PDA

View Full Version : Vray vs. other Render Enginges


isildur88
12-18-2010, 06:40 AM
Hi everyone,

I am working at a large architectural firm in Switzerland where two of us work exclusively on 3D visuals. We use Max for modeling/texturing and Vray for lighting and rendering.

We've never really had an issue with the software until now, the IT specialist came to my workplace yesterday and said that one of our bosses is not satisfied with the outcome of a specific project which my colleague worked on.

He went on and gave me a small promotional sales brochure from Fryrender and asked if I could please look into it.

The question is if anyone here thinks that Fryrender, or any other Render-Engine will or could give better results than Vray, from a general point of view. I'd like you to treat this question independetly to any specific image/material/project, since that's the way it was presented to me.

I would really like to have your personal opinion and/or answer to this question.

My personal, naive opinion is that all render engines work with the same set of lighting and material algorithms which are supposed to be physically acurate. There should be no big difference in the quality or realistic closeness from one to the other, as long as the user knows his or her way around the program and has a solid knowledge/understanding of how lights and materials work in real life, but please correct me if Im wrong.

Thank you all in advance.

playmesumch00ns
12-18-2010, 09:43 AM
My personal, naive opinion is that all render engines work with the same set of lighting and material algorithms which are supposed to be physically acurate. There should be no big difference in the quality or realistic closeness from one to the other, as long as the user knows his or her way around the program and has a solid knowledge/understanding of how lights and materials work in real life, but please correct me if Im wrong.

Yup, you're wrong :)

Renderers like VRay and mental ray cheat a lot in order to speed rendering up, while renderers like fry and Maxwell try to stay closer to physical correctness. This difference is commonly called biased vs. unbiased rendering, although those terms aren't accurate.

In "unbiased" renderers render times are typically much longer for a clean image and images may contain high-frequency noise if they're not left to run for long enough.

In "biased' renderers render times are typically shorter and images may contain low-frequency noise which flickers in animation if the quality settings aren't high enough.

Moreover, biased renderers ignore many light paths as their algorithms are specifically designed to consider a subset of the lighting integral in order to speed up calculations. Unbiased renderers typically consider all possible paths, leading to a more realistic image. The differences can sometimes be subtle, and how much they are apparent will depend heavily on what you're rendering, but it is true that a biased renderer can never generate an image as physically correct as an unbiased renderer, simply because it ignores so many paths that light can take (caustic reflections from glossy surfaces is one of the main ones I notice).

In short, "unbiased" renderers will be able to produce a more realistic image in less setup time than biased renderers, however your actual render times will be (a lot) longer. Proponents of biased renderers (who typically work in arch vis or product design fields) argue that the reduced setup time more than makes up for the longer render times.

If I were you I'd try fry or maxwell out and see if it fitted my workflow. There are pros and cons to any renderer, what works for you is completely dependent on exactly what you're rendering.

isildur88
12-18-2010, 12:45 PM
Thank you playme for your quick, thourough, informational answer.

What you are saying is that even if I take Vray and crank up to the highest possible settings, I will never get the same quality of reality-closeness, that I would get with a so-called "unbiased" render-engine?

I disagree, but I would really have to do some painful testing in order to back it up.

Have you playmesumch00 done this type of testing?

I'd be very grateful if anyone else who's (also?) done this type of testing could join the discussion and give your opinion. Thnkx again in advance.

MikeBracken
12-18-2010, 01:27 PM
Please dont take this the wrong way....but if your boss was not pleased with the quality of Vray, then you guys probably need some more experience with it. Vray is MORE than capable of producing photoreal images/animations. Just check out Neoscape, Studio AMD, etc, to see the quality of Vray in enxperienced hands.

Regards,
Mike

mister3d
12-18-2010, 01:42 PM
I saw fryrender rendering the simplest things and I think Vray cannot do this. I know Vray for many years, and I never saw such a result from it. It seems to me the earlier builds of maxwell looked more realistic, but that may be just my seeing. I think you should try demos and show us the result.

isildur88
12-18-2010, 07:01 PM
Please dont take this the wrong way....but if your boss was not pleased with the quality of Vray, then you guys probably need some more experience with it. Vray is MORE than capable of producing photoreal images/animations. Just check out Neoscape, Studio AMD, etc, to see the quality of Vray in enxperienced hands.

Regards,
Mike

no problem Mike, I totally agree with you on this, I also truly respect Vray as one of the best render-engines available. Specifically because you can control/determine the exact amount of bias needed. So yes, this is the type of answer I was actually looking for, in order to back up my own personal opinion which is the same as yours.

Anyone else whod like to agree or disagree, please do, youll be helping out a lot to clear things up.

isildur88
12-18-2010, 07:37 PM
I saw fryrender rendering the simplest things and I think Vray cannot do this. I know Vray for many years, and I never saw such a result from it. It seems to me the earlier builds of maxwell looked more realistic, but that may be just my seeing. I think you should try demos and show us the result.

Hi Mister3d, trying out demos and comparing the results would definetely help to find the answer. But theres one factor which will make things very complicated, and its the fact that in order to make a fair comparison Id have to have the same amount of knowledge in both packages (i.e. Vray AND Fryrender) in my case I have never used Fryrender, so it would be difficult to make a fair testing and comparison, since Id probably need a few weeks if not months or even years in order to really know how Fryrender works, and make a fair comparison. In the case of Vray it has literally taken me a few years to learn and understand all of its components, so I dont think it would be different with any other render-engine.

So thats the reason I opened this thread, in order to collect the information which I'm hoping other people allready have on this topic, before doing all the testing/learning on my own.

mister3d
12-18-2010, 07:56 PM
No, you don't need to make a complex testing, test the main properties. Even if you test metal vs metal you will see the difference.

isildur88
12-18-2010, 08:37 PM
No, you don't need to make a complex testing, test the main properties. Even if you test metal vs metal you will see the difference.

Interesting, I might just get a round to it, but please if you or anyone else has additional information and wants to share it, Ill be VERY thankful.

for example, does anyone know which method of approximation is being used in Fryrender in order to resolve the rendering equation (http://en.wikipedia.org/wiki/Rendering_equation) (i.e. Radiosity, MonteCarlo, MLT, etc.) and more importantly, does anyone know for sure if any of the existing methods is more accurate than the other, assuming the use of maximum computing resources at hand for each.

thnx

mister3d
12-18-2010, 09:05 PM
http://www.randomcontrol.com/fryrender-tech-specs
http://www.chaosgroup.com/en/2/vray.html

thethule
12-18-2010, 09:11 PM
Phew...there is a lot of bulls**t being spoken in this thread....:argh:

isildur88
12-18-2010, 09:27 PM
Phew...there is a lot of bulls**t being spoken in this thread....:argh:

hi there, you might be right, Im not quite sure though. would you mind explaining or exposing your own personal understanding on this topic, or was that all you have to say?

thethule
12-18-2010, 09:46 PM
hi there, you might be right, Im not quite sure though. would you mind explaining or exposing your own personal understanding on this topic, or was that all you have to say?

That's all i have to say really. it was really directed at you, sorry, i should have been more clear. Im not interested in the slightest in getting into a software discussion to be honest.

But to say (not you) that even in the simplest tests, Vray was not capable is just silly. Its one of the best selling and most capable renderers on the market.

mister3d
12-18-2010, 10:41 PM
That's all i have to say really. it was really directed at you, sorry, i should have been more clear. Im not interested in the slightest in getting into a software discussion to be honest.

But to say (not you) that even in the simplest tests, Vray was not capable is just silly. Its one of the best selling and most capable renderers on the market.
My mate is an archvizer, and he showed it to me. There was a really great difference. I believe my eyes.

isildur88
12-19-2010, 06:15 AM
it was really directed at you,

?

I agree with you on the fact that Vray is more than capable
but whats with "directed at you" ...? wheres the rest of the bull?

thethule
12-19-2010, 11:06 AM
?

I agree with you on the fact that Vray is more than capable
but whats with "directed at you" ...? wheres the rest of the bull?

Oops! I meant to type wasnt, not was. Lol
Sorry

isildur88
12-19-2010, 11:13 AM
thanks for clearing up.

mister3d
12-19-2010, 05:09 PM
Isildur88, it doesn't matter who thinks what about renderers until you see tests. I hope you do some. And maybe, even show the result.

InfernalDarkness
12-19-2010, 07:23 PM
Mister3D is wise, trusting your eyes is a big factor. But also: trusting your money. Time is money, and the reason Vray and mental ray are so huge in Arch/Viz is simply because of speed and customization and optimization-ability. Vray even moreso because it actually works, but MR is just as capable on paper.

All these renderers are capable of photoreal quality in the right hands. But good luck trying to render print-resolution in Fry or Maxwell on a tight deadline. With MR and Vray you can customize things so extensively and literally "pick your rendertime", which sometimes is very, very important.

@isildur88: We'd really have to see the project, although you're likely under NDA, to judge for ourselves its quality. There could be other factors involved - a butthurt client, an otherwise cool boss gone pissy for a day, etc. This happens to me constantly in my work - and my quality has only improved with each project. It could be as simple as "unrealistic building probability". Sometimes I lay my tile textures out in a way that could not actually be constructed with any durability or buildability - since I'm not a tile setter in real life, I have to make "educated guesses" often when I'm on a deadline.

So it might not even be your fault, or your team's. Could just be "part of the game"! If you can share your work at all, perhaps we can help critique stuff and help you learn.

That said, my work-deadlines are generally mere hours, from the site measuring (which I do myself) to modeling to texturing/lighting to rendering in one work shift. so I rarely share my arch/viz here because it's quick and stark, not the gorgeous fully-worked scenes you see here in the galleries all the time. I would consider it "speed-arch/viz".

And perhaps your works require such sloppiness too, for the sake of time. Often it's better for me to have something to give my boss and clients, even if it's not perfect, than nothing at all (making them think I'm not working hard enough).

isildur88
12-19-2010, 09:33 PM
thank you man for sharing your opinion,

a lot of truth behind this last post. Its comforting to know/realize theres others out there living up to the uncomfortable "parts of the game" we get involved with.

I can relate to your description of quick/stark, speed-arch vizualisation, and maybe, just maybe thats the whole deal with this post here in the first place...

more on that after testing Fryrender.

cheers!

noouch
12-19-2010, 10:53 PM
Having been working with VRay for about 5 years now, I can definitely say it is one of the most advanced renders out there for 3DS Max. It's hard to say if FryRender would be an improvement without seeing the project in question.

I've also done a few things in Maxwell (yes it's true that the earlier builds looked better, something to do with the tone mapping, weighted sampling, and early cutoff of rays as far as I recall), and I think the main reason why it and Fryrender look so much better is that the renderer won't allow the artists to get "wrong" results: Fresnel reflections for all dielectrics, pixel-perfect GI, etc. whereas VRay allows you to cut corners in all sorts of ways in favor of speed. Another major factor is that in Fry AFAIK you can only use mesh light sources, so fantastical computer graphics concepts like point and spot lights are eliminated. It all boils down to being forced to think more like a photographer. (I'm also assuming you already have a fully linear color workflow)

A huge disadvantage (and probably another advantage towards photorealism) using Fry and Maxwell is you lose the ability to use 3DS Max procedural maps like Noise, Mix, Color Correct, and so on. I feel this is often overlooked, and a huge reason I still prefer VRay above all else.

Have you looked in to VRay's progressive path tracing workflow? I believe it will offer you improved GI solutions.

Also, VRay 2.0 can render unbiased using Bi-Directional path tracing.

metamesh
12-23-2010, 12:58 AM
been using vray for years, and is one of the best solutions out there ( for me ) also with vray you can set up a PPT scene or even a unbiased solution ( this just in vray 2.0 ) that should give you the same solution as maxwell/fryrender. Even just using the standar "universal settings" will give you and awesome result. I dont think it has much to do with the renderes used, but with the person behind it...Ive seen awesome renders in both vray and maxwell/fryrender and also crap renders in both vray and maxwell/fryrender...

azozel
12-23-2010, 04:40 AM
It's all about the talent behind the render. I've seen scan line render look amazing. But as far as ease of use and quality vs render time it's Vray for the (EPIC) win all the way! :bowdown:

xdennisx
01-10-2011, 02:35 PM
one thing that renderings from an production renderer like vray might miss is a decent post-production. color correction, dof, bloom, maybe a lensflare sometimes... maybe add a subtle grain to add some more "analogue feeling"... it's not just the render engine alone...

noouch
01-11-2011, 11:53 AM
one thing that renderings from an production renderer like vray might miss is a decent post-production. color correction, dof, bloom, maybe a lensflare sometimes... maybe add a subtle grain to add some more "analogue feeling"... it's not just the render engine alone...

VRay has very nice dof and glare. Things like flares and color correction are best left to compositing tools, and even glare doesn't truly belong in the renderer IMO.

CGTalk Moderation
01-11-2011, 11:53 AM
This thread has been automatically closed as it remained inactive for 12 months. If you wish to continue the discussion, please create a new thread in the appropriate forum.