4 posts • Page 1 of 1
Vray seems to be the go to app for archviz so I've started looking into it a bit. I think Vray RT would be a huge help in learning the program by having the ability to tweak a dial and see what it's doing on the fly. But I don't really understand the deal with Vray RT.
It seems that Vray RT got announced at Siggraph last year as the hottest thing since sliced bread. It was awesome because vray is awesome but it had the added benefit of being super fast because they were using GPU rendering, which looks to be the future of speediness.
So I started looking into the Vray RT web page and it specifically says that it uses CPU rendering, NOT GPU rendering. So they showed something at Siggraph that they don't actually offer? That seems kind of slight of hand. The reasoning I'm seeing for sticking w/ CPU rendering is so that the RT rendering is the exact same as the production rendering. Supposedly you can't produce the same image through the GPU as the CPU. I'll just have to take their word on that I guess.
Am I getting this all about right?
Your mixing thing up my friend. If i'm not mistaken, what was shown in Singraph was a prototype for FUTURE Vray RT versions to show what they would be able to do with the power given by the GPUS. The problem, for not beeing released, is due to GPU coding right now beeing in his infancy, the only available language is CUDA for nvidia (and that would put ATI buyers out) and OpenCL had just been release. That's why they choose to use the multicore processores (and let's be honest, is much more easier to have 4 quad core cPUS than 4 top Quadro grafics cards).
Vray RT is still pretty damn fast and it's a blessing for the vray users, showing materials in real time as you edit/change them avoinding to do test renders is great. The main problem is that you need a copy of vray too to take full advantage of vray RT.
Hope it helps
Thanks for the info, that's interesting. As a current Maxwell user I certainly am not in a position to knock the speed of Vray, let alone Vray RT. The demos I've seen are quite compelling.
I'm also a user of unbiased engines, but i have to admit that if i knew vray well enough i would probably just use that. The big advantage for unbiased is the litle time is takes to set things up, wihtout needing to know too much technical stuff, but in return the render times are ridiculous big. Now the problem is that, in this CG business, we're always learning and evolving our skills and knowledge, with time beeing allways against us, so often we reach a point that we want the same quality (almost) with a fraction of the time and that's when we users, that ironicly always ask for simplicity, start studing and using more complex stuff that fit our new goals...
I have 2 friends that everyday remember me that no matter how good quality my unbiased render is, their Vray and mental ray stuff will allways look better (I'm talking really top stuff here). They don't even lose much more time than me setting a scene up (experience and material libraries compensate that) and in render times i don't have a chance either, especially if we're talking big print resolutions...Nowdays i'm using more and more biased methods for render and evolve my photoshop skills...some stuff is a pain but it's starting to pay off
4 posts • Page 1 of 1