Maxwell at Vismasters

[Devin Johnston]

I thought you may be interested to hear about Mike V's presentation of Maxwell at the Vismasters convention. He gave his talk during lunch on the first day, it was a packed house and he had everyone’s undivided attention. The table I was sitting at was very interested to see this software demo, many of them had heard about Maxwell and I was surprised to find out that only a few had any knowledge of NL's past history. Not wanting to spoil the show I thought it was best to let NL's software speak for it's self so I didn't bring up any of their past problems. I found that I was actually nervous anticipating what Mike would do and because it was the first live demo of Maxwell that I had ever seen. Mike started out the presentation by showing how simple Maxwell’s render controls were and to demonstrate this he compared it to Mental Ray. Once the Mental Ray's render settings were fully open it was easy to appreciate how simple Maxwell is in this regard. I would have called this a successful comparison if it hadn’t been for the low whispers at the table I was at. Several of the people there were obviously Mental Ray users and were quite upset that Mike was "bashing" Mental Ray in that way. In Mikes defense I don’t think he intended to insult everyone in the room who uses Mental Ray but I had the distinct impression that they weren’t to happy about his insinuation that they were all stupid for using such a complex GI engine. His point was that since you didn’t have to spend all your time tweaking GI settings you could devote that time to perfecting materials and in general making the entire scene better. He proceeded to demonstrate how easy it is to set up an exterior daylight scene, and he dropped a few spheres in and applied a few simple materials to them. Most everyone was impressed with this demonstration; even the previously humiliated Mental Ray users couldn’t deny that Maxwell seemed to be extremely easy to use. Toward the end of the presentation Mike had about 3 or 4 Maxwell engines open at once, each was rendering a different scene. As he went through them to close them down it was obvious that there was still quite a bit of noise on the more complex scenes. After he was done the people at the table had many questions but most were about the noise that never seamed to clear from his renderings. As you can imagine once I explained how long it took to get print resolution images, or heaven forbid animations out of it they weren’t as enthused as they had been. In the end most of the users were impressed with Maxwell but were apprehensive at the thought of such long render times. I think for NL it was a good showing of their product, but the presentation didn’t really touch on render times which is Maxwells single biggest weakness, but in that environment I wasn’t surprised that it was skipped over.

[Crazy Homeless Guy]

hey maxer. i was at your table. i was very impressed with Maxwell. ....besides the GI, the dynamics of the materials were impressive also.

 

i can't say that i was offended that he was criticizing GI engines for being overly complicated, i was more surprised that Maxwell would let itself be marketed in this manner.

 

....but i think the single biggest weakness that you pointed out is more than enough to make it completely useless for my daily work flow.

[Ray]

Hey Maxer,

 

I wasn't at your table, but I did hear Mike say that the biggest complaint about Maxwell was the long render times. He also said that the longer render times were balanced by the short (and almost non-existent) setup time thus equalizing it with other render engines. I'm not supporting it one way or the other, just reporting what I heard. I'll leave it to others to decide if he succeded or not.

[Crazy Homeless Guy]

He also said that the longer render times were balanced by the short (and almost non-existent) setup time thus equalizing it with other render engines.

 

i appreciate the nonexistent setup times, but as to saving you time.... it depends on what environment you work in. we were making changes on a project this morning that needed to be rendered and out the door by noon. even if i saved time early in my work flow by not tweaking settings, i might have wound up being screwed in the end.

 

i am all for a rendering engine that is this easy, but until either the software is highly optimized, or computers become 20 times faster at rendering, i will not be able to use it.

[Vince Paske]

I didn't pay too much attention to the Maxwell guy, my table was having too much fun and I wasn't very interested. On the other hand - the mental ray break out session with Chris Bullen was really great and I'll be experimenting with that when I become unburied with work.

[Devin Johnston]

Hey Maxer,

 

I wasn't at your table, but I did hear Mike say that the biggest complaint about Maxwell was the long render times. He also said that the longer render times were balanced by the short (and almost non-existent) setup time thus equalizing it with other render engines.

 

Hey Ray, you’re right what Mike said is true from a certain point of view and that is really the misleading part of it. If I were rendering a scene like the one he showed in his demo with the metal or wooden spheres in it I could probably render out a perfectly clear screen resolution image in 4 hours like he said. The problem with Maxwell is the larger the resolution is and the more indirect light there is the slower it gets. For instance I could render out an almost noise free exterior scene of a complex building with lots of glass using cooperative rendering and 10 dual core CPU's in less than 15 hours. Now if I tried to render out an interior scene with lots of shiny and blurred surfaces after 15 hours and 10 dual core machines my image would still be pretty noisy, in reality it would probably take more like 24 hours to reach an acceptable level of noise. If you work in the prototyping industry then Maxwell will serve you well because your working with small objects usually at lower resolutions and lots of direct light, that's not the case for architectural spaces. So while what Mike said about setup times is accurate he just glossed over the render time issue giving you a best case scenario which would never apply to arch viz.

[Ernest Burden III]

My biggest issue with the MWR session was that I wanted to talk to the people I was sitting with. On the other hand, it was a way to allow a sponsor to show off their stuff without cutting into conference time.

 

I missed the start and end of the demo, so I had to work out that it was Mike Verta from the face (minus the Darth Vader mask) and presentation style.

 

I think it would have been more effective with a presentation more focused on arch-vis, even with a simplistic demonstration scene. The rendered sphere is so 1989. At least it wasn't a chrome sphere on a chessboard. Maybe I missed it, but a discussion about the glass issues with the light-path render method would have been good.

 

The conference needs sponsors, so as long as what they are offering is relevant to our work, I'm willing to listen. But the sponsors should also listen to us to be sure they are approaching us in the best way, and not attacking one another. If I want to see a fight, I'll go to a hockey game.

[BrianKitts]

I got to hand it to him, he sold the product well..... I was even sold on it, and if I didn't have the knowledge about it's render times I would have ran home from the conference uninstalled vray and bought maxwell.

 

What got me the most is how he said at one point, "oh I just rendered this out this morning in about three hours".... referring to an image that took up maybe 60% of the width of the screen.... running on a projector meaning he was at most probably 1024res width making his image, like 600 pixels wide???? As maxer mentioned he didn't make any reference towards print res... which is pretty much why when he was done I spun around at the table and just said what a bunch of BS.

 

I've been trying to convince alot of the people in our office to present our stuff using projectors and go paperless.... for two reasons, one enviromental, and two I'd rather render out 1200 pixels wide than 3000 if I know they are only projecting. Under this circumstance maxwell could actually become a possibility for production, but if it's going to paper....forget it.

 

I won't dwell on the bashing style of his presentation since it's been mentioned already, I am just curious to know what next limit said after they heard about how he presented and if he tainted their reputation..... oh wait, nevermind

[Devin Johnston]

I'm pretty sure Mike was at Siggraph with Juan and Victor and I would think that they know him pretty well since he has been involved with Maxwell from the beginning, so I doubt his presentation style or anything he said came as a surprise to them.

[Fran]

I think Mike probably chose to contrast Maxwell against mental ray mainly because that is his other production renderer and he knows it well. Mental ray is also one of the more complicated engines regarding settings.

[gfa2]

I got to hand it to him, he sold the product well..... I was even sold on it

 

I actually felt the same exact way. I found myself saying...maybe there is something to this program, but then there are those render times, which I also cannot deal with. The people I was sitting with were all Vray users, but knew nothing of Maxwell...I hesitantly filled them in about the render times and told them to make sure they research that before they made any final decisions.

[Adehus]

I've always been really bothered by the idea that quick setup time makes up for long render times. Think about it like this- how many times do you model your entire scene to your satisfaction, map it to perfection, and then perfect your render settings without actually, y'know, rendering? I know I don't. So, you render, see 20 things you need to change, and then re-render. And re-render again, right?

 

At this point, any time you've gained from a relatively quick setup on one small part of the entire process (that never struck me as being too painfully difficult anyhow) has been utterly obliterated by literally hundreds of hours you could be waiting for your renderings to clear up enough to evaluate them!

[Adehus]

Regarding speed, it should be noted that it is possible to introduce bounce control into Maxwell... it was there in the alpha version and it *dramatically* sped up the rendering speed, with a seemingly imperceptible loss in light quality, etc. However, NL has refused to introduce it into any subsequent versions, stating that it's not 'unbiased'.

 

This seems ridiculously academic to me... ever had a client request an 'unbiased' rendering? Of course not. Now, have they ever asked for it yesterday? You bet!

 

NL rejects our reality and substitutes their own.

[Devin Johnston]

I've always been really bothered by the idea that quick setup time makes up for long render times.

 

In this case I have to point out that the amount of work and time it would take an intermediate user of Mental Ray or Vray to get to the level of quality Maxwell can provide in a few mouse clicks can be painful. I don't think Maxwell is looking to bring in expert users of these programs; their target is going to be those newer users who don't want to spend time learning complex GI settings in order to get nice renderings. My only point is that NL should be more up front about render times and not dismiss them as the price you pay for using Maxwell.

[Adehus]

I don't think Maxwell is looking to bring in expert users of these programs; their target is going to be those newer users who don't want to spend time learning complex GI settings in order to get nice renderings.

 

'Easier' is fair, 'faster' is utterly ridiculous. I bill by the hour.

[AJLynn]

I'm not at all convinced by the arguments about it being hard to set up other renderers anymore. Looking at mental ray 3.5 in Max/Viz, the sun is a no brainer, the FG presets are easy, switching between FG and photons+FG is easy, the image sampler settings are easy, the sun and sky and exposure control are easy, interior lights are a heck of a lot easier than Maxwell with the Erco drag-and-drop interface. The Arch&Design materials are easier to use than the Maxwell materials. The results are very good after not much practice, and without really doing any optimizing it's a lot faster. With more practice, you get better and your optimizations get better, but it seems like a very simple thing for beginners to use.

[Devin Johnston]

Ok now that we've go some Mental Ray users here let me ask since I don't know, can a new user with a week of MR experience produce a photorealistic rendering in a 24 hour period of time?

[AJLynn]

Is this a total newbie or somebody with some rendering background? Somebody who's transitioning from some other renderer like Vray, fR, etc., should be able to do great stuff after a week.

[Devin Johnston]

This person would be coming from a scanline background so this is their first experience with either MR or Maxwell.

 

My first experience with GI was with Final Render, I'd say that after a week I was just beginning to realize that I had no idea how that program worked. It took a few months before I felt comfortable enough using it to put it into our production pipeline. When I transitioned over to Vray I had a much easier time because those programs are very similar. Maxwell is very different from any of those programs, for exterior scenes there is virtually no tweaking necessary which I can't say is true for either Vray or Final Render. Saying something is easy doesn’t make it so, it's only easy because you understand what your doing and are good at doing it.

[Ernest Burden III]

there is room for a lot of render engines in any production environment. a good artist will use the one that suits the job at hand.

 

I couldn't agree more.

[Devin Johnston]

I'm not even trying to say that Maxwell is the perfect engine for everyone, I'm just going off of what Mike said in his presentation, Maxwell is easier to set up initially that any other engine out there except maybe for Fry. This fact isn't debatable in my opinion however just because it's easy to set up doesn’t mean it's perfect for every job and I never said it was.

 

When does ease of use equal loss of functionality?

[Ernest Burden III]

just because it's easy to set up doesn’t mean it's perfect for every job and I never said it was.

 

I certainly didn't think you meant that.

 

Ease of use is important. I got used to Lightscape, and it was a shock to be going back to all those damned settings in Cinema4D's engine, and FinalRender. Next will be vray. I'm happy to have Fry and Maxwell for that ease of use. But once you get a good set-up with an engine with all those buttons and sliders, you can get by pretty well.

 

What we have here is a situation where we have a lot of really good options. None are perfect, but these programs are all capable of producing professional work. It's been a long time in coming, and I've worked my way through that entire time. Things are good.

[BrianKitts]

Three years ago when I started vray, and probably even more so in it's even earlier stages, it was no where as easy and efficient as it is today. That's why all software develops over time, only apple realeases stuff that works straight out of the box (joking......well, partially) Maxwell is still in it's development stages where it's not fully ready yet Maxwell is like the young cocky kid who just jumped in to the ring and tried to box with ALI cause it thought it could.... and now they look silly cause they got smacked pretty hard.

 

It's like maxer said, if they'd just acknowledge their short comings I don't think people would be so judgmental against the product.

 

You know the funny thing is in a couple of years all of these engines will be ironed out, and they will be like cars.... you have plenty of choices, they all work and it's really just a decision of personal preference. (I'm just guessing) In time Maxwell will probably catch up in the game in we'll laugh at all these maxwell threads thinking man did they sure go about it all wrong... but it all works now.....(maybe)

[AFK_Matrix]

Well I was one of the lucky winners of an envelope under the seat And guess what I won..... A free copy of Maxwell hehe. I have never used Maxwell and am one of those 'noobies' as you say that hasn't used Mental Ray or Vray in there work (i use Light Tracer with scanline) soooo. I am very interested in all of your comments as I really do want to learn Mental Ray & Vray and the list of settings for them is intimidating. So i will have a play with Maxwell when i get it and I will also be getting to know Mental Ray/Vray. It should proove interesting

[Adehus]

It's like maxer said, if they'd just acknowledge their short comings I don't think people would be so judgmental against the product.

 

This really says it all... NextLimit has conducted themselves in a such a way that the product becomes secondary in many people's minds relative to the attitude of the company that produces it. They've made it clear that they don't really care about their customer's needs and concerns, and they're getting hurt because of it.