OctaneRender™ Standalone 2.11.1

[abstrax]

Abstrax,
My scripts directory is properly set. Still, Octane crashes when the above mentioned scripts are closed or exited. I even tried the Octane Windows 32 bit version and it does the same thing. These crashes also occurred in version 2.10.

Ok, thanks for the feedback. We will look into it tomorrow.

[geo_n]

http://render.otoy.com/features.php
Object Visibility Options (On/Off for shadow casting and camera visibility)

Does this work on standalone?
It doesn't work in lightwave plugin.

These two rectangle shadow should not be there. They're reflection cards with cast shadow off.
They're showing up in the render and alpha. The ground plane is matt shadow material






Seems it is related to this
http://render.otoy.com/forum/viewtopic. ... ty#p201250

"The shadow visibility doesn't work, if checked or unchecked the shadow is always visible for the visible or invisible object."

[maxuser2]

Guys, viewport autoupdating problems still here. It's really annoying to work with hires (6000x6000 and more). Octane keeps updating vieport even when minimized in task bar which takes a lot of PCI-E traffic/delay on some systems. I got this problem on 2.10 and it's still here. It makes microfreezes in system (sound, UI - all) when, for example, 8192x6144x24(?)bitxRGB = 432(!)MB goes from gpu to system memory every couple of seconds. It makes no sence when window is minimized. Until 2.10 only active (maximized) octane window kept updating and system was very comfortable to do any kind of work with minimized window.
I think this problem could be even more significant on multi-gpu systems, especially with low-speed 1x-4x PCI-E raisers.

[funk]

I found a bug with the EXR saver. Tonemapped EXRs are being save with the wrong gamma (washed out).


1. Start a newproject
2. Add a rendertarget
3. Change environment texture color to (0.5, 0.5, 1.0)
4. Save a tonemapped EXR

The EXR is saved with the wrong gamma (it seems to be burning in a 2.2 gamma so the image looks washed out).

[Zay]

When using multible obj's connected to a geometry group, sometimes when refreshing one of the obj's will hang Octane when you click on the Geometry Group. It hangs when the scene is compiling and the render don't start. Compiling the scene never finishes. To fix it I have the disconnect one of the OBJ's to the Geometry Group and then reconnect it again to get the rendering started.
Unfortunately this seems random so might be difficult to bug hunt.

[stratified]

I found a bug with the EXR saver. Tonemapped EXRs are being save with the wrong gamma (washed out).


1. Start a newproject
2. Add a rendertarget
3. Change environment texture color to (0.5, 0.5, 1.0)
4. Save a tonemapped EXR

The EXR is saved with the wrong gamma (it seems to be burning in a 2.2 gamma so the image looks washed out).

Hi Funk,

I cannot reproduce that issue here. The images look the same in my viewer and in Octane:

OctaneRender and my viewer look the same

Can you post a screenshot of the image while you're viewing it and the .exr file in question? What program are you using to display .exr images? Could it be that the viewer itself is applying gamma correction (i.e. correct to 2.2 - default display gamma)?

cheers,
Thomas

[funk]

Can you post a screenshot of the image while you're viewing it and the .exr file in question? What program are you using to display .exr images? Could it be that the viewer itself is applying gamma correction (i.e. correct to 2.2 - default display gamma)?

I'm using 3 different programs to view the EXR: Photoshop, XnView and djview. All 3 display the tonemapped PNG correctly, but the tonemapped EXR doesnt match.

As you mentioned, the viewers ARE applying a display gamma of 2.2 (or sRGB in photoshops case) but this is correct behavior, because EXRs should be in LINEAR space.

Here is a screenshot. I have also attached a zip containing the OCS and the saved tonemapped PNG/EXR files.

[funk]

Here is a better test. I have set the octane camera response to LINEAR and the gamma to 2.2 so we can compare the tonemapped and untonemapped images.

I have attached a screenshot. The project files and saved images are also in the zip.

As you can see, the untonemapped EXR is being saved linear (1.0) so photoshop and djview (also xnview which I didnt add here) display it correctly. The tonemapped EXR is being saved with a 2.2 gamma burned in, so it displays washed out.

EDIT: I have also tested this in After Effects CS5.5 and have exactly the same results. The tonemapped EXRs appear washed out, while the untonemapped EXR and PNG are correct (screenshot added)

[abstrax]

Here is a better test. I have set the octane camera response to LINEAR and the gamma to 2.2 so we can compare the tonemapped and untonemapped images.

I have attached a screenshot. The project files and saved images are also in the zip.

As you can see, the untonemapped EXR is being saved linear (1.0) so photoshop and djview (also xnview which I didnt add here) display it correctly. The tonemapped EXR is being saved with a 2.2 gamma burned in, so it displays washed out.

EDIT: I have also tested this in After Effects CS5.5 with linear workflow set up and have exactly the same results. The tonemapped EXRs appear washed out, while the untonemapped EXR and PNG are correct.

Yes, that's correct: If you save an image tonemapped, well, then it's tonemapped, i.e. has the camera response curve applied to it. If you select a camera response curve other than "Linear/Off" the result will be non-linear. Most response curves have a gamma of ~2.2. I tried to remove that "implicit gamma" a while ago, but got blasted for it and quickly reverted the change (see http://render.otoy.com/forum/viewtopic.php?f=24&t=41404).

EDIT: The problem is that all HDR viewers assume that EXR images are linear and apply a tonemapping to sRGB on top of the already tonemapped result we saved which is wrong of course, but the viewer has no way to find out if an EXR image is already tonemapped or not.

[funk]

EDIT: The problem is that all HDR viewers assume that EXR images are linear and apply a tonemapping to sRGB

Thats correct. Every application I have tried assumes the EXR is linear. Photoshop, After effects, djv view, xnview

EDIT: I also just tested NUKE (personal learning edition) and got the same results.

So, in conclusion if the major compositors (AE and NUKE) assume linear EXRs, I think this should be fixed in Octane (only tonemapped EXRs need to be fixed).