Vray DOF Effect

[phizikl]

Any chance we can see a vray dof effect in the future? The incamera usually takes very long, and cebas final dof really isnt cutting it either. Using a zdepth map and a post package effect is OK, but what about integrating a post internal vray effect, which uses 3d information AFTER render has been completed.

Quicksilver does DOF pretty much instantly using your gpu. Isnt there any way to get a quick Great looking V-Ray DOF?

[vlado]

Hm, the QuickSilver DOF looks great to you? In any case, most of the current compositing applications have excellent DOF utilities, which are vastly superior to anything we can code. I would prefer that we spend time on the raytracer itself.

Best regards,
Vlado

[instinct]

Hm, the QuickSilver DOF looks great to you? In any case, most of the current compositing applications have excellent DOF utilities, which are vastly superior to anything we can code. I would prefer that we spend time on the raytracer itself.

Best regards,
Vlado

Quoted for agreement!

[phizikl]

well, with great i meant, considering it took 1 second to render that image with DOF, yes its f*kin greaT!

[3DMK]

Apparently the DOF in 3rd & the Seventh was all done in post with After effects plugin "LENSCARE".

Now if that doesn't show you how good DOF can look in post!!!!......


Yes VRAY DOF is the best [always has been..always will be] BUT it takes a lot of time & then you have to consider rendering thousands of frames in an animation.

Unless you have a massive render farm or budget then DOF in post is almost compulsory for animations.

Cheers

[phizikl]

jea i know, there are good post DOF effects, and vray dof is the best
i was just mentioning my wish, for a QUICK DOF inside max, one that doenst calculate it a thousand times, but just blurs it accordingly, the two options right now are very slow inside max, or render out an comp together, both take longer than having a quick preview,

even the PREVIEW button of the standard max cam already does what im looking for, only that it does it in the viewport, and not in the render,

my wish is to have that effect after the render, it takes a few seconds to calculate at the most, and already gives good impression of how dof could enhance the image

[3DMK]

Yes...you do have a point!!....It would actually be a nice feature!

Especially if it's based on the physically correct Vray PhysCam. You could then use it as a guild for how much DOF to do in post.

Good idea!

[phizikl]

wooohoooooo!


glad someone understands me ^_^

[tammo]

Setting up good looking DOF in post is not easy.

You have to setup and render a much larger additional image sequence with the z grayscale depth information.
To get a decent quality you have to double your targeted output resolution and turn OFF filtering (anti-aliasing)
otherwise the edges of the ojects have wrong depth and look bad.

A quick workaround would be to include a "double resolution" button in the render elements rollout.
Then you would be able to get a good z-path in one render-job.

[phizikl]

render large image sequence, and double resolution?

never heard that one before, but whatever it is you mean, it is +1 for a good quick DOF effect inside max ^_^

(and i dont think the double resolution button will save you much time, when you have to render 3840x2440 pass but no seriously i have never heard of rendering double resolution for depth pass)

and this is only to aliase your image? i dont get that process,


i can do post blur with normal resolution its not a problem, but its hard to have a middle sharp point and front blur and back blur, i havent been able to do that with one z-depth pass....


would be great if vray offered a way to blur out the noise (fast) DOF from the physical cam, then you could render just quickly your dof, and z-blur it out quickly, to get a feel for what the final image would look like

[tammo]

Here is a detailed explanation about the AA problems with post DOF using Z-depth grayscale images:

"The restriction is that the depth buffer cannot be anti-aliased.
If the depth buffer were anti-aliased,
any gray values in between the black and white areas would indicate wrong depth values and would thus result in artifacts.
Conversely, when there is no anti-aliasing on the depth buffer
but there is anti-aliasing on the image,
then those two images don’t match exactly,
which results in more or less visible artifacts.
One way to deal with the second problem is to render in doubled
resolution, apply Depth Of Field and then resize back to normal resolution.
In order to reduce the rendering time for the raytracing,
it is possible to decrease anti-aliasing on the image by the same factor the resolution has been increased.
For example, if 16 times oversampling is enabled with a doubled resolution only 4 times oversampling is needed.
If you already rendered out your image in normal resolution it is acceptable to
increase image size in After Effects and render out only the depth map in doubled
resolution.
Then use that z-buffer to apply Depth Of Field on the bigger resized image."

(From the lenscare manual)

In addition to those problems there is also the missing pixel-data behind the foreground objects
that would be needed in order to generate a correct Depth Of Field effect.

A "Vray Quick-DOF" could get such data from the 3d scene but no post-DOF could get around that problem.


The question is if such a quick-DOF would really be much faster then the real DOF because it had to produce so much extra data.

[Neilg]

even the PREVIEW button of the standard max cam already does what im looking for, only that it does it in the viewport, and not in the render,

You cant do this to a render - it actually moves the camera. To get this effect vray would have to render your scene 12 times or more.

[phizikl]

You cant do this to a render - it actually moves the camera. To get this effect vray would have to render your scene 12 times or more.

jea i thoguht about that as well,
but we do have the 3d information

what im asking for is actuall this:

can chaosgroup come up with a better solution than cebas did with finalDOF

in max we can still work with the 3d depth information (just like cebas does), work with it after render is finished...
pixologic actually calls these 3d pixels, pixols, that have xy and z information




would be cool if vray started working with 3d pixols for efficient post effects

[vlado]

We probably can, but we do not have the resources for it right now. Also, like I pointed out, there are much better solutions out there, than we can ever hope to code.

Best regards,
Vlado

[phizikl]

jea ok, i understand ^_^