View Full Version : Caustics + Shadows
slimpy 06-11-2002, 09:36 PM Here is a simple simulation of a vitamin-E pill. The accuracy of the caustics is actually quite amazing. However, there is a problem with the way Cinema 4D calculates the shadows...
The real pill casts a solid dark shadow - with the caustic spot at the center. If I understand the physics correctly, this occurs because the objects refraction bends and concentrates the all of the light at one point.
As you can see, the Cinema 4D version doesn't do this and seems to calculate the transparent shadows normally. Any suggestions?
http://www.eyewinder.com/temp/pill.jpg
|
|
ThirdEye
06-11-2002, 09:53 PM
try unchecking the transparency of shadows in the light-shadow settings... and then if it's too dark use the opacity regulation for shadows ;)
slimpy
06-12-2002, 12:21 AM
Cool, thanks! I guess there could be potential problems if other objects like flat panes of glass are in the scene (which would require transparent shadows). But it's a good start...
Hopefully, Maxon will add exclusion/inclusion settings for lights & objects in XL 8.
slimpy
LucentDreams
06-12-2002, 02:53 PM
see this is where I still have a problem with this exclusion ting,how would that help here. if you exclude the light your not going to have the caustics either. I thinnk part of the issue the overall thickness of the material. basically, when I tested some ottles here, the colour in the shadow is definitely less then in a typical render, but if you use bhodinut fresnel, it takes a lot more of the colour out from the outer edges of the shadow. as well, what is the colour of your lightsource in the rel world test. I can guarantee it ain't white. thsi may sove some issues once you colour it a bit blue. and add a little blue to the shadow too.
slimpy
06-12-2002, 04:02 PM
Here is a new rendering with shadow transparency turned off and shadows reduced to 70% opacity. I think the shadow & caustics look much better.
http://www.eyewinder.com/temp/pillnew.jpg
That looks much better. You can't have more light going in the pill than is coming out.
To Kai: Light exclusion would help because you could have one light for the pill, which only the pill sees, and another for the rest of the scene, which the pill doesn't see.
Jeff Carns
06-12-2002, 04:50 PM
After looking at your image I really thought the caustic effect was way too strong, so I went over and pulled one out of the bottle and put it on the counter. If anything the real world pill has MORE caustic effect and less shadow (depending on the lighting of course) than your image. The shadow appeared more as a small ovoid ring around the large caustic effect. Also the particular brand I was looking at appeared clearer and more on the yellow side.
Not sure where this is going. Just wanted to add some real world comparisons and observations. Good work though.
Jeff
Caravaggio
06-12-2002, 05:07 PM
After pulling out a pill myself I agree the 2nd one loks better as far as sunlight goes. But I'm a little confuseed by MJV's statement about more light coming out than going in. Is this in the concentrated part or what?
slimpy
06-12-2002, 05:12 PM
Kaiskai,
I don't believe that SLA fresnel can modify the way C4D calculates it's shadows. C4D still does a straight projection from the light source - through the object. When SLA fresnel is added, the only result I've ever had is a shadow with a darker ring around it's perimeter. If you can give me some suggestions for fresnel settings that prove my theory wrong, I'd be very greatful.
In support of the inclusion/exclusion, I created a quick example where turning off transparency (my best solution so far) can cause problems. I added a bluish light & shadows as you suggested. The ball creates a pretty realistic shadow & caustics, but the thin sheet (which has depth BTW) should let most of the light pass through. If I could restrict the current light to just the ball, I could create another light (with transparency on) and restrict it to only the sheet.
Saying all of that, I understand your point. I shouldn't have to turn off transparency, play with exclusions etc. to achieve a realistic image. But IMHO, the problem lies in the fact that 3d renderers take so many shortcuts in emulating real world situations - we have no choice but to take bizarre & unrealistic countermeasures.
slimpy
http://www.eyewinder.com/temp/ballandglass.jpg
My point was that you can't have more light coming out one side of the object than is going in. The caustic represents the light going in that is focused coming out, so other than the caustic, there should be no other light coming out. Therefore, if this scene had for example only one light source, the shadow of the pill would be almost black, even darker than it is now. Removing the color helped the illusion, but the only surefire way I've discovered to do such an object is to use a second invisible (invisible to the camera using a tag) object which casts more shadow. In this way you can have a darker shadow, or a shadow with a strong falloff on the edges, using fresnel, or one that is completely black.
slimpy
06-12-2002, 05:28 PM
Jeff,
I'd have to agree with your observations. Perhaps the thin gelatin coating on the real pill (which I didn't bother to model) causes the caustic to spread. Maybe I'll try a radiosity version with more transparency & a lighter colour...
LucentDreams
06-13-2002, 04:30 AM
what I meant by using fresnel, is that it would take some more colour out of the render. Basically allowing more black (or whatever colour your shadow setting has) to show than the colour. but I do find from my realworld test with a bottle (maybe the botle makes the difference) is that you should have a little colour in the shadow anyways.
As for your explanation of Light exclusion, I am catching on. I still will personally avoid this type of thing, but I can see where it can amke a difference.
slimpy
06-13-2002, 05:45 PM
Kaiskai,
IMHO, the bottle would be a significantly different case. It has a much thicker shell material, and it's contents can have widely varying refraction indexes. In any case, SLA fresnel should definitely be used ;)
Here is my final rendering with colour & transparency adjustments, seam bump added, and a bit better lighting. I tried to do a radiosity version of it, but had to call it off because the area shadows & caustics were already killing my machine :hmm:
There is still the problem with the caustics being too focused... Anyone have ideas about how to enlarge & blur them?
http://www.eyewinder.com/temp/pill_final.jpg
LucentDreams
06-14-2002, 07:41 AM
lloks good, as for the caustics, only played with them when I first got 7, haven'y used them since. Sorry
Jeff Carns
06-14-2002, 03:25 PM
What settings did you use for the caustics?
slimpy
06-14-2002, 06:47 PM
Jeff,
Here are my general settings. The spotlight mentioned below doesn't actually produce any visible light (there is a negative version of it in the scene), it just creates the caustics & shadows. An omni light is used to illuminate the rest of the scene.
BTW, the last version I posted took about 1.25 hours to render - Ouch!
Spotlight:
- Area Shadows (Transparency Off, 75m Width, 90% Samples)
- Surface Caustics (100% Energy, 500000 Photons)
Material:
- Caustics (100% Generate, 100% Receive, 25m Sample Radius, 100 Samples)
Render Dialog:
Surface Caustics (150% Strength)
CGTalk Moderation
01-13-2006, 08:00 AM
This thread has been automatically closed as it remained inactive for 12 months. If you wish to continue the discussion, please create a new thread in the appropriate forum.
vBulletin v3.0.5, Copyright ©2000-2012, Jelsoft Enterprises Ltd.