OctaneRender® 1.025 beta2.57 release [obsolete]

[Daniel79]

Thanks for this update but, what is lacking for the instances?

Great

[ROUBAL]

@Rico : The Deep (depth channel also named Z buffer) is often used in compositing to add some mist in a scene, or to drive the DOF, or modify the brightness or color according to distance, simulating atmospheric damping effects.

[kivig]

Pass is very cool. But for full happy need specular and diffuse (:

This is not practical with unbiased render engines, and is not needed, provided you update your workflow...
I have had contact with quite a few small to medium studios who are using unbiased rendering with new workflows which depend on making the materials physically correct in the first place, so post production mixing of specular, diffuse, glossy, etc... Is not required anymore.

Radiance

Photographs too look physically correct, but we do use Photoshop 99.99% of the time to make them.
Also, not mentioning that masking of reflection portion is often impossible in 3D, it is impractical to re-render whole sequence for slight alteration of reflection opposed to doubling rendering time with each pass.

I often try to convince my art director to do whole job in single 3D scene, but in the end it is my laziness that speaks and post-produced result always ends-up better and much more efficient.

Edit: I do understand, that Octane has no such thing as "Specular" and reflection might be impossible to separate from diffuse, given the amount of interactions. But there might be some tools, like special materials, to help the artist to at least simulate those passes artificially. And there are options, like material that ties out alpha to it's brightness etc. Or making the whole scene "multy-layered", where user has more than one workspace to assign materials to the same scene and can switch between them without making separate projects (one scene could fake diffuse pass, etc.).

Those might be extreme solutions, but are no more than just examples. In the end the artist cares about the result, not if the pass was rendered with button that actually said "reflection" or faked around.

[mbetke]

Maxwell Renderer also features full Layer support for every tiny bit you can alter in Photoshop including a very streamlined workflow with After Effects and other tools.
So its definitly possible to include. Octane is a very young product and did a great progress. I'm a very inpatient person but learned that codings seems to take its time hehe.

For me I always was one of the users whoch do most stuff in 3D and render out a frame to get things looking right. But things like zdepth are very handy now because for compositing its essential if you want to do some things the easy way.

[Faizol]

Maxwell Renderer also features full Layer support for every tiny bit you can alter in Photoshop including a very streamlined workflow with After Effects and other tools.
So its definitly possible to include. Octane is a very young product and did a great progress. I'm a very inpatient person but learned that codings seems to take its time hehe.

For me I always was one of the users whoch do most stuff in 3D and render out a frame to get things looking right. But things like zdepth are very handy now because for compositing its essential if you want to do some things the easy way.

Fully agree wrt to render layers. Arnold does that. Mental Ray has its own form of "unbiased mode" that also fully supports render passes. This is definitely a must if Octane is to be a world class production renderer.

[ribrahomedesign]

thanks Roubal for explanation , would need to see some finals or tutorials to understand how to use it
Rico

[mjsarfatti]

great update this one!

just two things:

1. very unpractical to set the correct z-depth lightness the current way (we have to go all the way to the imager, thus altering the exposure of the real render), a slider right in there would be the most common-sense thing
2. it would be great to be able to save the channles automatically as layers of a psd file or something

and a question: why do you say we should save the image as an untonemapped EXR? i tried saving it as png and it worked just fine.

[jamnique]

great update this one!

just two things:

1. very unpractical to set the correct z-depth lightness the current way (we have to go all the way to the imager, thus altering the exposure of the real render), a slider right in there would be the most common-sense thing

How about using the focus picker?

2. it would be great to be able to save the channles automatically as layers of a psd file or something

and a question: why do you say we should save the image as an untonemapped EXR? i tried saving it as png and it worked just fine.

Probably because EXR is 32-bit ?

[roeland]

great update this one!

just two things:

1. very unpractical to set the correct z-depth lightness the current way (we have to go all the way to the imager, thus altering the exposure of the real render), a slider right in there would be the most common-sense thing
2. it would be great to be able to save the channles automatically as layers of a psd file or something

and a question: why do you say we should save the image as an untonemapped EXR? i tried saving it as png and it worked just fine.

You can save the channels as PNG, but then you lose a lot of precision. This could be a problem for the position and z-depth, because these channels can have a big range of values (including negative values for position). To display these channels the position channel is scaled and shifted to fit between 0.0 and 1.0, and the Z-depth is scaled according to the exposure value of the tonemapper, and may be clipped if this value is too large.

If you save as PNG, you get these tonemapped values. Also in 32-bit PNGs having only 8-bit values may not be precise enough for some.

If you saved untonemapped EXR you get the original values, for instance if you save the position channels and the visible point is (-5, 2, 10), then these values will be saved in the EXR file, rather than some values between 0.0 and 1.0 , and if you save the z-depth channel there is no clipping.

For the material ID and the normals saving as a regular PNG file is probably no problem.

--
Roeland

[vipvip]

great improvments but a agree with mjsarfatti suggestions: zdepht should be driven differently and a multilayered PSD outpout would be so great !...