Hello all,

I have David's compositing tutorial, but I was wondering if there was a way to render in layers not everything, but so that you could take the background, or the foreground, depending on what you want to focus on, and do some DOF softening effects in 2D in post? But still have shadows and such calculated? So I wouldn't need every light and shadow broken out, I'd just decide if I wanted to soften the background or sth. Like a 1-3 pixel Gaussian blur animated over time might look nice, keeping my subject in focus, and it would be a lot less render intensive.

thanks,

Seana

PS. What I mean is, I would like to render out in layers that are kind of "grouped" I don't need every light and shadow, etc broken out. I would need to define the background, foreground, subject. Also I have to check the tutorial, but I would need to do the post effects in 2D in a Targa sequence, so I think that's just a matter of reloading the EXRs?

Seana,
I may not be following your question entirely.
You certainly can render different elements of your animation out and then composite them back together right in Animation:Master.

The best way to find out though is to experiment.
Use of the Alpha Channel will be required extensively.

Try a project with three characters at different locations in the scene.
You can even have another choreography with nothing but a background but keep in mind the possibility of different lighting.
Render each of these characters and backgrounds (or other objects) separately.

Then bring them back into one Choreography as Rotoscopes or layers.
You can add Post Effects to the images and see the results in your renders.

or


You can skip this step now and just do this via v12s Compositing as well.
Once you have everything in place you can experiment with various Post Effects to adjust each characters Blur, Lighting and more ON THE FLY!

With the exception of v12s Compositing this is not a new aspect of Animation:Master so we should get the word out.

Rodney

Hi Rodney,

I think you have it mostly. I thought that I would like to render out things in layers one I have my lighting set and so on so that I could then apply softening effects to backgrounds and such to get some DOF effects in post that were less render intensive than doing them "up front." While I know AM can do them, if I have layers of 2D images, that also gives me the option to use other tools where appropriate for my setup. It's a question of working within time and resource limitations. So, I'd like the character and associated cast shadows to be on layers, and then the background objects could be on their own layer, alpha buffers on of course. I would then have the flexibility to make adjustments in post without resorting to masking, because _that_ can be more grief than it's worth.

I hope that makes sense. Would it be best to handle this in AM Composite? I need to review how to render things out in layers and still have shadows and such (i.e., not just "hide" things.)


S.

I think he is talking about this: In some programms, you can render in different layer/passes/whatever.
For example you can say: "Please just render spectaculars out, leave everything else behind.
After that, please render only shadows, nothing else.
...and so on.

If you render all these things out in different images with Alpha-Channel, you can go to a post-programm like AE or whatever and compose them together...
So you get more control over you images and if something has just a wrong spec-size, for example, you dont have to render the geometry again, just the specs. Or even better, you can manipulate the specs afterwards through a 2d-Pain-Programm, like Aurora etc.

It has some advantages... I am not familar what v12.0's composite can do, but I think it does quite the same, at least lights are handle separate.

S. Gross has written a plugin, which may help you, too. It is called Passes.
-> www.sgross.com

And as far as I can tell, is currently for Windows only. Seana's a Mac user.

Cannot afford to guess much these days. I can only go with what A:M can do related to the query.
There are entirely too many things A:M cannot do. Our focus needs to be on tangible solutions. I do now what you are thinking of here and v12 Compositing really presses us into new areas of rendering in passes. I need to investigate the possibilities myself to see what is available.

As for Steffen's 'Passes' plugin... that might be quite useful. I'm sorry to say I haven't used it though.

One really needs to experiment. Have to... got to... need to... experiment.
That is really the best way we learn.

Shall I start or do you prefer to get the idea going?

Experiment here... its the appropriate forum!
Lets render stuff in layers!

Best,
Rodney

Hello guys,

I'll see if I can set up to try a single still with a sort of "background" and "foreground/subject" group layer. Maybe two background layers so there can be a little less softening in the middle distance. My reason for doing things in a 2D compositing application in post is simply because when you end up with 2D images anyway, devoting all that 3D processing power at the outset isn't really necessary. I'm not at all saying that A:M can't do it, I know it can. I am just saying that for me I have to optimize my pipeline. It's enough to wait for each frame to render out in HD 720p. And I tried to do Greg Rostami's "better antialiasing" solution (Blur set to 0% as a post effect) and she quit halfway through one frame. So I'll do a 1 pixel Gaussian blur in post, possibly over two layers of the same sequence with opacity adjusted on the upper layer to fake "sub pixel" softening.

I'll see what I can figure out. Maybe David Rodgers will see this and add 2 more cents. Oh, and sorry for writing about this over American Thanksgiving. Ours in in October.

best,
Seana

Seana,
I often do post effects in such programs as CorelDraw and CorelPhotopaint. They are my 2D weapons of choice. I am using them less and less these days and with the introduction of the newer Compositing in A:M I expect to convert a lot of my pipeline to A:M more and more.

There are many reasons to do something in post but the two that make the most sense to me are 1) Speed 2) Quality. There is a third, Familiarity, that I 'm starting to close in on with A:M's Compositing. I'm finding that while now it may be quicker to do some things in post outside of A:M with a little bit of preparation I can cut even more time by keeping the project in A:M. Open EXR is a big part of this solution but I haven't explored it enough. Heck, I'm still trying to master A:M's Post Effects!

Little by little my familiarity with A:M's 2D capabilities is increasing and making part of my workflow (the 2D compositing in Corel) obsolete.

One thing that is really nice is that by rendering in layers you have OPTIONS.
If you want to continue in A:M you can.
If you want to continue in another program you can.
You can even use both methods together.
And you control the solution.

Its nice to have options.

Rodney

Yes, it is best to have options, which is why I want to know about layered rendering. I would generally just move to sth. else from A:M if the render times proved too ponderous (I'm on a Mac, I think there are certain things that the rendering doesn't like, though since 12.0k it does seem a lot faster), and I could get to point B faster just by treating something as a 2D post effect in another way. I used to use Painter and After Effects a lot. Now mostly I just use AE to composite layers for QT movies.

I use Mirage for almost all my 2D stuff now.

I just have to figure out the best methods for setting up a render with the kinds of conditions I've described. I'm still playing with it, but I'll get there.

regards,


Seana