Nils Hammers
61
AE Expressions Coins
43
After Effects Coins
142
COW Coins
2
Solve Coins
Forum Replies Created
-
Woah, apparently the answer was on the surface.. thank you for bringing it up, Dan!
I remember reading about it in the manual, wondered the real application, seldom saw it mentioned in tutorials. Copilot was none the wiser.
Thank you, thank you, thank you Master! -
Hey John,
Why would that be a problem?
Actually, as the light’s position went behind the camera, I was unable to generate any ray in any method, even I’d like to do so. Thus, all my example images are light-in-front-of-cam cases, even I am looking for the shots where the light source is off-frame.
-
Hey,
Thanks for your input, Graham.
If I got your idea correctly, then I’ve already tried that. It gave similar results to what can me seen in Lightburst example image. There – please, correct me, if I’m wrong – the raygen takes the matte from what’s seen in current frame from the camera i.e. skies. The problem arises when layers close to cam begin to intersect those in the background (see in another example street_from-2D-matte attached: traffic light front of bg houses, notice the circled areas), which in motion makes the effect 2D-ish. Besides, the edge also cuts the matte.. I could extend frame upward a little, however, it won’t remove the elephant from the room, as the source has no depth data.
If I use some custom matte instead, then the rays cover all the objects equally, no depth. I could possibly use the depth matte to composite those rays over the scene, yet it won’t generate the shadows from those objects, flat again.
Seeing the examples, can you suggest what am I missing? Or maybe whole another approach?
This far I’ve tried, with Shine on adjustment layer I can generate some sort of radial volumetric shadows rays, not the light rays.
-
UPDATE: Testing even further, I found out, that the affected part is the Effects animation, whereas the transform animation is correct: separate, underlying layers unaffected. See attached AEP for demonstration and own tests, as the remedy for this is still unknown.
-
I need the AcceptLights: OFF 😀 Apparently, this option is not there in A3D, yet.
Otherwise, I am using several layers of nesting to build complex pseudo 3D environments. What you say about mixing renderers in this scenario, at least until recent, won’t work – as soon as you collapse transforms, the comp is rendered by the engine of parent composition. Similarly, I can’t use AdjustmentLayers. Regarding the Classic3D, it is too basic, e.g. compare its camera and material properties to what’s in C4D.
And now, this issue too. I am in a Limbo – Adobe dumped what was working earlier, meanwhile, what could work is not made yet or is glitchy. *Sigh*
-
Hey David,
Thanks for your input.
I experience the same behavior also with Advanced3D renderer, not Classic3D.
Yes, the AdjustmentLayer can do the “trick”, however, after that, the 3D depth sorting is lost, and you cannot do anything more complex without making a split-layer-hell in the Timeline, e.g. I have some 60+ of such blocks for pavement borders, plus camera motion.
-
So, the problem seems to run even deeper and is unrelated to expressions.
Apparently, if I nest several copies of the same 3D comp with collapsed transformations, then the very top layer influences the rest of the copies, e.g. changing its time value (animated timeRemap, expression, or simply shifting the layer over time), makes all the underlying layers within the 3D container, look as the top one.
From there, all what I experienced earlier.
Am I missing anything essential about collapse transformations + time attribute? Such behavior seems fundamentally abnormal, however it should be tested by the other users, as it might be a local glitch.
Here I add a AEP file made of solids alone, see yourself by shifting the “MOVE ME” layer on timeline back and forth. You can also try enabling randomization expressions in *T, *S, *E compositions, which will override the existing color animation and make it more obvious.
Please, report back if you experience similar behavior.
-
UPDATE: Weirdest of all, soloing the single box layer changed its rotation, where I also have similar expressions, based on timeless random seed. It is like soloing suddenly impacted the layer’s index OR enabled the execution of particular expression. I can’t tell which one is true, as the details are small, but scene – too large.
Also, soloing everything was similar to not having solo’s at all, it is some certain number of visible layers that enables certain number of expressions, like there’s a limit, which seems illogical.
Can many randoms be too much, so that the software misbehaves, and expressions don’t work too?
I have some 60+ of those boxes there, so I needed them random to look organic.
If anyone has an experience with stress tests where AFX fails, please share – I’d gladly avoid such pitfalls in the future.
-
2D parallax shouldn’t be so hard to achieve by much more simpler means, aren’t you over-complicating the solution? Unless it is 3D, where the speed is calculated based on camera parameters, for 2D it is you who decides how much the second layer is moving faster/slower than the first one, i.e. same
value + layer1.position * 1.2 ;
may suffice.
If you necessarily want to use Dan’s expression, then the dirty solution might be to have it in every layer and multiply-by-expression the Layer speed control slider’s value instead. Just choose which layer will set the “default” speed, and link all the other layers to it with different factors. -
Select both, and, it is Speed graph for position and rotation you want to manipulate for a nice slow-in. Since 2 is farther away, its positional speed is higher than that of 1. With both selected you can compare the speeds, make the speed at first frame higher, and at the last frame 0. Similarly adjust the rotation.
