#354637
It's only a general enquiry. Is it just me, or is Maxwell for SketchUp standalone is not suited for interior design? I don't see any options to create lighting, other than to create emissive material and the new IES material from Maxwell. I do interior designs and how I must say, I am practically putting Maxwell aside when doing fully enclosed interior design (shops in mall, enclosed interiors, KTVs, etc) and using other renderers instead.

I know that someone would just post a reply and tell me to use main Maxwell Engine 2.0 to create the interior design, though that is not the point I want to bring up in this thread. It's about the product "Maxwell for Google SketchUp Standalone", which you guys have marketed "it contains its own render engine designed for and fully integrated into SketchUp" in the product page. Doesn't that mean MxfSU should be having its own built-in lighting options instead of relying on interior lighting on the main Maxwell Engine 2.0?

Yep that is the main concern here, it costs $95 to get the MxfSU standalone and I be very sure no interior designers would even consider it after trial. It pains me to know that Maxwell didn't consider the market needs of interior design when it came to developing the SU version.

Not trying to complain here, but just voicing my opinion. I think next MxfSU should focus more on lighting options development (omni lights, point lights, spot lights, rectangle lights, etc).
Last edited by DiditzZz on Sat Apr 07, 2012 5:20 pm, edited 1 time in total.
#354638
Those other light types will not happen in Maxwell -- regardless of whether it is the Stand-Alone Plugin or the Full Render Suite. For that matter the lighting options are pretty much the same between the 2 options at this point... so there is very little advantage in purchasing the Full Render Suite when it comes to setting up emitters.

Most of the light types you mentioned can either be easily replicated by the existing emitter materials applied to geometry or are by their nature Biased and thus will never gain a foothold in Maxwell (which is unapologetic in it's stance on Biased vs Unbiased lighting... IES being a grey area).

What exactly is it about emitters that creates such a problem? I really want to understand this so maybe I can deal with it in future tutorials... I mean, I know all of the Biased light-types but I fail to see their inherent superiority to the Maxwell way of doing things. :? What am I missing?

Best,
Jason.
#354639
Hi HL, thank you for the kind response. Your opinion is very much appreciated.

If you are talking about the use of emissive material, in my personal experience I came to conclusion that emissive material generates significantly more noise to the interior. I was doing a pub design using my regular renderer engine, which I'll usually also try to optimize the interior for Maxwell rendering. The results were, too noisy. When I did interior design with natural light-lit ambience (usually residential with lots of windows), interiors would render nicely from few hours to up to 10 hours (sometimes more) of render. Which this isn't the case with fully enclosed interior (like the pub I was designing).

The Pub plays on linear cove lighting design alot so I applied emissive effect, and the noise was unbearable. I guess it is also the flaw in the use of 'emissive' material which also happens in other renderers (same noise, normally). They'll recommend using a fake light source to properly lit the interior (eliminating noise), hence the "biased" style. Though I must say, it is not always the case like that. One render engine also has unbiased render process, and the workaround still needed me to create an invisible lighting to help the 'emissive' lights to lit the interior.

I am not even sure my English and/or my overall experiences on using emissive materials plus the lighting will fully explain the concern that I have. There aren't many examples as well in terms of enclosed interior rendering with Maxwell, or even tutorials to support that. If only there are interior designers experienced in Maxwell, I'd be happy to be friends with them and exchange infos. I am sorry I don't have any renders to back up my claims yet. After all, I started this thread to voice my opinion and it would be very much appreciated if experienced Maxwell users can put in what they have to say.

Cheers.
#354640
I am certainly no expert in Arch-Viz/Interior Design rendering -- however I too am very interested in this subject because it has come up repeatedly when new people try Maxwell... I have to admit I am at a loss as to why they have difficulty with emitters since it seems to be a very straightforward concept to me. But I never considered that the real issue is not setting up the lights but getting faster results.

On that count I have to say that Maxwell can be very problematic -- I can help with individual scenes to find problems (my focus is usually to get the materials squared away first before considering other problem areas... mostly because I find materials (particularly downloads from the Maxwell Resources website) to be a major culprit most of the time. However sometimes the needs of the scene simply requires longer rendering time and there is no (Unbiased) way around it.

I guess the first thing that should be said is having the right camera settings is very important to completely closed-in interiors -- since the lighting levels are generally very low compared to similar scenes with daylight. An EV of between 5-8 is more correct but I often find people are trying to use EV values closer to 10, which means their emitters are often unrealistically strong.

Another issue is obviously the reflective surfaces -- I see all too often users trying to render scenes packed with unrealistically reflective surfaces (roughness 0)... which of course increases noise since there is so much light bouncing off every surface (as caustics)... one thing that can help here to to open part of the geometry to "empty space", meaning create holes for the bouncing light to escape through to help noise resolve quicker.

Finally, (and this is the big one) white surfaces (often in the form of walls) can easily have unrealistically high reflectance 0 values -- anything above RGB 225 is asking for problems, but then if you use an additive layer over anything close to that value you can invite extra noise.

I would look over all those things very closely before assuming Maxwell emitters cannot clear the scene fast enough -- because often Maxwell is much faster than it appears (due to user error).

A case could be made that system could be made more user friendly to avoid those issues -- but as of now the only resort we users have is to educate each other on the best ways to clear the noise quickly based on our experiences.

BTW, the easiest way to test your scene to see if the emitters are really the problem is to do a clay render -- everything but the emitters will render as the default material and any issues with lighting should become crystal clear. If the clay render clears quickly then you know it is a problem with your materials somewhere.

Best,
Jason.
#354641
Interesting insights, I must contribute an extra time to test on that.

Currently this KTV + Pub project is underway, so I can't afford to set aside extra time to test all this. Once the deadline is met, I will dedicate an extra time to test all these, as well as upload the render plus the skp files for you to test. Meanwhile, your advice should be useful at least.

Anyone of you who have tried or achieved a nicely lit enclosed interior render, share if you can. It would be deeply appreciated by the community here. And lastly thank you in advance.
#354642
Let me try to clarify what I think is a misconception running through these posts...
DiditzZz wrote:I don't see any options to create lighting, other than to create emissive material and the new IES material from Maxwell.
The plugin is first Maxwell, second for SketchUp -- the point being: its being standalone does not change the fact that in Maxwell, emitters are always geometry-based. The only other lighting in Maxwell is environmental -- physical sky, skydome, or IBL, and the plugin being standalone does not change this.
DiditzZz wrote:Doesn't that mean MxfSU should be having its own built-in lighting options instead of relying on interior lighting on the main Maxwell Engine 2.0?
...
I think next MxfSU should focus more on lighting options development (omni lights, point lights, spot lights, rectangle lights, etc).
Building on the previous point, it can be inferred that the only way to accomplish this would be for the plugin to create geometry-based lights for you, which would be rendered by the engine exactly the same as if you had built them yourself. That is, it could give you the impression that you were using omni/point/etc lights, but in reality, no new functionality would have been introduced. That is how it works in plugins for applications which have these types of lights -- during export, the plugin reads the parameters of the host application's omni/point/etc light and uses them to build meshes & materials to mimic the behavior of those lights.

I say mimic, because there is no way to precisely duplicate the behavior of such 'lights' in Maxwell, since it is not generally based on any concrete physical behavior. In such applications, then, the recommendation is always against using such lights, since an automatically-generated approximation will never be as highly optimized as a purpose-built light. In fact, automatic translation in these cases is provided mainly as a default fallback, existing to prevent the situation in which a new user opens their scene, renders it in Maxwell, and asks: "Why does my scene have no lighting?" After the person has gained an understanding of how Maxwell works, though, it is never recommended that they would continue to rely on the automatic translation.
DiditzZz wrote:If you are talking about the use of emissive material, in my personal experience I came to conclusion that emissive material generates significantly more noise to the interior.
Again here, in Maxwell, there is only environmental lighting, and geometry-based lighting. Hopefully this clarifies things a bit -- in terms of lighting, there is no inherent limitation in the standalone plugin; its render engine is the same engine contained inside of the Maxwell Render application. Its limitations revolve around extra functionality which is not made available to the plugin, or which has no sense in the context of Maxwell Fire: Multilight, render channels, and such. The actual render engine, though, is the same.

The key, then, is to understand that any image you have seen rendered with Maxwell could just as well have been rendered with the the standalone Maxwell for SketchUp plugin, provided that it were physically possible to construct the equivalent scene using SketchUp. Since your statements indicate that you do not perceive this to be true, the question becomes: how? As Jason has outlined, the answer to this generally comes down to gaining an ever deeper understanding of how to optimize scenes in Maxwell -- minimizing the number of triangles used for emissive geometry, making sure to use realistic scale and camera setups, optimizing the materials used, and so on.

I hope you find this information to be useful.
#354644
Yep, very! I will try to understand deeper about how all these system works. I'll try to solve and if I have anymore concerns I'll post it here. If I can somehow achieve what I want, I might even make a WIP tutorial for it. Thanks a bunch, JD!

Cheers...
#354736
As a question to go more in depth with the emitters in Maxwell for SU:

1) As said in tutorials if I create a component with an emitter material of say 200w, then the component is considered as one unique light source and this entire geometry is equal to 200w. but if I copy 10 times this light component and then groups them, is it 10 lights of 200w each one or not.
2) does the scale of the geometry of this component changes anything about its lighting power. I mean light sources are generaly small objects even though the bulb around is big...
thanx to help
#354742
1) Each copy of the component will create its own mesh (the plugin disables instancing for entities with emitter materials), so 10 meshes X 200W for 2000W total. If you explode all ten, join their faces into a single group, and apply the same emitter material to the group, you will then have one (discontinuous) mesh, producing 200W.

2) Output power remains constant, but power density changes. I often use the example of spreading 10W of power over a surface the size of a football field: regardless the size of the surface, you have 10W output, but the camera probably will not be able to detect it, visually. Emit the same 10W from a surface the size of a grain of sand, and it will appear to shine brightly. It's still 10W output in either case.
#354885
Thanx for your answers JD Hill and to make things clear about this subject.
If I asked this is because I often get to push my lights to max to make them appear in my scene
I explain: I get an house scene in daylight. I put a 100w light, say in the bedroom. When rendering light do not appear at all. to make it appear I get to push it to very high non realistic values... I know you will thats because daylight is so much more powerfull than a 100w light etc... But test it for yourself and you will see it is absolutly wrong. In real life architectural observations human eye will almost always notice when there is a light shining in a room or not.
#354889
Are you absolutely sure about that? If so, the core engine developers would be interested in testing your scene, because the basic intention of Maxwell is that it be physically accurate. What you might want to do is to check how your 100W light renders with the environment disabled, noting the exposure settings you need to use, and then compare that to what is required when the scene is brightly lit by the environment. Here, it is necessary consider the factor of render time as well -- are you giving the engine sufficient time to resolve all of the light in the scene, or is the relatively quick resolution of the bright exterior leading you to judge the result before the interior light has been sufficiently calculated? Lastly, it is possible that your perception of what should be visible does not square with what would actually be visible in an equivalent real situation, using a real camera, and a real 100W light source. Keep in mind that we are not concerned here with what the human eye would see, but with what would be exposed on the film or sensor of a camera, since this is what Maxwell intends to simulate. The alternative to this suggestion is that the engine is, in fact, inaccurate, in which case, as I say, the core developers would definitely want to take a look, because it would go to the heart of what Maxwell is supposed to be.
#354893
is it possible that your light normals are inverted after export to maxwell?
I had this problem sometimes with exported emitters from sketchup (and only with emitters). I had to turn the face normal upwards to make it shine down...
#355116
No it is not an inversion of normals.
Solving time for interior lights...hum :? It can be an argument...However I thought time was mostly controling the overall quality of the image not the intensity. Camera vs the human Eye? There are differences of course but when configured to a neutral position it should simulate human eye and not getting to be so different.
#355123
It is not that it controls intensity, per se. Think of it this way: say that you put the camera inside of a box, which had only a pinhole in one wall to provide light. This would clearly take a long time to render, but we would not say that time was controlling the light intensity, or that there was any inaccuracy involved; clearly, it is only that because of the nature of probability, it takes a long time (i.e. many computations, of which only a few yield a useful result) to find the light in the scene.
Will there be a Maxwell Render 6 ?

Let's be realistic. What's left of NL is only milk[…]