Darby Edelen
Forum Replies Created
-
[Andrew Shanks] “if it is a problem key, you can do custom log-linear conversions to build up a better exposed plate for pulling the key”
Teach me O master…
=)
-
[KirAsh4] ”
Now if I want the pictures to be “flying” in, like from behind the camera, past it and then onto the wall, am I better off doing that as a pre-comp on each wall, then create a new comp with the two put in?”Yes, I would say so. You’ll want each pre-comp to have two 3D layers (assuming only one picture is flying on), 1) the wall with no rotation or animation and 2) the picture flying onto the wall. Then in your final comp make sure that the pre-comps are 3D layers (rotated on their Y axes as the solids/walls were originally) and “Collapse Transformations/Continously Rasterize” is enabled for them. This allows the animation of the picture flying onto the wall to maintain its 3D transformations, otherwise the pre-comps would look flat.
-
[The Crestomizer] “The only thing I can come up with is Using spherize and CC cylinder.”
Which is also very limiting because the 3D representation of the cylinder exists on a 2D plane, you’ll have to animate positions and such inside of the CC Cylinder effect. In addition it doesn’t put a top or bottom on the cylinder =O
Best to work with a real 3D app in this case.
-
[jessieddogg] “it says up to 8 lights are supported with this card”
Are you using OpenGL fast previews? Are your results satisfactory with this method? I find Adaptive Resolution works a lot better for me.
-
[justinkhwong] “Would this pose a problem?”
Yes, but you might be able to at least fake a glow if the guy is sitting pretty still in the chair with some masks and an adjustment layer. Luckily a glow is usually pretty diffuse, so nobody should notice if it doesn’t emit exactly at this guy’s edges. The skin changing would be a more difficult thing to pull off with masks…
-
Yup, sounds like you’re pretty much on the right track =) You’ll need two 3D solids rotated on their Y axes and a camera.
-
Steve is right, you’ll need to pre-compose the color shifted solid.
After Effects has its own logical order of applying/considering effects on layers (which doesn’t always make sense to the user). Essentially AE is looking at the ‘footage’ (the solid) behind your Hue/Saturation effect to grab the color. If you pre-comp the layer (make sure you ‘Move all attributes into new composition’) AE will consider the pre-comp as footage after all the effects have been applied in the pre-composition.
-
I don’t know if I’ll be able to help much, as color spaces are one area that I’d definitely like to improve in myself.
My understanding of the log vs. linear issue is a little underdeveloped, but essentially film negatives have a much higher dynamic range than video. If we imagine black in video as having a value of 0 and white as having a value of 1 then film can store values at least as bright as 13. In addition to this, film negatives have a logarithmic response curve to light. This basically means that film is much more sensitive to differences in light in the 0 to 1 range and much less sensitive, say, in the 12-13 range (asymptotes anyone?).
Cineon (and .DPX files as I understand it) are 10-bit color and can store 1024 unique values per channel (2^10). If these 1024 values were distributed linearly you would have about 75 values for the 0 to 1 (video) range and 75 values for the 12 to 13 (overbright) range. This would not be ideal. Film is not as sensitive to this overbright range and our eyes are not as sensitive to this overbright range. So the distribution of color values in Cineon (and .DPX) follows a logarithmic curve as well: giving the darker 0 to 1 range more data than the overbright ranges.
This is a good thing as it more faithfully represents the native color of the film negative. However, you should never composite with a log file as the operations that we are accustomed to (blurring, opacity, blending modes) will behave much differently in log color space. It is ideal to composite in a linear color space. This is where my understanding begins to break down some.
A linear color space is one in which the gamma is 1.0. Compositing done in normal video space (where the gamma isn’t 1.0, more like 2.2) is actually not representing color values the way that film would.
There’s a lot of information on this phenomenon out there, most of which I haven’t read and so I feel uncomfortable getting too deep into it here. Most of what I know comes from Mark Christiansen’s excellent “After Effects 7.0: Studio Techniques” in the chapter entitled “Film, HDR, and 32 Bit Compositing,” but even that is a primer on the subject.
Without being able to tell you the precise workflow, I believe the way you’d want to approach this is as follows:
Have them provide log .DPX files (to properly preserve the film negative’s color). Convert these to linear color space. Composite.
And that’s as much as I can say on the subject honestly, I wish I could give you more specific details but I haven’t been through this process myself. This is a topic I’m very interested in exploring, but I’ve barely begun to.
I hope more people have input on the subject as I’m interested in learning as well.
-
[Steve Roberts] “You’re supposed to convert the HDV files to an intermediate codec.”
Correct me if I’m wrong (haven’t had the pleasure of working with HDV yet)… But wouldn’t this be an ideal situation in which to use proxies?
-
This seems like a problem that would best be described with pictures…