Sorry for the constant rework of 5 or 6 scenes, I need this workflow for contrast the evolution of the camera and to fix a lot of things. More over I will proyect some things more... 40 min. 16 samples per pixel. If the value of the luxes to light the scene are too high there are some problems in my...

6.60 samples per pixel.
Insolit.

I discover why for the cathedral scene the layers mix bad. I used to set up the lights as surface dependency (lux), and it is better for some circumstances to set up in nits unit... 15 minutes. I prefer this approximations. It let me work faster. Maybe 10 hours in 8K is great almost always in black ...

I found a double exposition (it is called so?). It is due the position of the softboxes.

A recreative and a traditional development (better (more traditional), closer to output). 33:30 hours with the machine at 74%. View 3. Elevators and other minor things not present yet... (but closer). Mie cumulus noise more easy to ride than with indigo camera. Retro ilumination very hard if the sun...

Cloud shader working. Mie cumulus.

O-SS, view 3.
Recreative solutions with 28:30 hours.

A play, 24 hours (is dirty still). I introduce real tanulated spectrums for the glasses and coatings (invented) works very well. I introduce a milimeter thickness filter (dichroic coated glass from galinette). I hope to give some solutions (not perfect for sure). When the sun rounds near the cam, th...

The retro iluminated scenes has a very high contrast, this happens in real life though...
Or is part of the process...

Working in the cloud shader, without succes. Now it has an envelope dependet in normals of the mesh but i can´t organize it yet... 4 hours with the machine at 74%, and with the standard parametes of metropolis it get slow... Beauty ... This kind of expression///impression is ... reminds me another t...

PrötöStudies ClöÜds.

Henyey-Greenstein phase function. Is a startpoint for a shader that evanescence in the border of the polygons. Is a bit early to say, actually affect at all the clouds and not only in the border...

A bit "buggy". Seems that the best practice with metropolis and optics is to leave the standadrd parameters of the algorithm. Is fundamental a value of o.4 in the large mutation probability parameter. Is a lot noisy, but is the only way to acquire good lens flares. Look at this disaster. Lens flares...

Pse, pse...
I lost the position and setup of the lights...
30 hours.

Sometimes, aperture is present... In this case, in a half of hour process, seems to "homogenized with the incoming light to the image"... Is a non ortogonal ortographic view of a photographic perspective... PrötöStudies Cölëctör view 2.jpg Yes, aperture diffraction present but a little weak. Look at...

PrötöStudies Cölëctör.

7:15 hours, 600 spp... Sigma objective is really fast. 15 hours, 1000 spp... Canon is more complex.