Depth canvas, occlusion, tutorial or software?

Pickle Renderer

I have an idea that I can composite a scene by rendering the background with a few characters, then render the background with a few other characters, etc. then merge them all together to make a crowd scene, if I render depth, maybe there's some software that understands depth and will occlude characters correctly behind each other when compositing.  After a little googling I found Fusion 9, which apparently does this kind of thing.  Fusion 9 (free version) fails to start on my box however, sitting at a black splash screen forever.  I have to kill it from Task Manager.

Are there any tutorials involving compositing IRay canvases in software like GIMP?  It needs to use depth info to know how the scene is constructed...

Gordig

By "occlude", do you mean in the sense of dynamically blocking light? If so, I wouldn't think a depth map would be the best way to do that. You can assign groups of characters to different beauty canvases and render those one at a time, which I believe will accomplish what you're after (capturing the correct lighting, DOF etc. while cutting down render times). I also wouldn't recommend rendering the background multiple times; whether you include the background with the first group of characters or render it out separately, successive renders should only include new elements.

Sevrin

This video gives a bit of an overview of compositing assets using canvases.  The tutorial shows Photoshop, but these canvases can be opened with The GIMP and be manipulated in a similar way.

To do what you're talking about, you would need to layer your foreground characters layer above your background characters and background.  There's no real need for a depth canvas, since you can either use DoF in Daz Studio or apply a blur to the background layers.  Photoshop lets you use a depth canvas as a mask for the Lens Blur filter, but I don't believe the GIMP offers that option.

There are a bunch of GIMP compositing tutorials on YouTube you may want to check out.

Pickle Renderer

Unfortunately I've erased any and all Adobe software from my PC (I rage-quit their products for good as they're so bad). 

 

So to be clear, what I mean is if you have some character X sitting down, and there's a table blocking the visibility of their legs, is there a way of compositing the two images such that they can be rendered separately and in the composite image the pixels of the table are in front where the Z depth of the image at that pixel is < the character's pixel (relative to the camera), and the character pixel is rendered where the Z depth at that pixel is < the table.  You know, standard video game Z-buffer occlusion.  I think I need to render out the depth and composite them with a < rule. 

 

I will continue my Googling then.

 

(also I thought perhaps I wanted to continue to capture shadows and bounce lighting from the set, so each set of characters is rendered in the scene, not alone).

 

lou_harper

Pickle Renderer said:

Unfortunately I've erased any and all Adobe software from my PC (I rage-quit their products for good as they're so bad). 

 

So to be clear, what I mean is if you have some character X sitting down, and there's a table blocking the visibility of their legs, is there a way of compositing the two images such that they can be rendered separately and in the composite image the pixels of the table are in front where the Z depth of the image at that pixel is < the character's pixel (relative to the camera), and the character pixel is rendered where the Z depth at that pixel is < the table.  You know, standard video game Z-buffer occlusion.  I think I need to render out the depth and composite them with a < rule. 

 

I will continue my Googling then.

 

(also I thought perhaps I wanted to continue to capture shadows and bounce lighting from the set, so each set of characters is rendered in the scene, not alone).

 

I'm not entirely clear what you're saying, but it sounds like something you'd do with canvases. MartinJFrost has a nice explanation here how he uses them: https://www.daz3d.com/forums/discussion/466351/ultrascenery-new-territory-commercial/p66

Photoshop has a duplicate layers feature that is useful compositing images rendered separately. If you render first the guy, and then the table with the same camera, using duplicate layers you can place the table at exactly the right spot.