Rendering

[antonio_frias]

When I read all the talk about the new release of the Maxwell Renderer, I just wonder if all you people ever heard about Vray or even Mental Ray? Don't we get excellent results with these and other renderers with less trouble?

[AJLynn]

I've heard of those but some people on the Nextlimit forums told me they were biased, and I only want to use a render engine that's open to all ethnicities.

[Ernest Burden III]

And vRay isn't available (yet) for Cinema4D.

 

Antonio--can you explain your avatar? I see the Canadian flag, but you are in France, and another flag that I do not recognize. Just curious.

[Christopher Nichols]

I've heard of those but some people on the Nextlimit forums told me they were biased, and I only want to use a render engine that's open to all ethnicities.

 

Very funny... but it brings up a serious question in my mind...

 

How many users out there can actually explain what "unbiased" means? If they can, can they explain why it is important? Now it may be just me, but an unbiased result doesn't really serve any true advantage in the world of rendering for archviz or vfx or anything like that. It would seem that its cost (time), is a much greater advantage. It would be like buying a car (my favorite analogy) and paying $15,000 for the manual, or $150,000 for the automatic. I say $150,000 because it seems that Maxwell is around 10 times slower.... with simple renders taking 15 hours.... just based on my observations.

[AJLynn]

Actually, what I like about it are the "beta love" and the concept of a "camera simulator" instead of a "render engine" - though the latter is academic and the software is certainly not yet at the "useful enough" stage and won't be until my computer is 4x faster and there is a certain level of understanding between me and the materials system. I figure it will take a few years for the first to happen, but the second should come when I have more time to practice with it, NL makes a few changes and users start posting MXMs.

[Ernest Burden III]

For me the appeal was the camera analogy. I have a background in photography and it makes sense. Shadow map biases, ray tollerances, those do not. Lights expressed in watts that behave as you would expect them without coding, materials that are simple and work. That was the draw. I'm not sure where we've ended up, though.

[bricklyne]

Very funny... but it brings up a serious question in my mind...

 

How many users out there can actually explain what "unbiased" means? If they can, can they explain why it is important? Now it may be just me, but an unbiased result doesn't really serve any true advantage in the world of rendering for archviz or vfx or anything like that. It would seem that its cost (time), is a much greater advantage. It would be like buying a car (my favorite analogy) and paying $15,000 for the manual, or $150,000 for the automatic. I say $150,000 because it seems that Maxwell is around 10 times slower.... with simple renders taking 15 hours.... just based on my observations.

 

It might have something to do with the fact that the light quality produced by Maxwell, is rather difficult to replicate with Vray and Mentalray, without the benefit of years upon years of mastering, learning, and then tweaking tweaking and tweaking ad nauseaum. My guess would be that that's where the unbiased factor comes into play, seeing as it removes the guesswork as to how a certain scenario would actually look in the real world. Since Vray, for example, takes a lot of shortcuts and cuts quite a few corners, the user is forced to have a good understanding of natural light dispersion and distribution in order to be able to fill in the blanks left behind by those "shortcuts". Here's probably where experience comes in, and is essentially what Maxwell allows users to side-step in a shorter time with their "unbiased" philosophy.

So in a long and roundabout way to answer your original question, most users (and by users I mean CGvizers) may not technically or semantically know what 'unbiased' means with regards to explaining or defining the processes involved, but they certainly are aware of, and appreciate the significance of it's value in simulatiion and 'realism' in CG.

 

It probably also doesn't hurt that Maxwell ( the plugin->render-engine workflow, as opposed to the Studio workflow) is considerably easier to learn and master than Vray. Or that it's available for more platforms and software (without having to bridge through 3DS MAX), than Vray currently is. But of course, if Vray had already released their stand-alone by now, that would be a moot point.

 

Your car analogy makes no sense whatsoever with regards to comparing monetary cost to time/renderspeed. Most people who buy Vray, do so largely because of they place a higher premium on faster renders than, say, ease of learning and use or maybe render quality; by reasoning that the first factor (speed) will over time allow them to overcome the other two (ease of use and render quality). Maxwell buyers on the other hand place a higher premium on Render/Light and ease of use, than one would typically do on speed.

If two users each buy Vray and Maxwell and are each able to come to a level of mastery and overcome the respective shortcomings of the 2 software, to allow them to produce near-identical levels of high quality photorealistic images, how then, can either one of them objectively discredit or devalue the other's cost and or methods of being able to do what they have done with their chosen and preferred resources?

 

Nextlimit certainly have their failings and flaws as a software company and Maxwell has a bit of a ways to go before being considered a formidable render solution, but the fact that a whole host of people are willing to put up with a lot of the crap that happened during the development process to get a quality render engine, or perhaps the fact that a lot of the other more established renderer developers are already beginning to include concepts introduced by Maxwell (physical sky, physical camera, physical sun etc. etc) does intimate that there just might to something to the whole 'unbiased' and 'easy-to-use-or-figure-out' deal.

[Devin Johnston]

The sad fact is that they have been slowly moving away from the product that most of us liked which made us purchase the Alpha/Beta. Ease of use is out of the door, speed is horrible, and now we have Studio which no one really wanted in the first place. I'm not sure how long it's going to be before Maxwell is actually a useful tool like Vray but it's not going to be found in V1.

[Christopher Nichols]

For me the appeal was the camera analogy. I have a background in photography and it makes sense. Shadow map biases, ray tollerances, those do not. Lights expressed in watts that behave as you would expect them without coding, materials that are simple and work. That was the draw. I'm not sure where we've ended up, though.

 

Oh you guys are in for a surprise then.... Did any of you see the VrayPhysical Camera at Siggy last year? Imagine what a year can do.

[Ernest Burden III]

Did any of you see the VrayPhysical Camera at Siggy last year? Imagine what a year can do.

 

I believe it was you that showed it to me. So just what has a year brought us?

[Christopher Nichols]

I hear you Clarence... you make a lot of sense. I come form a different perspective. I come from a work of renderman where you have to know a lot about light just to be able to fake everything including GI... and you have to teak think like specular exponent, and noise lacunarity. Vray is often too simple and does you have to give up a lot of control, much in the same way that you talk about the jump from Vray to Maxwell. The big issue with something like Maxwell in the business that I am in is that it would be insane to use as no one would have the tolerance for that kind of render time. Generally most renders need to take no more than 1 or 2 hours a frame at 2048x1558 (film res)... an be free of any artifact or noise (film grain needs to be added in post to match the plates.

 

It seems to me that your argument looses its advantage over time since eventually people will pass that learning curve.... well at least that is my hope, since I teach Vray.

[BrianKitts]

this is directed at those who have broken into the maxwell ring for I haven't....

 

but it has my curiousity over the level of control that the end user has in setting up the scene. From my understanding is that it does a great light setup and the end user has to do very little to get a good result. Whereas vray would require a little more knowledge and control to achieve the same results.

 

So here's my question, what happens when something doesn't come out the way you want to in maxwell. Do you have enough control to be able to tweak the way things are???... such that you can do in vray....or do you just say "looks good this way anyways, hope the client likes it"

 

If you revist the car analogy, sure you can buy an automatic and start driving it, and get pretty good at it.... but someone who's mastered a manual will beat you everytime off of the line....

 

any thoughts?

[bricklyne]

.....

 

It seems to me that your argument looses its advantage over time since eventually people will pass that learning curve.... well at least that is my hope, since I teach Vray.

 

 

That's quite a liberal assumption ( that people, or more bluntly, 'average' users, will eventually pass that learning curve). When I talk of learning curve, I take it to not only refer to the technical aspects of using a software (learning the tabs, sliders, functions, etc etc), but also to encapsulate the more aesthetic and esoteric aspects of rendering. By this, I mean the small types of tips and tricks used by expert renderers, or 'masters' as they are sometimes referred to, which gives their images that extra 'zip' and the almighty 'je ne sais qui'. There's obviously an artistic facet to rendering which, as I've already alluded to, requires an keen capacity to perception and to discerning and understanding how light diffuses naturally, and which, as a result, is an invaluable asset to any renderer anywhere. In addition to the fact that this can't easily be taught for obvious reasons, there's also the matter of with what degree of ease the render software used allows the artist to translate those observations and perceptions or simulate reality, without the benefit of years upon years of experience.

 

Maxwell's methodology, ( by virtue of the unbiased principle) makes all that moot, since it takes into account things that average users will tend to forget, neglect to do, or just simply will not know about light distribution, in reproducing images; a lot like lightscape used to years ago. And unless Vray where to alter their methodology ( and indeed they did, somewhat, with the introduction of PPT), to facilitate a means of processing which would essentially take away all the guesswork (or experience) from the user in simulating reality, then it's hard toimagine how one will ever be able to teach people the finer points of daylighting or diffusion without diluting the other aspects of what needs to be learnt.

 

Granted the rendertimes are a killer, and naturally, for someone in your profession, there's no way it even approaches practical usability. But even under the best of hypothetical circumstances, and assuming that Maxwell were somehow to be 10 or 20 times faster than it currently is ( matching the speed of Vray for example), it would still be ill-suited to a lot of the things you do due to fact that, in terms of 'faking' stuff, Maxwell isn't as forgiving or allowing as Vray is. And faking is a large part of CG/VFX. Consequently, even under the best of circumstances, I still wouldn't imagine that someone from your background and perspective would be best suited to evaluate its value; quite simply because, it will never quite address your specific needs in the ways that you're accustomed to. Nonetheless there still exists a market for whom its strengths and benefits would be almost immeasurable in their daily work.

 

If only they were to get their act together.

 

The idea is certainly solid. Unfortunately, the execution leaves a lot to be desired. .........

[bricklyne]

So here's my question, what happens when something doesn't come out the way you want to in maxwell. Do you have enough control to be able to tweak the way things are???... such that you can do in vray....or do you just say "looks good this way anyways, hope the client likes it"

 

Against better judgement of answering a leading question.......;No, you stop the render and re-do what you need to get what you want to see. When Maxwell starts a render you get a preview window akin to what Lightcache does in Vray with the "show calc. phase" box checked. So with MAxwell when you start a render you know within minutes whether or not it will converge to what you want. It's not exactly as if the process is hidden from view as it is being executed.

 

If you revist the car analogy, sure you can buy an automatic and start driving it, and get pretty good at it.... but someone who's mastered a manual will beat you everytime off of the line....

 

Based on what series of experiments, races and/or surveys?

 

Let's not even get into the obvious inlain presumption you seem to be making that an average automatic buyer and driver will be a marginal, mediocre or average driver at best, whereas your average Manual buyer will almost always be a natural born expert driver.....

[AJLynn]

I think what he meant was that all other things being equal, the degree of control a manual transmission gives you will allow you to beat somebody with an automatic in a drag race, given that both of you are experts.

 

But anyway...

 

Maxwell does give you some ways to correct things on-the-fly. You can change the film and shutter speed, and the gamma and burn, on the fly to get your exposure right, which is very useful, and with Emixer you can adjust the levels of the lights in your scene, though I hear that crashes the software. So does the glare feature - I saw somebody in the know, I think it was Mihai, blame this on a memory leak. And you have the preview render, which is a very fast way to judge what's going to happen.

[BrianKitts]

sorry if i was leading.... and perhaps I didn't phrase correctly....

 

I know that maxwell is a progressive rendering process that gives you a preview and then goes the distance, that much i've seen, I just don't know how much control you have.

 

From my understanding maxwell is great because it is able to create a solution for all the things that you would otherwise manually set in a rendering engine such as vray.

 

I know you can tell it a location and time such as set the scene for madrid @ noon.... and so on .... but say the light's secondary bounces aren't as strong as you want them to be, bouncing as much light as you would want into recesses.... can you override maxwells predetermined settings to get a desired effect. The hype I hear about maxwell is how amazing it's lighting solutions are, but I don't hear much about it's advantages for creating customized setups.

 

Again I reitterate I'm a vray user with limited maxwell knowledge hence I'm just curious to it's capabilities.

[Devin Johnston]

I know you can tell it a location and time such as set the scene for madrid @ noon.... and so on .... but say the light's secondary bounces aren't as strong as you want them to be, bouncing as much light as you would want into recesses.... can you override maxwells predetermined settings to get a desired effect. The hype I hear about maxwell is how amazing it's lighting solutions are, but I don't hear much about it's advantages for creating customized setups.

 

Maxwell isn't set up in a way that would allow you to modify the Physical Sky but you can adjust fstop and shutter speed. The theory behind it is that the physical sky produces a given amount of light as specified by the time and date you put in. There is no adjusting it just like there is no adjusting the sun to produce more or less sunlight in the real world. The parameters for adjustment are your camera, emitters, and materials you create. Maxwell now allows you to hide emitters so in the scenario you have created there are no secondary bounces to adjust and you would either have to adjust camera settings or create a fill light.

 

I suppose you don't hear anything about creating custom setups because most of the people using it find that its lighting solutions don't need them because the lighting for all intents and purposes is perfect. In other words there is no need for shortcuts because the lighting solution will converge to the correct solution every time if you give it enough time to do so.

[BrianKitts]

that makes sense.... there is control, its just a different set of parameters in the equation allowing to adjust the end product. i appreciate the articulate explination.

 

who knows maybe maxwell could change the current paradigm for how we go about setting a scene when they get their rendering times down a bit!

[xxfm5bassistxx]

I believe Maxwell is the new wave of render engines just has so much more to add. Once other companies start to see the effect of Maxwell they will start to produce "unbiased" engines and have alot better render times. Maxwell is too good for it's own good but at this time the end(render times and quailty) doesn't justify the mean(not having to set up bounce lights and such) at this point of the program.

 

Brandon

[Christopher Nichols]

I believe Maxwell is the new wave of render engines just has so much more to add. Once other companies start to see the effect of Maxwell they will start to produce "unbiased" engines and have alot better render times. Maxwell is too good for it's own good but at this time the end(render times and quailty) doesn't justify the mean(not having to set up bounce lights and such) at this point of the program.

 

Brandon

 

Well this is my point. All of these "advantages" that Maxwell offers. Camera, lighting via wattage, physical skys and sun, etc... can all be achieved with a "biased" (aka adaptive/importance sampling) rendering engine. The user interface and setup also has the least to do with the renderign engine core as well. What is the point of having an unbiased engine? I think that is the issue, people don't really know what it is and assume that all the maxwell "philosophy of lighting," can only be done with an unbiased engine. That is simply not true.

[manta]

Well first of all, giving someone a camera won't make them take a good picture, even though the camera is well aware of all the the things nature produces...

 

It will make it easier for someone who already knows what they're doing, but that doen't mean that he/she will be happy with what they get...

 

secondly, I use Vray's PPT to do previews, its a very handy feature, and is supposedly unbiased when adjusted correctly... so here I have an image created using PPT and the same image using faster methods, my question is why are they so close, I was kinda dissapointed when I first used PPT, thinking wow, this is going to be so cool, and so much more real than the last one...

[manta]

Here is the file, if anyone wants to dazzle me with a maxwell render, and is free to anyone else who wants a mac mini model, I think I pretty much eliminated all traces of vray, so it should load without problems if someone doesn't have vray...

[antonio_frias]

I was born in Portugal and lived there until the age of five and then my parents imigrated to Kitchener, Ontario where I lived and finished grade ten. My parents then returned to Portugal where I finished high school and a course on architectural drafting. After my military service I went to France where I've been living and working as a CAD designer ever since. Don't ask if I feel Portuguese, Canadian or French, I don't know the answer myself. I hope that explains my choice of avatar

[antonio_frias]

Thanks Christopher, its thanks to you that I started trying to work diferently with Vray and I must say that I am starting to see the results

[antonio_frias]

Hey Brian, I agree 100% with you