iZ3D are the makers of the popular iZ3D 22” 3D monitor, and have made waves with their recently announced working relationship with AMD. Vadim Asadov is the Chief Technology Officer of iZ3D LLC, and he shares some insights on their upcoming driver development and company plans.
1. It’s been a very exciting year for iZ3D, and a lot of it is credited to your well received stereoscopic drivers. For those unfamiliar, can you elaborate on what they do?
Do you have anybody here who is unfamiliar about iZ3D driver??? :-)
Our driver does two simple things:
Generates left and right image from game or application.
Converts them to different type of 3D outputs.
Of course, there is a lot of complicated mechanics inside the driver, like hooking DirectX calls, making double rendering, shader analysis, 3D settings control, auto-focus etc.
2. Can you tell us about some innovations you’ve implemented to make these drivers successful?
Features are not the reason of our driver success. The reason is that we put gamers approach as a base of driver development. Everything we are doing and we are adding is just idea of our guys who are gamers, and our customers who are gamers too.
Let me share with you what we are going to add in to 1.10. First, we corrected a few bugs in racing games, Red Alert, and some others. Shader analysis mechanism is improved and will take less space in memory and will work faster (we are doing benchmarking now). Profiles list will be sorted alphabetically as requested by many users. A new stereoscopic output will be added for free (left and right vertically split on one screen). Clipping control will come to Control Center, and user can turn it on/off and define width of black stripes. We’ll add server for statistics file that will help our tech support to work with users in case of problems.
We are going to add a 3D desktop function that will make a 3D wallpaper. And 3D wizard will help you to reach best settings in each and every game acting in-game and see results of your action on the fly. Interlace output will be optimized to increase the quality by using better interlacing methods.
3. Are there some consistent challenges you have been working to overcome, and is there hope in sight to overcome these issues? Do you have some ideas?
Driver is living in-between of games, operational system and graphics card - this is very complex environment. Because our driver is universal and we are serving all Windows OS and all GPUs and almost all DX8/9 applications, this is our biggest and constant challenge. At technical level two biggest issues we have are shadows and external software conflicts (anti-cheat, anti-virus sometimes). We assume our new architecture we are going to have in version 2.00 will help to overcome these issues.
4. When can we expect to see DirectX 10 support introduced? Will we have some working samples in the near future?
DX10 development is started and first development iteration is done. That means that we have version which can serve standard set of DX10 sample which are provided as part of DX10 SDK. These samples illustrate some specific functions in DX10 and all games are combination of such functions. Planned DX10 development has five iterations, thus few months from now can be considered as realistic.
5. How are you finding development compared to the DX9 model? Is DX10 driver development more complex?
DX9 was our first development and we walked through the dark with errors. We made a lot of code re-factoring since that time to improve code quality. For DX10 we have bigger and more experienced team (we doubled amount of our developers). We also implemented new development procedures. Finally, even DX10 itself is new, planning was done much better and as an example first iteration was finished one week earlier than was planned.
6. Your drivers currently support both AMD and NVIDIA GPUs. However, I think it’s fair to say that you dipped your toes in the red more than the green by going on a press tour with AMD to promote stereoscopic 3D gaming. Are you still going to support NVIDIA graphics cards?
Supporting both GPU is absolutely necessary from my point of view. 3D market is young and ANY barrier like "AMD only" or "nVidia only" is not good for market development. We may have business benefits (or economic benefits) for AMD or nVidia users, BUT at basic level both GPU has to be supported at 100% level.
7. What technical challenges would you like to see overcome through your working relationship with AMD? Do you think you will be able to accomplish the same benefits out of NVIDIA GPUs?
First of all we hope that new architecture of our driver will help us to optimize everything having tight technical relationship with AMD. Current architecture of our driver is made, let call it "non game" way. This means that AMD activity to optimize game running will not affect our driver. With new architecture we hope that every optimization at game level will give us optimization at driver level as well.
Another point I need to underline is corporate segment. As you know our driver can serve OpenGL Quad Buffer mechanism. If AMD will provide QB support then I think FireGL line of cards can join to corporate 3D segment.
8. Non-DLP LCD Shutter glasses are important to several vocal MTBS members. They just love them. The same holds true for head mounted displays that are dependent on pageflip technology (e.g. Vuzix VR920). You have tried to support this through your drivers, but there is a synchronization limitation that is preventing this from working properly when the frame rate gets too low. Do you think your relationship with AMD can impact this?
I hope so. Question is how long it will take? This is question of rendering architecture or an access to GPU internal operation. Having in mind that we are thinking about 1.10 alpha before the end of 2009 I don’t think AMD will help here except we’ll invent way around this issue.
9. Vadim! Do you use a webcam? There have been some webcams coming out that support stereoscopic 3D streaming. I’ve only seen anaglyph so far. What’s your take on an S-3D webcam working with the iZ3D monitor?
Yes, I have webcam on my laptop (Acer Aspire 7520G) and iZ3D connected to it, but it is not 3D webcam :-( We talked with Minoru about possible cooperation. I hope these guys will bring their product to the market soon. We have a lot of applications in mind :-) Good thing is that if we understand correctly Minoru will work with iZ3D MPC automatically without any problem.
10. How do you think people will respond to an S-3D webcam? Will we need special settings like low convergence for bad hair days, and high convergence for good hair days (convergence refers to the pop-out or out of screen experience)?
It is hard to predict the future but search for "xxx webcam" :-) I sure people will find the best way – I can imagine 3D webcam worldwide doing geographical web broadcasting as an example, 3D Youtube and many other interesting things.
11. While well received, iZ3D is working hard to further reduce the ghosting or crosstalk between the eyes with its monitor. How are things progressing, and what does Russian customs think of your new glasses?
Custom rejected them finally, but you know we always can invent something to solve the problem :-) We made internal test in US and it was kind of positive. We hope to make one more in Russian lab this week and make more samples for external beta test then. If external test will be fine we can start mass manufacturing.
12. I don’t know about you, but I get very frustrated by some S-3D hardware reviews because the critics aren’t using the equipment properly. It’s a challenge because grasping the concept of how stereoscopic 3D works can be difficult for some. MTBS has had success with our own non-proprietary settings guide for first timers (when the guide is read!). Do you have some ideas of your own on how to overcome these challenges?
Yes, you are right - this point is very important, especially having in mind mass market as our biggest goal. We made special 3D wizard (it is close to final test now). This wizard will act as overlay to help users to find best 3D settings for each game. So, you can play the game and 3D wizard will help you to adjust 3D setting depending on content you have on the screen. I hope it will be part of 1.10 release.
13. Let me paint you a picture. The U-DECIDE Initiative has been very successful so far in capturing the opinions of gamers and consumers about stereoscopic 3D technology. It has been actively promoted to gamers and consumers by AMD, iZ3D, MTBS, The Game Creators, and Blitz Games Studios. It has also been featured by countless media outlets including gamesindustry.biz, The Inquirer, Gamecyte, and several more. The initiative captures opinions on everything from 3D in video games to 3D in feature film directly from customers. The industry will be watching as MTBS reveals its preliminary results at the 3D Entertainment Summit on December 1st - a summit that has James Cameron as the keynote. Do you think the industry is truly prepared for what consumers have to say? Why or why not?
I don’t think we need to consider the industry - it depends on each specific company and/or group in industry. As businessman I think that talking with customers is only way to make business in modern world. And "talking" for me is two-way communication. If industry will not talk with customer it will die before it’s born.
14. At the 3DX Film Festival in Singapore, Jeffrey Katzenberg predicted that in 5 to 7 years, all movies will be made in 3D. Do you agree with this? Why or why not?
Let say this way: I can imagine this. But let’s consider this: not all movies have significant 3D part (like drama) then only reason to watch them in 3D would be just natural feeling, but you have 3D glasses ... thus you need so light and comfortable 3D glasses to keep natural feeling from movie non destroyed. Big work for chemical and plastic companies :-))
15. I understand you have some additional company information to share.
I’d like to share the news that all team improvement we made was also related with investment deal we closed recently. We got investment from Russian biggest investment bank Troika (http://www.troika.ru/eng/start.wbp). Artyom Yukhin, the leader of Troika fund, is also 3D related person - he is creator of most successful 3D face/body recognition technology. 3D scanners of his company Artec (http://artec-group.com/applications/broadway-3d-scanner-2.html) are another side of 3D market.
16. Some of our members have important questions they would like to ask you. Here goes!
Budda: I wanted to ask about the rendering of the actual output image for S-3D, and whether drivers can play a part with this suggested idea. It seems each pixel of the display is normally rendered as if it is a single point within the scene. Instead, is it possible for each RGB subpixel to be displayed as distinct red, green or blue points in the scene - as opposed to merely being displayed as the RGB components of a single pixel point. This may subtley assist in the 3D visualisation of the scene, as it would create extra image detail at the same resolution of the existing display.
Actually we work at subpixel level and it is not clear for me how we can improve 3D using idea described above
Yurithebest: Will you continue to manufacture the 22" iZ3D monitor when the 26" model is released?
yes - question is how long? If depends of sales for both models.
I. Would stereoscopic-aware 3D applications benefit from having real-time control over separation, convergence, mono trigger etc. at the API level, or is it ok for them to simply follow basic conventions outlined in available stereoscopic programming guides?
I assume having API is always good thing. We are going to make it for our driver, but it takes time and we are under constant pressure of new version of driver releasing. I hope we can do API in 2009 with our new bigger team.
II. Would there be any benefit of providing application developers with an industry-wide application profile or manifest format that would replace OEM-specific driver profiles?
May be, but it is always business issue first and technical issue them. We are talking with few big application developers, but first question for them how many additional money we can make selling 3D compatible apps.
III. If Direct3D API/runtime were stereoscopic-aware, would it allow for optimizations that are not possible to implement at the driver level?
I think so, it can be different form of menu, special form of interface, 2D over 3D elements.
IV. Have you tried to approach Microsoft/Khronos and talk about supporting stereoscopic at the core level of the API?
Yes, we did - but MS planning cycle is much longer than our cycle and we have not chance to be in OS 7 or DX11 :-(