NVIDIA gobbles up ray tracing software maker RayScale
By Donald Melanson 
posted May 22nd 2008 5:01PM
posted May 22nd 2008 5:01PM
Details are pretty slim on this one at the moment, but PC Perspective is reporting that NVIDIA has snapped up the Utah-based start-up RayScale, a maker of ray tracing software. As PC Perspective points out, the move is made particularly interesting in light of some recent comments made by NVIDIA CTO David Kirk, who mused about the merging of ray tracing and rasterization into a "hybrid renderer" of sorts, but didn't offer any firm details on the company's plans at the time. No word on a price tag for the acquisition just yet, or any other details for that matter, but it seems like things should be getting a bit more official in the next few days.
second!
Fail?
Isn't this even more interesting considering how Intel was breathing down Nvidia's back just a while ago? They were talking about how raytrace rendering will scale nicely across multiple cores and offers better visuals to boot over raster rendering engines, then they had some Quake demo running at 35 fps raytraced which was visually impressive. Nvidia got very alarmed and launched their counter-volley.
Is this a move by nVidia to have a just-in-case rendering solution?
I think the fight with Intel is EXACTLY why this is happening...
IIRC, nvidia acquired Blue Moon Rendering Tools (BMRT) and then optimized it for their own hardware and released it as Gelato. It seems like nvidia will likely do something similar here with Rayscale's software (Lightnow).
The funny thing is, when it was acquired, BMRT was a popular (free) alternative to Pixar's Renderman software. Lightnow hasn't even been released yet, and the young upstart Rayscale has nothing else on its resume. If I were an nvidia shareholder (which I am, BTW), I would be really anxious to see how much nvidia paid for this yet-to-be proven tech.
'interesting in light of some recent comments'
Is it just me that see this as a poor joke? As for this, i made an earlier comment about how this was a fight nvidia were gonna bring on, but this was before it actually happened, i have see this happening.
Interesting... Looking forward to developments
Your mom is second
By the time pac gets to the end of the rayscale logo he will be able to eat ghosts.
thats why engadget put it there....... a joke.
NVDA has been too flaky lately. i dumped it and made a little since buying at $18.76
Om nom nom nom....
Personally, I prefer WAKKA WAKKA WAKKA, but that's just my opinion.
Awesome. I was going to say something to that effect.
John Carmack has talked a little about hybrid rendering and how he wants to use it in idtech 6. Virtualized geometry, theoretically infinite geometric detail, a bunch of words you can't understand...
This was in an interview specifically about Intel's ray-tracing stuff.
http://www.pcper.com/article.php?aid=532&type=expert
The bits you can decipher are pretty interesting.
They did another interview with Cevat Yerli (from Crytek) on the same subject. Neither Carmack nor Yerli seemed all that convinced that ray tracing is the wave of the future. Not yet, at any rate.
Yay raytracing!
Can't wait to actually see this in games, no matter how good the raster engines get, raytracing will always look better.
Interesting - NVidia already has a pretty impressive raytracer in Mental Ray, which it got when it acquired Mental Images late 2007. What are they going to do with a second ray tracer?
@Carl Vitullo
It's way more complicated than that. As most developers in the field have been saying, ray-tracing is not the end-all of real-time 3D graphics development. Ray-tracing, ray-casting, photon mapping, radiosity, global illumination, et all will probably have some part in it, but so will different rasterization techniques. Rudimentary ray-tracing type algorithms are very inefficient for real-time graphics. The process is too slow and computes a lot of precise information that just isn't actually needed to achieve ray-tracing-like realism. There are a lot of tricks and approximations that can be done that your brain will never notice. Sort of like compressing raw RGB video to H264. With a high bit-rate, you would never know the difference between the two because the extraneous information that is tossed out isn't even visible to the eye.
the bottom line is that a hybrid approach is going to be the only way to achieve new generation visual quality in real-time.
Maybe they're next
http://www.caustic.com/
Rasterization is only efficient to a point. Scanline has a linear performance. Raytracing is logarithmic. The more effects and geometry you have in a scene, the better a pure raytrace system performs over a scanline system.
I have the feeling the mental images purchase was aimed more at Reality Server. mental ray has it's own rasterizer as well.
However, it would be interesting to see if there is a way to use the best of all worlds for games and film.
Gotta love Engadget's sense of humor. :D