Compositing with depth passes in AE

[Todd Kopriva]

You probably want to start by reading “About 3D Channel effects” and then check out the 3D Channel effects.

———————————————————————————————————
Todd Kopriva, Adobe Systems Incorporated
Technical Support for professional video software
After Effects Help & Support
Premiere Pro Help & Support
———————————————————————————————————

[Walter Soyka]

In addition to the resources Todd pointed out, Frischluft Lenscare is a really nice third-party DoF plugin.

Walter Soyka
Principal & Designer at Keen Live
Motion Graphics, Widescreen Events, Presentation Design, and Consulting
RenderBreak Blog – What I’m thinking when my workstation’s thinking
Creative Cow Forum Host: Live & Stage Events

[Tobias Andersen]

Thanks for your replies.

It seems like the articles you refer to talks about extracting the depth value from a rendered sequence if the depth is rendered into a .zpic file in the same directory as the .pic files. I think I can get this to work, but I actually had the two depth passes rendered out to .pic files as simple grayscale images.

Is this possible to use in After Effects or do I need to render the depth from Softimage as .zpic files?

[Tudor “ted” jelescu]

Yes- if you use the grayscale images you rendered as mattes for the footage composited on top and then adding blur or color correction.

Tudor “Ted” Jelescu
Senior Compositor/VFX Artist
Bucharest, Romania

[Tobias Andersen]

Not sure if I can see what you mean – won’t I lose the depth information if I just use them as mattes? I guess I need to use some sort of depth matte?

[Tudor “ted” jelescu]

The whole idea of rendering out a grayscale depth pass is to be able to use them in compositing applications that can not interpret depth data. So, it’s the same idea as in a displacement map- white is the closest to camera, black the farthest. By compositing the footage on top of itself and using the grayscale images as mattes for applying effects you can apply blurs, and the objects that are far away are affected the most, thus creating the illusion on depth of field. Also, with color correction you can reduce saturation and tint the objects in the back.
Hope this makes it more clear.

Tudor “Ted” Jelescu
Senior Compositor/VFX Artist
Bucharest, Romania

[Tobias Andersen]

I see it is useful for depth of field etc and applying effects taking depth into account. Perhaps what I’m looking for is not possible.. I have two objects A and B in two separate passes and sometimes object A is in front of B and sometimes the other way. Isn’t it possible to composite those two correctly? If I do it inside Softimage with its composite tool I use a Zcomp node and plug in the two passes and their corresponding depth – then it outputs the correct result.

[Tudor “ted” jelescu]

Read the links Todd provided earlier.
Here’s an extract from the adobe site (link below):
“3D Channel effects read and manipulate the additional channels of information, including z-depth, surface normals, object ID, texture coordinates, background color, unclamped RGB, and material ID. You can layer 3D elements along the z axis, insert other elements in a 3D scene, blur areas in a 3D scene, isolate 3D elements, apply a foggy effect with depth, and extract 3D channel information for use as parameters in other effects.

If you convert a layer with a 3D Channel effect to 3D and view the layer from anywhere but the front and center, it doesn’t appear as expected.

To display depth and other channel values for a pixel in the Info panel, apply a 3D Channel effect, and then—with the effect selected—click the pixel in the Composition panel or Layer panel using the Selection tool.”

https://help.adobe.com/en_US/aftereffects/cs/using/WS3878526689cb91655866c1103a9d3c597-7bd8a.html#WS3E78212A-C902-4d74-9FE6-C0E98E602AAA

Tudor “Ted” Jelescu
Senior Compositor/VFX Artist
Bucharest, Romania

[Tobias Andersen]

I have read the links but it doesn’t really explain much about how to use it, it’s more general about the depth information. I have made zpic files and extracted the depth – that works fine and I can get the depth information when clicking at specific points etc., but how to continue from there? I’d also prefer to use the grayscale images if possible.

I’m sorry if I have misunderstood or overlooked something in the articles but it’s not too clear to me 🙂

[Tudor “ted” jelescu]

From what I understand, you have 2 objects that are going one in front of each other at various times rendered out form your 3D software, together with a depth matte. Are the objects rendered out separately and you want to composite them as individual layers in AE? If so, you may need to do some masking to reveal the proper one in front of the other and it may prove a pain. Best way to do it is to render out both objects in one pass, with the depth matte. Then you won’t have to worry about not being able to have the right one in front, and you’ll be able to use the z-depth information with the AE plugin to composite.

Tudor “Ted” Jelescu
Senior Compositor/VFX Artist
Bucharest, Romania