The C4D Renderer vs The Rest!


#1

i’ve been reading this thread…prman vs mray
http://www.cgtalk.com/showthread.php?threadid=33132

and for awhile now i’ve been thinking how does the c4d renderer stack up against the rest…i mean i love the results from the c4d renderer…but i’m not exactly pushing it! the discussion in parts is over my head…but very informative…and worth a read…alot of the discussion is based around film makers needs which is only a part of the equation but obviously still still relevant…but does affect there requirements from a renderer as they are limited by time and the quantity or frames…rather than quality…often having to fake results rather than use them in anger.

what do we have? what don’t we have? what do we want? why can’t we use other renderers like other software seems to support…eg maya/xsi supporting mentral ray etc…and max brazil.

part of the discussion mentions that prman is open and hence flexible but also more complex as a result of this flexibility…they also mention prman stability…

i would just like to know what peoples take on this is…presumably its holding c4d back…or is it?

alot to answer i guess but hopefully something to get you all thinking…what you like or don’t like…and how we stack up.

cheers:beer:


#2

There are a lot of new and old renderers out there these days. Vray, Brazil rs, finalRender, Mental Ray, Prman… The only 2 renderers i’d take instead of C4D renderer are Mental Ray and Prman, the others just wouldn’t justify a replacement of C4D default engine (with AR of course). C4D’s raytracer is second to none and its GI engine is (more or less) the same of Cebas finalRender. What do Mray and Prman offer more than our renderer? Basically a serious and really fast displacement, a fast hi quality motion blur and the possibility to code shaders. Do you really need those features? Well i don’t, but if you do you should consider buying Maya or Softimage|XSI. Now the price: Mental Ray is 2000USD per licence, Prman is at least twice that price, do you really need those features now? :wink:


#3

i don’t think it’s holding cinema back any.

maya always had a terrible render engine yet it never held it back. The addition of Mental Ray more recently hasn’t actually helped it gain any ground as far as i can see.

Cinema is more likely to gain ground based on the fact that with cinema net render module you get infinite licenses. Mental Ray and PRMAN cost abou $5k per extra node, that’s a huge expense.

The quality of all of these render engines is very high. Cinema’s and Final Render share a lot of their Core. What Cinema’s currently lacks is SSS, MicroPoly Displacement, fast reliable depth of field and fast reliable motionblur. There are of course third party shaders for SSS and it has DOF and Motionblur, but at the moment their quality and speed are no match for MRay or PRMAN. In pretty much every other aspect the cinema render engine can match.

The main reason that Maya, XSI and now Max have MRay is that it’s cheaper… yup, you heard me right. For Avid, Alias and Discreet it’s cheaper to purchase the rights to license the render engine than to pay to have in house development of their own render engines. Of course it’s not cheaper for the consumer… (especially if they run a network)… of course we wont mention that without MRay their render engines were downright abysmal in the first place… :wink:


#4

you should look at all the great renders done in Brazil, I think it’s the second best there is right behind PRman :thumbsup: … and of course I love cinema’s but, you know how good LW’s is also, although it is much slower…


#5

Originally posted by sebek27
you should look at all the great renders done in Brazil, I think it’s the second best there is right behind PRman :thumbsup: … and of course I love cinema’s but, you know how good LW’s is also, although it is much slower…

Brazil isn’t really anything special, Mental Ray beats it any day of the year, just look at Panic Room or Fight Club or…


#6

well what are we talking about here in terms of quality?

Output? or feature lists?

I mean if you put stalhberg in front of some amiga thing with 10 year old software he would still beat the pants of most everyone.

I think its a little ambiguous to say this is best this isn’t.

What are we talking about when we say best? best output or what?


#7

Then add the user knowing exactly what they are doing with their render settings as well etc.


#8

Don’t include in the discussion the ability to use the software or the skills of the artist. And let’s stop saying the software doesn’t matter and Stahlberg blah blah blah, we’re not all Stahlbergs and if software wasn’t that important we’d all use Bryce. We’re talking about features, stability and speed with huge scenes, output quality.


#9

Originally posted by sebek27
you should look at all the great renders done in Brazil, I think it’s the second best there is right behind PRman :thumbsup: … and of course I love cinema’s but, you know how good LW’s is also, although it is much slower…

You know, as a user of both, I originally felt exactly the same way, but I’m not so sure now. Initially, my first impression of C4D’s render speed was incredible, but the moment I needed to do camera based animation, I came to a different conclusion. It is horribly flicker prone in Geometry AA mode, so I started experimenting with Best AA, Threshold and Min/Max levels to try to eliminate it as well as persistent jaggedness along all varieties of edges and AA errors through transparencies. What I found was, for acceptable quality levels (not perfect, but tolerable), I needed to bring down the Threshold to at least 2% and be at 1x1/4x4 Min/Max level for the bare minimum. This increased render times by over 800%. I’d really prefer to set the Threshold to 0% because much like LW’s ‘Adaptive Sampling’, I don’t like having the renderer arbitrarily decide what does and doesn’t need AA, but this pushed renders up to over 1200% slower and was simply unacceptable if I ever wanted to get this 4000 frame animation finished within a reasonable time.

So, I’m not so quick to bash LW’s renderer speed any longer when I start factoring in various quality issues. I’m finding that it can depend on a variety of things and isn’t as clear cut as I had originally thought. Surprisingly, I’m actually finding myself favoring LW’s IBL as well, but perhaps that is best saved for another discussion.

-Tronam


#10

i must say that i really like the c4d-ar engine because it is very fast. i never have seen a comparison or rendertime chart of the other renderers out there. yeah brazils radiosity looks great but that does c4d´s advancedrender, too. yeah you must make good settings but c4d´s renders just look awesome.

there are just two things i think the render engine lacks: other objectives (eg. fisheye, architecture perspective correction - those which finalrender stage1 offers) and a real dof (making all other posteffects real, too).

btw. i don´t understand what fr´s hyperradiosity does. am i right, that it could be faked by using dirtynuts with radiosity?

:shrug:


#11

HyperGI is just non-flickering object GI, it can’t be faked in C4D so far.


#12

greetings, sadbatu! :wink: always meet you anywhere.

I guess it`s primary relative to the very-own skills how ‘perfect’
a rendering looks.
More, than on any program-features the industry wants to know us.

I´ve seen a few really amazing blender-renderings for example,
made by a raytracer I was formally making jokes about.


#13

I am hopeful that future revisions of the Advanced Render module will include the HyperGI technology from FinalRender-1. That would be a great boon to C4D for GI animation.

-Tronam


#14

hi cornel:wavey:

tronam: ain´t the ‘camera animation’ selector doing just this - non flickering gi-animations?


#15

Originally posted by sad
[B]hi cornel:wavey:

tronam: ain´t the ‘camera animation’ selector doing just this - non flickering gi-animations? [/B]

Yes, but just for camera movements, not for object movements.


#16

THere shouldnt’ be any doubt the PRman and Mray are the two best renderers out there, no ifs ands or buts. As mentioned however to render on a network is going to cost an arm and a leg.

There is also the matter of usability. Mental Ray has a lot of special control through coding that interfaces like LW and C4D’s don’t really offer, I mean one can code plugins, but Mray originally (and Prman) was developed around that workflow. This offers a lot of control but o the basic user rendering is extremely difficult. Imagine spending as much time learning how to make an object transparent and cast a transparent shadow in them as you do making photorealistic glass in C4D. The learning curve is pretty steep particularily for a newb, they tend to be more ideal for studios collaborations.

As for renderers like Vray, Brazil, FR, LW, EI, and C4D’s, they all have comparable features and quality. Some have things like SSS and Micro poly displacements, while others offer things like Multipass support. Their ease of use and learning curve are far more user friendly particularily to newer users or individuals as its easier and quicker to render, they also tend to offer cheaper network rendering.

How does C4D compare to renderers on the same level, well in terms of multipass its the best that have seen, cebas offers a comparable PSD manager plugin for Max that combined with FR definitely makes a powerful team, but look at the price of the two together plus max.

Some features like SSS and micro poly aren’t available to date, and currently scene motionblur is the only really usable option for cinema as well, our AA is very good, speed is definitely comparable if not better and seems to get some speed increases in most updates. Our post effects system is decent but definitely has room for improvement, particularily in mixing of post effects and the ability to preview changes in post effects without rerendering like some others do.

Hyperradiosity is just a stupid name for a new hybrid of GI in which the samples on objects are stored and reused in the next frame, it really only increases speed and quality of animations for stills its not that big of a feature can’t really be faked at the moment.


#17

uh, that´s not very nice of this checkbox:)


#18

wow…some response in such a short time…wasn’t sure you people would feel so passionately about something like this…i guess i miss read.

firstly…in the thread i linked to…mray and prman only seem to be the two renderers in contention and finalrender and brazil don’t really come out that good…obviously thats an opinion but if parts of cinema renderer are based on final render then aren’t we punching slightly above our weight?

as for shader support…surely this is only a good thing and would add significantly to cinema kudos and functionality be that film or architecture or whatever…an open shader system even better especially if its stable as with the likes of prman. dunno i can only see this as a major major plus…ok maybe i’m not at an ability level to write my own shaders…but at the moment i don’t even have the option.:shrug:…well not totally true but you get where i’m going…or do i completely misunderstand what all this is about…feel free to correct me…its why i asked the queston!

they mention brazil cos of a few features or ability to access a feature easily from say a list or whatever input method it uses…but this is not full control its selecting an option.

motion blur on every frame, dof, and also adaptive sampling as tronam points out majorly eat into render speeds…
but surely we should try to benchmark the best feature set there is…be that what render or software developers want to offer us or what is available through research…whatever we should only benchmark from the best…
whether thats a 300 dollar program or an 8000 dollar prman licence…also as thirdeye points out stability and ability to handle large scenes is a majorly important part of the softwares function.

it just strikes me as odd that companies seem to be going out and adding mental ray to there software features set…and we can argue the merits of whether there respective implementations work or not till the cows come home…but they implemented it…why? are we saying solely because their own native renderers where s**t or there are other reasons.

no one like ilm with the wealth of technical knowledge they have goes out and spends that much on licensing unless they see a return? what returns do these guys see that we don’t?

from a personal point of view the only thing i’m concerned about is quality and stability…but thats me i have my use…and the rest of you have your respective fields to think about…what do you miss…what do you want…why don’t we have it? not saying speed isn’t important though! heh heh

are we ultimately saying…its totally unnecessary for cinema to have external renderer support? cos that is a major direction maxon would be taking…i’d call it a revolution or a future failure? which is it gonna be guys…and why? where do we go from here…do we go to support prman/mray…whatever…or do we go for revolution…or do we go for watered down half arsed hack…don’t forget we are the customers and to a large extent determine the development of this software…the choice is yours…just make sure you choose?


#19

Why Maya and Max included Mental Ray?

  1. their default renderers are embarassing
  2. Maya implemented it because its major rival (XSI) comes with Mray PERFECTLY integrated.
  3. Max implemented it because its major rival (Maya complete) comes with Mray for free.

Don’t forget THIS: it took years and years of work to a lot of Avid/Softimage engineers to have a perfect connection to Mental Ray, don’t expect anybody coming one day and saying “hey our integration of MRay is great”, i wouldn’t buy a packet of peanuts from him. Prman is a whole other story, generating a .rib file is something really different, i hope C4D will be connected to Prman one day.


#20

As far as I know the maya man plugin is supposed to be better then the maya builtin output to prman. Fact is the best bridges from XSI maya, max to PRMAN are developed by a thirdparty not Pixar or the 3D app developers. We have a 3rd party rib exporter being developed, only way its goint to get there is if people test it and offer requests suggestions. The only ones I would ever want to see brides to are PRman and Mray, but personally I can’t afford either myself, thats my hopes for studios, do not there are also lots of free renderman compliant renderers out there.

The others I just see no reason for, FR and Brazil while they may have a few features we don’t, don’t have multpass and an AE bridge that ours does and their prices are far more then C4D and AR put together not to mention additional render licenses. Theres just no feasible reason imo.

What I"d really like though is a good npr system, I’m not a big photoreal guy in the first place.

Oh, for those that want good DOF, C4D’s is basically the same as the DOF plugin in cebas optics suite for max, its not going to get a whole lot better (some maybe but not a lot) If you want a good one send a port request to evasion 3D, they have the best in the industry imo, its raytraced which means it would be slow, but the qualityis absolutely amazing.