Backlit Flag

[ARTECONI-CGI]

Hi All,

I am trying to render a flag backlit against a backplate and having a hard time trying to reproduce the translucent property of the cloth. It has to be detailed and photorealistic so no easy shortcuts will do.

I have already tried few approaches but none of them serve the purpose completely.

The flag is a single plane, turbosmoothed and zbrushed and I achieve a good quality with the 2 sided material. I am using the same texture for the front and back plane. The back one is just a bit more orangy and bright since it simulates the translucency given by the sunlight hitting it. The 2 maps are mixed with an opacity map derived from a cloth. All good so far if it wasn't for the fact that this flag folds, twists, bends and develops along the shot and disappear in the horizon.

The unnatural effect comes from the fact that I imagine this flag should become brighter and less saturated while it travels towards the sun. The truth is that it remains dark on some angles and doesn't seem to apply a logic in the way the translucency works. Often the front and the backside swaps because of the twisting of the cloth but the transluceny effect doesn't follow and make sense according to the light position.
I agree that the brightness and saturation might be addressed with some haze / volumetric effect but what I can't grasp with the 2 sided material is the flag getting hit by the light and transmitting / spreading along the cloth.

So I thought I could achieve the effect using either a vray material with refraction glossiness / translucency or a fastsss2.
In that case I will have to shell the material since the refractions takes the normals into account and behave differently depending on the twisting of the cloth. Fastsss2 also needs thickness.
I gave it a go but didn't have time to explore it completely. What bothers me is the fact that I still want to have an opacity map to make the cloth more or less see through and if I am using a shell on it it might be hard to see through the details.
On top of that render times with a shelled material with refraction glossiness set to .5 might be very slow. Maybe fastsss2 could do?

Has anyone dealt with something similar?

Thanks.

Giacomo.

[vlado]

Do you have a photo of what you want to get and where you think the material fails?

Best regards,
Vlado

[ARTECONI-CGI]

Hi Vlado,

Thanks for the reply.
I have actually solved the problem rendering in passes and comping it in post. It was the easiest solution.

Regards,

Giacomo.

[JezUK]

Be nice if you shared what you wanted to achieve and how you achieved it (in pictures)

[ARTECONI-CGI]

Hi Jez,

Thanks for the interest. I am not allowed to post any image. Anyway I figured out that I was trying to achieve too many effects at the same time and some of them were cancelling each other. I was trying to keep it translucent, transparent and reflective at the same time. At the end I have rendered out few iterations of the model with different lighting / reflection conditions that have been masked out in Photoshop later by the retouch artist.

Regards,

Giacomo.

[Art48]

Generally I am pretty sure VRay can achieve such a result at least physically accurate. When you want to do it unaccurate that is the only point where I could see VRay fail - but then again I don't know the look you were going for. Usually it's a user error (not saying it is in your case, just generally)

[ARTECONI-CGI]

Hi Art48,

Thanks, It's definitely a user error.
It was not my intention to be physically accurate.
I am using light cache and irradiance map that are both biased solutions.
I didn't have a hdri but was just trying to guess a similar vraysun position to match the backplate.
Also the 2 sided material is not physically correct but a good and fast solution to achieve translucency without having to wait for too long.
The Fastsss2 Material as well is not physically correct, isn't it? What vray does compared to other renders? It makes things happen. No matter what, physically or not, it delivers.

Giacomo.