PDA

View Full Version : Why Can't I See My Glass?



steamthunk
02-28-2009, 08:29 PM
Hi. Trying to do my take on this old tachometer and I'm stuck on the glass. No matter what I do I can't seem to get the surface to render. In fact, if I turn the glass pane in front of the dial off it makes no difference in the result. It's like it wasn't there at all.

I'm using the dielectric node and the material looks fine in the node editor (1-3 absorption; 1.5 refraction). I have the Render options to render shadows, transparency, reflection, refraction occlusion, and using 2-sided area lights. The tach sits inside a white box with the normals flipped so they're facing inwards and the top is open to the "sky". The sky is the standard gradient backdrop. The camera sits just above the origin pointing straight down.

The glass is a thin subpatched disc that just sits on top of the dial face.

I feel like there's just some silly thing I'm missing but I can't figure out what.

toby
02-28-2009, 08:42 PM
The one thing that you're missing more than anything else is some dirt and smudges.

Your angle of view is so straight-on that you'll hardly see any reflection or opacity either, they'll be so even across the surface as to be undetectable. The amount of reflection at this angle is extremely low anyway, due to the fresnel curve of glass/dialectric.

colkai
03-01-2009, 02:13 AM
Lack of environment? How clouds / broken up is your sky?

rsfd
03-01-2009, 07:33 AM
Hi,
by looking at 90 degrees onto a (flat) glass poly, you will - as toby mentioned - not get much reflection (in your setup it is only the sky zenith color that will be reflected).
As I understand, you already gave a thickness to the glass disc, that is important for refraction effects, you could also give a slight bend to the glass, which would enlarge the sky area, which is reflected (so the glass "sees" a bit more of the sky gradient - not only the zenith).
Are the advanced reflection settings for the glass material set to ray traycing+backdrop with GI on?

steamthunk
03-01-2009, 07:45 AM
Ack... the problem is that a gradient backdrop ala Proton's dielectric video was insufficient for the glass and (I suspect) camera angle. I'm speculating that since the object faced straight up it would only see an even color and not have anything to reflect.

I added an abstract wallpaper as an Image World backdrop, removed the sides of my box, and got the smudging through reflection of the environment. :thumbsup:

My only other minor question concerns adaptive sampling (0.1) and anti-aliasing (3). Turning it on seems to really smooth out the speckled weathering look on the dial frame I really liked when I had AA=1. Without AA of course the edges are too jagged. What's the best way of handling this?

steamthunk
03-01-2009, 07:56 AM
Hi,
by looking at 90 degrees onto a (flat) glass poly, you will - as toby mentioned - not get much reflection (in your setup it is only the sky zenith color that will be reflected).
As I understand, you already gave a thickness to the glass disc, that is important for refraction effects, you could also give a slight bend to the glass, which would enlarge the sky area, which is reflected (so the glass "sees" a bit more of the sky gradient - not only the zenith).
Are the advanced reflection settings for the glass material set to ray traycing+backdrop with GI on?

Yeah the glass has volume. Interesting strategy of convexing the glass. I'll have to keep that in mind. I think my problem was that the glass being recessed into the frame and facing straight up wasn't seeing the zenith at all.

I didn't have the advanced reflection settings enabled with my previous post. I turned it on and monte carlo radiosity for this one. Seems pretty close although the extra light brings out the weathering a bit more. :)

rsfd
03-01-2009, 08:59 AM
I think my problem was that the glass being recessed into the frame and facing straight up wasn't seeing the zenith at all.

Well, I thought the tachometer lies flat on the floor facing the zenith.
If the camera is looking perpendicularly onto the (flat) glass, then it is the camera which is reflected (in real world) and everything just exactly behind the camera in 3D world (as the camera does not render). It's that incidence angle = reflection angle thing.

toby
03-01-2009, 02:59 PM
Ack... the problem is that a gradient backdrop ala Proton's dielectric video was insufficient for the glass and (I suspect) camera angle. I'm speculating that since the object faced straight up it would only see an even color and not have anything to reflect.
Exactly, it's reflecting one (almost) solid color, which wouldn't be noticeable



I added an abstract wallpaper as an Image World backdrop, removed the sides of my box, and got the smudging through reflection of the environment. :thumbsup:
Note that you must use a light probe (looks like a picture of a chrome ball) with Image World, not a spherical map, or it will be badly distorted.


My only other minor question concerns adaptive sampling (0.1) and anti-aliasing (3). Turning it on seems to really smooth out the speckled weathering look on the dial frame I really liked when I had AA=1. Without AA of course the edges are too jagged. What's the best way of handling this?
AA3 with adaptive 0.1 is not very much, if that's blurring your detail I'd guess that it's pretty faint. Try increasing it's contrast/strength. Which reconstruction filter are you using?