OctaneRender® for 3ds max® v1.0 beta 3.03[Test]

[MaTtY631990]

There seems to be bug when renaming glossy and specular materials from their index no. of the material. Reopening file causes it to reset.

[Karba]

Octane cam is not shaking anymore. So that is fixed. Thank you.

But cameras still not working properly!
Now camera target distance is not updating itself when rendering out animations. (when using target as a focal point).
It works in Octane view, but with standard "Render" button it doesn't. Seems like it reads the target distance at the frame "Render" was pressed, but doesn't update it for next frames? When camera moves away from the target everything gets blurry. That happens with both - Octane camera and standard cameras.
Could you try to fix it asap? Could be an easy fix cause it's working Ok in Octane view.

Suv

Fixed

Please redownload.

[boris]

not sure if this is the right place to post, but i rendered a test animation with octane max 3.02 and stumbled over this by chance. i am not sure if this is right DOF behavior.
the cube in the foreground is out of focus but another cube in the reflection seems to be focused right. i am to lazy to make a real world setup for that so maybe some experienced photographer can confirm that the resulting image must look like that?
(i cropped the frames to the interesting area:)

only for clarity here a frame further in the timeline:

cheers
boris

[abstrax]

not sure if this is the right place to post, but i rendered a test animation with octane max 3.02 and stumbled over this by chance. i am not sure if this is right DOF behavior.
the cube in the foreground is out of focus but another cube in the reflection seems to be focused right. i am to lazy to make a real world setup for that so maybe some experienced photographer can confirm that the resulting image must look like that?
(i cropped the frames to the interesting area:)
...
cheers
boris

Yes, it is the correct behaviour. You ca try it out with a mirror: When you focus on an object visible in a mirror, you are focussing further then the mirror itself.

Cheers,
Marcus

[gabrielefx]

another issue:

distant camera now creates rayepsilon issues.

I often tweak the viewport (Nitrous) clipping to avoid triangles degeneration when the camera is too much distant from the object (or vice-versa). I noticed that in 3.03 I have to tweak differently the viewport clipping (I'm not speaking about the render camera clip).

The 2.58f is much faster (don't mind if the zoom is slight different in the two tests) and doesn't show any triangles degeneration.

The 3.03 is much more slower (about 1.5X) and shows corrupted triangles. Tweaking rayepsilon doesn't help.

The camera is far away about 28.000 millimeters to simulate an isometric view. For mechanical visualizations (no massive use of instances) 2.58f is the only version that works.

[abstrax]

another issue:

distant camera now creates rayepsilon issues.

I often tweak the viewport (Nitrous) clipping to avoid triangles degeneration when the camera is too much distant from the object (or vice-versa). I noticed that in 3.03 I have to tweak differently the viewport clipping (I'm not speaking about the render camera clip).

The 2.58f is much faster (don't mind if the zoom is slight different in the two tests) and doesn't show any triangles degeneration.

The 3.03 is much more slower (about 1.5X) and shows corrupted triangles. Tweaking rayepsilon doesn't help.

The camera is far away about 28.000 millimeters to simulate an isometric view. For mechanical visualizations (no massive use of instances) 2.58f is the only version that works.

Roeland just did some work in that area. Could you send us (Andrey/Karba) the scene to allow us verify that these kind of problems are solved?

The speed difference is indeed quite high. Would be interesting to see how much the difference is compared to 3.01.

Thanks,
Marcus

[suvakas]

Fixed

Please redownload.

Awesome !
It works now. Thanks a lot for the fast fix !

Suv

[abstrax]

another issue:

distant camera now creates rayepsilon issues.

I often tweak the viewport (Nitrous) clipping to avoid triangles degeneration when the camera is too much distant from the object (or vice-versa). I noticed that in 3.03 I have to tweak differently the viewport clipping (I'm not speaking about the render camera clip).

The 2.58f is much faster (don't mind if the zoom is slight different in the two tests) and doesn't show any triangles degeneration.

The 3.03 is much more slower (about 1.5X) and shows corrupted triangles. Tweaking rayepsilon doesn't help.

The camera is far away about 28.000 millimeters to simulate an isometric view. For mechanical visualizations (no massive use of instances) 2.58f is the only version that works.

Ok, we couldn't reproduce the slow down here. We tried it with a GTX 590 and with your scene there is a slow down of ~7% comparing beta 2.58f with beta 3.03. When we use a GTX 590 and GTX 690, we already get ~160Ms/s on beta 3.03, which is more than your 4 GTX 580. So something is fishy with your set up. Did you use the graphics cards for anything else?

Cheers,
Marcus

[Karba]

2 gabrielefx
Some texture is missed in your archive. Can you send it to us? So we will be sure we have exactly the same scene.

[merid888]

thank you marcus, i'm download right now and using into an interior scene
regards