在談到ATI的解決方案之前，我們先來了解一下目前在PC Windows的環境下，我們是如何來進行3D遊戲(或3D電影)的應用？由於3D立體影像的顯示與顯示卡有密切的關係，而且必須以LCD Monitor或以DLP Projector來顯示。因此nVidia遂制訂了一套專屬(Proprietary)的「3D PC」標準，也就是使用「nVidia 3D Vision」的配備：專屬的LCD Monitor、專屬的3D眼鏡與專屬的3D Driver，在PC (或Notebook PC)上來進行3D的應用。但自從nVidia於2008年開始大力推廣其「nVidia 3D Vision」的「3D PC」平台以來，由於其定位太狹隘與匪夷所思的行銷策略，終究無法成功地大量舖撒其「3D PC」平台(其詳細原因分析，請參考本部落格的另一篇文章 -- 「3D PC」何去？何從？)。
其實只要對3D產業發展有所認識者，應該不難看出目前各種3D標準正在制訂(或已制訂)，ATI當然也看出來這個發展趨勢；也就是說，ATI根本不會去開發什麼3D眼鏡與3D顯示器，因為3D眼鏡一定會由「3D Ready」的顯示平台(如120Hz 3D TV或120Hz 3D LCD Monitor)廠商提供，消費者只要買了「3D Ready」的顯示平台，根本就不用擔心要去那裡買3D眼鏡來搭配；就如同現階段，買了「3D TV」都會附贈有3D眼鏡一樣。至於3D Games方面，ATI認為目前他們已有iZ3D與DDD的3D Driver(或Middleware)可以支援被動式3D眼鏡，未來ATI將會免費提供「Quad Buffering」技術給iZ3D、DDD、遊戲開發廠商與任何其他軟體公司，以支援配備有HDMI 1.4a或是Display Port 1.2之「3D Ready」顯示器與其主動式3D眼鏡；換句話說，「Quad Buffering」技術是ATI顯示卡與支援主動式3D眼鏡之「3D Ready」顯示器(如3D TV或3D LCD Monitor)中間的橋樑。
很明顯地，ATI支援3D電影與3D遊戲使用3D眼鏡的解決方案，其實就是架構在現行既有的「3D Ready」顯示平台上，它只提供「Quad Buffering」技術並扮演橋樑的角色，這就是ATI在今年Game Developers Conference (GDC)所提出的「Open Stereo 3D Initiative」的構想，其最終目的就是希望藉由「開放3D」平台來吸引各專業之資訊廠商加入其陣營，大家良性競爭可以為消費者創造更多的選項，而結果ATI可以大賣其顯示卡(因為「Quad Buffering」只支援ATI顯示卡)。如此一來，對於採用專屬(Proprietary)封閉式規格的「nVidia 3D Vision」肯定會造成不小的衝擊。
3D films and games with glasses from ATI before Christmas
Soon NVIDIA will no longer be alone in offering a 3D film and gaming solution with stereoscopic glasses. ATI's alternative is on the way and pulls the rug from under NVIDIA's 3D Vision system. In contrast to NVIDIA's proprietary system it proposes an open standard. What's more ATI seem to think the future of 3D lies more in films than gaming.
Prorietary vs Open
A proprietary solution is limited to the technology of a single company. With Nvidia 3D Vision, you need Nvidia glasses, an Nvidia graphics card and a display (monitor, TV, projector) that has been validated by Nvidia for everything to work. A parallel in the software world is Microsoft with Windows.
On the other hand, an open standard is free and community based. ATI is promoting this solution. The company hopes to be able to work on 3D standards (in particular HDMI 1.4a and DisplayPort 1.2) and open their solution to anyone who wants to participate. You will nevertheless need an ATI card though! To draw another parallel with software, here the idea resembles Linux.
Updated: July 22, 2010
Two US representatives of AMD were happy to respond, Shane Parfitt (SP), Product Manager Infinity and Stereo 3D and Terry Makedon (TM), Product Manager of Catalyst.
|Shane Parfitt||Terry Makedon|
Here are our questions in bold and the answers we were given.
What is ATI's current position regarding 3D? Do you already have products on the market?
SP: There are 2 parts to the answer of that question. The first is that we do already support stereo 3D in some fashion. What we support, have been doing so for a number of years, is what's considered the older technology, solutions that involve passively polarised glasses and line-interleaved displays and checkerboard displays. The recent 3D hype has been around 3D shutter glasses.
3D technology with active glasses is indeed the most popular today.
SP: So the response to that is that yes we are working on it and we do want to enter this market. This year at the Game Developers Conference (GDC) in San Francisco, we announced what we call an open 3D initiative. This signifies that we are interested and do want to be involved in this market. With Open 3D we are building an ecosystem of partners that share our open approach to this market. So whereas other people are pushing proprietary or closed standards, we want to develop open standards.
Does this mean you won't be offering a proprietary solution like Nvidia, but will be launching an open technology?
SP: That's correct. The open standards that we're looking at supporting are the HDMI 1.4a and DisplayPort 1.2, which both have Stereo 3D support built in. With respect to glasses, we're supporting the idea that the glasses can be synched to the display or driven by a USB emitter connected to the host (computer). We're not trying to reduce options down to one solution. We want there to be competition in that respect. We are however working on the idea that glasses can be synched up to the display and display vendors can supply those glasses.
This means that in theory, ATI technology could be compatible with Nvidia glasses?
TM: No, they block it. The difference is that Nvidia has glasses, monitors and software that will only work with Nvidia equipment. Their package costs $200 for the software and $500 - 600 for the monitor. Our approach is that you will never see a pair of AMD glasses. You'll see glasses and monitors from many other companies with a panel shipped with those displays that simply says: Works with ATI/AMD.
When will we be seeing this??
TM: We can't talk about exact dates but we believe it's going to happen in the second half of this year.
SP: We don't have permission to announce the exact date but it's going to be in the next few months.
Right now, if I have an ATI graphics cards, what are my options for gaming and watching films in 3D?
SP: It depends on what you want. Today, you can do the older technologies with line-interleaved displays that are already on the market. Those are considered low-cost 3D options.
TM: You can go to a company called DDD.com (Dynamic Digital Depth). They sell a software package for $50 which a user can download today for use with anaglyph glasses, one blue and one red, which are available for free and you can play games and watch films with that software.
How about iz3D (Editor's note: who also offer ATI compatible solutions)?
SP: iz3D is another segway into this discussion of middleware. Part of our open stereo initiative is working with these middleware companies to provide that gaming experience. Now these companies already support the anaglyph technology and line-interleaved displays with passive glasses. We've been working with them for some time to give them access to our drivers to provide those formats into our drivers and graphics hardware. In the future we are also going to work with them on what we call quad buffering. The quad buffer is a key element for the new technology that supports shutter glasses. We have given them access to the quad buffer that we're working on in our drivers. So when we launch our updated stereo 3D solution then all these companies can be interoperable with our drivers and our quad buffer.
If I already have an ATI card, will it be compatible with this technology or will I have to buy another card?
SP: It depends on how old your current card is. We have planned support on existing cards for the updated solution but that gets cut off at a certain line.
Which generations will be compatible?
TM: We definitely expect that the 5000s will be compatible and we expect that mid/high level 4000s will also be compatible. We don't know this for sure yet. We would not be interested in releasing a solution that would stutter or drop frames.
|Generic ATI Radeon HD 4870 X2 2 GB||Generic ATI Radeon HD 5850 1 GB||Generic ATI Radeon HD 5870 1 GB|
Some people believe that although Nvidia currently communicates a lot on Stereo 3D, in the end the future will be with 4K. What's your point of view on this?
SP: I guess it comes down to a personal opinion of Stereo 3D and where you think the market is going. We do believe that Stereo 3D can add to the user experience and the realism of games and films. The graphics hardware that we and even our competitors have now is definitely powerful enough to drive a very good 3D solution, but the question of whether this will have a mass market appeal comes down to the end user and whether they will accept this technology and invest in it, pay the premiums to invest in it in their homes. It all kind of hinges on whether they'll be comfortable wearing the glasses.
TM: The other thing that I would add to that theoretical question is that from what I've seen from our customers, and by that I mean the big OEMs (Dell, HP, Acer and so on), is that they believe that the future of Stereo 3D is more in film content than gaming. We have not seen a huge amount of mainstream gaming interest in 3D but we certainly have seen a lot of - possibly Avatar related - Hollywood type movie interest in 3D. So I predict that OEMs will be shipping a lot of systems for netbooks or desktops that are labelled as being 3D video capable.
Does this mean that your solution will also be compatible with televisions?
TM: As long as the television is HDMI 1.4a, 100%.
SP: Once again, we're going after that open standard. So if the 3D television supports the HDMI 1.4a standard, and most of them will, then it will be compatible with our solution.
Are you currently working with games developers like Nvidia is with its The way it is meant to be played programme?
SP: We don't have a specific marketing programme like Nvidia's The way it is meant to be played for Stereo 3D. Right now our engagement with games developers is behind the scenes. We have a team of people whose specific job is to help games developers. Again we've already talked about middleware for conversion of games into stereo 3D and the next step is to get these games supported natively. This is the same type of engagement where we're giving them access to our quad buffer, so that they can add Stereo 3D to their games natively. That's how we're supporting the ISPs at the moment.
TM: More importantly that's for future games. The games of today don't need such support to function. We just need to engage with the middleware companies such as iz3D and DDD, to make sure that our drivers work well with their software. It's very important to understand that what the user needs to do is get this software from these third-party companies.
Now, Nvidia controls the process from the hardware though software to the monitor and yet still sometimes has implementation problems. Do you not think that your solution, where you only control the graphics hardware and are dependent on third-party middleware, might lead to additional complications?
TM: I would argue the exact opposite. I would argue that Nvidia is not a software development company. Nvidia is a company that sells graphics cards and they do all the additional developer work to enhance sales of their graphics cards. I would argue that all these other companies whose business and profits depend on the creation of software, will do a better job. We're leaving it to the experts. We're not going to create this software and try and sell it to users. We're going to let these guys compete. There's two of them right now but there could be ten in the future. One thing that I can say though is that our total cost of ownership, the total cost of our solution will be cheaper than Nvidia's.
So pricing is one of your main selling points?
TM: Absolutely, we will be a lot cheaper.
Do you have anything else to add?
SP: Our solution is based on an open standard. This means that we're basing our project on HDMI 1.4a and DisplayPort 1.2. Then there's also the possibility of proprietary DVI monitors that work with a pair of that company's glasses. We're open to that. That monitor would be a specific monitor built to work specifically with our graphics cards.
It's a shame that the solution won't work with current 120 Hz monitors/projectors using the DVI port as this hardware doesn't have HDMI 1.4a or DisplayPort 1.2 support..
TM: Again, these monitors are proprietary. We're not saying that there can't be or won't be any proprietary DVI monitors. I'm talking about being able to purchase any monitor in the store and plug it in. If you've got HDMI 1.4 or DisplayPort 1.2 then this is what you've got.
TM: Any HDMI 1.4 or DisplayPort 1.2 monitor will work. For DVI monitors it depends on the vendors. We're not going to build any and put our sticker on it, like Nvidia does. If any of these companies want to do that, then they certainly can. A very good test of this market is to see if big companies such as Samsung, ViewSonic, Hitachi and others build such monitors. I don't think they're going to. Why, because they're not selling too many of them. The one question you can ask and I don't have the answer, is: How many of these packages with glasses has Nvidia sold? I believe the answer is not very many.
I have one myself.
TM: But you got it for free I bet.
No, I bought it to see how 3D effects would work over the long term.
TM: As a user what do you think?
For now, the lack of films limits the experience to gaming. When games are well made, such as Batman Arkham Asylum, Metro 2033 or Just Cause 2, immersion is a real plus and very nice to use. When this isn't the case and you come across bugs, it becomes very annoying.
TM: I think that the really important thing for the success of gaming in the future is that games be made for 3D. Just like Avatar was built for 3D. These games will need to be built for 3D, especially for scenes where you've got the overlay in front of you and that objects float in the air and the effects add to the enjoyment of the game. Current games that are made for PS3 or Xbox and that someone adds 3D effects to aren't bad but aren't mind blowing or incredible.
|Batman Arkham Asylum||Metro 2033||Just Cause 2|
TM: As Shane was saying, the cost of it is incredible. If you have to go and buy a special monitor, glasses, software, the market becomes very small very quickly. If you already have a monitor on your desk and you can get a solution for $50 - $60 - $70, then that looks much more attractive.
SP: And then if you have proprietary solutions that only work with certain graphics cards, you further cut the market in half.
I really like the idea of working on an open technology. That said, sidelining current owners of 120 Hz products by concentrating support on HDMI 1.4a and DisplayPort 1.2 is problematic.
SP: You know, with the 3D Vision programme, those monitors are locked out of working with ATI graphics cards. Nvidia locks out any cards other than theirs.
TM: There's a pin on the DVI connector which sends the signal to check the presence of an Nvidia graphics card. This is why those monitors won't work with us.
So you're saying that this block isn't simply at a software but at a hardware level?
TM: Absolutely yes. If you look at the DVI connector, you'll see one of those pins. A DVI connector has 11 pins. Actually only 4 or 5 of them are needed for the image information. The others are for optional data transfer (Editor's note: the number of pins given by Terry is approximate and no doubt concerns a single link DVI. For 3D connection with a 3D Vision kit, you need a 24-pin dual link DVI connector). That's the only way they can justify that $200 package for the glasses. If it was open to competition, the price of the package would drop very quickly to $40 or $50.
I hope that it will when your solution comes onto the market.
TM: I guarantee you that it will. And if it doesn't, then you can call me one year from now and we can have another discussion.
OK, I've made a note of that!
Thanks again to Shane Parfitt and Terry Makedon for having replied so clearly to our questions. We're impatient to see the results of this development!