Protoestudios

[yonosoy]

Hello burnin.

I agree with you, my way of proceeding has led me to always prefer a procedural workflow and to investigate ISL.

A complete synthesis must dispense with photographic textures.

For me, the fundamental thing is the use of an orthographic camera to leave the aspects of light, perspective, etc. to the camera model.

For example, coatings are 5 times thicker than they should be. The lens model is an approximation, spectrography is invented, (with Canon data, for example, etc...).

That is a good question.

I am a bit heavy-handed in claiming the term "synthetic photography" for this procedure. I think the difference lies in "synthesis image" versus "synthetic photography."

Physically, the process is photographic.
In fact, is an electronic synthesis of a physic photographic behaviour.
ETC, excuses...

Greetings.

[yonosoy]

Atmospheric simulation, instead of captured window or a HDR map (captured window is a special map (but fits better)).

The same goes with the IES file for the lighting, I prefer to model the lamp, etc...

You are right with the question...

I was made some test with clouds in atmosphere that agree completely with the term synthetic, but I didn´t process it like this serie...
Also the ALL test, lens flare test, ...

Synthetic because the electronic process, photography because the nature of the set-up.

For me Indigo is a light simulator (versus render engine) ...
Is this a "render"?
I feel it closer to a simulation or re-creation...

A recreative process!

When I was a child the video games explode. Recreative games.
This is an adult version of this excitment about to play...

[yonosoy]

Seems that I play all the day! Like a kid!

Also the term "synthetic photography" is claimed for another purpose actually...

[yonosoy]

And the last.
Near 300 hours in total. Is only a game about art...
From Kosuth to Mondrian in a recreation and deconstruction of a chair...

[yonosoy]

PrötöStudies (chäïr).

[yonosoy]

Hello burnin.
I'm thinking about your question...

About textures.
Displacement and roughness better than color. Complementary maps almost always come from the manipulation of color photography. These fit better.
The composition of different mix maps with photoshop brushes or other various possibilities...
We cannot abolish these resources because the possibilities of ISL are currently limited. But you have to be aware of not doing photo rendering. To fake the minimum possible, but keep it in mind.
Understanding that a completely procedural scene would be optimal, it is about "lying" as little as possible.

As for lighting, it is clear.
Atmospheric simulation with clouds (birds, what do I know, ...).
The Sun is a synthesis, a specific and scientific emission spectrum. The same as the atmosphere, a mixture of gases in a spherical vertical profile. The same as the mie cumulus possibilities for water vapor from clouds or fogs (a luxury exclusive to indigo). Currently the speed of this system is quite good.
For interior lighting, filaments. Gases are not possible, only the surface emits. Light with planes... I prefer to "naturalize" the situation and use, for example, a soft box with its textile materialized as a diffuse transmitter. Should we represent each strand? (YES!, one of these days... procedurally).

Atoms and electrons, physics and chemistry, color and sound!
Small spaces to start, the Earth, the Solar System, ...
Proportions, universal constants, ... instead of infite decimals.
Quantum to experiment,...
Obviously we are still in its infancy, although a good GPU can achieve a good image in 5 minutes that deceives the eye...

For example, it mixes the basis of Earth and lunar simulations (based on satellite maps) with a procedure to achieve the appropriate detail. This solution would be perfect, because it leaves open the possibility of procedural planets (widely researched in games) and includes the possibility of representing this planet while preserving the essential features that characterize it...

Synthetic is the obtaining of the document (what you represent does not exist), it is photography when it is simulated with a camera.

If you has performed any kind of tests over this area, I want to see it! If you don´t want to post it, send me a PM.

Greetings...

[yonosoy]

It must to be understand, I love kklors work, and is pure photo based rendering... there are more options. You can lie a lot and make sense of it. Fantasy requires a bit of it I guess. To lie sometimes has to be with imagination (this must be understand too).

Synthetic photography is a beautiful awaking area, pure investigation (art|science|technology|art)!

With this quality, it only could be achieved with Indigo (almost from my part, I don´t know another tool capable and I was tried some).

[yonosoy]

I feel that this difference is remarkable and deserve an own name...
If synthetic photography is not appropiate, sorry for the inconveniencies...

I need by the moment certain life based textures, but when there is a capable procedural texturer, I switch to it and feed the proyects with it. No repetitions and visible patterns, moire, ..., etc, ...

20 minutes (with metropolis bidirectional, done in a rush in a GPU). speed 0.476 M/samples second.
14 hours, speed 0.330 M/samples second.

[yonosoy]

In terms of efficiency, ...
In terms of satisfaction, ...

[yonosoy]

If this difference is appreciable with uniform absorption and cauchy, just imagine with the complete real tabulated spectrums of the lenses and coatings.

[yonosoy]

Sigma 56mm, F-stop 16, 8K, 43 hours, 1500 spp.
Aperture difraction of the silver seeds present.

I´m not agree with the human blood material, reflexions in the specular material are too weak and SSS is too dense (or the real tabulated absorption of the hemogobline is out of range...).

The displacement of the plate is wrong, want to change it for code that don´t need UV´s.
I feel that I could do it better with the table, actually is a displaced plane floating in an enclosed room... (this one is the "fault", fast investigation + a bit of lazy).
I want to define more precisely the soft boxes for ilumination.
Finally, I want to test some variations of composition and the Canon objective.

The text in the image... I know, but I like it.
Maybe I will make another shoot.

[yonosoy]

Other fight is the develop of the tonemapped EXR.

There are some LUT that camoflage the optic noise a lot, there are some beautiful ones.
But It is better for me to take the EXr from a linear profile and perform some minor things like curves, exposure, ... little things.

12 hours. Is quite good for me. Still will work a little more this situation (stones less rounded, lighting tests, maybe I test some singular modeled clouds, etc...).

20 hours.

[yonosoy]

These are the ilumination possibilities for clouds, atmosphere and mie cumulus that I talk about. No noise. SS 4. 6K.

Protostudies O-SS, View one.

Two accidental shutdowns of power in the PC.
20 hours.
30 hours.
Probably I will go deeper more over.

In version 4.6.0 and upper, the light direction of the sun is modify by the cloudbox, a little bug.

Downsized to 25%. Same promotion and development.

[yonosoy]

A study with singular clouds...
Near to launch the final serie.

In this case there are only the textures of the vegetation, and could be substitute for plane colors for this perspective.
Is a so little lie that I allow it. The fundamental thing is that there is no an "environment map" and there is no a "thin lens perspective camera" ruling the scene. This is a giant step for a synthesis...

Are you agree?

[yonosoy]

PrötöStudies O-SS.

Zenith 73º.
Canon EF 24 mm.
6K downsized to 3K.
35 hours.

Final render. Will fix more things in the nexts shoots...

Seems that if atmosphere has precedence 1 the bug is less present.

Black and white development SIlver Efex, Kodak Panatomic X.
Color development of the image reintegrated to darktable, except the final refinement of exposure (displacement) and text in photoshop.
The grain is perfect, no presence of any kind of noise.

In love...

View 1.

.
View 2.
28 hours.