Very first impression

[Devin Johnston]

Just so you know CGArchitect will ban anyone using pirated software, if you are using it don't post it here, if your not then have a nice day.

[Bongo51]

Trust me, no one has been more critical of Maxwell's speed deficiency than me. As Maxer said, it can be useable under some circumstances, but for interiors it simply isn't commercially viable.

OTOH, I wouldn't discount it completely, even for interiors. If you want the absolute best quality you can get, time be damned, I don't think VRay or any of the others can touch Maxwell.

I use Cinema 4D Advanced Render (AR) and finalRender-2 in addition to M~R. Hopefully I'll be adding VRay soon, as there's an effort afoot to build a CinemaVRay connection.

In terms GI, AR and fR-2 have similar quality, but fR-2 is much faster. AR is a faster and IMO better looking raytracer.

So, maxwell's worst enemy in my opinion is actually Next Limit. They have single handedly misrepresented and mis'branded' this puppy.

Maxwell is a problem solver... an artist's dream... a glorious liberation to those photographers/3D artist who want to blend the two disciplines! It's cg tool capable of meeting the demands of a photographer. As a photographer and Lightwave animator, my life has thus far been about how close I could bring the two together. Maxwell makes this even MORE possible.

-

however....

-

It really doesn't do much to help with the following:

1. High pressure, time related or goal related projects. - that light simulation comes at a high price. Animation using Maxwell is DIFFICULT at best. with any high polycount scenes you are looking at several minutes/frame simply to generate the mxs file (if they still call it that)... AND each frame can easily be up to 1 GB of data... I'M NOT KIDDING. So animation in any real professional capacity is OUT OUT OUT.

-

2. Flexible. Not. Nadda. NEVER has been. The whole LIGHT SIMULATOR thing pretty much takes care of that. If you want flexible... Lightwave, Max or Maya. wanna render out everything at once? reflection, shadow, alpha, garbage matt, zbuffer, etc. etc.. well maxwell can do alpha right? but there is the fact I only have 200GB of space left... that's 300 frames or so before I run out of room. Hmmm... so a few seconds of animation then archive/delete... rinse repeat.. again.. not production pipeline friendly.

-

3. "I want to render 12 gazillion polygons of interior space but I only have 12 hours"... why is it so NOISY AND SLOW. well my young padawan... repeat after me. LIGHT SIMULATOR... run that through the noodle for a few minutes. Then do it again. Next Limit is pushing it as the solution for all of us. How... sorry...no... still laughing... err... ok.. I'm ok now. Maxwell is a few things, but a general purpose tool isn't one of them.. yet. Of course, if you really wanna see slow, try setting up an interior render in Lightwave with full radiosity/caustics of anything in the general vacinity of talking about 100 million polygons and see what happens. By the time you have a grainy maxwell PHOTOGRAPHIC image, (by that, something you could reduce to 100x100 and email to a client and get directions) and by this time you may just see the tip of the upper left corner of the wall... no... wait... it's too bright... quit that... no... it's not quitting... is Lightwave frozen?

-

Personally... I don't get the "it's so slow" thing... maxwell is anything but slow, that is, for what it's accomplishing.

Can you imagine the marketing buyline if they targeted the true maxwell audience?

Maxwell. The light Simulator. Designed exclusively for real world simulation. Where price is not an issue, and time is something you have oodles of! All you pro photographers, physicists, unt ubergeeks, will swoon! Maxwell Render, providing the most realistic scenes known to man, with a speedy average render time of only 60 hours for a 2000x2000 pixel image. Looking for professional rendertimes? Not a problem, you have lots of cash right? Yes you do! We have lots of licenses, and you will need at least 20. A render farm is the only way to fly. If you want poster sized renders in under a week a render farm is the way to be!

-

Oh yeah... I can just HEAR the money rolling in.

-

Maxwell should NOT be in the same ring as Vray or Brazil and other biased renderengines. They are pro tools for a certain job. They are excellent at what they are designed to do. IMO maxwell can't compete with them in many key areas, such as speed and noise (at production speeds). On the other hand, if that were maxwell's true job, then yeah, it would be awarded the suckiest of suck trophy... but it's not. It's a CG camera. And damn if that isn't a hard concept to sell to anyone who hasn't already spent hours looking through the viewfinder of a CAMERA camera. It's slower then all the other kids... it's noisier then all the other kids... but really... in the end. If it's a scrap where photorealism is the only parameter... well it's bloody nose time for all the other kids.

-

As a pro renderer for an architect where it's the ONLY renderer... oooo bad call.. but as a SECOND dedicated renderer where it's only task is to do those HIGH END photographic comps... or to simulate materials and products with the utmost of clarity and faithfulness... I know of NOTHING else. Which is bluntly why I've stuck with it so long. To render out something which will look stunningly like something which rolls off the factory line is very much something of value. This is it's true strength.

-

Next Limit has pulled a few things quickly and at the last minute out of the dark anus where they were previously storing all of our tech support emails... but in effect the renamed product "light simulator" may yet be of some use to us. Though in the understandable goal of wanting to stay in business they have politely neglected to mention that it was practically useless for any production work where a render farm of 100 machines isn't a given. (which means $25,000 of maxwell render) So... production IS possible for studios... not so much for peeons like me. Anyone purchasing Maxwell thinking it's a cure-all for what ails you needs to find a different doctor.

-

they are tracking us...

not this ship sister.

oh, in case I've offended anyone. well. that's never my intent. honest.

[AdamT]

Maxwell should NOT be in the same ring as Vray or Brazil and other biased renderengines. They are pro tools for a certain job. They are excellent at what they are designed to do. IMO maxwell can't compete with them in many key areas, such as speed and noise (at production speeds). On the other hand, if that were maxwell's true job, then yeah, it would be awarded the suckiest of suck trophy... but it's not. It's a CG camera. And damn if that isn't a hard concept to sell to anyone who hasn't already spent hours looking through the viewfinder of a CAMERA camera. It's slower then all the other kids... it's noisier then all the other kids... but really... in the end. If it's a scrap where photorealism is the only parameter... well it's bloody nose time for all the other kids.

-

As a pro renderer for an architect where it's the ONLY renderer... oooo bad call.. but as a SECOND dedicated renderer where it's only task is to do those HIGH END photographic comps... or to simulate materials and products with the utmost of clarity and faithfulness... I know of NOTHING else. Which is bluntly why I've stuck with it so long. To render out something which will look stunningly like something which rolls off the factory line is very much something of value. This is it's true strength.

Yes, I think that's it in a nutshell. Okay, a pretty big nutshell.

[Fran]

Excellent points all around Ian.

[CHE]

I am writing this from the arch-viz point of view. Users with backgrounds other than architectural (like product design for example) might think otherwise. Please read this with an open mind.

 

I have been doing some test with the 3dsmax plugin and I have to say that I am very disappointed with this release. There are some good improvements like the multilight option and the more interactive preview. However, there are too many things that need to be fixed or finished. Maxwell V1 uses more memory than the other builds and there are reports of possible memory leaks on top of that. You need at least 2GB of Ram to be able to render something. The manual doesn't go over the mayor updates in depth, like the new stand alone material editor (MXED), so it takes a lot of testing to understand it. We have been told by the A-team to avoid using the so called "beta materials" within 3dsmax and instead use MXMs for everything which are not fully integrated into the plugin neither material editor nor view port feedback in 3dsmax for example). The MXED has its own cluster of bugs. It only supports RGB maps, the material preview is very grainy, clipmaps are not working as expected (not to mention the time it takes to setup the material), bump maps are not working, etc. I have not even tried learning Maxwell Studio because I can only expect more of the same and it would be just a waste of time.

The A-team obviously didn't do their job to the fullest so it is up to us to continue testing V1 in the hope that NL will listen this time. Devin is right, we are the B-Team whether we like it or not.

 

I really hope that NL delivers some patches soon. Until then, I will continue to use the last Beta built for all my production work for now.

[AdamT]

Agreed. This was all entirely predictable after they decided to release v1 without a real RC. In essence v1 *is* RC1.

[Bongo51]

CHE, yeah I come from the photographic still image point of view... so maxwell kinda justifies it's existence regardless of how much of a quirky inlaw it is. I can only imagine my frustration if I had bought Maxwell V1 for $1000 and discovered it's true state of development at the expense of whatever job I bought it for. It really is misrepresented online... but that's nothing new. Next Limit released V1 as RC1.. I don't know if you ever had the pleasure of using RC1 but the installer didn't even work properly. And they called that A FINAL VERSION... MAXWELL V1. Personally I don't think the current V1 really is a full 1.0 product, but I imagine if they delayed any more there would be either firings or layoffs. If it were my product I would make sure there was one of those content rating warnings on the 'virtu-box' stating clearly. "contents not suitable for homo sapiens; hybrids or big-brained aliens recommended"

-

Adam... what are you saying? I don't think anyone has ever commented on the size of my nutshell before. Though I do tend to ramble when it's really late... and I've had way too much 'coffee'.

-

Fran, thanks

[Devin Johnston]

Personally I don't think the current V1 really is a full 1.0 product,

That's the understatement of the year, I don't even know if you can call it an RC because it's still missing a lot of features and from what I can tell is hasn't been optimized yet.

[Adehus]

Personally I don't think the current V1 really is a full 1.0 product, but I imagine if they delayed any more there would be either firings or layoffs

I'm starting to think they were damned if they didn't release, and they're damned now that they did. Setting aside personal feelings about the company or the product, if you just look at the level of activity on the NL forum, it seems to me that Maxwell may have fallen short of it's promise one time too many. There are very few new images being made, and those that are shown are not equal in quality to the beta work. Could be because the app is inferior, or could be because the best talent has given up on Maxwell (or was banned).

 

All in all, with users like Zuliban giving NL the finger and forum activity nearly where it was prior to v1.0, I can't imagine that NL is a financially healthy company.

 

And man would I love to know how many $1000 licenses they've sold!

[Bongo51]

adehus, yeah... Next Limit's belief that "this is our forum and our politics" is a an ideal way to drive anyway anyone really capable of using Maxwell professionally. It still makes me laugh to think how much of their current user gallery on maxwellrender.com was generated in the golden age of ALPHA.

maxwell is a worthwhile technology to have, but I hope between horrid customer service, manipulative selling and advertising, product mis-branding that it wont be the end of maxwell. If maxwell fails I land blame squarely on next limit's management... not the software. But I guess that could be said about any excellent product that failed due to the incompetence of those building it. If anyone is familiar with the story of the brickland (sp?). Maxwell is the software equivilant in so many respects it's creepy.

[AdamT]

There are very few new images being made, and those that are shown are not equal in quality to the beta work. Could be because the app is inferior, or could be because the best talent has given up on Maxwell (or was banned).

All in all, with users like Zuliban giving NL the finger and forum activity nearly where it was prior to v1.0, I can't imagine that NL is a financially healthy company.

And man would I love to know how many $1000 licenses they've sold!

 

It's probably a combination of a number of things, but yeah, I'd say dilution of the talent base is the No. 1 cause--and probably NL's biggest marketing blunder. In addition, the complexity of the material system, screwed up plugins, and memory limitations are all major hindrances to producing good work with v1.

 

I've seen a few good *potential* renders over there (potential because they're too noisy), but there's a high percentage of almost-laughable pics too.

 

As always, the person using the tool is more important than the tool itself.

[_PopArt]

Dont forget that the new v1 makes a lot more noise and is slower than the beta, and this is probably due tot the new material system.

 

MLT and path tracing was from the beginning very long and more so difficult to clear away noise especially at higher sample levels.

 

Adding on such a complex & sophisticated material system can only slow down and compicated the tracing and sampling algoithms, and make the much more difficult with all of the scattering calculations.

 

v 1.0 is ambitious with alot of really innovative things, but my impression is still that the core has been borked. They tried to get too much out of a good thing:)

[Devin Johnston]

I'm working on a rendering for a client now using V1, and I'm finding that anything beyond a simple material with one BSDF layer is very difficult to set up. I also think that V1 is missing something that the beta had in terms of image quality but what "it" is I can't say.

 

When I purchased Maxwell I thought it was going to be the answer to all of my problems, but it turns out that it's caused more problems than the software I'm using now. There was so much potential and NL has flushed most of that down the toilet and given us cheap tricks like glare to satisfy us. I'm finding that I need 40 or 50 high power computers to render out an interior scene over night which is just a ridiculous use of resources. I can't even use it to do animations because of the long render times but even more daunting are the incredible file sizes the MXI's generate. I think I may just buy a copy of Vray and see for my self which one is actually better.

[Adehus]

maxwell is a worthwhile technology to have, but I hope between horrid customer service, manipulative selling and advertising, product mis-branding that it wont be the end of maxwell. If maxwell fails I land blame squarely on next limit's management... not the software.

 

So true- if they'd have just realized the niche they should occupy and marketed Maxwell accordingly, they'd probably be in a much better position than they are right now.

[Adehus]

There was so much potential and NL has flushed most of that down the toilet and given us cheap tricks like glare to satisfy us.

 

What... cheap tricks from NextGimmick™?

 

Errr... I mean, NextLimit.

[runrun]

Maxwell Render is an awesome idea and concept. It's a great piece of beta software. But, unfortunately the release was poorly executed.

[FlytE]

Ive been looking into the possibility of getting maxwell recently, not as a substitute, rather as an addition to vray. I consider vray a great renderer and would happily use it for 90% of stuff I would do. The role I see maxwell taking is one of high end production - the moneyshot - the big cahoona.

 

So with this in mind can I play devils advocate and run a question/scenario by you.

 

A client comes to me with a brief and I estimate all things considered, he will get a final image in 2 weeks times. After modelling I could use vray, taking several days (lets be fruitful and say 3) to tweak and render, tweak and render, until i get what a result im happy with. On the other hand I could take maxwell, not have to spend that time tweaking and testing, instead leaving it on for 36 hours to render and also have a result i am happy with.

 

At the end of the day, what is the difference?

 

I mean im soaking up this debate like a sponge because i am trying decide if maxwell is a decent enough renderer to use. So im going to the galleries all over the place and seeing some terrific maxwell rendering.... and wondering what the problem is? Obviously someone is turning out these renders effectively. Can this rendering time not be figured into a project timeframe?

[FlytE]

Oh i forgot to add this link. Some lovely images these people do.... Although much of their work is conceptual, im wondering if they experience workflow difficulties with Maxwell:

 

www.rendertaxi.de

[Devin Johnston]

Ive been looking into the possibility of getting maxwell recently, not as a substitute, rather as an addition to vray. I consider vray a great renderer and would happily use it for 90% of stuff I would do. The role I see maxwell taking is one of high end production - the moneyshot - the big cahoona.

 

So with this in mind can I play devils advocate and run a question/scenario by you.

 

A client comes to me with a brief and I estimate all things considered, he will get a final image in 2 weeks times. After modelling I could use vray, taking several days (lets be fruitful and say 3) to tweak and render, tweak and render, until i get what a result im happy with. On the other hand I could take maxwell, not have to spend that time tweaking and testing, instead leaving it on for 36 hours to render and also have a result i am happy with.

 

At the end of the day, what is the difference?

 

I mean im soaking up this debate like a sponge because i am trying decide if maxwell is a decent enough renderer to use. So im going to the galleries all over the place and seeing some terrific maxwell rendering.... and wondering what the problem is? Obviously someone is turning out these renders effectively. Can this rendering time not be figured into a project timeframe?

 

Your single biggest problem with Maxwell is going to be its speed. For you to produce a 3000x2550 image of an interior will take anywhere from 50 to 100 hours on a single PC using all 4 licenses. If you have access to a render farm you can produce this image over night but it will take a minimum of 25 machines and 7 Maxwell licenses to do it.

[Fran]

... After modelling I could use vray, taking several days (lets be fruitful and say 3) to tweak and render, tweak and render, until i get what a result im happy with. On the other hand I could take maxwell, not have to spend that time tweaking and testing, instead leaving it on for 36 hours to render and also have a result i am happy with.

 

At the end of the day, what is the difference?

 

From my experience, the new material system and compromised emitters that come with M~R V1.0 pretty much level the playing field as far as tweak-savings go. The time you save not messing with GI controls is completely lost while you tweak the materials and try to compensate for the unrealistic results you get from the emitter presets. So, once you figure all that out, you still have a mammoth render time.

 

BTW, the rendertaxi render you reference was done with the beta, which is a different render engine from the current release.

[Adehus]

It's really difficult to explain the pros and cons of Maxwell without using it, but I'll do my best:

 

*Not knowing how long a rendering will take- Times can vary widely. Noise can be absent from a majority of a rendering very quickly, and yet never disappear from other areas (ever). This seems to be worse with v1. Interiors in particular are very time consuming. It's not unusual to hear of 200+hr render times for large format interior images. Try to imagine what you'd do if a client saw your rendering and asked for a few 'minor' changes before he showed it to his client the next day?

 

*Setup time/tweaking- The upside is that setting up a scene is quite easy, the downside is that if you did something incorrect (like excessive bump on a texture map) it can take a very long time for a rendering to clear up enough for this to be evident. I tend to do at least 4-5 renderings before getting my settings the way I want them using Maxwell. As you can imagine, this is excruciating.

 

*Bugs- they're everywhere. Plugins for many applications are very problematic, and Studio is not a welcome addition to my workflow. General consensus seems to be that the beta version was the best overall, and that image quality and render times are both worse in V1. (note that the Rendertaxi images, like most of the more successful images, were done with beta).

 

*NextLimit- I'm sure you've read the criticisms... what can I say, they're true. Don't buy the software if you have any expectations of being able to rely on the developer for support or updates (or even basic honesty).

 

----

 

For me, it just doesn't work, but I'm generally working on projects that require at least 5-6 views in a 1 week/2 week timeframe. Some seem to be able to make it work for them, though it seems that enthusiasm among archvis users is dwindling.

[FlytE]

very interesting. I still dont really understand how they could have gone to a worse place from the beta though. Surely overall there is an improvement?

 

Plus, while 100 hours seems rediculous on the face of it, aim to have a scene finished by wednesday night, and you will have it ready for the monday morning.... *shrugs* This of course doesnt eliminate the error/change factor that you speak of ... which admittedly is an all too real factor than you have to take into account no matter what software you are using!

 

So the basic recommendation here is that if you are interested in maxwell and feel like braving an overall innefficient rendering engine, you need to use the buggy beta version and still prepare for extremely long render times. Hah!The prognosis doesnt seem too good...

[Devin Johnston]

Actually one of the members on the Maxwell forum just did a very thorough and scientific analysis of the beta vs. V1. His conclusion was that V1 clears out noise 4 times faster than the beta does, and although there is some question as to his methods of getting to this answer most people are in agreement with him on the speed. My recommendation is if you have the money and don't mind software that is still in development even though the makers of it claim it's a finished product, and you can deal with the still long render times, give it a try.

[Fran]

Devin, I would not use those tests as conclusive evidence that V1 is faster at clearing noise in dielectrics or indirect lighting situations. The test was specifically a direct light/studio product shot type render of metal surfaces arranged on a diffuse surface.

 

My comment hoping that someone would do such an indepth study of dielectric and indirect lighting situations was completely ignored.

[Adehus]

Devin, I would not use those tests as conclusive evidence that V1 is faster at clearing noise in dielectrics or indirect lighting situations. The test was specifically a direct light/studio product shot type render of metal surfaces arranged on a diffuse surface.

 

I second that... indirect lighting situations certainly don't seem faster to me. IIRC, several users have done tests that seem to support that, though they weren't anywhere near as complete as ThomasAn's tests.