OctaneRender™ 2020.1 RC1 [superseded by 2020.1 RC2]

[KeeWe]

Since no devs seems to care to answer in the specific thread, let me ask it here: are there any plans for headless rendering in the near future?

[rajib]

Thank you Otoy / Dev Team.

Finally my scene that always had a OOC memory issue since 2018.x is rendering under 2020.1 RC1. It has been a long wait.

2020.RC1 OOC memory usage.

Regards,
Rajib

[SSmolak]

Finally working NVLink for 2080ti is great but please don't stop working on it further to make memory pooling fully implemented and support it for textures too not only geometry.

[SSmolak]

Will be there any new updates to Hosek&Wilkie sky model like support for independent light intensity ? There was discussion about it some time ago.

[Kalua]

Thank you, Otoy for this pre production release! It feels nice so far.
One question: Is this <<Spectron area/sphere primitives>> feature implemented in every light material? Or is it additional to the standard lights used so far?
I ve been looking for it in the C4D plugin version but found nothing.

PS: I think I found it. Will it work with shapes other tah square and spheres? Disc for intance?

[jimho]

Another bug report:
A scene I am testing is not working with the network rendering for this 2020.1 RC1, if only rendering on the master, it works, but once turn the network rendering on, the slaves's Daemon will keep restart then master's standalone just crash.
I can run this same scene on 2019.1.5 very stable together with slaves, also quite suprisingly, it seems faster on the 2019 ithan the new 2020, on 2019 it is arround 6-9 Ms/s yet on 2020 it is about 3-7 Ms/s (when both without slaves).
with RTX OOC NVlinks all off...

within the scene I have ornatrix fur and Forest pack scatters and environment fog, it seems these three kind of things make the trouble, because if I turn the fog and the scatter off, the scene runs well with slaves.

[funk]

I'm doing some tests with the sun models and noticed there is a lot of noise in reflection/refraction on glass.

I can't find a way to clean it up either.

[karl]

Will be there any new updates to Hosek&Wilkie sky model like support for independent light intensity ? There was discussion about it some time ago.

Yes, there is a new "Sun intensity" parameter for daylight environment (this affects all models). This was just added today and will be in the next release.

[SSmolak]

Will be there any new updates to Hosek&Wilkie sky model like support for independent light intensity ? There was discussion about it some time ago.

Yes, there is a new "Sun intensity" parameter for daylight environment (this affects all models). This was just added today and will be in the next release.

Thank you very much !

[wallace]

I'm doing some tests with the sun models and noticed there is a lot of noise in reflection/refraction on glass.

I can't find a way to clean it up either.

glass_noise_sun.png

I had a look into this and it seems like it's a general path tracing problem, it's not really good at tracing indirect rays and hitting the sun.

This is because the sun is relatively small compared to the rest of the sky, and so your indirect ray would hit the bright sun with really low probability density, thus the contribution of such hit would be really really bright. This generates bright fireflies on rays that do hit the sun, and coupled with dispersion it generates extra color noises on top of it. The problem is worse for glass due to 0 or little roughness, so there is no direct light to help reduce the variance.

The short term solution would be to increase the sun size and perhaps use a different kernel such as PMC.