Slashdot Mirror


NVIDIA Shows Interactive Ray Tracing On GPUs

MojoKid writes "During SIGGRAPH 2008 in Los Angeles, NVIDIA is demonstrating a fully interactive GPU-based ray tracer. The demo is based purely on NVIDIA GPU technology, and according to NVIDIA the ray tracer shows linear scaling during rendering of a complex, two-million polygon, anti-aliased automotive styling application. The article reproduces screenshots from NVIDIA's demo. At three bounces (rays being traced as they bounce three times through a scene), performance is demonstrated at up to 30fps at HD resolutions of 1920x1080 for an image-based lighting paint shader, ray-traced shadows, reflections and refractions running on four next-generation Quadro GPUs in an NVIDIA Quadro Plex 2100 D4 Visual Computing System." Meanwhile reader arcticstoat passes on Intel's latest claim that rasterisation will die out the next few years, possibly in favour of ray tracing.

260 comments

  1. Beautiful by AKAImBatman · · Score: 5, Interesting

    Wow, those screen caps are gorgeous. I hope this finally puts to rest the idea that rasterizing with upteenth number of features added in can compete with the image quality of Raytracing. While rasterizing may have a number of competitive features, it's hard to get the same level of specularity, reflection, shadows, shading, and other features so nicely demonstrated by this demo.

    The genius of what NVidia is doing here, I think, is that by using their existing GPU architecture, they create a path by which Raytracing can be phased in as a technology without removing the support and investment in current rendering pipelines. This is a bit different from Intel's goal, which appears to be a cutoff between the old and the new.

    Another interesting point is that this demo is currently capped at 3 casts per pixel. Which means that the scenes shown could look even better than they already do. Shadows could be softer, reflections could be more complex, and inventive scenes could be created to make for interesting styles of gameplay. (e.g. Fighting in a hall of mirrors.) If 3 casts/pixel is the baseline, then NVidia is setting up a vast new territory for graphical improvements. Each increase in casts/pixel will increase the realism of the scene. Thus graphical quality becomes a matter of raw horsepower. A market that I'm sure NVidia would gladly be interested in opening up.

    Funny how things change, eh? :-P

    Actually, I doubt NVidia has changed its position by very much. They're probably making a smart business decision and ensuring that they ride the wave of Intel's hype. If Intel *does* succeed in convincing the market that Raytracing is the future, NVidia will be ready to compete rather than cede the market.

    1. Re:Beautiful by Colonel+Korn · · Score: 5, Interesting

      They're probably making a smart business decision and ensuring that they ride the wave of Intel's hype. If Intel *does* succeed in convincing the market that Raytracing is the future, NVidia will be ready to compete rather than cede the market.

      It's great for nVidia that they can do this with their chips, but I don't think this was done primarily for tech purposes. I think you're close to the truth when you say they can ride Intel's hype, but not quite spot on. I think this is meant to break Intel's growing ray tracing hype machine, not come along for the ride.

      "Look, we can do now what you say you'll do in two years, and we can do it WAY better than you will be able to then, but on our current tech."

      I can't imagine anything could be more effective at ending the "Intel will crush nVidia with ray tracing" meme that's been affecting nV stock.

      --
      "I zero-index my hamsters" - Willtor (147206)
    2. Re:Beautiful by qoncept · · Score: 3, Funny

      I suppose, but that guy could certainly use some tire-wet. Those sidewalls look awful.

      --
      Whale
    3. Re:Beautiful by WilyCoder · · Score: 2, Interesting

      "Another interesting point is that this demo is currently capped at 3 casts per pixel."

      You mean three bounces per pixel.

      Also, from TFA:

      "running on four next-generation Quadro GPUs in an NVIDIA Quadro Plex 2100 D4 Visual Computing System"

      Sure this is impressive, but they are pushing kilowatts to get this kind of performance.

    4. Re:Beautiful by hr.wien · · Score: 4, Interesting

      They are gorgeous? Really? I think they look distinctly average. The lighting calculations look very simplistic. Yes, the shadows and reflections may be pixel perfect, but that just doesn't matter that much. You usually can't tell they are anyway. The same scene rasterized with a simple cube map for the car's reflection and some proper shadow maps would look much better. Not to mention run faster.

      And "graphical quality becomes a matter of raw horsepower"? This unlike in rasterization then?

    5. Re:Beautiful by ChronoReverse · · Score: 4, Insightful

      What on earth? How are those overly shiny objects beautiful in any way?

      The technology is probably better than that but the actual screenshots are distinctly ugly for this day and age.

    6. Re:Beautiful by AKAImBatman · · Score: 3, Informative

      Yes, the shadows and reflections may be pixel perfect, but that just doesn't matter that much. You usually can't tell they are anyway.

      Sure you can. A human instinctively knows when something looks "right" or "wrong". And one of the reasons why rasterization is capped is due to lighting problems. Lighting technology has improved significantly in the last decade, but still not sufficiently to compete with raytracing. Raytraced lighting will look more natural to an untrained viewer.

      And "graphical quality becomes a matter of raw horsepower"? This unlike in rasterization then?

      Rasterization is heavy on hardware features to improve the quality and performance of the scene render. e.g. Blending, pixel shaders, z-buffers, etc. Ray tracing is a far simpler operation on the hardware side, though it still behooves the software side to improve the number of objects tested for rendering. (Nothing new there.)

    7. Re:Beautiful by mypalmike · · Score: 4, Informative

      Every bounce casts a new ray, so "3 casts per pixel" is an accurate description.

      --
      There are 0x40000000 types of people: those who understand 32-bit IEEE 754 floating point, and those who don't.
    8. Re:Beautiful by Stalus · · Score: 1

      Shadows could be softer

      They only mention three bounces per pixel, and do not mention samples per pixel. They also don't give any detail about their light sources. If they maintain point light sources, the shadows will always be hard no matter how many bounces or samples they take. If they allow area light sources, and sample it properly, they can do that with a single bounce. And you can do some pretty nice stuff with three bounces.

    9. Re:Beautiful by Enderandrew · · Score: 3, Insightful

      Intel's last demo was running on 8 GPUs wasn't it? On those were GPUs designed for ray tracing I thought.

      I like Nvidia's approach to use existing architecture, and I agree with the poster above who says this is a much better method for consumers.

      I disagree however with Intel saying rasterization is dying any time soon. Intel and Nvidia can't produce these effects with reasonably priced hardware, and even when the hardware becomes affordable, we still need games designed for this, and then a few years for the technology to be accepted by the masses.

      I say rasterization sticks around 3-5 years.

      --
      http://blindscribblings.com - Tasty pop-culture in conceptual fashion.
    10. Re:Beautiful by Anonymous Coward · · Score: 0

      Yes, the ultimate in image quality is clearly the domain of ray tracing. That's why all the film vfx guys used ray trace based renderers.

      Oh wait...

    11. Re:Beautiful by BiggerIsBetter · · Score: 2, Insightful

      They're nice work, but c'mon, how about some demo VIDEO instead? I'd love to see the full effect of the reflections in the Windows and chrome of the wheels, and the way the lighting moves...

      --
      Forget thrust, drag, lift and weight. Airplanes fly because of money.
    12. Re:Beautiful by Anonymous Coward · · Score: 0

      What on earth? How are those overly shiny objects beautiful in any way?

      You seriously ask that when some of the most popular games in existence feature overly shiny space marines? I mean, SHINY SPACE MARINES? You're worried about the shininess of a car? Did I mention, SHINY SPACE MARINES. As in, the guys you'd expect to be grimy, sweaty, and loaded with cloth. But they're not. They're SHINY.

      TFA shows a car. A CAR. That's shiny. And yet you're going to complain that it's too shiny? Yeash.

    13. Re:Beautiful by hr.wien · · Score: 5, Interesting

      Sure you can. A human instinctively knows when something looks "right" or "wrong".

      Yes, and amusingly reflection and refraction (two of the main benefits of ray tracing) are a couple of things that humans generally can't tell if are being fudged. As long as it's in the ballpark it's enough to fool the human eye. Pixel perfection is way beyond what's required.

      And one of the reasons why rasterization is capped is due to lighting problems.

      What does that mean? Capped?

      Lighting technology has improved significantly in the last decade, but still not sufficiently to compete with raytracing. Raytraced lighting will look more natural to an untrained viewer.

      Why? Details please. What exactly is more realistic about tracing each pixel through the geometry than drawing the geometry directly in the appropriate pixels? The underlying lighting calculations are the same either way, meaning they will both look the same, so the only real concern is speed for any given scene.

      So if we look at speed ray tracing only has a real benefit in reflection and refraction, but that's not really a winning argument because, as I said, people can't really tell if it's 100% accurate anyway. I sure can't. Ray tracing can also do accurate soft shadows relatively easily, but the ray count required makes that completely unrealistic in real time for the foreseeable future. Shadow maps will be faster either way, and look 99% as good.

      Ray tracing is simple to implement, but so is cracking a password using brute force. That doesn't make it the best solution.

    14. Re:Beautiful by Austerity+Empowers · · Score: 4, Insightful

      I think if you look closely you'll see that they used materials very sparingly. The man behind the curtain (IMO) is that they're dedicating all their GPU and memory bandwidth to ray tracing computations, at the expensive of traditional raster manipulations.

      Who cares? Well, I think if you're playing a game where you are free to run where you like, you may care.

      I agree, nVidia is showing that ray tracing doesn't scare them at all. And when it's ready to happen, it will. I disagree that it's ready to happen any day now.

    15. Re:Beautiful by Anonymous Coward · · Score: 5, Informative

      The screenshots look relatively ugly because of the hashed-together-demo quality of the environment textures. But it's not a texture demo, it's a raytraced lighting demo.

      Bare in mind this is ray tracing at a very rough and ready stage, but the potential is enormous. If you want to see the sort of effects it can achieve, check out some professional 3DSMax/VRay renders.

      There's a nice render here for illustrative purposes.

      That's just a single frame with high quality textures, but it surely shows the potential.

    16. Re:Beautiful by Oasiz · · Score: 1

      Cmon, can't you see? You can live those 'glorius' 90's pre-rendered backgrounds in realtime ! ;)

    17. Re:Beautiful by jacquesm · · Score: 4, Informative

      It isn't. More casts per pixel means that more rays are cast at slightly different angles through each pixel, and those are then averaged to yield the actual pixel colours.

      Three bounces per ray simply means that a single ray can bounce three times before it's colour values are known.

    18. Re:Beautiful by jacquesm · · Score: 1

      I think you're missing the shadows bit completely, and they make a *HUGE* difference in what is intuitively right and what is definitely computer generated. I have a hard time with some ray traced images to tell them from the real thing (usually they're just a bit too perfect), but rasterization, no matter how well done feels 'wrong' right away.

      Raytracing simply comes much closer to how things actually work and so has a much better chance of appearing natural.

      Once things start moving those differences reduce, but they're still quite present.

    19. Re:Beautiful by Thought1 · · Score: 1

      The genius of what NVidia is doing here, I think, is that by using their existing GPU architecture, they create a path by which Raytracing can be phased in as a technology without removing the support and investment in current rendering pipelines. This is a bit different from Intel's goal, which appears to be a cutoff between the old and the new.

      Intel tried the same thing with IA64, while AMD went the x64 route. Guess which approach will win this one, too? (:

    20. Re:Beautiful by Dekker3D · · Score: 1

      okay, i don't think those people so opposed to raytracing actually get what it means. please, do prove me wrong.

      anyway, raytracing current scenes is one thing. raytracing things too complex for a rasterization-based engine is another. we could have detailed scenes using all sorts of transparancy without worrying as much about z-fighting and other detail limits.

      then again, the only things i know about raytracing are based on non-realtime renders, so i may be way off.

    21. Re:Beautiful by phantomfive · · Score: 3, Insightful

      Yeah, exactly...it's like Web2.0 of the graphics world....sure, the buttons are shiny, but that doesn't make them look any better. That is why people like John Carmack have suggested that full on ray tracing isn't the way to go.

      I absolutely do not understand the issue fully, but here is my take: in the early 90s, ray traced graphics looked way better than anything else. You could render a ball and people would say, "wow, that is so cool!" it took 30 minutes to render, but it could be done.

      Since then, 3D rasterization has come a long way. With texture mapping, commodity 3D graphics hardware, pixel shading, alpha blending, etc, we have games that look really, really good, without ray-tracing.

      Now ray-tracing is starting to become possible in real time, and I guess people are remembering how good it looked in the 90s and thinking it must still be the holy grail of graphics. In theory it's a good idea, render everything the way real light does.

      The ultimate question has to be: does it look better? Or is there another way we can use that processing power that will make the graphics look even better? My guess is that ray-tracing is a technique that will be useful in some ways, and will be mixed with techniques we already have now. Much like today we use 2D texture maps on 3D objects, and it looks good.

      --
      Qxe4
    22. Re:Beautiful by samkass · · Score: 1

      One thing I noticed is that the curves and shiny surfaces look really, really smooth but that the straight lines look aliased and fake. The railings look like they were drawn in with a line tool.

      --
      E pluribus unum
    23. Re:Beautiful by hr.wien · · Score: 2, Insightful

      The reason ray traced images often look very realistic isn't because they're ray traced, but rather because they are done offline and as such can take ages to do the calculations required for a realistic lighting model. What can be done given enough time doesn't matter. What is efficient enough to be feasible real time is.

      Raytracing may be "closer to simulation", but that's completely irrelevant when the performance isn't there for the quality of the output. Can you honestly tell me you think the shadows in Nvidia's demo look better than the ones in, say, Crysis?

    24. Re:Beautiful by jacquesm · · Score: 1

      I think the answer is yes, take a look at the mirror in the green car demo shot (bottom right, at high res), that's a pretty impressive following of the contours of the car by the shadow, a 'straight' computed shadow would never ever do that, compare this shot:

      http://www.gamershell.com/static/screenshots/9451/264393_full.jpg

      from crysis, where the shed shadow meets the terrain contour.

      That's exactly the kind of thing I meant.

    25. Re:Beautiful by mrsteveman1 · · Score: 3, Funny

      Well, once they perfect digital eyelash rendering, I'll be sold.

    26. Re:Beautiful by mopower70 · · Score: 0, Offtopic

      No... pussy LIKES shiny tires. And whatever the pussy likes, the pussy gets.

    27. Re:Beautiful by hr.wien · · Score: 1

      I'm not "opposed to raytracing". I'm opposed to the commonly held belief that raytracing automatically equals realistic graphics and that it somehow scales better than rasterization. It may if you completely ignore acceleration structures, but if you do that both ray-tracing and rasterization will run at non-realtime framerates for any moderately complex scene.

      What matters is what gives the best result using the available computing power. Right now that is rasterization, and it will stay that way until the hardware is fast enough to make the extra work usually involved in making rasterization work not worth the effort in terms of quality.

    28. Re:Beautiful by Hairy+Heron · · Score: 1

      Oh wait...

      That must be why most of the 3D CGI in films sticks out so much.

    29. Re:Beautiful by AKAImBatman · · Score: 2, Interesting

      What is efficient enough to be feasible real time is.

      Yet in the case of Raytracing, "efficient enough to be feasible" means real shadowing, not the shadow-map crud we see in most video games.

      Can you honestly tell me you think the shadows in Nvidia's demo look better than the ones in, say, Crysis?

      I can and I will. Shadow maps used in games like Crysis beat you over the head with depth perspective. i.e. "Look, there's a shadow! Now you know how far off the ground the helicopter is!" While that's nice and all, only the objects that you map and define to have shadows actually cast a shadow. The subtle interactions of the environment and the actors are usually lost. It's almost like someone took a bunch of stage props, covered a stage with styrofoam sand, then threw all the props on top. Thus the actors stand out from their environment.

      With ray tracing, the shadows are correct. Period. If a monster picked up that NVidia car and threw it through the air end over end, you'd see all kinds of shadow interactions happening that simply won't be visible on a shadow-mapped engine like Crysis.

      Heck, you don't even have to get that complex. The car casts shadows upon itself. Which already makes it look more realistic on a closeup than anything an existing game engine can do.

      In addition, ray tracing can scale a hell of a lot farther when it comes to realistic lighting and shading. With more rays not only comes softer shadows, but multiple shadows, reflective surfaces, and other "realistic" shadowing techniques that can't be touched by today's shadow map engines.

    30. Re:Beautiful by hr.wien · · Score: 1

      The same scene using shadow maps would look more or less the same, but in addition you would have compute power left to actually make the shadow fuzzy like it is in real life. The hard shadows you get in ray tracing without sufficient light traces look anything but realistic. Especially in outdoor scenes.

      I'm not sure I understand what you mean about the Crysis pic though. Looks fine to me?

    31. Re:Beautiful by Phat_Tony · · Score: 1

      Funny how things change, eh?

      I think Nvidia would have preferred a different previous naysayer for you to link to.

      I heard the title of their talk at SIGGRAPH was:
      "Practical Real Time HD Raytracing With Nvidia Hardware: Suck It, John Carmack."

      --
      Can anyone tell me how to set my sig on Slashdot?
    32. Re:Beautiful by hr.wien · · Score: 1

      Heck, you don't even have to get that complex. The car casts shadows upon itself. Which already makes it look more realistic on a closeup than anything an existing game engine can do.

      Bullshit. Crysis has self shadowing, just like any other moderately recent game. The screenshot you posted is in one of the lower quality modes in which it may be disabled.

      Crysis also has shadows from a number of light sources simultaneously, but the number is limited (4 IIRC) for performance/quality reasons. I suggest you look up what shadow mapping actually is before you spout off nonsense.

      Rasterization has moved forward since Quake 3 you know.

    33. Re:Beautiful by jacquesm · · Score: 1

      I think the fact that it looks fine to you is more or less where the problem lies :)

      The contour of the shadow of the shed looks completely unnatural to me (not to mention very poorly anti-aliased). I wouldn't bitch about it too much though, because for one it's a game and it was generated on the fly by hardware not to be compared to that of TFA.

      Sure, if you spent as much time as you wanted on that image using the very best of the rasterization algorithms without any optimizations for performance that would impact the image quality in a negative way you could make it (much) better, but it couldn't hold a candle in terms of photo-realism to a similarly unconstrained ray traced image.

      That would be comparing apples to organges though, because even at that level I'd expect the ray tracer to burn through at least an order of magnitude if not more computing cycles.

      The only way to really tell is to take the exact same scene and lighting model, run both algorithms, then compare the results side by side, and in those comparisions, ray tracing wins hands down, at the expense of a vast increase in computing resources used.

      From what I know (which isn't much) even high end shops like Pixar only recently started using ray tracing, before that it was all rasterization using procedural textures and fairly complicated lighting models.

    34. Re:Beautiful by ShenTheWise · · Score: 1
      Intel's hype wave is powerful indeed. Nvidia has never claimed that raytracing is *not* the future.

      This "new way" vs "old way" crap was entirely fabricated by Intel's hype machine as an attempt to polarize the audience, and the press can't help but the chug down the coolaid.

      The fact is that Nvidia has delivered great graphics for over a decade while Intel's *existing* integrated GPUs are still unusable after years of promises.

      I am not buying it. We will be playing 2015's games on an Nvidia (and maybe AMD) GPU.

    35. Re:Beautiful by AKAImBatman · · Score: 1

      Bullshit.

      Back 'atcha.

      Crysis has self shadowing, just like any other moderately recent game. The screenshot you posted is in one of the lower quality modes in which it may be disabled.

      It's not disabled. Look at the upper right corner of the ship, just to the left of the right wing. You will see a minor shadow cast by the ship onto the ship. Look at the superstructure to the left. The overhang is casting a shadow onto the lower part of the tower.

      Both look like crap.

      Crysis also has shadows from a number of light sources simultaneously, but the number is limited (4 IIRC) for performance/quality reasons.

      And you think the NVidia demo wasn't limited? The sun-lit effects of the NVidia demo don't look like there are very many light sources to me. Yet the results are easily superior. Reflections off of glass, internal shadows, internal reflections, realistic refraction, etc.

      Crysis still looks like crap in comparison.

      I suggest you look up what shadow mapping actually is before you spout off nonsense.

      I suggest you pay more attention to what I wrote. Existing game engines can't hold a candle to what NVidia has demonstrated here. Shadow mapping can do a lot better if you're willing to spend the cycles, but then you get into a computational frenzy that starts to make raytracing look pretty darn good.

      Rasterization has moved forward since Quake 3 you know.

      And Raytracing has moved forward since the days of Myst. What's your point?

    36. Re:Beautiful by Spatial · · Score: 1

      Raytraced lighting will look more natural to an untrained viewer.

      Looks like like every other raytracing demonstration I've seen: a bunch of reflective surfaces and mirrors that are impossibly clean and flat, with fairly poor textures. Most things in games aren't and don't need to be shiny, yet if you take the reflective surfaces out of this demo it would look awful. I'm not impressed.

    37. Re:Beautiful by hr.wien · · Score: 1

      I think the fact that it looks fine to you is more or less where the problem lies :)

      Okay then, please humour me. Which general area of the picture should I be looking at? There's a lot of shed.

      The only way to really tell is to take the exact same scene and lighting model, run both algorithms, then compare the results side by side, and in those comparisions, ray tracing wins hands down, at the expense of a vast increase in computing resources used.

      Whoo there. That depends highly on the scene in question now doesn't it? If you're talking about the ray tracer's dream of the reflecting ball on checkerboard, then yes, obviously. But if you take a normal real life scene, the result is far less clear cut.

      I also don't see why you need to impose the "same lighting algorithm" requirement. That will automatically bias towards one or the other technique. What matters is the resulting image quality (and to some extent the work needed to make it happen), not how it's achieved.

    38. Re:Beautiful by brunascle · · Score: 1

      keep in mind that's not a real crysis shot. for 1, that scene was never in the game. 2, it's dated a year and a half before it was released. 3, it looks like crap. look at the missiles at the left. that's not crysis.

    39. Re:Beautiful by hr.wien · · Score: 1

      It's not disabled. Look at the upper right corner of the ship, just to the left of the right wing. You will see a minor shadow cast by the ship onto the ship. Look at the superstructure to the left. The overhang is casting a shadow onto the lower part of the tower. Both look like crap.

      Ah, so when you said game engines can't do it, you really meant they can, it just doesn't look very good?

      Well, that is purely down to shadow map resolution and power limitations in the filtering of them. You'd hit the similar limitations using ray traced shadows if you try to do them in real time. Which is why Nvidia has used a single light source trace in their demo. Add more and it doesn't run fast enough. Do the shadows in Toy story look bad for example? They are shadow maps too.

      But either way, if you honestly think the shadows in the Nvidia demo looks closer to real shadows (not necessarily more accurate) than the ones in Crysis, I'll leave you to it. I strongly disagree though.

    40. Re:Beautiful by hr.wien · · Score: 1

      you really meant they can, it just doesn't look very good?

      Meh, going back to your original post I seem to have misinterpreted your statement the first time around. That was in fact all you said. My apologies. My point still stands though. For the quality, shadow maps are more efficient, hence yield more realistic results in real time.

    41. Re:Beautiful by nabsltd · · Score: 1

      Yeah, exactly...it's like Web2.0 of the graphics world....sure, the buttons are shiny, but that doesn't make them look any better.

      Basically, ray-tracing might be to games what Vista is to operating systems.

      It's prettier than the previous version, but the gameplay/OS functions aren't any better (and are sometimes far worse). There are so many games now that can quite comfortably be called "just another first-person shooter" that when something truly original comes out (e.g., Portal), it's huge.

      I can't recall anybody saying "well, Portal would have been good if the graphics had been more realistic."

    42. Re:Beautiful by circusboy · · Score: 1

      unfortunately more than one piece of glass is not one of them. for an example, 3 bounces means 1) from the camera to the near glass surface, 2) from the front of the glass to the other side of the glass, to 3) the next thing it hits. if the next thing it hits is a piece of glass, you're screwed, no bounces left.

      one thing to remember is that there are many places where the number of samples is limited. in most ray tracers, there is a certain number of samples per pixel, this can vary based on the anti-aliasing method. (one renderer I used to work on used to sample up to 256 times on a single pixel, best.anti-aliasing.ever. for each sample (ray cast) when the ray hit the first surface, there was minimum one ray per light in the scene, more if there were area lights, depending on the size. if the ray hit a reflective or refractive surface, the another ray (or set of rays) would be cast from that point. at which point another set of light rays would be checked for every light in the scene. and so on and so on.

      the difficulty with area lights in a ray tracer is that there is always noise. for example if a light is big enough to require say 4 samples minimum, which isn't very big, and it is casting a shadow, any ray that strikes a surface in the penumbra cast by that light may or may not see the full light source. it then averages the light value that it receives. in order for the light source not to cast what would appear to be 4 distinct shadows, it has to randomly vary the part of the light that is sampled. the possibility is that one pixel's sample could return 4 light samples all in shadow, the pixel right next to it could return 4 pixels in full light. you then have to anti-alias that discontinuity by one of a variety of methods, throw more light samples, throw more pixel samples, blur the resulting image, or some combination of the those (and others.) but you can see where the problem is likely to increase render times exponentially.

      most 3d renderers in the post production world (PRMan, Mental Ray etc. (though mental ray is a hybrid, and renderman is not a ray-tracer)) take a minimum of 4 samples in each pixel, one at each corner. good, (but certainly not all) ray tracers, in the interest of efficiency will fire a minimum of two samples per pixel, and then add more and more samples in problem areas. any discontinuity will cause more samples. texture discontinuity, lighting, edges etc. to answer Hairy Heron's remark below, it's usually the texture filtering done to prevent this aliasing that's the first subconscious giveaway to a 3D CGI in a movie. smooth gradients are generally the fastest things to render in a ray tracer that works this way.

      --
      -- it's ridiculous how many people misspell ridiculous... (damn, damn, damn...)
    43. Re:Beautiful by hr.wien · · Score: 1

      It shows what can be done given (relatively speaking) unlimited time. Unfortunately that has no bearing on what is feasible in real time, as the Nvidia demo handily demonstrates.

    44. Re:Beautiful by AKAImBatman · · Score: 1

      Ah, so when you said game engines can't do it, you really meant they can, it just doesn't look very good?

      Well... more or less. :-P

      More specifically, I mean that they can't internally shadow the way you'd expect real objects to shadow. When a soldier moves his leg far forward during a sneaky walk, you don't see a shadow of his weapon and arm cover his leg. That is realistic internal shadows. Placing shadows on obvious overhangs is just more beating the player over the head with depth perspective.

      Not to say that the artists don't do a good job. The models and textures have a lot of shading already built into them that hides the lack of, say, neck shadows. Similarly, the use of standard shading based on the angle from the light source hides many of the missing shadows. But at the end of the day, you simply don't have a number of shadows that should be there. And those missing shadows contribute to the instinctive "wrongness" of the scenes.

      You'd hit the similar limitations using ray traced shadows if you try to do them in real time.

      Yes and no. You'd miss a lot of shadows based on secondary or tertiary reflections (e.g. the sun bounces off the glass of the building, which then strikes the car, which creates interesting shadows as well as reflections), but the primary shadows would all be correct.

      Secondary shadows have issues anyway as dust from the air has "dirtying" effects that are computationally expensive to figure out.

      Do the shadows in Toy story look bad for example?

      Toy Story is a difficult comparison as they did not try to render in real time. Rendering was done on expensive render farms. As a result, you did get a lot better shading and shadow effects. (e.g. Woody's nose cast a shadow on his face.) None the less, these shadows were a long way from realistic. Scenes were far too strongly lit, with shadows missing all over the place. Shadows were added were it was artistically pleasing, but artistically pleasing does not equate to "realistic".

      That's why Toy Story was done in the style of a cartoon; because the Renderman technology of the day could not produce convincingly realistic scenes. But it could compete with the animation industry, where almost all the work was artistic anyway.

      But either way, if you honestly think the shadows in the Nvidia demo looks closer to real shadows (not necessarily more accurate) than the ones in Crysis, I'll leave you to it. I strongly disagree though.

      Fair enough.

      BTW, apologies if I got heated. I now remember why I stopped posting so much. I'm way too excitable. :-)

    45. Re:Beautiful by Bombula · · Score: 1

      Wow, those screen caps are gorgeous.

      Really? I was just going to say they looked pretty crap to me. I mean, they'd be OK in real-time on the X-Box or PS3 I guess, but certainly they aren't photorealistic and couldn't be used in film CGI. The car itself is fine, it's a nicely done model (Bugatti Veyron?), but it essentially has no textures - it's perfectly smooth. Maybe that's the point of this demo. But it doesn't look real by any stretch. Uncanny valley, big time.

      Worst of all is the lack of lighting and shadows in the surrounding environment. What good are reflections on a mirror-smooth car when it still looks like it's floating in mid-air over a blurry grey blob instead of actually casting realistic shadows on the road?

      Harumph.

      --
      A-Bomb
    46. Re:Beautiful by robthebloke · · Score: 2, Insightful

      The only way to really tell is to take the exact same scene and lighting model, run both algorithms, then compare the results side by side, and in those comparisions, ray tracing wins hands down....

      When rendered at 60fps in a fast paced car game, are you going to ever notice those minute details? Would you even care?

      For the record, I think the quality of the images are pretty poor. Just look underneath the car at the back wheels (pic bottom left). The shadow colour is uniform, there's no texture, no detail, no ambient occlusion, the shadow edges are sharp, etc. For another example, take a look inbetween the 2 signs - now look down just a touch. Pretty horrible imo. Now move down just a little bit more till you get to the traffic cones. That red is most definately wrong. It's the same colour as the rear car lights. Due the distance it is from the camera, there should be some saturation fade on that colour (Notice that the traffic lights are all green, and that green is identical on all instances - regardless of distance).. Now look at the building directly behind the 'No Trucks' sign. See how there's a graduation in colour saturation from top to bottom? WTF is that about? It looks like a nasty fog hack rather than a high quality ray traced renderer.

      The quality for everything in those images (apart from the reflections) is truly awful. It looks no better than than something you could do with openGL's fixed function hardware. That basically is the problem with ray tracing - It's not a technique you should be using everywhere, but use in places it's needed.

      From what I know (which isn't much) even high end shops like Pixar only recently started using ray tracing, before that it was all rasterization using procedural textures and fairly complicated lighting models.

      It still is all scan line rasterisation (it's a REYES renderer). The renderman spec recently added a trace command that you can use within the prman shaders, but it's only ever used for those specific edge cases where rasterisation isn't going to be up to the job (i.e. reflection and refraction).

    47. Re:Beautiful by aztracker1 · · Score: 1

      When you consider they're only doing three rays:pixel, on 4X a modern workstation GPU, with barely passable game-play rates... It's not so close.. given the graphics industry increases at about the same rate as the CPU industry, I'd say this is 2-3 years out from being at a playable level on high-end GPUs for gaming.. and another 2 after that before it's common for mid-range computing.. just the same though, it's pretty cool.

      Cooler still, is even at the current level, with maybe an 8-way GPU, it could mean real-time rendering acceptable for HD cartoon/animated series. I think real-time Toy Story 2 rendering is something to push for... Even if it doesn't make it into gaming, this is a real boon to the animation industry, being able to do more than wireframe animations at real time.

      --
      Michael J. Ryan - tracker1.info
    48. Re:Beautiful by hr.wien · · Score: 1

      When a soldier moves his leg far forward during a sneaky walk, you don't see a shadow of his weapon and arm cover his leg. That is realistic internal shadows. Placing shadows on obvious overhangs is just more beating the player over the head with depth perspective.

      But you do see that. :) Crysis has completely uniform shadow mapping. Everything casts shadows on everything (unless it's been specifically disabled by the artist). All light sources are completely dynamic as well, so you can take any one of them and move it about and the shadows will update accordingly. This is exactly the same as ray traced shadows, the only difference is quality and power requirement. I have a feeling you may be thinking of projected shadows which do indeed have a few of those attributes (they suck).

    49. Re:Beautiful by robthebloke · · Score: 4, Informative

      I say rasterization sticks around 3-5 years.

      I used to hear exactly the same things being said by the ray-tracing evangelists in the FilmFX industry 15 years ago. Rasterization is still the primary techinique used for any film you care to mention, and I'm almost 100% certain it will still be the primary technique 30 years from now.

    50. Re:Beautiful by jacquesm · · Score: 2, Insightful

      the low shed central in the picture

      the 'reflecting ball on the checkerboard' is a technology demo to show basic principles, it's not a realistic scene.

      In a realistic scene *everything* has a shadow, and every bit of the image interacts with almost every other, making the 'model' (if there is such a thing) a one off for every camera viewpoint and for every object movement. There is no way that you're going to make your model that complicated for a rasterizer. A ray-tracer sidesteps that model complexity issue, complexity emerges from the scene and camera viewpoint, it is not hardcoded to an arbitrary level of precision or specified in the scene.

      For computational speed you can limit both models (the one by limiting the scene description, the other by limiting the number of rays cast through a pixel and the number of bounces / splits), for a given amount of horsepower available the rasterization model wins if you're prepared the go the distance and spec your scene.

      The nice thing is that all those 'goodies' (shadows, reflection, transmission) come for free (as in you don't need to specify them beyond the basic physical properties of materials). So, the model spec for a ray traced scene of a given accuracy should be less complicated than what you need to do to a rasterizer before it will generate the same image.

      As an example, a field of grass would need a shadow map for every blade, and one for every partial transparency. In the case of a fractally generated tree or landscape that can get tedious real quick.

    51. Re:Beautiful by jacquesm · · Score: 3, Insightful

      From what I understood (correct me if I'm wrong!), the movie 'cars' was actually done using a ray tracer, which for pixar was a first.

      As for the comments wrt picture quality, yes, I agree with all the comments, but there is some stuff there that would be pretty hard to copy with a rasterizer, and I would expect the quality to dramatically improve if/when they decide to pursue this further.

      The impressive thing is not really how well the pig dances, at this stage the impressive thing is that the pig dances at all.

    52. Re:Beautiful by WilyCoder · · Score: 1

      And its misleading too.

      You can fire 3 primary rays through the pixel with NO bounces occuring.

    53. Re:Beautiful by hr.wien · · Score: 2, Informative

      As an example, a field of grass would need a shadow map for every blade

      No, no, no. You have one shadow map per light source in most implementations (or a cube map for a point light. Depends on the algorithm used). You render the scene (depth only) from the light's viewpoint into this texture, thus finding the world space coordinate of the front most geometry (the shadow casters). Then, when you render from the camera viewpoint you look up each pixel's world space coordinate in this texture to determine if it's behind a shadow caster or not. If it is, do nothing, else calculate light contribution for the pixel.

      Again, look at Crysis. Every leaf of grass casts a shadow.

    54. Re:Beautiful by fitten · · Score: 1

      The NVIDIA 'demo' also only 3 bounces deep... that's not going to give a very good image at all in comparison to an image generated with 100+ bounces... of course, this was done so they could get "up to" 30 fps in the demo.

    55. Re:Beautiful by robthebloke · · Score: 1

      That must be why most of the 3D CGI in films sticks out so much.

      Wrong department. You want to be blaming the compositors - not the rendering technique.

      For any CG element to be comped into a live action plate you've got a few things to do:

      1. When filming, accurately measure the lighting conditions. Typically this is done with white or chrome spheres, then later someone has the fun job of stiching together 6 or more photographs into a single environment map to attempt to simulate the lighting conditions.
      2. After filming, someone has the fun task of match-moving the camera to attempt to create an exact copy of the motion in the 3D app. This is typically done by hand, though there are a few app's out there that can help.
      3. The 3D texture artists and the shader writers then have to work together closely to be able to create a decent looking skin (or whatever the CG element is) material that's going to look realistic on it's own.
      4. Hit render at this point you get a CG element. However that will never match the original film plate. Lets say you are putting some object under a tree, it won't have any of the shadows of the leaves. So the next thing that happens is that some poor goon has to go off and create a shadow model to build into the render process to be able make the object fit the scene better.
      5. Eventually the CG plates are sent to the render wranglers for final production renders.
      6. Then the compositor gets hold of them, will adjust the colours/brightness and modif the lighting model as much as possible on the 2D plates until the fit is as good as possible.
      7. Finally that will be sent to be signed off by the director. If he signs it off, fine. If not, the whole process repeates itself until done.

      If any of those elements are rushed, or for some reason happens to be a pita, then it's very easy for the CG to stand out from the background. Unfortunately in the film FX industry, those shots are typically needed about 3 weeks ago; there often isn't the budget available to perfect the shots 100%; and more often than not you'll get a producer who says "We're paying a lot of money for these shots, so we want you to brighten up the 3D elements so that the audience notices them!" (which more often than not, is the reason for sucky visual FX).

      So basically, Ray tracing won't help you at all for any of that. Infact typically it's mathematical accuracy is actually a hinderance to making believable shots - you ideally want a pretty fudgable renderer - and normally that's Pixar's prman (which is not ray traced).

    56. Re:Beautiful by niteice · · Score: 1

      JPEG compression.

      --
      ROMANES EUNT DOMUS
    57. Re:Beautiful by fitten · · Score: 1

      Not really, to me... they look very plastic/fake. Of course, NVIDIA is only rendering at 3 bounce depth, which will make the quality not-so-great anyway, but that's because they needed to be able to run the demo at "up to" 30 fps.

    58. Re:Beautiful by hr.wien · · Score: 1

      Again, look at Crysis. Every leaf of grass casts a shadow.

      Grr, no that's wrong. They only receive shadows. There's nothing technically preventing them from casting shadows though. It's probably disabled for performance reasons.

    59. Re:Beautiful by Anonymous Coward · · Score: 0

      John Carmack and most of other graphics programmers disagree. But no, by all means if you want to buy 4 $500 video cards so you can run something that doesn't look as good as games that can run on 1 $200 video card be my guest.

    60. Re:Beautiful by jacquesm · · Score: 1

      well, you sort of made my point for me, *EVERYTHING* casts a shadow, for every lightsource (in fact for every lightray that hits the object)...

      Not something you determine or decide to swich on / off per object.

      It's precisely in situation like that where for instance light filtering through leaves would look immediately believable in a ray tracing rendering but utterly wrong in a shadow mapped renderer.

      Thanks for correcting my idea on how shadow maps (should) work, I figured they were on a per-object basis, but per lightsource makes more sense (less computationally intensive, same effect).

       

    61. Re:Beautiful by Hairy+Heron · · Score: 1

      Wrong department. You want to be blaming the compositors - not the rendering technique.

      No, it's very much the right department. When the vast majority of the rendered CGI doesn't even look photorealistic, there isn't anything better compositing is going to do.

    62. Re:Beautiful by hr.wien · · Score: 1

      Now the leaves in Crysis do cast shadows since in those cases missing shadows would be immediately noticeable. :) And you would do the same kinds of optimizations in a ray traced engine if you wanted it to run real time. Barely noticeable details like grass shadows are fair game when optimizing for performance, and culling all that geometry before tracing would probably help (though maybe not as much as in a rasterized engine).

      There's always going to be trade-offs like that in real time graphics, ray traced or rasterized. It's all about fudging it just enough that you can get away with it, you just have to fudge different things depending on the technique. Dynamic geometry is for instance a real pain in the neck in ray tracing since you have to rebuild (parts of) your acceleration structure every time something moves. This overhead can get quite significant in a dynamic game world.

    63. Re:Beautiful by afidel · · Score: 2, Interesting

      No, for illustrative purposes of what ray tracing is capable of you have to look at something like this, it's very very close to being indistinguishable from a photo. That took 21 hours on a P2-350 (465 MFlops) so on a GTX280 which is ~3,000 times faster it would take about 45 seconds to render, not exactly a playable framerate. We're a few doubling of transistor count away from being able to do photo-realistic ray-tracing at playable framerates.

      --
      There are 4 boxes to use in the defense of liberty: soap, ballot, jury, ammo. Use in that order. Starting now.
    64. Re:Beautiful by Enderandrew · · Score: 1

      By that I meant, rasterization is sticking around at least 3-5 years even if ray tracing were on the horizon, yet Intel claims rasterization will disappear very soon.

      --
      http://blindscribblings.com - Tasty pop-culture in conceptual fashion.
    65. Re:Beautiful by ravyne · · Score: 1

      Are you looking at the same screens as me? Because the shots in the article look like utter shit. Its no wonder nVidia wants you to believe that ray tracing is 10s of years away; its because they can't do it.

      They're doing *just* 3 bounces, and they run at "up to" 30fps at 1920x1080 resolution. The shadows are hard, and the lighting is flat, everything looks washed out, and its due to the lighting (ie, the rays) not the textures.

      And to what glorious setup do we owe this marvel? Just *4* of their highest-end, workstation-class Quadro GPUs (which probably run 1-2 grand apiece, or about $450 apiece for consumer equivalents.) All this is interfaced to a pretty serious machine, of course... Color me impressed. No seriously, but don't forget to pick my jaw up off the floor first *rolls eyes*.

      ATI's Cinema demo was far more impressive, and looks awesome by comparison. It was running at 1280x720 (roughly 1/2 the pixels as 1080p) but was running at around 60fps with anti-aliasing -- virtually running at the same resolution as the nVidia demo, but twice as fast...

      And to what *insane* amount of AMD graphics hardware and muti-socket, multicore AMD processors do we owe this ungodly power to? A *single* Radeon 4870 and a standard workstation.

      Don't believe me? Look for yourself:
      http://www.youtube.com/watch?v=BzquM5Td6bM&NR=1
      http://www.tgdaily.com/content/view/38145/135/

    66. Re:Beautiful by AKAImBatman · · Score: 1

      But you do see that. :) Everything casts shadows on everything (unless it's been specifically disabled by the artist).

      If that's true, then where are they? I have yet to see a gameplay example where a soldier casts a shadow on himself. You can see a soldier's shadows on the ground, plants, buildings, and in other areas, but never on the soldier himself.

      Crysis has completely uniform shadow mapping.

      I don't think that means what you think it means. A uniform shadow map is a shadow map that's effectively fixed regardless of perspective. This is inferior to raytraced results.

      This is exactly the same as ray traced shadows, the only difference is quality and power requirement.

      You keep repeating that, but it's simply not true. Shadow maps only compete with raytracing if you spend the computational power to project a shadow for every polygon in the scene for nearly every frame. Game don't do that. It's far too expensive.

      There's a good discussion on the Crysis shadowing technique over on the OGRE3D forums.

      I have a feeling you may be thinking of projected shadows which do indeed have a few of those attributes (they suck).

      Shadow maps ARE a form of projected shadows. That's why they're also referred to as projective shadowing.

    67. Re:Beautiful by hr.wien · · Score: 1

      If that's true, then where are they?

      How about this one? The torso of the guy in the middle of the shot casts a shadow onto his left leg. No you don't get shadows in every nook and cranny with shadow maps due to aliasing (you need to bump the map resolution a lot to remove that), but you don't need that to get fairly decent results.

      Crysis also has screen space ambient occlusion (oh noes! more fudging) which makes up for a lot of the details lacking in the shadow maps. The over all effect is in my opinion much better than the pixel perfect ray traced shadows in the Nvidia demo. That looks about the same as the stencil shadows you saw in Doom3. Horrible stuff.

      I don't think that means what you think it means.

      Yeah, my terminology may be off. What I meant was shadow mapping that works uniformly across the entire scene, of which perspective shadow maps are one kind.

      Shadow maps ARE a form of projected shadows.

      Again, my terminology is probably off. I was thinking of per object projected shadows which project a shadow from the object onto the ground, but nothing else.

    68. Re:Beautiful by miyako · · Score: 1

      That isn't actually a very good example because most of the realism from that particular image comes from the caustics (IIRC that image was created as a demonstration of POVRays caustics capabilities), and caustics drastically increase render time for a scene. Adding caustics and global illumination will quickly take the speed down from 30fps to .3fps. Don't get me wrong, real time ray tracing will be awesome, but it will be a long time before we can do real time ray tracing with that kind of quality.

      --
      Famous Last Words: "hmm...wikipedia says it's edible"
    69. Re:Beautiful by AKAImBatman · · Score: 1

      When rendered at 60fps in a fast paced car game, are you going to ever notice those minute details? Would you even care?

      Actually, you care MORE about the shadows then, not less. Shadows provide visual cues that transmit everything from depth to orientation to backface information. e.g. When a car suddenly swerves, you see the result in the shadow just as much as you see it in the object itself. When cars are passing each other, the shadows cue their relative position regardless of the perspective. And as they pass light sources causing the shadows to swing around the vehicle? Well, that's just cool! :-P

      For the record, I think the quality of the images are pretty poor.

      There's no denying that the scene is overly simplistic. But that's okay. It shows off the key advantages of raytracing. Adding additional texturing, detail, soft shadows and other features are all possible, and in fact probable. But this isn't a 10 million dollar game. This is a simple demo. And for what it is, I think it looks great. :)

      And for the record, ambient occlusion is not what you want in a raytraced scene. Such occlusion is a form of illumination "cheating" that gives decent, but not spectacular results. There are far better shading models available for ray tracing.

    70. Re:Beautiful by hr.wien · · Score: 1

      Oh, and none of the shots you posted look anything like Crysis in game. Especially the characters look out of place. I'm guessing they're pre release renders. (Raytraced! :P)

    71. Re:Beautiful by The+Gaytriot · · Score: 1

      It still is all scan line rasterisation (it's a REYES renderer). The renderman spec recently added a trace command that you can use within the prman shaders, but it's only ever used for those specific edge cases where rasterisation isn't going to be up to the job (i.e. reflection and refraction).

      I've thought that they should implement the same thing in games.
      I remember looking at screenshots of Quake III and IV rendered with ray-tracing and thinking, wow that looks like shit. Then I saw some screenshots which included lots of shiny reflective objects and it looks incredible.

      I think ray-tracing is a big step up from poorly done specularity to mimic reflection, it would be cool to have that in a game without being too expensive performance wise.

      --
      Srsly u guys. U guys, srsly.
    72. Re:Beautiful by Creepy · · Score: 2, Informative

      hmm - I was a bit underwhelmed, myself. First of all, graphics people have done GPU ray tracing on small static scenes has been around since about 2003 or 2004. The real limitation then was available memory. This is a standard ray tracer (no photon mapping), and highlighting an optimal model (a reflective one, not a diffuse one). They mention it scales linearly, which means they replaced the fixed function pipeline and used only shaders.

          The scene shown is 2 million polygons (about the baseline of this generation's hardware) but with a single light source and hard shadows. Three reflections is adequate to get a decent looking scene (I've done good looking scenes with 2 reflections, but it depends on the scene).

    73. Re:Beautiful by Goaway · · Score: 1

      Lighting technology has improved significantly in the last decade, but still not sufficiently to compete with raytracing.

      Completely and utterly wrong. Raytraced lighting looks extremely unnatural, due to the lack of handling indirect light. To get lighting that actually looks realistic, you need to use global illumination methods, and those are so incredibly expensive, the only real way to do them is to pre-calculate them, for rasterizers and raytracers both.

      A raytracer with global illumination will look far better than a raytracer without.

    74. Re:Beautiful by Goaway · · Score: 1

      With ray tracing, the shadows are correct.

      Once again, wrong. Raytraced shadows will not take into account reflected light, and will therefore be far too uniform and flat.

    75. Re:Beautiful by SpectreHiro · · Score: 1

      Where are my mod points when I need 'em? You, sir, deserve another +1 Informative.

      --
      You can't win, Darth. If you mod me down, I shall become more powerful than you could possibly imagine.
    76. Re:Beautiful by Goaway · · Score: 1

      You know, there's hardly anything in that picture that couldn't be done just as well with a rasterizer and a bit of trickery, and would run an order of magnitude faster.

    77. Re:Beautiful by Goaway · · Score: 1

      That's quite an optimistic estimate for rendering speed with global illumination! That'd have to be one extremely optimized algorithm to achieve .3 fps. Frames per minute or per hour sounds more realistic.

    78. Re:Beautiful by VeNoM0619 · · Score: 1

      Wow, those screen caps are gorgeous

      Yea.. thats all photoshopped all the reflections are all wrong...definitely photoshopped...

      --
      Disclaimer: I am not god.
      We may not be created equal
      But we can be treated equal.
    79. Re:Beautiful by Goaway · · Score: 1

      That must be why most of the 3D CGI in films sticks out so much.

      It most certainly is not.

      Raytracing is not some magical algorithm that creates photorealistic pictures. It's merely a different way to make a rough approximation of solving the rendering equation, and it is not a particularly good one. It handles shiny spheres on infinite checkerboards quite well, but it is really not very useful for most real-world scenes. And that is why it is not used much.

      Most people who don't know much about computer graphics have a romanticized image of it, but this is basically just superstition.

    80. Re:Beautiful by Goaway · · Score: 1

      No, just a lack of supersampling.

    81. Re:Beautiful by SpectreHiro · · Score: 1

      From what I know (which isn't much) even high end shops like Pixar only recently started using ray tracing, before that it was all rasterization using procedural textures and fairly complicated lighting models.

      From what I understand, Pixar's renderer (PhotoRealistic RenderMan) is still primarily a REYES renderer, but since Cars, has supported optional ray-tracing for certain effects such as reflections. This is an important point that I think a lot of folks don't understand or are overlooking... You can mix different rendering techniques. You can have a scan-line rasterizer as your primary rendering technique and also have raytraced shadows and reflections. It's an amazing world we live in, isn't it?

      --
      You can't win, Darth. If you mod me down, I shall become more powerful than you could possibly imagine.
    82. Re:Beautiful by Anonymous Coward · · Score: 0

      Welcome to Slashdot, home of fail.

      Step 1: Ignore parent context

      Step 2: Find horrendous edge case that's discussed elsewhere and isn't even the topic of discussion

      Step 3: Post the word WRONG like you know what you're talking about

      Step 4: ???

      Step 5: FAIL!

      With this patented Slashdot formula, you too can fail on a regular basis!

    83. Re:Beautiful by ardor · · Score: 1

      Yet again, somebody thinks Raytracing is a magical wand which automatically improves image quality.

      Here's the deal:

      1) Raytraced shadows do NOT look real. Better than shadowmaps, but not real (they are too sharp). Casting multiple shadows to soften it is expensive, and causes banding if done poorly.

      2) Raytracing does NOT improve lighting. It makes shadows easy, yes, but the lighting equation(s) remain the same. Pure raytracing takes into account only direct lighting; indirect lighting is left out, and belongs to the area of global illumination techniques, which are more or less orthogonal to the direct rendering method (meaning they can be used with both rasterization and raytracing).

      3) The only real speed gain of raytracing is in the secondary ray scenario (e.g. indirect stuff like refractions, reflection, shadows etc.) For primary rays, the rasterizer is better, BECAUSE the rasterizer is nothing else than a special case of raytracing where the actual tracing can be optimized away and the coefficients between the vertices can be linearly interpolated. It is totally stupid to use raytracing here. Example: a grey concrete wall. Rasterization delivers perfectly correct results here easily. So, use rasterization when possible, raytracing when necessary. Rasterization will always be faster than raytracing for these cases, and we are still far away from the point when the actual speed gain is negligible.

      4) The X million triangles per second counts are irrelevant in games. (And, consumer 3D cards are typically used for games.) Games favor less complex models, but with complex materials. There are subtle but important benefits from texture mapping vs. high-detail models, one being the fact that details applied to textures are much less affected by aliasing, while high-detail models tend to flicker intensely, unless very high AA levels are employed. Also, raytracing has significant issues regarding cache coherency, even with bundled rays. Rasterizes just fetch and interpolate all the time; cache coherency is trivial here.

      --
      This sig does not contain any SCO code.
    84. Re:Beautiful by SpectreHiro · · Score: 1

      And for the record, ambient occlusion is not what you want in a raytraced scene. Such occlusion is a form of illumination "cheating" that gives decent, but not spectacular results. There are far better shading models available for ray tracing.

      AFAIK, ambient occlusion is often used in conjunction with raytracing. While raytracing can provide shadows for every given light-source in a scene, it generally fails at dealing with ambient light and its occlusion. There's certainly no good reason not to use these techniques together, and a bit of AO can do a lot to help define subtle surface details that standard traced shadows do not.

      And to be perfectly honest, raytracing is just as much of a cheat as any of the other techniques mentioned in this whole thread. It's more physically realistic only in the most abstract sense, and doesn't provide any of the more interesting lighting effects currently being used in high-end rendering.

      Regardless, raster rendering is good enough for Pixar, and for most purposes, it's certainly good enough for me.

      --
      You can't win, Darth. If you mod me down, I shall become more powerful than you could possibly imagine.
    85. Re:Beautiful by afidel · · Score: 1

      If you can use a raster renderer to give that good of reflections on the street then you are a better artist/programmer than just about everyone in existence because I've never seen one that comes close.

      --
      There are 4 boxes to use in the defense of liberty: soap, ballot, jury, ammo. Use in that order. Starting now.
    86. Re:Beautiful by mikael · · Score: 1

      Another interesting point is that this demo is currently capped at 3 casts per pixel

      That's three ray bounces.

      The first ray goes from the eye point through the image plane (the resulting image) into the scene. The second ray goes from the reflection or refraction of the first intersecting surface into the scene again. The same with the third ray. This allows the renderer to draw the reflection of the car from the plate glass windows on the other side of the street.

      If you want more rays per pixel, that is multi-sampling (or anti-aliasing) which will do the effects you describe.

      This may be done using a hybrid rasterizer/ray-tracer. A simple ray-tracer will rasterize the resulting image by scanning each pixel column-by-colum, row-by-row.

      That could be optimized by rasterizng each triangle (forming the first ray) then doing the reflection/refraction calculations with the rest of the scene.

      Interesting to note the car is a fixed color with a reflective component. This will probably be of benefit to the CAD industry who like to generate ray-traced views of designed geometry.

      --
      Vintage computer adverts: http://www.vintageadbrowser.com/computers-and-software-ads
    87. Re:Beautiful by Anonymous Coward · · Score: 0

      The raytracing algorithms used to render that frame are several orders of magnitude more expensive than what they're doing in the NVIDIA demo. Which btw does not look very good, they could make it look much better with traditional rasterization, but it is nevertheless a very interesting development that will most likely find use TOGETHER with traditional rasterization.

    88. Re:Beautiful by Anonymous Coward · · Score: 0

      In what way is that close to being indistinguishable from a photo?

      It looks like digital art to me. The coloring and clarity makes it look like it was painted in photoshop.

    89. Re:Beautiful by TyFoN · · Score: 1

      Sure this is impressive, but they are pushing kilowatts to get this kind of performance.

      Maybe this is their way to solve the performance issues with Qt4/KDE4 in Linux ;)

    90. Re:Beautiful by ozbird · · Score: 1

      A human instinctively knows when something looks "right" or "wrong".

      For example, where is the reflection of the photographer?

    91. Re:Beautiful by jensen404 · · Score: 1

      Toy Story 2 isn't ray traced.

    92. Re:Beautiful by tyrione · · Score: 1

      Something truly needs to be approached from a different perspective to produce equivalent results without requiring 4 massive GPUs to produce it.

      To make this level of raytracing a reality there has to be more thought and redesign into how we go about it. This pure brute force approach is dying for a challenge that uses a fraction of the KWh and silicon.

    93. Re:Beautiful by Pseudonym · · Score: 1

      I hope this finally puts to rest the idea that rasterizing with upteenth number of features added in can compete with the image quality of Raytracing.

      Any time anyone says this, keep this in mind: All Pixar films before Cars used no more than 6 minutes screen-time worth of raytracing.

      (And, BTW, it's no coincidence that NVIDIA chose a car for this demo. Cars have curved reflective surfaces, which is one of the very few things in the world that are hard to fake convincingly with scanline rendering. It's much easier to fake them convincingly with raytracing.)

      Another interesting point is that this demo is currently capped at 3 casts per pixel. Which means that the scenes shown could look even better than they already do.

      Indeed. They could, for instance, use rasterisation for "primary" rays and reserve raytracing for secondary rays like most high-end software renderers do. That'd give you the speed of rasterisation where it counts and the illumination model of raytracing where it's needed.

      --
      sub f{($f)=@_;print"$f(q{$f});";}f(q{sub f{($f)=@_;print"$f(q{$f});";}f});
    94. Re:Beautiful by DerWulf · · Score: 1

      the definition of ray tracing is that it's "brute force". This real time renders are still spectacular. A while back it was consencus that ray tracing would never be done in realtime and now we already have prototypes.

      --

      ___
      No power in the 'verse can stop me
    95. Re:Beautiful by DerWulf · · Score: 1

      I think so too. Average, really. I suspect this demo wasn't done for end consumers though. It just shows that real time raytracing might be feasable in a few years which is very important to know if you are planing on doing a new game engine sometime in the future. Look at this:

      http://en.wikipedia.org/wiki/Image:Glasses_800_edit.png

      It's hard to believe that's a rendered picture. In a few years game graphics will look like *this*. That's the signifcance of this demo.

      --

      ___
      No power in the 'verse can stop me
    96. Re:Beautiful by DerWulf · · Score: 1

      Yes, real time ray tracing is not currently feasible. I think we all knew that. This demo is a peek at the future. You know how everyone wonders what to do with the bazillion cores that CPU vendors are planning bring out? Raytracing might be it for games. Ray tracing, in addition to being more realistic is also a lot easier to implement and very easy to paralize. I think it will be a relieve for developers to get *all* lighting and shadow on *all* objects with one relatively simple alorithm.

      --

      ___
      No power in the 'verse can stop me
    97. Re:Beautiful by moongha · · Score: 1

      From what I understood (correct me if I'm wrong!), the movie 'cars' was actually done using a ray tracer, which for pixar was a first.

      You're wrong. Renderman has options for ray tracing, it's still using it's rasterization for most stuff. Doing everything with raytracing would be incredibly inefficient, and wouldn't look any better.

    98. Re:Beautiful by savuporo · · Score: 1

      It could be just me, but i am seeing a lot of annoying aliasing ( look at the rims for instance ), and that is in still images, must be pretty awful when moving.

      --
      http://validator.w3.org/check?uri=http%3A%2F%2Fwww.slashdot.org Errors found while checking this document as HTML5!
    99. Re:Beautiful by mdwh2 · · Score: 1

      Ray tracing, in addition to being more realistic is also a lot easier to implement and very easy to paralize. I think it will be a relieve for developers to get *all* lighting and shadow on *all* objects with one relatively simple alorithm.

      Note that current rasterisation algorithms are also very easy to parallelise (consider that a GTX 280 has 240 stream processors). It's also possible to get all lighting and shadow on all objects with one relatively simple algorithm - the reason shortcuts are taken is because of performance. But general shadowing and dynamic lighting algorithms for rasterisation have been used for a few years now.

    100. Re:Beautiful by mypalmike · · Score: 1

      Nobody said primary ray.

      --
      There are 0x40000000 types of people: those who understand 32-bit IEEE 754 floating point, and those who don't.
    101. Re:Beautiful by Mattsson · · Score: 1

      And "graphical quality becomes a matter of raw horsepower"? This unlike in rasterization then?

      In both cases, visual quality is limited by the quality of the 3D models, textures and "special effects".

      In a traditional 3D engine, though, you also have a maximum level of visual quality achievable by that particular engine. Once you have a card that let you run everything at maximum quality settings, more performance will only give you higher frame rates, until the monitor becomes the limiting factor.
      If you're raytracing, you can add more ray-casts if you got more performance, making the scene more and more realistic, and you can keep increasing the amount of rays indefinitely.
      There will of course be a decreasing return once we're getting closer and closer to a "perfect" raytrace of a particular scene, but ten years after launch, a raytrace game will still be slowly increasing in visual quality.

      --
      /.Mattsson - My native language is not English, so please don't whine over linguistic errors. (That's lame anyway...)
    102. Re:Beautiful by aztracker1 · · Score: 1

      I was referring to level of quality, not the specific technique.

      --
      Michael J. Ryan - tracker1.info
  2. What's the power consumption on that rig? by Rogerborg · · Score: 3, Funny

    Al Gore would like to have a word with you about just how important it is to beat photo-realistic hookers to death.

    --
    If you were blocking sigs, you wouldn't have to read this.
    1. Re:What's the power consumption on that rig? by Anonymous Coward · · Score: 0

      haha what Al dislikes even more is that the Bugatti Veyron in the photos gets about 3 mpg.

    2. Re:What's the power consumption on that rig? by maxume · · Score: 1

      He doesn't mind if you do it, he just wants you to call it a privileged and to pay someone for it.

      --
      Nerd rage is the funniest rage.
    3. Re:What's the power consumption on that rig? by Anonymous Coward · · Score: 0

      Have you seen's Al Gore's setup?

    4. Re:What's the power consumption on that rig? by Rogerborg · · Score: 1

      Maybe a hooker trading scheme? You can beat a ray traced hooker to death if you can offset it against someone beating an 8 bit bitmap hooker to death?

      --
      If you were blocking sigs, you wouldn't have to read this.
    5. Re:What's the power consumption on that rig? by Anonymous Coward · · Score: 0

      Eliot Spitzer might have a problem with Al Gore.

  3. Re:That's a nice canned post ya got there by Anonymous Coward · · Score: 5, Informative

    The kind with a star next to his name, obviously.

  4. I've got an idea by initdeep · · Score: 0, Troll

    How about they focus instead on correcting the issue with their mobile chipsets which they are studiously avoiding making right with any of the people who have had issues.

    1. Re:I've got an idea by Abcd1234 · · Score: 1

      Yes, because no corporation can possibly two do things at the same time...

    2. Re:I've got an idea by LingNoi · · Score: 1

      Apparently they're not which is the GPs point.

  5. hmm. by apodyopsis · · Score: 1

    I had a tenant once, who stated that the first use of any new multimedia technology was always for, shall we say, *adult* services. In the sense they drove DVD sales (multi viewpoint, stop rewind), CGI, and so on.

    gah. the mind boggles what the animation crowd could do with this.

    On a similar note, when I did my BA years ago my dissertation was on the veracity of digital images, when there is no real referent to accompany the photograph, and the loss of credibility of images in the onslaught of photo realistic fakes.

    Well, it was inevitable that technology would catch up and we will no doubt need to add "video realistic" to our dictionary as well.

    Just give me silver halide and call me a Luddite.

    1. Re:hmm. by mopower70 · · Score: 1

      Wow. If only we could limit nonsequiturs to three bounces a post.

  6. don't quit your day job quite yet by Ancient_Hacker · · Score: 5, Informative

    The devil is in the details. Ray tracing with glossy surfaces is relatively easy. But if you want to simulate real-world textures like orange-peel, bark, hair, or skin, things can really slow down.

    1. Re:don't quit your day job quite yet by Anonymous Coward · · Score: 0

      Hey man, I don't know what kind of weird stuff you're into, but I do perfectly fine without any orange peel, bark, hair or... oh wait.

    2. Re:don't quit your day job quite yet by MadnessASAP · · Score: 1

      Bark isn't too hard to pull off at least form a technical standpoint, from an artistic standpoint it can of course be a bitch to pull off but that's a different matter. Hair is a bit trickier but you can usually get decent enough looking results. Orange peel has a bit of Sub Surface Scattering (SSS) but in most scenes you can again get away without it.

      So that leaves skin, for a long time realistic skin was almost the holy grail for CGI given that SSS and multiple layers are a huge component of skin but now for pre-rendered images the technology and the horsepower is largely in place to pull it off. Of course this is about RT raytracing so I can expect that it will be quite a while(on a technological timeline) before your average desktop computer can render skin in les then a minute much less 30 times a second.

      --
      I may agree with what you say, but I will defend to the death your right to face the consequences of saying it.
    3. Re:don't quit your day job quite yet by Ambiguous+Coward · · Score: 3, Insightful

      Umm, no. I admit it's been a while, so my memory may be off, but I distinctly recall that procedural textures in ray-tracing are really, really easy, and add almost no (necessary) overhead to speak of. If you can find a way to do those things you mentioned with a procedural texture (those cases you provided are the textbook examples of how to do procedural textures, mind you) then you can almost certainly do them easily and cheaply. Any graphics course will have you rendering textured oranges inside the first week of the ray-tracing portion. Rainbowed CD undersides and the cool microscopic rings on the underside of a brass pot the day after that.

      I don't remember the implementation details, as it's been many years since I tried, but it's easy, and doesn't add any real overhead to speak of. Yes, when running procedural textures, you CAN make them heavy (it's a procedure: it'll do whatever the hell you want) but by no means is that a requirement.

      Which do you think is worse:

      • Loading a 10MB texture image that only works at certain resolutions, and eats up 10MB of space, or
      • Loading up a 1k code segment that generates the same (or similar) texture at any resolution, on the fly, and only generates the texture for the exact pixels you're looking at?

      Not to mention that if anything even close to the support given to the current texture models is given to procedural texture models, they will almost instantly outpace the current options and limitations.

      Did you ever notice that the early-generation ray-tracers supported procedural textures long before they supported the "regular" texturing model? There's a reason for that.

      -G

      P.S. Yes, I'm fully prepared for nit-picking you-used-the-wrong-word-here responses, so fire away. :P

      --
      Their may be a grammatical error, misspeling, or evn a typo in this post.
    4. Re:don't quit your day job quite yet by Ambiguous+Coward · · Score: 1

      Everything else I said above left aside, I think what I'm really driving at was this one statement:

      ...if anything even close to the support given to the current texture models is given to procedural texture models, they will almost instantly outpace the current options and limitations.

      -G

      --
      Their may be a grammatical error, misspeling, or evn a typo in this post.
    5. Re:don't quit your day job quite yet by Big_Breaker · · Score: 1

      At least in the beginning I imagine skin features like you descibe would be pre-rendered into the texture as some sort of efficiency compromise. It's not like raster methods do a great job right now anyway. Most games are still going further into uncanny valley rather than climbing out of it.

    6. Re:don't quit your day job quite yet by caywen · · Score: 1

      I'm admittedly only very journeyman in my understanding of RT (I've written some basic RT's). How would they slow down? You can process bump maps pretty easily - it's still just adjusting the diffuse light value based on a tweaked surface normal.

    7. Re:don't quit your day job quite yet by oliverthered · · Score: 1

      that's all fine and well, but it sure looks a lot sexier (proper curved surfaces) than rasterization which is what it's competing with

      --
      thank God the internet isn't a human right.
    8. Re:don't quit your day job quite yet by MobileTatsu-NJG · · Score: 1

      Which do you think is worse:
      Loading a 10MB texture image that only works at certain resolutions, and eats up 10MB of space, or
      Loading up a 1k code segment that generates the same (or similar) texture at any resolution, on the fly, and only generates the texture for the exact pixels you're looking at?

      If you're running more lines of code per pixel with a procedural than you are by simply looking up a memory location for an rgb value, how is that going to be 'no overhead'?

      Maybe I'm misunderstanding a fundamental concept of how raytracing works, so I won't be ashamed or argumentative if you can correct me on this, but I don't get how raytracing will make procedurals run nearly as fast when the renderer is going to have to ask a shitload of questions per-pixel about what that color is supposed to be. Do forgive me, though, I'm only really thinking of real-world use of procedurals that'd actually compete with high res textures, not simple marble patterns.

      --

      "I like to lick butts!" by MobileTatsu-NJG (#32700246) (Score:5, Informative)

    9. Re:don't quit your day job quite yet by Anonymous Coward · · Score: 0

      I don't think "pre-rendered" subsurface scattering really counts as subsurface scattering.

    10. Re:don't quit your day job quite yet by fractalus · · Score: 1

      Well, for starters current 3D engines don't always do just a simple texture-mapping anyway; they can be doing environment-mapped reflections, bump maps, and so on in their fragment shaders. This is roughly the same level of complexity.

      On top of that, though, is the issue that when you ray-trace, you only generate textures for pixels that actually appear on the screen. When you rasterize, you don't always have your polygons ordered front-to-back so you end up rendering a pixel's texture only to have it replaced by another polygon's texture within the same scene. This is redundant work that raytracing eliminates.

      The big downside to procedurals is that you're limited by your algorithms; they can produce infinite, non-repeating textures but you have to be a bit of a mathemagician to come up with completely new ones. Artists, on the other hand, are much easier to come by...

      --
      People are never as simple as their stereotypes. This applies equally to Christians, Muslims, and Emacs-lovers.
    11. Re:don't quit your day job quite yet by MadnessASAP · · Score: 1

      Sorry I meant non-realtime as in rendering that takes minutes or even hours per frame.

      --
      I may agree with what you say, but I will defend to the death your right to face the consequences of saying it.
    12. Re:don't quit your day job quite yet by Anonymous Coward · · Score: 0

      The GP is probably talking about subsurface scattering. That can't be done with texturing alone, procedural or otherwise.

    13. Re:don't quit your day job quite yet by Ambiguous+Coward · · Score: 1

      The way I see it, if resources were devoted to processing procedural textures instead of storing texture images, then that, at the very least, puts the two on an even playing field.

      What follows is train of thought, totally unorganized:

      Yes, RAM is cheap and thus you can store large amounts of textures in it for a relatively low-cost. But, in general, REALLY FAST RAM (like cache, or better) is NOT cheap, and so you DON'T store textures there. You store it in the cheap, "slow" RAM. On the other hand, if you generate the texture on the fly, procedurally, in the processor (GPU in this case), then that's FAST. I have absolutely no doubt that a GPU on a graphics card is more than capable of generating a texel as fast or faster than it can be used by the rest of the pipeline. Yes, this depends on the procedure used, of course. But assuming we'll continue to blend multiple textures and do all kinds of different texture operations, I'm equating a single, flat, procedural texture to a single, flat, texture image. Blending of multiple textures (procedural or otherwise) happens later.

      That being the case, there's no reason that, if graphics cards allowed procedural textures to be generated in the pipeline, that they couldn't easily outpace pre-rendered textures.

      Even with the current texturing models, the renderer is still asking a shitload of questions per-pixel about what that color is supposed to be. The trick to making it work right now is just that there is a lot (and I mean an obscene amount) of support in the pipeline for making those questions get answered fast. With a similar amount of support for procedural textures, a lot of opportunities arise.

      I just can't stop thinking about other rendering situations that are vaguely similar to procedural textures. Take vector vs. raster graphics, for instance. Photoshop (or Illustrator) is able to render high-quality vector graphics just as well as high-quality raster graphics. In some case, actually, vector graphics are significantly faster, even though the surface has to be calculated just to find out "where" it is. On the other hand, raster graphics are simply buffers in memory. In this case, however, Adobe's done a lot of work to make the two comparable.

      I'm totally rambling at this point, and I apologize for that. Long story short, pre-rendered textures are just as complicated as procedural textures, it's just that right now, renderers only put their horsepower towards pre-rendered textures.

      Imagine what your CPU could do, rendering procedural textures to memory (instead of to your screen.) Insanely complex textures, hundreds of them, blended in insanely complex ways. You could do some pretty amazing crap like that. Even the relatively "slow" GPU could do that. And that's all you really need. Once it's rendered a texel to memory, it can be used in the pipeline like any other texture source. That gets us so far beyond simple marble patterns, it's not even funny. :)

      -G

      --
      Their may be a grammatical error, misspeling, or evn a typo in this post.
    14. Re:don't quit your day job quite yet by Anonymous Coward · · Score: 0

      I won't argue that it can't be done (because it can), but look for what you don't see in the demo.

      In the first 2560x1600, have a look at the rims. They look terrible.

      While procedural shading is possible, I'm not seeing any. Look at the road - since they've got a texture map, adding a bumpmap would have been trivial, right? But you're not seeing it, which means that there's some reason for it not being there. I'm betting the reason is that 3D textures require a higher level of antialiasing, which would slow the demo down even more.

      There's no attempt to create quasi-realistic lighting, either. If they've got a realtime raytracer, you'd think they'd give a bounce or two to skydome lighting.

      So - other than reflections - the demo doesn't really show off the sort of things which are a strength of raytracing. To be honest, most of the scene looks worse that what you could get from a scanline renderer.

      Given a choice between this and a scanline renderer using reflection maps, I'd choose the scanline in a heartbeat.

    15. Re:don't quit your day job quite yet by Anonymous Coward · · Score: 0

      Because that's how GPUs work. The procedures (up to typically 256 instructions on older cards and more on newer ones) can run in like 1 cycle on the GPU (this must take an assload of silicon, but GPUS have it...) If you're not running a procedure it just runs a NOP there.

    16. Re:don't quit your day job quite yet by gaggle · · Score: 1

      Indeed.

      Plus, graphics are usable when they scale back to what consumers actually have. I agree with the article that this sort of processing power will be available in retail product in a couple of years.. but that doesn't put it in the hands of anyone. How about adding an additional, oh I don't know, four+ years before it's actually available in a broad userbase? Because raytracers don't scale back to current gen hardware, and then what? It's already a huge hassle supporting DX9 and DX10 renderpaths, I can't imagine (m)any games dedicating themselves to two completely radically different rendering techniques on top of that.

      And that's assuming this nVidia techdemo was all it took to make a game. Like the parent says, there's a lot of different materials in the world and raytracing ain't exactly the magic bullet for all of them.

      Raytracing will come along at some point, but it's not like we're facing some imminent blitzkrieg.

    17. Re:don't quit your day job quite yet by Josef+Meixner · · Score: 1

      And what exactly do you think is different for rasterizers? Ever heard the term "pixel shaders"? That is exactly what can be used to do procedural texturing on GPUs now. At that level there is no difference between the two systems. A rasterizer will have to evaluate the function per subpixel and so will a raytracer. As soon as additional environmental information is needed, a raytracer has the advantage that you only need to fire additional rays, a rasterizer needs more programming effort. But neither is very cheap.

      A rather simple example are silky surfaces. They require a high amount of tests into the environment as they result from a surface which scatters light into many directions (microfacet model). To calculate such a surface you send rays into the scene from each pixel where you hit such a material. The density of rays you need depends on the surface roughness.

      It is rather unimportant if you do that sampling by sending rays or by environment mapping. Both have there advantages and disadvantages.

    18. Re:don't quit your day job quite yet by Big_Breaker · · Score: 1

      I totally agree that it wouldn't count in an academic sense - IE achieving realtime subsurface scattering.

      I just meant rendering human figures that don't look like they are made out of plastic. Maybe on close inspection or in rare circumstances the trick becomes apparent but in practical terms the effect works. Sometimes that's the best you can do and it is better than nothing. I am think of some of the texture swapping that used to be done before shader programs. Didn't Quake do something like that to implement lighting effects? I seem sto recall a color pallet limitation.

    19. Re:don't quit your day job quite yet by Ambiguous+Coward · · Score: 1

      This is a really good point, and I'd completely forgotten about pixel shaders. That's pretty funny, actually. I was saying, "we need this!" and lo and behold, we already have it. So, there you go. Be aware, though, you've only supported my previous statement/response to the GGP, in that realistic textures in ray-tracing need be no more costly than realistic texturing using the current model. You did what I failed to, though: you pointed out precisely why this is already the case. :P

      -G

      --
      Their may be a grammatical error, misspeling, or evn a typo in this post.
  7. Of course Intel would say that... by Anonymous Coward · · Score: 0

    It takes a lot of CPU power to raytrace at that level, and hey, what's that Intel is selling?

  8. What a waste of resources by rogerbo · · Score: 5, Insightful

    I just looked at those pictures and then checked a high res shot of Gran Turismo 3 Prologue on a PS3:

    http://o.aolcdn.com/gd-media/games/gran-turismo-5-prologue/playstation-3/22.jpg

    I don't see enough of an improvement to increase GAMEPLAY in any significant way. The reflection maps and shadows that are created by the current rasterization tricks are good enough that you suspend disbelief.

    I'd much rather the increase in GPU power be used through a GPGPU API for artificial intelligence, advanced physics simulations, fluid dynamics, flocking behavior or other things which could really add to gameplay.

    A few extra reflections and slightly softer shadows???? I won't even notice and neither will the average gamer.

    1. Re:What a waste of resources by jamie · · Score: 4, Insightful

      I kind of assumed the big win was that game development gets easier. If your game is rendered by ray-tracing can't you spend more time on building the models, lighting and gameplay and less on fine-tuning rendering tricks?

    2. Re:What a waste of resources by Anonymous Coward · · Score: 0

      Problem is we are reaching the design limitations of the current technology, whilst ray tracing is currently primitive it has a bigger future then the current tech.

    3. Re:What a waste of resources by Big_Breaker · · Score: 4, Informative

      The gameplay improvement is in deformable physical environments. Combined with mainstream physics engines, raytracing would allow for a sea-change in gameplay by allowing interactive gaming environments.

      Raster methods rely on a bunch of tricks, many of which need to be precalculated for static maps. The most obvious example is binary space partioning tables. This leads to very static feeling environments that disallow interaction beyond doors of various types and moving platforms.

    4. Re:What a waste of resources by pak9rabid · · Score: 1

      I'd much rather the increase in GPU power be used through a GPGPU API for artificial intelligence, advanced physics simulations, fluid dynamics, flocking behavior or other things which could really add to gameplay.

      I think you're missing the purpose of what a graphics processing unit is for.

    5. Re:What a waste of resources by fluffykitty1234 · · Score: 1

      Hmm? I don't think so. BSP trees and other tricks like this allow you to only draw what's visible. When you are ray-tracing you would still need a way to prune away geometry that isn't visible, or it would take forever to find ray itersections.

    6. Re:What a waste of resources by caramelcarrot · · Score: 1

      Uh, you do know that in order to do ray-tracing efficiently, you have to use data structures like bsps, kdtrees and octrees? I.e. Ones that you don't want to have to rebuild every frame because something moved. I get your point that you won't have to precompute lighting for radiosity, but even raytracing often using radiosity for global illumination.

    7. Re:What a waste of resources by jacquesm · · Score: 1

      Better yet, you have to have a way to prune that which is not reflecting into an area that is visible!

      And then there's all kind of colour-bleed effects, they can also come from objects that are out of the directly visible area. Light sources, particles in your simulated atmosphere that light can scatter off and so on. There's a lot that goes into deciding which parts of your scene are 'active'.

    8. Re:What a waste of resources by bmajik · · Score: 0

      I think the big benefit here is that scenes will be considerably more realistic, especially under motion.

      GT has never done an _authentic_ job at anything. Most of its vehicle physics and eye-candy are faked, and in the case of vehicle dynamics, badly.

      In the GT shot you've posted, the cars appear to have reflective surfaces, interacting shadows, etc. But the background doesn't. Why not? Becuase people look at cars.

      Now look at the nVidia shot. Note that you can see the sign and bridge geometry reflected in the glass of the building. What are the odds that, if you had to explicitly program for this behavior, that a developer would bother to render accurate geometry reflections into individual panes of glass on a building.

      That's computationally expensive, it takes a lot of programmer protein power, and it's just not something you'll see done with game-style renderers.

      But raycasting and other light-simulation techniques are interesting precisely because they aren't optimized for any particular application (i.e. cars that look wet in dead boring backgrounds).

      I think that as scene rendering and lighting becomes more general (instead of relying on tricks for certain desired effects), scenes will become over all more realistic, and things will start happening that nobody expected or planned for.

      The "shadows" in that GT shot are awful. I don't want a vague, aliased, dark-map cast uniformly off the front glass. I want the reflection of the building to be cast over the windshield. I want the curvature of the glass to be apparent because the reflection has distorted over its contours. I want the reflected building to be translucent because when looking at a reflection on a peice of glass, there's a mix of reflection and transmission. Since the reflected building it itself covered in glass, i want to see reflected reflections on the windshield. Since light energy is lost at each reflective stage, i expect there to be more image attenuation between the 1st and 2nd reflected images.

      And I want all of it to move and change realistically as the car moves (and the building doesn't).

      It's not that this level of accuracy will make racing more fun (or that it would affect me anyway since i use exclusively in-car views (although it would be cool to see glare/reflection across the glass modelled properly!), it's that the world will become more accurate and immersive _in general_ in ways that developers didn't plan for.

      Procedurally trying to acheive this effect is going to be painful. Shooting light energy into a completely described 3d world and letting the chips fall where they may is easier and more accurate. The issue is that it's never been plausible for real time applications.

      --
      My opinions are my own, and do not necessarily represent those of my employer.
    9. Re:What a waste of resources by Big_Breaker · · Score: 1

      I see your point and the point of fluffykitty. I guess my intuition is that the tricks are piled on so thick in your typical raster based FPS engine that in order to simulate lighting, reflections and transparency effects certain things, like the environment, need to remain static.

      Raytracing addresses some of these shortcomings of raster so that these tricks need not be used, a more ground up approach. That way the environment can be more dynamic as its features are recalculated in real time.

    10. Re:What a waste of resources by roystgnr · · Score: 1

      I think you're missing the purpose of what a graphics processing unit is for.

      No, he's going beyond the purpose of what a graphics processing unit was originally for, and looking ahead to what General Purpose GPU computing is going to be for. There's nothing in the GPU that requires it to operate on polygons; the silicon is there to do parallel stream processing, and streams of geometry/texture/lighting/etc data are just one of the ways to use that.

    11. Re:What a waste of resources by Anonymous Coward · · Score: 0

      Raytracing does easily:
              - Sharp reflections
              - Sharp shadows
              - (oh yeah, procedural textures too) (* although these rasteriser cards now have "shaders" which support that)

      Not so easy still:
              - Soft shadows and reflections

    12. Re:What a waste of resources by hardburn · · Score: 1

      Both nVidea and ATI^H^H^H AMD are pushing GPUs for use for general computing tasks. The sheer number of transistors dedicated to processing means that even though they're tuned towards graphics, they can still be faster than a CPU for many tasks. Even when they're slower, they can be better on a price or energy usage point of view.

      --
      Not a typewriter
    13. Re:What a waste of resources by Traa · · Score: 1

      I'd much rather the increase in GPU power be used through a GPGPU API for artificial intelligence, advanced physics simulations, fluid dynamics, flocking behavior or other things which could really add to gameplay.

      You would have loved the AMD talk on advanced rendering techniques at Siggraph. Using DirectX 10.1 techniques they demonstrated a gaming scenario with 3000 characters (froblins, don't ask), each 1.2 million triangles at the highest level of detail, that demonstrated flocking behavior, AI, path finding (using fluid dynamics), animations and interaction....all on the GPU.

      Very impressive stuff, great use of the GPU.

      Should be a large paper about this on the AMD website somewhere, I'll try to find it for you.

    14. Re:What a waste of resources by Traa · · Score: 1

      The paper is called March of the Froblins

    15. Re:What a waste of resources by Anonymous Coward · · Score: 0

      GT3 is "best of breed" of current rasterization and the raytracing is a pre-alpha-technology-mockup.

      Even comparing the two together speak tons for the future of raytacing.

      current rasterization tricks are good enough that you suspend disbelief.

      I don't know anybody who would mistake current rasterization games for real world images.

    16. Re:What a waste of resources by Anonymous Coward · · Score: 0

      I think those GT3 shots were pre-rendered, probably on a raytracing renderer. They didn't just take the in-game assets and rendered it using the game engine in hi-res.

    17. Re:What a waste of resources by rogerbo · · Score: 1

      In the GT shot you've posted, the cars appear to have reflective surfaces, interacting shadows, etc. But the background doesn't. Why not? Becuase people look at cars.

      Now look at the nVidia shot. Note that you can see the sign and bridge geometry reflected in the glass of the building.

      Who cares? When I'm playing a racing game the fact that the buildings aren't reflecting the surroundings accurately doesn't detract from my gaming experience at all.

      They could have put a reflection map on the buildings using a rasterisation technique and 95 percent of people would never notice any problem.

    18. Re:What a waste of resources by CaseyB · · Score: 1

      Most of its vehicle physics and eye-candy are faked

      As opposed to what? Putting real cars inside your monitor?

    19. Re:What a waste of resources by bmajik · · Score: 0

      Who cares? When I'm playing a racing game the fact that the buildings aren't reflecting the surroundings accurately doesn't detract from my gaming experience at all.

      I suppose that would be fine if all games were racing games and all gamers were you.

      --
      My opinions are my own, and do not necessarily represent those of my employer.
    20. Re:What a waste of resources by bjorniac · · Score: 1

      Until someone does it right, and you do: Imagine you walk towards a room with a light on inside, then the light goes off and you can see the reflected image of someone standing behind you in the now dark glass. At the moment, of course, this can be directly programmed, but with ray tracing you will see a lot more emergent gameplay - a metal door can become a mirror to see around a corner, a flashlight can be bounced off a surface to temporarily blind someone/lose their night-vision. Again, this can all be done right now, but it's clunky and forced but soon it will be dynamic. The changes that accurate physics made (in say, Half-life 2) were huge to emergent gameplay - I could throw a grenade behind a box to get the goodies in it to come to me, instead of fighting my way over to it etc. It's up to the developer of course to work out how to make environments that take advantage of it, but with a bit of imagination they should be able to do some cool stuff.

    21. Re:What a waste of resources by bmajik · · Score: 0

      I see you were compiled with -Wpedantic today.

      What I should have said is, "as someone with a fair bit of race track experience, I find that the authenticity of the simulation of the vehicle dynamics presented in the Gran Turismo series of games is lacking compared to other racing games and simulators I have experience with"

      More concretely, the GT cars appear to respond to inputs differently than I would expect, to the point that it is frustrating at times.

      All simulation games have a model of reality that dictates the behavior of the cars in response to given stimuli. Based on my experience with the GT series, their models exhibit problems which stick out to me. A notable problem is the propensity of the cars to retain front tire turn-in grip when you are performing maximum braking, or worse, have locked up the front wheels entirely. If you lock up the front tires and then dial in a giant bundle of steering lock, the car sails straight. Sims like RACER are very in-your-face about this (on certain cars)... in GT i've "gotten away with it" when I felt I shouldn't have.

      On the other end of the spectrum, keeping a base model BMW going in a straight line at high speed on the Nurburgring in GT can be challenging. Given that I've driven the Nurburgring __in real life__ and had no such difficulties, there is clearly something fishy going on there.

      In the latter case, I think this may partially be an issue of Logitech DFP input mappings or similar. The braking+turning stuff is much more likely to be an issue of insufficient dynamics model in the game. You cannot even explain it away with obnoxiouisly high tire coefficients since a locked tire should not be turning a vehicle. IOW - its not that certain simulation parameters are wrong, it is almost certainly the case that the simulation is wrong.

      --
      My opinions are my own, and do not necessarily represent those of my employer.
    22. Re:What a waste of resources by tgd · · Score: 3, Insightful

      The people writing the game, in most cases, are buying a library from a 3rd party.

      There's not much of a gain to be had from that.

      We used to joke about realtime ray tracing being two years away when I was in college.

      Fifteen years ago.

      The problem is, its always slower than rasterizing. You can get faster hardware, but as soon as you do people want bigger textures, higher resolution, more polygons and suddenly once again raytracing is too slow.

    23. Re:What a waste of resources by mdwh2 · · Score: 1

      I guess my intuition is that the tricks are piled on so thick in your typical raster based FPS engine that in order to simulate lighting, reflections and transparency effects certain things, like the environment, need to remain static.

      Lighting, reflections, transparency (and shadows, come to that) can all be done dynamically these days. Is there are particular thing you are thinking of that still requires static items with rasterising?

    24. Re:What a waste of resources by Big_Breaker · · Score: 1

      I guess the latest Id engine shows that there are no particular limits to raster as you suggest as long as the optical effect can be pushed through a shader.

      It is certainly the case that 3d gaming environments are typically static, and except for bleeding edge gaming engines, they have to be.

      If rendering isn't the reason for static game worlds, why aren't there more?

    25. Re:What a waste of resources by Qhartb · · Score: 0

      Ray-tracing relies on similar tricks. You mention, for instance, space partitioning structures. Similar structures, like kd-trees, are the secret to raytracing's much-touted O( log(n) ) performance (where n = number of triangles).

      Less talked-about is the fact that if the scene moves, the structure needs to be updated or regenerated, and that making a structure that gets you O( log(n) ) rendering takes something like O( n*log(n) ). Note how real-time raytracing demos boasting 10,000s of polygons have no animation. The deformable environments you mention would be hell on a ray-tracer.

      Of course, there are ways to cheat this, but my point is that ray-tracing relies on its own bag of tricks. If performance is no object, the theory in a ray-tracer is simple and the code elegant. Real-time ray-tracing can be just as ugly as rasterization.

    26. Re:What a waste of resources by pak9rabid · · Score: 1

      Right, but with ray-tracing on their plate for the near future, I'd rather see them utilize their GPUs to accomplish that, then utilize it for other things. Work on rendering ray-traced scenes in real-time first, then deal with the other stuff (physics, AI, etc) later.

    27. Re:What a waste of resources by smoker2 · · Score: 1

      And 640k should be enough for anyone.

    28. Re:What a waste of resources by mdwh2 · · Score: 1

      Static geometry makes it easier to determine which particular parts of the world are visible from any given viewpoint. And this would be true for raytracing also.

      I mean, ask yourself why there aren't any raytracing games with dynamic worlds, if it's so much better?

    29. Re:What a waste of resources by rogerbo · · Score: 1

      sorry they were from gt5 prologue, not Gt3, and I have a PS3 in front of me and the reflections on the cars and the shadows look pretty close to that screen shot in real gameplay.

    30. Re:What a waste of resources by rogerbo · · Score: 1

      it's a matter of choosing where to spend the resources. From what I understand the increase in GPU power to move from current level rasterization tricks to the ray tracing in the Nvidia demo is enormous.

      Imagine two game developers, one decides to use ray tracing in a game, and it uses the full power of the GPU. One decides to stick with the best rasterization tricks and uses the remaining "left over" GPU power through a GPGPU API for AI, flocking, massive crowd simulation, fluid dynamics etc.

      Which game do you think will be more fun to play and offer more groundbreaking gameplay? The one with the soft shadows and real reflections or the second one I describe?

    31. Re:What a waste of resources by GordonS3 · · Score: 1

      The gameplay improvement is in deformable physical environments. Combined with mainstream physics engines, raytracing would allow for a sea-change in gameplay by allowing interactive gaming environments

      You must be a middle manager of some kind. "sea-change"?! Pah! What gaming environment is not interactive?! How exactly can we not use mainstream physics engines without raytracing? How can we not have deformable physical environments without raytracing?

  9. nAwesome by Anonymous Coward · · Score: 0

    I love this company. Always moving ahead.

  10. Re:That's a nice canned post ya got there by Colonel+Korn · · Score: 4, Informative

    Story submitted: 11:00am. Your post submitted: 11:00am. There's just no way in hell you formulated a response and typed out all of that in less than a minute. So just what kind of douchebag are you, anyway?

    Uh, maybe he looked at the story on the firehose.

    --
    "I zero-index my hamsters" - Willtor (147206)
  11. No Dirt... by griffman99h · · Score: 1

    ...there's still no dirt. I wont be impressed till I see some grime build up on a fender in realtime.

    1. Re:No Dirt... by Tubal-Cain · · Score: 3, Funny

      Head up the stairs (to ground level), down the hall, and out the door. You should be able to find a bumper to stare at for a few hours.

  12. Of course, but when? by MojoRilla · · Score: 5, Insightful

    Of course ray tracing, or one of its decendants, like photon mapping, will end up dominant. The question is when. Ray tracing is used now for rendering movies like Cars, which are probably pretty much state of the art for computer graphics, and would be used for things like PC games except that is so computationally expensive.

    As to when rasterization will be replaced, the short answer is not any time soon. The article's title is misleading. It says "Intel: Rasterisation will be replaced in five years", while Intel's ray tracing guru Daniel Pohl actually says "Looking ahead five to ten years from now, I believe that rasterisation will be used less and less in games". Big difference there.

    So, I think this will progress quickly, but we won't be getting rid of rasterization any time soon.

    1. Re:Of course, but when? by kripkenstein · · Score: 1

      I tend to agree. While ray tracing has a lot of potential, it is several years away from being usable on mainstream hardware, and in addition, during that time rasterization technology will further improve. So even if this demo will work on gaming PCs in 5 years, rasterization graphics may look amazingly better by that time.

    2. Re:Of course, but when? by pr0nbot · · Score: 1

      IThe thing I like about ray tracing is that it's more or less pure physics (as opposed to current GPU rendering is a clever box of tricks and hacks and approximations). So, you can implement a glorious renderer today, and just wait for the hardware to catch up.

    3. Re:Of course, but when? by Hoi+Polloi · · Score: 1

      Maybe the best course would be to selectively use ray tracing. Use rasterization for backgrounds and other low visibility or undynamic images and use ray tracing for the focus of the viewer's attention. That dragon you are fighting may benefit from ray tracing but the tree behind it in the far distance could just be rasterized.

      --
      It is by the juice of the coffee bean that thoughts acquire speed, the teeth acquire stains. The stains become a warning
    4. Re:Of course, but when? by Anonymous Coward · · Score: 0

      Just a nitpick: Photon mapping is a global illumination algorithm, not a rendering algorithm. You can use photon mapping with ray-tracing or rasterization, but it does not actually produce any sort of meaningful picture output on it's own. However, it is strongly based on ray-tracing (which is used to find the path a photon will take), so any sort of hardware advancement in ray-tracing will make real-time photon mapping closer to being a reality.

    5. Re:Of course, but when? by overd-ose · · Score: 1

      Actually, Pixar's RenderMan implementation employs the Reyes rendering algorithm to minimize the amount of ray tracing necessary. This technique tessellates high order surfaces into sub-pixel size quads (micropolygons) and rasterizes that geometry. Interesting stuff: http://en.wikipedia.org/wiki/Reyes_algorithm

      --
      i like grapes
    6. Re:Of course, but when? by JFMulder · · Score: 1

      Actually, I was at a presentation done by Pixar last year and they said that for Ratatouille, they ditched raytracing for rasterising. The reason why is that they realised early on that most of the scenes were taking place where there was a lot of metals and glasses with translucent liquids (like wine). They wanted to have nice effects, but rendering a single glass of wine took hours for a single frame. In the end, they decided to rasterize those scenes instead faking the caustics with environment maps and other raster tricks. I don't know if they used the same technique for Wall-E

  13. Re:"favour"? by oldspewey · · Score: 1

    You missed birchbark canoes, the Trailer Park Boys, real beer, Timmy's, the CBC, and back bacon eh?

    --
    If libertarians are so opposed to effective government, why don't they all move to Somalia?
  14. good but... by Bizzeh · · Score: 1

    this looks pretty nice, and its cool that nvidia are working on getting ray tracing methods down to an acceptable level of processing... but the problem i can see with this.... everything within all those images looks like its made from the same plastic that samsung tvs are made out of... everything is... TOO reflective.

  15. Seen it by Anonymous Coward · · Score: 0

    The article mentioned something about Intel doing this with a Quake 3 engine. I remember seeing some guys who built their own GPU using a FPGA that did hardware ray tracing using a modified version of the Quake 3 engine when it first went open sourced. They couldn't do anything real time because of the speed restrictions of an FPGA but I remember it looking as good or better than this. Good to know that the professional hardware developers are finally catching up with the hobbyists.

  16. Still waiting... by Ukab+the+Great · · Score: 5, Funny

    I want NVIDIA to come out with a card that gives boring DOOM clones intriguing plots and compelling gameplay.

    1. Re:Still waiting... by kenp2002 · · Score: 1

      I can fix it:

      East side cop finds his ex-wife dead and he's the primary suspect.

      Digging into the murder he finds his wife moonlighting with a weird cult operating out of the boiler room of a massive fast food corporation.

      They open a gate to hell, the cop gets possessed and proceeds to murder the entire city as a demon.

      Once the last soul is taken the gate seals up and the event becomes a cold case.

      30 years later: An ex-army gone PI chases down a filandering wife and her boyfriend to the ruins of an abandon metropolis in the western part of Brazil where something bad happened 30 years ago...

      --
      -=[ Who Is John Galt? ]=-
    2. Re:Still waiting... by Dan9999 · · Score: 1

      what makes me laugh is that there are some who modded this insightful instead of funny.

    3. Re:Still waiting... by Luyseyal · · Score: 1

      The truth is insightful.

      -l

      --
      Help cure AIDS, cancer, and more. Donate your unused computer time to worldcommunitygrid.org. Join Team Slashdot!
    4. Re:Still waiting... by i_liek_turtles · · Score: 1

      He then gets sent to fight in World War II for undisclosed reasons.

    5. Re:Still waiting... by Dan9999 · · Score: 1

      good point, but I took insightful as meaning that some would like to use their toilet to cook their breakfast.

  17. I wonder by ThePhilips · · Score: 2, Insightful

    Meanwhile reader arcticstoat passes on Intel's latest claim that rasterisation will die out the next few years, possibly in favour of ray tracing.

    I'd love that to happen.

    But reality is that several best games I have played were ... 2D.

    Intel, Good luck adding RT to 2D graphics. ;)

    RT in my experience is rather expensive - on end of development. Not all games manage to exploit all lighting models. And RT needs that even more than actual 3D graphics. It would take some long time for games to adopt it. On side of CADs picture is much simpler: they are easy to fork $$$ for good and fast rendering.

    --
    All hope abandon ye who enter here.
  18. What this will mean for games by Tibor+the+Hun · · Score: 4, Funny

    Like someone said a few months back, now all the games will be composed entirely of shiny balls, toruses, and checkerboards.

    --
    If you don't know what AltaVista is (was), get off my lawn.
    1. Re:What this will mean for games by CaseyB · · Score: 4, Funny

      Don't forget teapots!

    2. Re:What this will mean for games by Anonymous Coward · · Score: 0

      Like someone said a few months back, now all the games will be composed entirely of shiny balls, toruses, and checkerboards.

      You forgot teapots.

    3. Re:What this will mean for games by Spatial · · Score: 1

      And that damn baby head...

    4. Re:What this will mean for games by Anonymous Coward · · Score: 0

      Teapots are just special-case torus', so he's already covered
      (Yes, I realize it's a joke!)

    5. Re:What this will mean for games by Funk_dat69 · · Score: 1

      And bunnies!

      --
      FUNK!
    6. Re:What this will mean for games by not-enough-info · · Score: 1

      And Bunnies!

      --
      ---k--
      </stupid>
    7. Re:What this will mean for games by Anonymous Coward · · Score: 0

      Who modded this funny?

      He already said toruses.

    8. Re:What this will mean for games by YetAnotherLogin · · Score: 1

      He already covered that case. Even the Stanford bunny is equivalent to a torus.

  19. Not quite yet, I'm affraid by Moraelin · · Score: 3, Interesting

    Actually, I think that the biggest problem with animated porn isn't texturing-vs-raytracing, but the models and animations. Last time I ran into some adult Poser-fu, it looked all wrong in a massively uncanny-valley way, and not because of the texturing.

    And with the animations, well, I'd assume it's actually easier and cheaper to find a gal who'll bounce on a cock for half an hour for a few (thousand) bucks, than a highly skilled artist and animator who'll make that look natural.

    Plus, raytracing is IMHO entirely the wrong secret sauce there. Ray tracing works best for sharp, metallic/mirror reflections. Because then you can take each ray and reflect it as one ray. If you want to go diffuse, that's a lot more expensive with ray-tracing. Then you need to split each ray into sub-rays that reflect into slightly different directions from there. Same as anti-aliasing is done by calculating sub-pixels, basically.

    I.e., ray tracing looks grrreat and is the cheapest for shiny cars, crystal cups, and the like. Which is why everyone ray-traces cars and the like. It sucks for something like human skin, unless, of course, you want to make those humans look like polished shiny plastic dolls.

    So, well, I can't imagine that much need for it in porn at the moment. Unless, of course, you want to make a Transformers sex movie. Or maybe one with liquid- metal Terminators fucking. (Hey, they must have made some female versions too, right?;)

    --
    A polar bear is a cartesian bear after a coordinate transform.
    1. Re:Not quite yet, I'm affraid by Anonymous Coward · · Score: 2, Informative

      I.e., ray tracing looks grrreat and is the cheapest for shiny cars, crystal cups, and the like. Which is why everyone ray-traces cars and the like. It sucks for something like human skin, unless, of course, you want to make those humans look like polished shiny plastic dolls.

      Subsurface Scattering.
      http://en.wikipedia.org/wiki/Subsurface_scattering

      http://graphics.ucsd.edu/~henrik/images/imgs/layered_skin_model.jpg

    2. Re:Not quite yet, I'm affraid by Anonymous Coward · · Score: 1, Informative

      I imagine that all the examples use highly reflective materials to dramatically show it's accuracy in the number of bounces being made -- not because that's the only thing it's good at rendering.
      I'm excited about ray-tracing for specular reflections and subsurface scattering. Obviously you don't want to render everything with mirror-like reflections but virtually everything has some level -- however slight -- of reflection. Also good fresnel reflection [increasing reflection at increasingly-oblique angles] will help bring realism to renderings.

    3. Re:Not quite yet, I'm affraid by Anonymous Coward · · Score: 0

      a gal who'll bounce on a cock for half an hour

      You just hit my manhood straight in the... manhood. Thanks :(

    4. Re:Not quite yet, I'm affraid by Anonymous Coward · · Score: 0

      Or maybe one with liquid- metal Terminators fucking. (Hey, they must have made some female versions too, right?;)

      You haven't been keeping up - watch the Sarah Connors Chronicles.

  20. Pity they don't include an RPU by argent · · Score: 1

    Slusallek's RPU was doing real-time raytracing with about the same hardware as Rage Pro, three years ago.

    They could afford to put a couple of souped up RPUs in the corner of a GPU without noticing the "lost" transistors and get this kind of real-time raytracing on entry-level GPUs *today*.

  21. Back at you... by Anonymous Coward · · Score: 0
    So what if Intel thinks Ray Tracing is the future, I doubt it will happen anytime soon and I don't think the Larrabee will offer anything good for rasterisation now. It will be garbage just like there current GMA offerings.

    This is just Nvidia saying to Intel "If you think you can get ahead by your claims of ray tracing being the only future we'll show you we can do ray tracing and not only compete with you but we will crush the Larrabee GPU."

  22. I, for one, welcome our ray traced overlords by caywen · · Score: 3, Interesting

    I think raytracing is going to be the only way to go in 5-10 years. As poly counts go up, average poly size is going to drop below 1 pixel, and the overhead of tracing rays will match the overhead of rasterizing tiny polys + scene overdraw, etc. Of course, what I fear is that all the games will start feature too much reflection on every surface, because they can.

    1. Re:I, for one, welcome our ray traced overlords by Anonymous Coward · · Score: 0

      Poly counts aren't going to get that high because perf takes a nosedive when the polygons are less than 4 or 16 pixels.

    2. Re:I, for one, welcome our ray traced overlords by Anonymous Coward · · Score: 0

      I very much doubt we will operate with polygons once there is no rasterization, rather NURBS (or something similar), whereby the surface is more of an equation than a set of polygons.

      This means less strain on the modellers who don't need to tweak millions of polygons, and better quality output (you can zoom in infinitly on the edge of a curved polygon and it will still appear curved).

    3. Re:I, for one, welcome our ray traced overlords by Anonymous Coward · · Score: 0

      ... Of course, what I fear is that all the games will start feature too much reflection on every surface, because they can.

      What? I couldn't quite make that out, I'm blinded by the bloom and HDR.

    4. Re:I, for one, welcome our ray traced overlords by NeMon'ess · · Score: 1

      Games generally don't need smaller polygons, they need more objects. Natural areas need denser grass and ground cover to look better. Also more and better rocks instead of flat paths pretending to be gravel and dirt. Man-made areas need objects on the sidewalks or in rooms to look more realistic. Ever more detailed textures will really make the final difference. When asphalt or stone looks as detailed as the real thing, and not a blurred texture, that'll be a huge milestone.

  23. Re:That's a nice canned post ya got there by jacquesm · · Score: 4, Insightful

    You're just sore that you didn't get to annoy a lot of people with your frist post or gnaa rubbish.

  24. Scene search by caywen · · Score: 1

    I wonder how much control future ray tracing engines will give the programmer over scene search algorithms? If I'm doing a top-down terrain render, I'd want to have a super-simple O(C) first hit time (probably using a simple grid/bucket). If I were doing a space scene, maybe an oct tree. In either case, it sounds like a problem that can't easily be hardware accelerated.

    1. Re:Scene search by nobodylocalhost · · Score: 1

      You wont be using any space parsing trees other than n-d tree, of course you can adjust the levels of an n-d tree in order to specify the rendering complexity. And yes, there is already hardware acceleration available for n-d tree, it was implemented in the famed ray tracing chip. http://graphics.cs.uni-sb.de/~woop/rpu/rpu.html

      --
      Where is the "Ignorant" mod tag?
  25. Crysis now run? by TheDarkMaster · · Score: 1

    Maybe will be possible to play Crysis now

    --
    Religion: The greatest weapon of mass destruction of all time
  26. Is that you, Pete Shirley? by Aidtopia · · Score: 1

    I wonder if this is what Pete Shirley went to NVIDIA to do.

    The article and some of the comments here have oft-repeated myths about ray tracing. For example, ray tracing algorithms are generally simpler than rasterization algorithms, not more complex--though they do require more processing power.

    One commenter said this demo was limited to 3 rays per pixel. That may not be true. The article said each ray was limited to 3 bounces. That doesn't preclude firing multiple primary rays per pixel for antialiasing. From the images, though, it doesn't look like there's much antialiasing.

    I'm also disappointed to read that it's still done with bajillions of polygons. One potential advantage of ray tracing is rendering smooth, curved surfaces. Few of the real-timers seem to attempt this though. Too bad, the savings in memory for the model could improve the transform and rendering times.

  27. Re:That's a nice canned post ya got there by omnipresentbob · · Score: 1

    That, and he's a subscriber.

  28. Re:That's a nice canned post ya got there by ceoyoyo · · Score: 1

    I dunno. I just read a shareholder update from a privately held startup that I happen to know the real stories behind. That post sounds a LOT like it... enough spin to make you dizzy.

    Not that the screen shots aren't pretty, but they're not revolutionary (yet).

  29. Rack of Servers?? by mounthood · · Score: 1
    The article says its running on "four next-generation Quadro GPUs in an NVIDIA Quadro Plex 2100 D4 Visual Computing System (VCS)" which isn't out yet, but the S4 VCS is a rack of 4 servers that costs about $14,000.

    http://www.nvidia.com/object/quadroplex_s4.html

    --
    tomorrow who's gonna fuss
  30. Why a demo with a Volkswagon?! by G3ckoG33k · · Score: 1

    Why a demo a Volkswagon?! That Volkswagon Veyron is not even good looking in real life... :(

  31. Dollars $$$ by Nom+du+Keyboard · · Score: 1

    And this systems costs how much? And consumes how much power? And needs how many P/S's to operate?

    They may have put on a desktop what it took a supercomputer to do before, but it's still a very expensive desktop, which makes it interesting, but hardly useful to 99% of us.

    --
    "It's the height of ridiculousness to say for those 9 lines you get hundreds of millions."
  32. Re:That's a nice canned post ya got there by xerxesVII · · Score: 4, Funny

    Um, he's fucking Batman. He can do all kinds of stuff.

    --
    "We shall grapple with the ineffable, and see if we may not eff it after all." - Douglas Adams
  33. Here's a picture by Ihmhi · · Score: 3, Funny

    ...of the card that will be needed to run raytracing.

    I guess you micro form factor guys are kinda screwed.

  34. Did I miss something by Anonymous Coward · · Score: 0

    I'm so impressed at these INTERACTIVE SCREENSHOTS. Wow, I can look at them and they... uh, they look back at me?

    Show us the video or go home. We've all seen raytracing before as still images.

  35. that's barely ray tracing by Anonymous Coward · · Score: 0

    This is barely ray tracing. Hard shadows.

    They've limited the ray bounces to only 3 to get the 30 fps performance. Real ray tracing requires many ray bounces to achieve realism. Hollywood films have 400+ ray bounces per pixel. This is a heavily constrained demo to achieve decent performance on a $10K+ system. Nice try Nvidia.

  36. Polygons? by ishmalius · · Score: 1

    The whole point of ray tracing is to be able to model anything and how light interacts with it. Surely they mean the -equivalent- of that many polygons.

  37. NVIDIA doubts ray tracing is the future of games by heroine · · Score: 1

    http://games.slashdot.org/article.pl?sid=08/03/07/1659250&from=rss

    Quite a change in 5 months. Ray tracing is the way it should be done. No more hand coding reflections and shadows.

  38. That's nothing. Cell can raytrace a whole city! by Funk_dat69 · · Score: 4, Interesting

    Speaking of Ray Tracing...

    Check out this video showed at SIGGRAPH this week of the University of Virginia Rome model being ray traced in real time by a Cell Blade:

    http://www.youtube.com/watch?v=YZnbMWy9A0Y

    Nifty!

    --
    FUNK!
    1. Re:That's nothing. Cell can raytrace a whole city! by Anonymous Coward · · Score: 0

      Very cool. That would be six teraflops minimum, thank you. Has Intel released any estimation on Larrabee performance (in teraflops) yet?

  39. Re:That's a nice canned post ya got there by Anonymous Coward · · Score: 0

    I'd rather be fucking Rachel Dawes. Well, not now anyway. I never was into the necrophilia thing.

  40. 1920x1080 by Areyoukiddingme · · Score: 2, Insightful

    Is anybody else saddened by this horrible sneaking invasion of crappy HD resolution as the benchmark for renderers? I mourn for the loss of display resolution progress. Have people forgotten 2560x1600? Has anybody else noticed that 1200 vertical pixels has become vanishingly rare in monitors? If this is the future as brought to us by LCD panels, I'm really really not liking it. They blur and smear when things move fast, pixels go dead on them all the stinking time, and HD TV is crippling resolution because Joe Sixpack (and presumably corporate America) has no idea why more pixels is better.

    You can pry my 21" flat Trinitron CRT from my cold dead hands.

    1. Re:1920x1080 by Anonymous Coward · · Score: 0

      Your not the only one, but its not only resolution but dot pitch as well. I want a monitor where the pixels are small enough I don't need antialiasing. Sadly (except for the 15" 1920x1200 display on my laptop) all of the high resolution monitors also tend to be physcically large. I don't want a 30" monitor to get a resolution over 2k horizontal. Those monitors look like crap.

    2. Re:1920x1080 by Siridar · · Score: 1

      I beg to differ. I went from a 21" Trinitron (Sony GDM-520FD) connected via BNC to a Dell 24" LCD (2407-WFP) and the difference is like night and day. The CRT was the best of the best back in the day, but in terms of clarity, sharpness, and colour reproduction, the LCD gets my vote. Not having to calibrate the screen to different resolutions, no interference from my Logitech G25 (why, yes, I *do* have a lot of toys) and the massive difference in desk space (the Sony was 50cm from glass to plugs, and weighed about 30kg) are other plusses.

  41. Not impressive by Anonymous Coward · · Score: 0

    These screenshots look worse than many games that are already shipped and available on consoles. Raytracing has it's uses, but unless you just want a plastic looking car on a plastic road in a plastic city, there are better approximations of lighting already available that are more efficient.

  42. at $10K USD, not for games by markjhood2003 · · Score: 1

    You guys realize that the system being demoed here, the NVIDIA Quadro Plex 2100 D4 Visual Computing System, with 4 next-generation Quadro GPUs, starts at $10,750, right? http://www.nvidia.com/object/io_1218520087945.html

  43. a bug's life by hierophanta · · Score: 1

    i would have loved to see this hardware render the 10 second coke bottle clip in 'a bugs life'. IIRC that scene cost as much (in render time) as the rest of the movie on account of it being ray-traced. how many millions might Pixar save by moving to this?

    1. Re:a bug's life by peektwice · · Score: 1

      0. It wasn't available then.

      --
      Other than this text, there is no discernible information contained in this sig.
  44. LONG LIVE VOXELS! by mmarshall · · Score: 1

    ...rasterisation will die out the next few years, possibly in favour of ray tracing.

    I'm curious... what other than ray tracing could replace rasterisation?

    LONG LIVE VOXELS!

  45. The one thing I want by Chemisor · · Score: 2, Interesting

    > But reality is that several best games I have played were ... 2D.

    Damn right. And you know what capability I would really like to have on a card? Masked blit. That is the single most time-consuming operation on all 2D games. If you could copy all your tiles to the video card memory and then maskblit them onto the visible page (or blit and flip if it's too slow), that would really make 2D games smooth as silk and leave more CPU power for AI and real gameplay.

    1. Re:The one thing I want by Hektor_Troy · · Score: 1

      Simple question: What is "masked blit"? What does it do? I've no clue and Google wasn't helpful, nor was Wiki ...

      --
      We do not live in the 21st century. We live in the 20 second century.
  46. Re:That's a nice canned post ya got there by redstar427 · · Score: 1

    He lives in a different timezone, of course.

    --
    "Two things are infinite: the universe and human stupidity; and I'm not sure about the universe." Albert Einstein
  47. What does "fully interactive" mean? by j1m+5n0w · · Score: 1

    Without video, it's hard to know what they mean by "fully interactive". (Perhaps someone who saw the demo at siggraph can enlighten me.) Do they mean, interactive framerates with static geometry and a moving camera? Or do they actually move the geometry around in the scene? From the screenshots, I would guess that maybe the car moves and not much else. If that's the case, it's a impressive demo, but it doesn't address one of the hardest problems, which is how do you re-arrange the acceleration structures when anything moves?

    A good implementation of BIH ought to be able to handle a few thousands or tens of thousands of independently moving objects at reasonable framerates, but traversal costs are usually a bit higher than a SAH-based kd-tree. High frame-rates are neat, but if they had to spend an hour or two to generated an optimized tree, then their accomplishment is a little less impressive. (Which isn't to say that it's not impressive at all.)

  48. Rant about the website. by serviscope_minor · · Score: 1

    Why on earth do websites like this on insist on using Javascript instead of a normal href when you click on the images? It's like the last 5 years of tabbed browsing have passed these guys because it will only open the larger images in a new window, rather than a new tab. Way to go guys. Welcome to 2003.

    --
    SJW n. One who posts facts.
  49. Bounces, not casts. by iansmith · · Score: 1

    The "3 casts" is actually "3 bounces" in the article. It has nothing to do with anti-aliasing or shadows.

    They are talking about how many surfaces a single ray can refract through or reflect off of.

    With no bounces, you get no reflections.

    With one bounce, you can see the reflection of the room in a mirror.

    With two bounces, you can see the room reflected in a chrome sphere viewed in a mirror.

    With three bounces, you can see a tiny reflection of the mirror on the surface of the chrome sphere that is seen through the reflection of the mirror.

    With four bounces.. well you get the idea.

    Three is good enough so you would have to look REAL close to notice the limited reflections unless you loaded up CHROME_WORLD.WAD on your server.

  50. Ray tracing potential by religious+freak · · Score: 1

    I just feel the need to share. I was looking up the tech specs in wikipedia on raytracing and came across this. I suppose this is what is possible and what raytracing will look like in the future.

    http://en.wikipedia.org/wiki/Image:Glasses_800_edit.png

    Amazing

    --
    If you can read this... 01110101 01110010 00100000 01100001 00100000 01100111 01100101 01100101 01101011
  51. Finally! by thexile · · Score: 1

    Something that can run Vista and Duke Nukem Forever.

  52. masked blit by Chemisor · · Score: 1

    for (coord_t y = 0; y < m_Width; ++y)
        for (coord_t x = 0; x < m_Height; ++x)
            if (m_pPixels[y * m_Width + x])
                g_pScreen[y * g_ScreenWidth + x] = m_pPixels[y * m_Width + x];

    In real code you'd also have custom offset, clipping, and obsessive optimizations in the inner loop, but this is the gist of it. This makes color zero transparent, which is useful if you want to, say, draw a tree or the player sprite on top of your adventure game tile map.

    1. Re:masked blit by ThePhilips · · Score: 1

      No sweat, guys.

      DirectX D2D has that from the beginning. Do not remember the name for the feature (something with "key").

      OpenGL has that from the beginning. In 2D mode, when Z=0.0f, simple alpha mask using monochrome bitmap. Supported as feature and quite works fast.

      --
      All hope abandon ye who enter here.
    2. Re:masked blit by Chemisor · · Score: 1

      > OpenGL has that from the beginning.

      I am talking about hardware implementation on the video card, not the API. Yes, you can do masked blits in OpenGL and DirectX, but they are implemented in software, and are not nearly as fast as they could be.

    3. Re:masked blit by ThePhilips · · Score: 1

      Uhm... They ARE implemented in hardware.

      But of course they work on internal image presentation, internal bitmask presentation, etc. That's why you have bunch of fancy APIs you can use to write portable program without headaches.

      Problem is that, the loop in GGP post, simply cannot be implemented efficiently. That loop is slow and unscalable regardless where you would put it: CPU or GPU. That's why you have fancy internal presentations to allow chips to parallelize the operations. But since most expensive operation happens to be communication with video H/W, they pack whole bunch of operations into one big operation called "scene rendering."

      Again: that acceleration would be dog slow because communication over bus is slow (due to frequency difference between CPU and GPU - in worst case interrupt also might be involved). It can only be fast when you can glue lots of such operations together and communication with H/W is performed in few (per frame) requests as possible.

      That's why we have DirectX and OpenGL. Without the APIs it all would be even slower.

      P.S. In past I programmed PCI133 bus. It is blazingly fast - unless you send something byte by byte without memory mapped I/O (that's how our h/w guy did it in the beginning). I had cases when doing simple hand-shake (REQ->RSP->ACK) with device was taking 2-3ms. If communication is done properly, one can send about 2MB of data in the same 2-3ms time. IOW, accessing H/W capabilities directly would only make whole system slow due to high rate of small requests. The only solution is to talk with hardware with large requests with many operations included - precisely what modern video cards are doing for past decade.

      P.P.S. And the internal frame presentation actually optimized for RAMDAC (or whatever it is called on new DVI/HDMI cards). The device is responsible for converting pixels into frequency a display can make picture from. This is actually slowest of all operations in video cards: sending a frame to the display. Now it is also done in parallel - modern cards have enough spare video memory.

      --
      All hope abandon ye who enter here.
    4. Re:masked blit by Chemisor · · Score: 1

      > Uhm... They ARE implemented in hardware.

      No they are not. At least, either my nVidia 7600 card does not implement it. I have not heard of any other card that implements it either. 2D acceleration is limited to BitBlt, a monochrome blit (for printing fonts), and sometimes a scaling blit (which my card can do). A masked blit always had to be done in software.

      > Problem is that, the loop in GGP post, simply cannot be implemented efficiently.
      > That loop is slow and unscalable regardless where you would put it: CPU or GPU.

      What are you, a total n00b? The loop is completely parallelizable; each pixel operation is independent - if the value is not zero, you write it, if it is, you don't. You could do the whole thing in one cycle if you wanted to.

      > But since most expensive operation happens to be communication with video H/W,
      > they pack whole bunch of operations into one big operation called "scene rendering."

      And it's also why many cards (like my 7600) support DMA transfers of image data from the CPU to the GPU.

      > Again: that acceleration would be dog slow because communication over bus is slow

      If you read my proposal, instead of skipping every third word or so, you'd notice that I'm asking for a BitBlt-like function, which is done from one part of the video card memory to another. No communication is involved except for the command itself, which is the whole bloody point in the first place.

    5. Re:masked blit by ThePhilips · · Score: 1

      Again: that acceleration would be dog slow because communication over bus is slow

      If you read my proposal, instead of skipping every third word or so, you'd notice that I'm asking for a BitBlt-like function, which is done from one part of the video card memory to another. No communication is involved except for the command itself, which is the whole bloody point in the first place.

      OK, I'm n00b. I'm system developer n00b. And I also have seen real benchmarks of video hardware from decade ago (and also simple recent tests) where calling H/W acceleration was dog slow for all the reasons I have described in GP.

      I do not mind to be n00b, but unless you would try to do it by yourself once - program hardware a bit - your proposal is pretty worthless and make you look even worse than n00b.

      P.S. BTW, read in

      --
      All hope abandon ye who enter here.
    6. Re:masked blit by Anonymous Coward · · Score: 0

      Again: that acceleration would be dog slow because communication over bus is slow

      If you read my proposal, instead of skipping every third word or so, you'd notice that I'm asking for a BitBlt-like function, which is done from one part of the video card memory to another. No communication is involved except for the command itself, which is the whole bloody point in the first place.

      ROFL

      So, in your part of universe, two bitmaps would magically without overhead appear in video memory. Then command to perform the operation (again magically, without overhead) would be sent to video card. Then video card (that must be really magic for all it to work fast) would report back that operation had finished to your program?

      Provided you have some magic it would work and work fast. LOL

      Good luck finding the magic, you retard.

  53. Re:That's a nice canned post ya got there by Flodis · · Score: 1
    I thought this was kinda old news...?

    I seem to remember an article on slashdot from not too long ago about someone doing real-time raycasting on ATI hardware... (looking, looking... Can't find the slashdot post, but I did find it mentioned in an article...):

    "Watch out, Larrabee: Radeon 4800 supports a 100% ray-traced pipeline using DirectX 9":
    http://www.tgdaily.com/content/view/38145/135/
    Btw, the article actually mentions that the code works on nvidia hardware too, but that the lack of a tesselation unit makes it slower.
    If you look around for the guy behind it (Jules Urbach), you'll also find a number of youtube videos where the guy explains the tech and shows some demos.

    I'm not working with graphics myself, so I can't really tell if the tgdaily link is equivalent to what nvidia are showing off in this story, but if we look at the statement...

    The genius of what NVidia is doing here, ...

    ... it does indeed sound a like nvidia are trying to spin things.