Renderman vs MentalRay - a quick rendertest


#21

Hi daddyo, which version of PRman was that? the proserver or the RtM product?

Pete


#22

It was the pro version, I’m guessing, as the studio I worked for had a custom pipeline for it… I can tell you for sure we were using Maya 6 and SLIM, and some other interfaces built in-house.

The lighting TD put together the shaders, but I don’t think it was anything written from scratch, so I assume that Renderman is just crazy fast out of the box.


#23

Yep, PRMan is great, but this thread is about RfM I believe, a different product. Its a small subset of the full Renderman renderer that works as a plugin for Maya. It’s fast, easy to configure but not as flexible/powerful as MR. Personally, I think Turtle is better value for money. Dunno if RfM can generate exportable RIB files, can it ?


#24

Well, yes and no. It is about RfM, but the only difference to PRman is the interface and what you can do with it. The rendering engine is 100% the same. So what goes for one, goes for the other, too. With the addition that PRman is more configurable.
Same goes for MentalRay for Maya. It is an interface for the MentalRay renderer. As such you have a small subset of what is possible at your hands in a point and click way. I tested this point and click way. Not any nifty shaders and special moves (which you can do with both RfM and MentalRay for Maya by the way).


#25

I’ve done a bit of testing between the 2 renderers myself and I can confirm that these tests are a pretty accurate assesment. Renderman for Maya got a reputation that it was somehow crippled and not ideally suited for production. Honestly, I habe no idea where this rumor started. It cannot export RIBs and it cannot use any custom DSOs. It CAN use renderman shaders and you are allowed to compile them yourself. You can network render with it. Its $1000 PER COMPUTER, not per processor, like Mental Ray is currently.

A couple of additional things I dont remember this test mentioning:

Renderman for Maya renderers particles in software. Not even the native maya renderer can do this.

All the features of the renderer are set up to be easy to access and use. Subsurface scattering and GI are so easy to use its almost criminal. There is no fiddling with tesselation and any polygon n-gon surface can be a subdivision surface. Hair and fur render beautifully and fast.

Export times on almost everything I’ve tried to render are almost negligable. Mental Ray export times are atrocious and have the tendancy to throw memory exceptions on big scenes. Renderman scenes consume far less memory, in general.

Mental Ray’s rapid motion blur is faster than normal MR blur, but neither can even remotely hold a candle to renderman motion blur.

In the brief time my studio has been evaluating Renderman for Maya, our productivity has gone up noticably. Less time is spent fiddling and troubleshooting and more time is spent tweaking and fine-tuning results.

If you are a small or medium-sized shop then renderman for maya is just about the best available solution out there. IMHO, of course. :slight_smile:

EDIT: Im by no means a Mental Ray master, or a TD, but i know my way around the basics. I know MR can be tweaked to hell and back to sqeeze every ounce of render times out of it. But my shop consists of 3 animators. We are very very small and we are all artists, not TDs. Renderman for Maya can deliver results with minimal tweaking, and generally orders of magnitude faster than Mental Ray. In a large shop, priorities can be different, but in a small shop like mine, Renderman has found a permanent home here.


#26

Whats up Frenchy? You’re using PRMan you lucky bastage? … Hows work?

az


#27

man, stop talking and start doing some interesting test!
in your judgment i found many inaccuracies, but we are already trying to convince some other, so, no more words, make some test and try to make it right!!


#28

All I see is your whining in this thread. How about you come around with some decent criticism first dagon1978?

Obviously some people seem to like to get to their results fast without major fiddling and tweaking scenes to death. That is what Renderman for Maya promises and that is what it delivers. :deal:

If you guys think some tests where inaccurate (and we already heard that Rapid Motion Blur also looses against the Renderman Motion Blur), stop whining, download the evaluation version of RfM and do the tests yourself - I didn’t do it any different as you can see on the watermarks in my renderings.


#29

dagon is one of the most constructive contributors on this forum. i’d listen to him if i were you.

sure, RfM gets great results without tweaking - he used the tutorial scenes that come with the program. render as many teapots as you want - i don’t care, because none of my 3d projects involve flying teapots. i’m not interested in seeing what the tech-demo files look like, i’m interested in real world results. dagon is right… you should be making real world comparisons, not rendering RfM’s tech-demos in another renderer. That’s like saying “Maxwell is much better than MRay, because the scenes that came with it look much better in Maxwell.”


#30

What should I listen to? He didn’t make one statement of value in this thread. None of his ranting were backed up. He didn’t give any details what inaccuracies he ment. I can do without the meaningless blurps of your so called constructive contributor. What about his posts was constructive?

And this I completely don’t get. What is wrong with the test setup that came with the plugin? Did you even look at it before you started rambling? It doesn’t sound like it. Otherwise you would have seen that there is nothing Renderman specific in these scenes. It is a friggin’ poly object lit by a Maya directional light in most cases. The fur examples use Maya fur, nothing special here either. If I built the examples by hand it wouldn’t have looked any different. I just took the pot, converted it to nurbs, so both renderers could render it smooth (not using Rendermans subdivs). It doesn’t get more unbiased then that.


#31

bla bla bla… one more time words??
i made many tests with mantel ray, u can find it in this forum, just use the form “search” :smiley:
and i’m not here to make any more test with rfm, i’m waiting for your tests, where are your test with rasterizer? detail shadow map?
u dont know mental ray but u are still trying to convince us… with WORDS! yeah
bla bla bla :rolleyes:


#32

So? I could easily say, you don’t know Renderman and try to convince me with (as you put it) WORDS! yeah
What is it with this zealots that they get all fuzzy about THEIR baby? Why not just use the best tool for the job? Nothing more nothing less. yawn You bore me guys.

P.S.:Very strange everyone EXCEPT you MentalRay guys seem to apprechiate someone doing a little test (not a scientific paper).


#33

It’s not that strange… after all it is the Mental Ray guys who use Mental Ray, and your doing a comparison with… Mental Ray! If you were comparing VRay to Renderman and you compared RMs global illumination to VRay’s slowest form of GI, they would be vocal about that I’m sure.

Noone dislikes tests, they just like them more if they’re fairer. I’m not sure how this thread turned out so… defensive.

:hmm:


#34

Hey Alex. I think the reason for why people get easily upset about this topic is twofold, really: a) this has been done, discussed and tested a zillion times before (and EVERYONE knows each renderer has its own strengths), and b) because there is a border when “quick and dirty comparison” gets so quick and so dirty that it is not representative for anything.

That being said this little ‘test’, as you describe it, does not seem to be fair to most people because of the way it is conducted and as it does not even show anything close to a real-life scenario. Imagine a ‘test’ set up with an all reflective scene set up, rendered via sattellite on a couple of machines and compare that to RMfM - the conclusion to that little test would be something like this, if I wrote it like the one at hand: “MR for Maya does come right out of the box with no extra costs involved, whereas RMfM will set you back another 1000 for a single machine! Yet, MR kills RMfM in terms of raw renderiong speed by the factor of 10 easily. Hence, MR is - ‘in most cases’ - vastly superiour.” - load of crap, of course, but then why even do a test that really does not show anything?

The ironic part is, that most of the people attacked as Zealots in this thread do know a lot about both respective renderers AFAIK. Yet, such a test would be interesting for most, so let’s all calm down and come up with more representative set-ups and conclusions for those in need of a new render plug-in.

cheers


#35

…and u aren’t still doing any test, right? :wink:

ok, so i have to download RfM to convince you that your test is not appropriate? :rolleyes:

so, if u like word so much, lets say something… i’ll be able to use all my dual core processor with RfM?
yeah! but just 1 core! great, no? :wink:

i appreciate all tests here, but if u think to make a test, u can’t expect no one criticizes it, if is a bad test :wink:


#36

… i’ll be able to use all my dual core processor with RfM?
yeah! but just 1 core! great, no?

Just for the sake of clarity. Actually RfM can be multithreaded. I refer you to this section of the RfM FAQ-
“1) Node-locked licenses are “locked” to a single machine (or dongle). Node-locked licenses are “uncounted” meaning you are able to initialize an unlimited number of processors on the locked machine with one license.”

I realize you may have meant something more specific about how Mental Ray multithreads as opposed to how RfM does. Your quote just made it sound like they limited you to only one instance of it running at a time. It may not be practical in all cases to do so, just pointing out you can.

Grim

Quickly ducking away from the general topic in this thread.


#37

I don’t have anything against critisism dagon1978. I am all for it. But you didn’t criticise. You just came in ranted and went away again. What MasterZap did was criticising. He made a point and backed it up. I respect that and have replied to in in a good manner (at least I think so). I don’t count you pointing me to the search of the forum. This is a topic I started, so if you have a critique, you could at least back it up with a link, but you didn’t even do that.

And FYI I am rendering the exact same GI test with Motion Blur right now with Rapid Motion Blur. And I am still not too hot on MentalRay motion blur.
So what? I didn’t say anywhere you should not use MentalRay ever again in my article. I said it is a great renderer, but don’t use it in combination with motion blur unless you have to. Nothing wrong with that. You cannot expect any renderer to shine in any area.

But you (and I say it again) zealots jumped all over this one detail in my article. It seems you and all the others have completely missed that I also had good words for MentalRay.


#38

LOL
Yes, it is a stiff breeze in here. :smiley:


#39

its so stupid face the renderer each other -
both has its strengths and the field of work is different
for highend stuff noone will work with both there are prman and mr standalone
so save time and dont do such silly tests with no experience pls


#40

Can I use two processors/cores on a dual processor/core machine?

Yes and no. RenderMan for Maya will support a muti-process mode of operation. Presumably all Maya plug-in renderers support this mode, but Maya doesn’t make it trivial to utilize this mode. If you have lots of frames to render (and sufficient RAM) you can launch Maya from the command line or a spooling program and designate each invocation to process a subset of your total frame count. Now you can utilize the additional CPUs on your system with no additional software licenses required. This method works specifically (that is, only) with node-locked licenses of RenderMan for Maya.

So, for example, you could have one processor rendering frames 1-10 and the other 11-20, for example, using the one node-locked license for that machine. You cannot, however, use both processors in concert to render a single frame.

Can I get a floating license?

Yes. You can purchase node-locked or floating licenses. Floating licenses require a minimum purchase of five seats.

i dont criticizes your test, i criticizes your statements, u dont know mental ray, how can u do any statement?
i dont know RfM, so i avoid to do statements on it, especially with a test render like this one
but if u want u can keep testing flying teapots, no problem :wink: