software because it paid a lot of money and signed a contract not to public any information (or the source code) revealing any of them. It' just a very short-sighted view of the MPAA to make easy money.

LinuxPPC

[singularity]

I would certainly hope that kits are going to be available for LinuxPPC. Remember that all B&W G3's and all G4 computers ship with either a Page 128 or Rage 128 Pro, so most people running LinuxPPC on these machines have an ATI card installed.

As for everyone sayign that this weakens the deCSS case, I might agree, but there are still very strong arguments to be made - everyone who is not running a Rage 128 card under Linux.

Re:Unfair?

[Dman33]

Why is a big company like ati allowed to have the dvd specifications and develop software for it, while us, the norma everyday programmers not allowed to?

One word: GREED!

I am willing to bet that ATI must pay a pretty penny to the MPAA in order to write this stuff. Besides, they are 'licensed'. I would be willing to bet that even though this software is free, ATI must pay a royalty to the MPAA for each and every DVD capable video card they ship. That is also why you will never see any 3rd party software that is open source that will allow you to play a DVD regardless of the video card. (Ahm, DeCSS, cough) The game works like this:
They make to movies. They sell the DVDs for a reasonable price to the consumer. The carge the manufacturer a 'per unit' royalty to be a 'licensed' DVD player manufacturer. They charge a 'per unit' royalty for every piece of commercial software that is used to play dvd movies. And there you have it, "Mo money, Mo money, Mo money!".

For everyone that uses DVD, they pay the MPAA in 3 ways: 1, the disk; 2, the player; and 3, the decoder (software).

Mpeg-2... very interesting...

[Booker]

People have been begging Sigma Designs for information on their mpeg-2 playback engine, saying that there certainly can't be anything wrong with providing specs on that part of their DVD decoders. CSS is the only problem, right? And they come back and say "Well, but if we tell you how to do mpeg-2 playback, then you'll use css-cat to play DVDs, and that's illegal!" - which sounds like a strange argument to me... I think that basically, they're afraid that the MPAA will come down hard on them if they help open sourced DVD playback in any way.

But now ATI comes out with this same information... this is great. Hopefully others will follow. The whole thing just smacks of strong-arming by the DVD industry.

(See the "Question about MPEG playback on HW" thread on the video4linux archives - if you can find them. Red Hat - you host this list, why don't you archive it?)
----

Nice...except ATi's not the first one to do this..

[bdjohns1]

It's great that ATi's finally helping out by supporting their software in the open-source world...but they're nowhere near the first developers to open-source their DVD materials. Creative has supported their DxR2 cards under Linux since November/December 1999. (see opensource.creative.com) Creative's DxR2 card is purely hardware, so they can distribute a completly open-source implementation of DVD playback.

CSS is handled legally by the drivers - they just call the ZivaDS (the descrambler chip) and say "unlock the drive, unlock the title, decode the data". The driver code has no knowledge of CSS secrets - they're all contained within the hardware.

Right now, the support is good - VGA overlay is working for the most part, and IFO parsing (with the exception of selecting multiple camera angles) is done, so movies like The Matrix and Tomorrow Never Dies play correctly. Chapter searching works to a certain extent as well. There's a graphical frontend (gdxr2) available.

Unfortunately, there's no support yet for Creative's newest card, the DxR3 - it has some functions in software, which have not been released by Sigma (who manufactures the decoder hardware). Maybe ATi's release of binaries for the restricted materials will spur them along...

Oops, I'm an idiot.

[Booker]

I should read before I post...

Although the sample code will be released open source, the MPEG-2 driver will be released in binary format in order to maintain DVD security.

Since WHEN does MPEG-2 playback have anything to do with "DVD security?" Argh. The DVD Cartel must have some pretty big goons that they send around to their license holders....

----

Confusion...

[MinaInerz]

I see that a lot of people are confused by what video card manufacturers mean when their pretty boxes scream 'DVD Acceleration!' on them.

Most of the modern video cards on the market can convert from YUV --> RGB on the fly. This is fairly significant, because to do this in software is about 25% of the decoding process on a software DVD player. As far as I know, the only video card(s) out there that support this under Linux are the Matrox G200/G400/G400MAX cards, using the mga_vid kernel module.

There are a few cards on the market that support what is called 'motion compensation', which eases CPU usage by significant amount, also. ATi cards have, in addition to motion compensation, what is called iDCT (Inverse Discrete Cosine Transform) - which eases CPU usage by an additional 5-10%.

These features are not the same as what MPEG-2 decoders do - they do nearly all the work onboard the card, themselves, except a few things, like CSS. Sigma Designs has designed a chip that does CSS in hardware, so we might start seeing hardware decoders from them in the near future for Linux. In the meantime, the best cards to have for Linux are from Matrox. (The G400MAX SGRAM card is the best, actually, for X speed and DVD). Feel free to read the DVD Playing HOWTO for more information. Additionally, README.MGA in the X documentation talks a bit about these cards - "Makes extensive use of the graphics accelerator. This server is very well accelerated, and is one of the fastest XFree86 X servers." Because Matrox has released the specs for the cards, things are coming along real well for it (like Utah-GLX, etc.) Check the cards out!

"Human beings were created by water to transport it uphill."

Actually...

[Svartalf]

They happen to have most of a MPEG2 decoder in place in the chip. This, however, doesn't allow DVD play- just MPEG2 (which translates into SuperVCDs, etc.)

ATI Registered Developer Program

[BabyP]

You too can be a 'well known software developer', if you sign up for the ATI Registered Developer program.

ATI Developer Registration

-partap