CGTalk - Coverage Passes - how to use them correctly?

View Full Version : Coverage Passes - how to use them correctly?

pixlix2

01-19-2007, 04:49 PM

Hello everyone,

I'm trying to understand the theory behind coverage passes. I've did some tests in Fusion but could not get it to work the right way. It would be great if someone could shed some light on this matter from software independant point of view (or is a coverage channel support actually unique to fusion?).

I "just" want to understand what you can do to overcome the antialiasing problems that are introduced with z-Channels, especially when making depth composites or DOF-effects. Half of the people on cgtalk say z-channels should never be anti-aliased and the others say the exact opposite. I understand that coverage channels have been invented to solve this AA-issue but I hate "blackbox"-mechanisms (and even the blackbox seems to be broken on my side) and I would like to know if you can make use of coverage passes in other compositing apps, like Aftereffects.

btw. I'm using maya /mentalray to render out the passes

many thanks for your help

DaXiiD

01-21-2007, 11:15 AM

Hi!, i have the same questi , i am using max9 and vray. I have done some research and i haven't found one paper or tutorial that clearly covers the coverage issues and uses for fusion or any other app. Maybe someone could shed some light to the issue.

NeonMojo

01-31-2007, 11:37 PM

I also would like some light shed on this. I have found some info on here and other forums explaining what the coverage channel is for and what it is but none on how to utilize it the best way... Do apps that support this just have an on/off swith do use it or is there a certain way that it should be used?

Aruna

02-01-2007, 06:28 PM

It's been a while since this question was posted, and no one has tried to tackle it, so I'll give it a shot.

What you're calling coverage channels, I assume, is a channel that effectively separates the objects at different Zdepths, and gives a nice antialiased effect on them?

There are several reasons why Z Depth passes are not aliased. The primary reason is because your Zdepth is just pixel information, usually coded in the image via grayscale values. If you take a look at an image that's aliased, you'll notice that the edges (of the alpha, let us say) contain gray values as well. If you alias your Zdepth, you're going to get wierd artifacts because the calculations your software package is doing is calculating that aliased edge as Zdepth information. Coverage channels (which I've never heard of until this thread), try to fix this problem by specifying which objects are in the Z channel, and giving you an aliased pass to 'cut' the Zdepth with, after the fact.

During comp, it is beneficial to render out individual Z depth passes per object in your scene. This will make Z depth based blurs and focuses more visually accurate . Of course, there are other tricks when coming with Zdepth where you don't even need to do this. And rendering multiple Z depth passes can be time consuming. It's much easier to render a bunch of mattes that you can use to isolate objects in Z (These coverage channels).

There is another trick where you can alias Zdepth, but the cons outnumber the pros. It's much simpler to take one pass, cut it with an aliased alpha of the object, and comp it in.

I hope that explain things a little bit.

CGTalk Moderation

02-01-2007, 06:28 PM

This thread has been automatically closed as it remained inactive for 12 months. If you wish to continue the discussion, please create a new thread in the appropriate forum.