I've searched long and hard for a proper workflow for working with Alexa LogC footage in After Effects, for delivering in LogC, but it seems there's just no workflow to be had without some kind of compromise.
Here's the two I've found so far:
1) Use the LUT generator from ARRI's website to get the proper 3D LUT, and simply use that as a guide adjustment layer above all the other footage items in your comps, so you can see what it'll look like. This is fine for some projects that don't require a linear working space for proper color blending, but is unacceptable for comping in 3d elements rendered in a linear space. If I apply the LUT to the footage, then I have no way of getting it BACK to log space for delivery! The problem is, I have no way of generating a "reverse-lut". The parameters I'd LIKE to use for a LUT are: (url option doesn't seem to be working... :/ )
Alexa, LogC, Extended TO Video, Legal WITH a color space of Rec. 709. That last bit there is the kicker. It seems there is no way to generate the reverse LUT (video, legal to logC, extended) with rec. 709. That option is just gone. So what I've been doing is just using the 1D LUT, which DOES allow me to reverse it perfectly... but the color is wrong. So this solution is far from ideal
2) Set the color profile of the footage to "Universal Camera Film Printing Density", which works OKAY, but just feels very very sloppy as the color doesn't match up to the color of the footage after run through the proper 3D LUT... which on its own wouldn't matter, but it means that any other footage I introduce isn't going to blend together correctly because it's not correctly converting to a linear work space.
Has anybody out there found a proper workflow for this? Is there any sort of "Remove LUT" utility out ther somewhere? Or better yet, is there an Arri LogC color profile I can use?
Mmmm that's what I was afraid of. What he describes as his workflow is pretty much just what I outlined in #1. This would be perfect if I was rendering out LogC from Maya... but I'm not... nobody is. Therefore by using an adjustment layer above all the other layers, you're not only converting your logc footage to video space, but you're ALSO applying that same logc->video LUT to stuff that WASN'T logC!
What I don't understand, is why isn't there a "Remove LUT" effect, that just applies an inverse lut to the footage? I mean... that SHOULD be possible right? The inverse of the curve is still a function, right? I may just have to look into doing this myself...