PDA

View Full Version : RLA export --> AE



symko
11-08-2009, 02:32 PM
Has anyone done a RLA export from LW into After Effects? I want to add the DOF in AE using the DOF effect. In theory this seems like a great way to get fast DOF without having to calculate it in the LW render. My problem is I can't get it to work. I can export a RLA file, but when I play with it in AE, it's a mess. Weird artifacts, etc. Does this technique actually work?

dwburman
11-10-2009, 08:36 PM
As far as I know, the depth channel in the RLA and RPF exporters are broken, and I don't know a way to add a separate depth pass to the channel of another piece of footage.

If you use 'extract channel' (same place as the depth blur), you can see what LW is saving.

Personally, I think post-process depth of field based on a z-channel looks kind of funny with overlapping objects because the software can't see what's supposed to be behind the foreground object. That's not to say it can't be useful some times.

You could also render in layers with foreground, midground and background elements in separate layers.

You might also try the compound blur in AE with a separate depth pass controlling the blur.

toby
11-11-2009, 03:12 AM
I concur, I don't see any depth channel in the .rla files, Deluxe or regular 24/32bit.

blindsided
11-13-2009, 02:22 AM
I too have not been able to get this to work since lw turned 9, is there another option that does work, or is it being fixed in the new release?

toby
11-13-2009, 01:04 PM
Use Render Buffer Export to save the depth pass separately.

symko
11-16-2009, 07:02 AM
Thanks for the info. Looks like it doesn't work as cleanly as it should in theory. No big deal, it just seems like it would be a cool option to have.

toby
11-16-2009, 01:25 PM
ExrTrader can save depth and all kinds of other buffers in the rendered image file, it is pretty cool -
http://www.db-w.com/content/view/81/104/

JBT27
11-18-2009, 02:26 PM
ExrTrader is definitely the way to go here, and AE CS4 with the open-source ProEXR to extract the channels.

That said, and I'm putting this down to my inexperience with this stuff, I'm right now in the thick of adding fog and mist to a shot, using the depth channel from an ExrTrader output. It works well enough, technically, though I am struggling a bit with the alpha for the horizon, which is filled with trees, and somehow the fog doesn't look quite as soft or ethereal as it does when I use LW's own.

But that will be my fault, because I am getting more control this way, so I'm sticking with it until I figure it :)

One thing though: I can see that a nodal compositor is likely far better with this than AE's layered method. I'm creating alot of separate comps to create different effects, with depth, and the alpha, before I can start assembling the shot. Either I'm badly off-target with how to work with these outputs from ExrTrader, or AE is perhaps not THE way to work with them .....

That said, can't afford a nodal compositor right now, and I really like AE CS4 :D

Julian.

toby
11-18-2009, 03:31 PM
Just like LW, you can download Nuke (thefoundry.com) and it will run in PLE form, and there's free videos on the site too. Nuke blows away Shake and AE - if you can afford it. But if I started a business, there'd be no other compositing apps in the house.

Lightwolf
11-18-2009, 04:21 PM
It works well enough, technically, though I am struggling a bit with the alpha for the horizon, which is filled with trees, and somehow the fog doesn't look quite as soft or ethereal as it does when I use LW's own.
A depth channel usually contains no AA. You can activate it in the LW render tab. While not technically being correct, it can help with issues like that.

Another option it to just render a double size depth pass with no AA, use that to generate a double size fog plate and then scale down and merge onto your footage.

Even better would be to render everything double size. The main issue here is that you're applying all the effects in compositing on a per pixel basis, and the information used for them is per pixel as well (i.e. the depth buffer).
Double size and then scaling down is then a manual way to AA the effects added in the comp... if that makes any sense.

Cheers,
Mike

JBT27
11-19-2009, 03:57 AM
A depth channel usually contains no AA. You can activate it in the LW render tab. While not technically being correct, it can help with issues like that.

Another option it to just render a double size depth pass with no AA, use that to generate a double size fog plate and then scale down and merge onto your footage.

Even better would be to render everything double size. The main issue here is that you're applying all the effects in compositing on a per pixel basis, and the information used for them is per pixel as well (i.e. the depth buffer).
Double size and then scaling down is then a manual way to AA the effects added in the comp... if that makes any sense.

Cheers,
Mike

Thanks! Yes, it does make sense .....

..... if only I had the time and the render power - already working at HD! It is literally that horizon line which is jagged across four or five pixels all the way along, because of the trees ..... I might be being a bit anal, but if it looks wrong, well it looks wrong and needs fixing :)

OK - I'll try the AA on the depth channel, and I'm using straight alphas, which are working pretty well.

Julian.

Lightwolf
11-19-2009, 04:01 AM
..... if only I had the time and the render power - already working at HD!
The depth pass can be simplistic. Turn all objects to matte objects, disable any radiosity or other render options, double the size, just save the depth.

Granted, it's not as simple as just exporting the depth along with the main render - where you get it for free.

I'd try AA'ing the depth buffer first as well though ;)

Cheers,
Mike