Intro
Am back into Octane for a week or so. Until the following happened:
I actually had a long conversation with a renowned photographer. On the one hand I asked for advice on photography and especially (studio) lightning. Simply since he is an expert in lightning. The conversation was grand. After like an hour or so (which felt like minutes) it turned out he never really tried 3D, but is heavily interested. He is sure it will play an even more important part in future in regards to professional photography - having a heavy impact on his work.
So it came all down to a studio scene we agreed upon. A studio scene he just shot: about a wine bottle.
Agreement
I create a realistic render of a wine bottle in a lightning he suggested. He sends his work afterwards to me and we compare and discuss the final results.
Status
Well...so I modeled a wine bottle: nothing too difficult until I came to materials... but more about that later on.
Project: wine bottle
scene: the scene is lit from the backside through translucent material diffusing the emitted light
addition: behind the camera is a diffuse light material to provide indirect light to the front of the bottle.
Materials V1
Glass and wine material V1: While the glass and wine material was easy (as well as the modelling after reading this forum and modifying recommendations to my needs for fluids in containers), I came to...
Paper label V1: My first shot was a translucent diffuse material.
Test render v1: This one went heavily wrong. Esspecially the indirect lightning killed the scene and it took forever to clear up. The wine, the glass and the label looked horrible.
Materials Vx
I spare you with my futile attempts...
Materials V16
Glass V16: The glass is a specular material with scattering and absorption with a small structure due to making the glass. The glass color is only due to absorption. Add some scatter and you got a glass material which isnt painted green, but completely tinted green.
Wine V16: Similar to the glass the wine has been setup to include all those particles in win, which provide a slick scatter effect.
Paper label V16: The label with a thickness of 0.12 mm (normal light paper thickness) is set up of three materials derived from one basic paper material.
- basic paper material: a translucent diffuse material with scatter and absorption tested in bright sunlight. Additionally a bump map for the paper surface as well as an absorption map to handle the "crushed stripes" which happen from production of cheap paper (you will see those against the light, where paper has been pressed shortly before being cut). Add a absorption modifier with a slight red tint and you are good for the inside material, as well as a light color for the diffuse surface. Basically I found: the realistic approach is best.
- background and sides: the basic paper material as mentioned before.
- front: this is made of two materials. First the black parts are actually a lacquer made of a glossy material. The rest of the label is tinted paper with the tint as diffuse. Add in another bump map for the glossy parts (the black writing and border) and you are good to go.[/list]
Bottom line on paper: lacque print on paper must be its own material, tinting paper isnt easy either or else it looks either wrong on the diffuse (eg. on table) or it looks wrong against light (absorption and scatter due to the paper and the tint).
Remarks
- To add realism, I also have slight "wrinkles" on the label due to gluing the label to the bottle.
- Indirect lightning is a KILLER, especially with light going through loads of layers of translucent and specular materials. This becomes worse when the main focus (the label) relies only on scattered light from the backside as well as backscattered light
- Render times... dont ask 22K samples minimum until the scene lights up. Probably need 50K samples for a clean render.
- Loads of Postpro to getrid of the noise.
- "A trick" with a diffuse scatter emitter in front with like 5W to "cheat" the long render times. Turns out I dont need it past 30K samples but I know that only since today. Starting at about 15K samples the scene lit up seems it took that long for RR=auto (with increased exploration) to find those.
- The effects especially on the paper are almost invisible. Try something similar and you realize the issues.
- In a good mood I might make a vid with "realistic" paper in bright sun light; maybe even of the bottle. This only to show the paper to light interaction, which looks pretty cool in bright sunlight.
Now I am curious....
I just sent the E-Mail to the photographer and leave it to you to discuss my result for now. Please be frank and open - dont hold back with criticism. Did I overestimate myself? Or the possibilities of rendering/Octane? Maybe both? Materials or anything else went wrong? Be frank and keep in mind its meant to be photorealistic.