Daz 3D is part of 
Connect
DAZ Productions, Inc.
224 S 200 W, Suite #250
Salt Lake City, UT 84101
Licensing Agreement | Terms of Service | Privacy Policy | EULA
© 2024 Daz Productions Inc. All Rights Reserved.Daz 3D is part of
Connect
DAZ Productions, Inc.
224 S 200 W, Suite #250
Salt Lake City, UT 84101
Licensing Agreement | Terms of Service | Privacy Policy | EULA
© 2024 Daz Productions Inc. All Rights Reserved.
Comments
Rashad - Fun challenge Classroom? Your scene looks really nice. It almost passes as photo realistic and only when closely scrutinised there may be some minor inconsistencies detectable. Closed rooms are difficult to light. I tried a couple of things, like Obscure Light (one of David's ideas) with the radial inside and - removing the shadow casting for the walls - outside and a few other tricks but nothing satisfactory resulted. The best I can offer is a flat toon room. Apart from the sun shining through the window, there is only an HDRI from inside, not TA optimised because it is too small, and no shadow casting (no change but took 30 longer to render with IBL shadows). Changing the Pitch of your HDRI created different light. So I made a small HDRI with the zenith fully white fading gradually to half bright at the nadir. With adjusting Pitch between 90 and 0, the light could be balanced for the ceiling and the table tops. It rendered 12 times faster than yours (10 minutes at 64 rpp versus 2 hours in this size).
Above Rashad's original szene, half size. Below my toon version.
Nice experimenting, Horo.
Ik still like Rashad's version better, mostly due to the fact that your ceiling is very, very light. But I prefer your stronger contrast of the light through the windows on the floor. And I also like the warmer look in your version. What is missing in both, is the light from the lights. Perhaps a future addition?
Several years ago I was also concerned about the illusion of GI indoors. Here are my old renderings
https://www.daz3d.com/gallery/images/21040
https://www.daz3d.com/gallery/images/21041
Hansmar - thank you. Of course Rashad's version is better. I switched off soft shadows for the sun because my scene looks toonish and soft shadows on the floor would spoil the toon character. Maybe I'll take this scene again later and light it by the visible lamps. But here, it was using TA.
Slepalex - I remember your scene "Lunch at a country house" you showed at bryce5.com already over 5 years ago. The lighting is excellent.
Horo - both your renders are awesome, I like the toon one more.
Slepalex - wow great lighting, beautiful peaceful scene. I love it.
Horo - I appreciate the lack of soft shadows. Main issue I have with your toon one is the white ceiling, which looks weird to me.
Slepalex - wonderful scene and lights! Quite elaborate set-up. Must have taken a lot of effort.
Thank you, Hans. I do not remember. Given that all models were made in Bryce, it took a considerable time. The most unpleasant (given the difficulties in Bryce) is to make models. The most interesting is the layout of the scene and the lighting.
I certainly see what you mean. Making models in Bryce is a pain, compared to using real modeling tools. However, there is also quite some joy in succeeding with a challenging task, even in using a less than perfect tool. All the more reason to be proud of your work.
Here's another try using TA for Rashad's Classroom. I used Obscure Light for indoors with a Hypertexture as explained in my private video to David nearly 3 years ago. Very helpful, isn't it?
In a nutshell: the lights dome, fill (cube and sphere) that are set to TA optomization are seen as a radial (distribution can be 0, inward, outward have no effect). The only exception is the square parallel light, which is seen as a rectangular sheet with both sides lit, Y size can be set to 0, infinite width and direction control have no effect.
So there is a sheet near the ceiling and one near the floor, no Diffuse, TA optimised with a procedural gel. This gel has only hyper texture boosted Ambience and this creates the light. An additional parallel light but with infinite width and some diffuse shines down from above to create the shadows. Rendered premium with soft shadows, TA, Boost Light and 256 rpp in 84% of the time Rashad's render needed, there are only 3 lights, no HDRI. Yes Hansmar, the ceiling is again a bit on the bright side. The underside of the grey part of the "beams" on the ceiling is black, as is the underside of the window frame. I don't know why.
This effect is I think that funny little bug that crops up occasionally. If you were to tilt either of those black surfaces ever so slighty off being perfectly parallell with the ground - they would recieve TA light normally.
David - thanks for that, yes, you're right, I forgot such happens. I rotated the whole room a tiny bit and it looked fine.
Great tip, David.
I like this render, Horo - the floor is, for me anyway, off-putting, too distracting - perhaps, a plain texture might suit, as all those wonderful beams of light upon it are, somewhat, lost.
Jay
Jamahoney - thank you Jay. Well, I took Rashad's scene exactly as it is (even with the fog and haze, but without the HDRI) and only removed all lights and attempted to light the room with TA.
Where is Rashad's school work avaialble...sorry?
Jay
First page here, second comment there are the links-
Very nice, Horo. I think the ceiling is much better now. Not so extremely white; quite believable, actually.
Sorry again, Horo...could you link where Rashed's school setup is...cheers...
Jay
Jay- Thank you for your interest. A direct link likely won't be necessary. On the first page of this thread at the top of the second post you will find all the study materials. Don't be daunted by the long first post. There's a lot of stuff there so please have a look and please report back with anything interesting you discover.
Alexey- I've always loved this piece from you. Lights, models and textures. Brilliant as always!!!!
Hansmar- Its not obvious due to the way the overhead lights have been modeled, but there actually is light being emitted from those areas. There are thin panes within the light fixtures that make it difficult to see the bulbs. But the light is there shining more or less straight down.
Horo- Wow and wow!! Thank you for tackling this challenge! Results are promising especially in the lastest example. But lets take each individually.
In the toon version I can agree with Jay and Hansmar that the lack of shadows is really the only "toonesque" quality of the image. Otherwise it feels pretty legit.
The second version actually looks a lot like the first one but with several smart corrections. Let me state that the Obscure Lighting rig you've devised here is a bit of voodoo magic and I'd LOOOOVE to see the file you've come up with. This render feels much warmer in tone compared to my fake GI offering. Part of the reasoning for this is the lack of cooling blue hues being picked up from the incoming skylight. But it's also because TA is bouncing those warmer tones from the wood floor around the remainder of the space.
The ceiling in the second version is great. Still too bright perhaps, but we can forgive that. It must be a slightly different material than the white used for the walls. The render bug where certain faces of the default cube primitive dont receive any TA light is definitely making itself known. Glad to know a slight rotation is all that is needed. Rather not have to go back and remodel everything. If there was a way to somehow introduce some degree of blue influence I think I'd find it even more convincing. The first version I made only used indirect skylight, all the overhead lights were disabled. But for the upload I decided to add the overheads.
My fake GI version could be much smoother in terms of banding at higher quality settings, and it is at times like this where Obscure light might be the faster approach, even if not as accurate, it still produces a very pleasing image.
I assume scattering correction is disabled? The vertical walls seem to have that specular signature.
Honestly, this result you accomplished here is likely outside the abilities of most Bryce users, myself included. This is the type of education I was hoping to gain. Thanks soooo very much!
Anyhow, if ever the day comes when someone asks for input on the subject hopefully studies like this will help us better understand where TA still needs to grow. Many thanks.
Jay - it's the Advanced Faked GI for Interiors link
Horo - very nice render.
Rashad - thank you for taking the time to look at my experiments. Scattering Correction, Boost Light and Reflection Corrections are enabled. Maximum ray depth is 4, at 3 the glass in the door gets obscurred so 4 is the minimum. The yellowish warm tone is from the floor as you assumed, without boost light, the scene looks cooler, the colours are less saturated and look rather dull. Boost light does not only boost the light but also the saturation.
For a limited time, the file can be downloaded from this link https://horo.ch/guestxchg/OL_Classroom.zip (94 MB).
The catch with this obscure light method is that if the cube light extends the walls, a thin dark line appears where it breaks through the wall. If it fits or is smaller, a thin bright line appears. Here, both extend over the walls and I placed them so that the dark line isn't obvious - that's why I used two instead of one at half height of the room. The line is exactly in the cetre (if the cubic parallel light points up or down) and there is no difference whether it is left cubical or Y reduced to 0 to make it a sheet. A sheet is just easier to position.
Mermaid - thank you.
Here we go again. I rotated the beams, window and the door frames by Z=0.001 to get rid of the "alignment-bug". I found a light assembly left-over that I deleted. Then I still had a hypertexture shadow capture sandwich on the floor I forgot to delete from earlier experiments. I added two 2D faces with reduced transparency, one between floor and the lower flattened cube parallel light and one between the ceiling and the upper ligh sheet. With them, the brightness of the floor and the ceiling can be balanced. The upper face looking down and the lower face looking up is considered by TA to light the walls. Render time is now 30% (256 rpp), file size 58% and memory usage 53% of Rashad's original.
Thanks, Rashad, Mermaid and Horo...cheers
Jay
Horo, in my opinion, the scene is too overexposed. The aperture-shutter speed is too high. It is assumed that the scene is illuminated only by scattered light from the walls and surrounding objects, as well as from scattered light from the window. The sun's rays are too bright.
I see the disadvantage of TA in that it paints surrounding objects too much. If you paint the ceiling beams in red, the shadows on the ceiling will be red, but in life it's not. In TA it is impossible to regulate the saturation of shadows from surrounding objects. In life, secondary rays from all objects are mixed in space and have shades of gray. Only at very close distances can one see the hue of the neighboring object in the event that the latter has at least a small specularity. Also TA almost uniformly illuminates all corners of the room. Although the angles adjacent to the windows should be more dark, etc. The same applies to the shadows from the ceiling beams.
I prepared my version of the lighting of this room. This version is simpler, although it is also not ideal. In the next message I will make screenshots and explanations.
I used only Cube Dome Light. All lights in the scene including HDRI, I turned off, and also reduced the light from the sun. Render Normal AA, 3 hours 50 min.
Slepalex - you may be right about the brightness of the room. Rashad set up a challenge in his last comment on the previous page https://www.daz3d.com/forums/discussion/142591/rashad-s-thoughts-on-lighting-in-bryce-7-1-pro-in-2017/p6 to light this room with TA. I did not question the brightness of the room just took up the challenge and try to emulate his scene with TA. TA does indeed spread colour sometimes a bit far, even more so if Boost Light is enabled. On the other hand, without TA there is no spread of colour at all. So we have two opposing effects and neither is completely natural. In my experment, the walls take up too much colour from the floor. Without Boost Light they stay grey and all the colours in the room have much less saturation. The table tops in Rashads example seem a bit bright to me, in your example, the back of the chairs seem a bit bright, considering the only light comes through the windows on the right, and in my last render the white teachers desk is much too bright. Rashad's scene rendered on my computer in 93 minutes, my last in 28 minutes, both 800 x 445 px.
Yes, I agree that both methods have significant drawbacks.
I tried Cube Fill Light. This was even worse, although the chairs in the foreground were less lit, but near the boards were too bright. On the other hand, if you look from the camera on the side of the blackboard, the picture will be exactly the same. This is the result of a uniform distribution of point sources within the Cube Fill Light. As a result, a smaller number of point sources are closer to the camera. In Cube Dome, on the contrary, all sources are located on the surface of the cube and the main part is closer to the camera. Therefore, the backs of the chairs in the foreground are too much illuminated.
In any case, I'm in favor of looking for the simplest solution with different rendering methods. The easiest thing I could do was scenes with still lifes. But the interior is a bit different.
Alexey- Nice simplcity. I can agree that it is missing a couple of things. I look forward to your more detailed breakdown. Thanks for taking time to play with the study.
On the color bleed issue, I disagree with you to some degree. Depending on several factors there could be strong bleed especially if there are not other colors nearby to challenge it. I'll explain shortly.
Everyone----
Great discussion!! I'm not at home now so I cannot talk too much but there are a couple of things to keep in mind as we analyze.
The reason the desk tops are so bright In the original rendering is because there ARE lights active within the fixtures overhead. It is not obvious from this viewing angle but if you look directly upward you will see there are tube shaped lights active. They are 3D Fill Lights elongated into the shape of cylindrical bulbs. The result is that they automatically provide rather diffused key light as well as soft shadows onto the desks beneath. The overhead light fixtures are an entire tutorial on their own, but we'll get to that in a bit.
So for the record it needs to be clear to all participants that the original file has two major lighting influences. 1. The incoming skylight. 2. The bright overhead lights. One is cool colored, the other is warm. More on that later. Both of them provide a somewhat diffused light because none of them are emitted by what appear to be typical point-like lightsources. Skylight is obviously diffuse because the sky itself is so wide and in this case so is the light from the elongated bulbs, again due to the fact that it because it doesnt come from a single point. The third light influence in this scene is from the point-like Sun but surprisingly is not too major in that the sunlight is limited to making contact with a small area of the floor only. For accuracy sake I should have placed a low powered yellow/orange tinted spotlight facing upward to model the "bounce" of the direct Sunlight in that area toward the ceiling a bit. But it takes careful tuning to get it right.
Obscure Light has never been the most "accurate" way of representing indirect light in Bryce. It's based on counterintuitive hacks of a sort, or voodoo magic as I call it. The fact that David and Horo have developed it speak to the need we have for solving certain serious shortfalls with the current TA implementation , as great as it is compared to previous versions. Depending on the scene, it might provide a decent approximation of indirect lighting, especially if the key light source illuminating the room is paced at the center of the room. But more on that too at a later time. As Alexey observes, the result is pleasing but doesnt necessarily rise to realism. I am however still extremely impressed that Horo was able to wrest this much awesomeness out of it.
Accuracy would seem like a strange thing to consider when the lighting methods we are discussing are highly biased. TA is biased, and God knows that fake GI is biased. However there are different degrees of accuracy we can reach even knowing that complete accuracy is out of the question.
This is the point where Alexey and I tend to depart from one another. Alexey in my view seeks the sweet spot between simplicity and realism. He seeks the simplest possible route to an acceptably realistic result. For me, I am concerned only with acheiveing as close as possible to absolute accuracy, so I am not necessarily looking for simplicity. I welcome simplicity when it arises, but it isnt a specific goal. Thus, my lighting methods might seem to be a bit of overkill for Alexy and other users who think similarly, which is great as this is the whole point of conducting discussions. But if anyone ever wants to know WHY I've done something I can always come up with a good explanation for why I have done certain things for the sake of accuracy.
When i get home later I'm going to upload some critical concepts for rigging a faked GI, the reasons why and why not. This will help to train our eye so we can discuss an image for both its degree of pleasingness to view as well as its accuracy when compared to real life. See you soon!
Slepalax: - highlight on the floor is superb...well done.
Jay
Just found an error in the original file. In the ceiling light fixtures I've kept the 3d fill lights turned on but for some reason the cylinder models which represetn the physical bulbs are hidden. That could be part of why we dont see the rods of the bulbs, but then again maybe not. Anyway, it doesnt alter to overall lighting just a modeling issue. I am currently collapsing meshes so I can export the scene to Octane with the exact same camera angles and other things. Wish me luck!!
So far no one has noticed the EASTER EGG in the classroom scene. A set of early vegetation studies, tropical plants. Unhide the group "Plants Sources" and have some fun!!!
As is known, point light sources are located on the faces of the cube (Cube Dome Light).
Pay attention to the boundaries of the Dome Light in the top view.
The right (eastern) side is pushed back to the next room, for the light from her gave the feeling that to the left of the camera there is a large window that is not right.
The left (western) face is almost entirely in the wall and has no effect on the scene, except for those point sources that are in the window hole. This side of the cube simulates the light from the window. Perhaps more correctly is to push this side out of the room behind the windows. But too long rendering and I did not have enough time for experiments. This thought only occurred to me now. I've done dozens of manipulations and dedicated it to evening and night.
The lower (southern) border is outside the room and does not affect the lighting of the scene. Otherwise, with a linear falloff, the front part will be too much illuminated, as if the light comes from the chalkboard.
The view from the right shows that the lower and upper boundaries are near the floor and ceiling. I tried to manipulate the reverse gradient, but I stopped on a simple linear falloff.
To ensure the softness of the shadows, I had to increase the Quality to 2166. I manipulated the Bias = (-25) parameter to provide a balance of light from the floor, ceiling and walls.
Thus, most of the light enters the room from the side of the camera, from the ceiling and floor, which in life is not right, but in the form of a camera looks the most realistic.
I wanted to give a try to your indoor scene Rashad, too (thanks for sharing it). But to be honest I was confused by the huge amount of dome lights. Half an hour I tried to understand the sphere dome lights, which I feel now have no influence on the indoor scene here. But there are many cube dome lights which I could not figure out properly, which ones is used for what.
Therefore I had to make a big step back and simplify the lighting radically. I deleted all dome lights and added only one. Anyway for my matter of taste the original scene is a bit too bright. But my main point is the lacking of a fall off in brightness in the left part of the room, which is farther from the windows and from the light source. But maybe this is intended. Maybe the overhead lights do give light here and I missed that point. However here my example with one cube dome light. In my second example I tried to simulate the reflection of the floor and added a second cube dome light, that radiats upwards to give the ceiling and walls a little orange glow. My examples are far away from final versions (e.g. sometimes you can recognize any reflections of the light sources of the dome light) but it is interesting to play around and think about lighting.