OctaneRender™ 2021.1.2

[abstrax]

I have a question about a feature request I had for a while.
Is there any news on references (livegroups) of shaders? It would be great to have the option to live link to an shader on disk. So if a shader has to be changed in an project you can just publish the shader and all your files are updated.
At the moment we have to edit all the ocs files by hand. Is this something that is on the roadmap?

I also posted this in the request area:
https://render.otoy.com/requests/?qa=64 ... -ocs-files

let me know

Unfortunately there are no news regarding that. We (/I) definitely would like to see support for links to external assets (/ x-refs) in the future. Just had no time to start working on that.

[karl]

I think that something wrong is with Sun Intensity in Hosek Wilkie Model. As I remember in the first implementation there was possibility to completly turn it off for simulate ambient, cloudy lighting. Now only its shadow dissapear but Sun lighting is still visible and more noticeable. For example in Octane daylight mode there is way to completly turn it off.

Can you share an example scene that exhibits this problem? Everything seems to work fine in my tests. When I set sun intensity to zero the hard shadows disappear, the lighting gets much dimmer and more blue. The results from the Octane model and the Hosek-Wilkie model are approximately the same.

[john_otoy]

I think that something wrong is with Sun Intensity in Hosek Wilkie Model. As I remember in the first implementation there was possibility to completly turn it off for simulate ambient, cloudy lighting. Now only its shadow dissapear but Sun lighting is still visible and more noticeable. For example in Octane daylight mode there is way to completly turn it off.

Hi SSmolak,

I also tried comparing 2020.1.5 to 2021.1.2 and can't see any difference between the way sun intensity works. Sun intensity is controlling the sun strength as expected. Images and sample file attached.

Can you provide a simple example which works in an older Octane build, but not in 2021.1.2?

[SSmolak]

Please see. Increasing Turbidity will increase Sun intensity even if it is "disabled"

[aoktar]

Please see. Increasing Turbidity will increase Sun intensity even if it is "disabled"

There's not any issues on export parameters. Follow your issue on standalone topic.

[karl]

Please see. Increasing Turbidity will increase Sun intensity even if it is "disabled"

This is intended behavior and seems to work more or less the same between the Octane and Hosek-Wilkie models. When turbidity is high there is simply a lot of light coming from the sky near the direction of the sun (not from the sun itself).

The illumination from the daylight model is made up of light directly from the sun, and light from the sky. The turbidity controls the haziness of the sky - when it is set to a higher value, less light comes directly from the sun and more light comes indirectly via the sky as it is scattered by the haze. On a completely overcast day, 100% of the daylight illumination would be from the sky, because you can't even see the sun directly (the daylight models in Octane don't support turbidities that high though). In that theoretical case, changing sun intensity would do nothing.

If the sun intensity slider were to also control light from the sun that was scattered by the sky, it would simply scale the entire output of the daylight model (which is what the power slider already does). It is to be expected that scaling down the direct sunlight has less effect when less of the overall light is coming directly from the sun.

I think that the solution to your use case would be if Octane supported much, much larger values for turbidity (so you could actually make the sky overcast). That would give much more accurate results than using lower turbidities and scaling down the sun intensity (which doesn't make much physical sense, since if the sun is dimmer the sky should also be dimmer). Unfortunately, the mathematical models that the likes of Hosek and Wilkie came up with only model a narrow range of turbidities (in fact, the slider goes to 15 but the Hosek-Wilkie model actually stops at 10). We would need something like 50 (which apparently represents dense fog) but the Hosek-Wilkie model does not provide data for this. To quote their research paper:

We did not go beyond a turbidity of 10 mainly because the layering of very hazy atmospheres starts to be rather specific after some point: inversion layers, which would require additional parameters to control, can start to play a visually prominent role. We therefore restricted ourselves to a scenario of a gradual increase in haze up to moderate haze levels.

So, if the Octane daylight model works better for you, please use that There is a 2021 successor to the Hosek-Wilkie model (called Wilkie-Vevoda) but it doesn't appear to provide a wider range of turbidities.

[SSmolak]

The turbidity controls the haziness of the sky - when it is set to a higher value, less light comes directly from the sun and more light comes indirectly via the sky as it is scattered by the haze.

Thank you for great explanation. But Turbidity should works exactly as you described above and how it works in Octane Daylight Model. In Hosek-Wilkie it works inverted when sunlight is on 0% Probably I know why and this is limitation of Hosek-Wilkie which was developed for sunlight scenario where increasing Turbidity add Sun halo, glow and intelligently desaturate overall sky. Anyway Octane Model works great for ambient, cloudly scenarios. And even better because we have control over sky color. This can be simulated in Hosek-Wilkie if Sun direction is exactlty -90% ( directly to the ground )

[karl]

The turbidity controls the haziness of the sky - when it is set to a higher value, less light comes directly from the sun and more light comes indirectly via the sky as it is scattered by the haze.

Thank you for great explanation. But Turbidity should works exactly as you described above and how it works in Octane Daylight Model. In Hosek-Wilkie it works inverted when sunlight is on 0% Probably I know why and this is limitation of Hosek-Wilkie which was developed for sunlight scenario where increasing Turbidity add Sun halo, glow and intelligently desaturate overall sky. Anyway Octane Model works great for ambient, cloudly scenarios. And even better because we have control over sky color. This can be simulated in Hosek-Wilkie if Sun direction is exactlty -90% ( directly to the ground )

Yes, the issue in this case is that even though the light is coming from the sky, it's mostly coming from a small region of the sky near the direction of the sun, so it appears similar to if the light had been coming directly from the sun. This halo/glow is a little more prominent (which is probably more physically accurate) in the Hosek-Wilkie model than the Octane model.

[SSmolak]

As I know Vray introduced new sky model that enhanced sunset, dawn, night skies by moving sun or moon under horizon. I don't know if this is the new Wilkie-Vevoda - https://youtu.be/oEkJckzSDyY?t=58

Also in Corona : https://youtu.be/hCGNwCjn624?t=55

[john_otoy]

OctaneRender 2021.1.3 has been released: viewtopic.php?f=24&t=79371