robcat2075

Happy Birthday Dhar! I've had the pleasure of meeting Dhar in person when i've been out to SF and I can report that he is doing well these days!

As an experiment I used QuickTime Pro to convert a JPG numbered image sequence to Quicktime movies in Animation, Photo-JPG and H264 codecs and a TGA sequence and imported those into A:M (32-bit) I tried each one as a camera rotoscope in the default chor. My observation is that A:M doesn't actually load all the frames of these things just because you've imported them, it waits until you need to see a frame, then extracts it from original file and stores the decompressed image in memory for later use. All five formats worked as rotoscopes but the time it takes A:M to initially extract and display frames from Quicktime movies is longer than for image sequences. Scrubbing through an image sequence is fairly responsive but scrubbing through a Quicktime source is pokier until enough frames have been extracted and loaded. Photo-JPG seemed to be the fastest of the Quicktime codecs and H.264 the slowest. It is unlikely you will ever encounter a QuickTime in Photo-JPG, however. The only reason I can think of that you would convert a source to a Photo-JPG QuickTime rather than an image sequence for use in A:M is if you wanted to retain a soundtrack. Numbered Image Sequences are probably your best solution for rotoscopes and decals in A:M

Now an A:M report. Hopefully Steffen can look at that.

I presume the AO was "Transparent AO" off?

I haven't tried it recently but a simple QuickTime codec like "Photo-JPG" may work adequately for rotoscope purposes because it has no compression that depends on pre- or post- frames.

The eyelid/eyebrow structure always baffled me for a long time and I'm still pondering the best strategy for it.

Targa should work, that's what i always use. Did it crash or was it just slow? I suppose it's remotely possible whatever you are using to export your targas is doing it in a variant that A:M isn't familiar with.

David, I watched your AtTime tutorial and I think i am using it correctly. Good tutorial! Here is a demonstratin project that shows teh difference between a keyframed target and an expressioned target. ExpressionTest08_keyframe_vs_expression.prj In both chors the Vase has an expression to set it to the Y value of the Null at 1.5 seconds. In Chor "Keyframed" the null is keyframed in a conventional manner to float up and down. In this chor the Vase stays in a fixed place, as expected. In Chor "Expressioned" the null is on a Sin() expression to make it float up and down. In this chor the vase does not stay fixed and follows the Null.

My observation so far... AtTime works properly if the target it is getting a value from is something that is keyframed. If the target is something being moved by an expression, AtTime does not work correctly. I didn't save my working example where the target was on a Sinewave expression and i can't recreate it today. Maybe i dreamed it.

This will just transfer the problem to a different object. This would make it much simpler. Currently you get a circularity or a crash if you try to assign a value from the same property, even a past value.

Is the SSAO on the left?

Some nested "If()" statements along with some "AtTime()" statements could probably get a few things like that done. What would be an example of what you would want to do, Robert? Serg wants to get a random value captured less often than every frame Say, capture it on frame 0 and assign it to the Y property and then leave that value on that property for 5 frames. Then... Capture a new random value on frame 5 and assign it to the Y property and then leave that value on that property for 5 frames. And so on...

Basically i want to sample and hold a value until i need to sample it anew.

That is my first suspicion, but not supported by the case where we are sampling from a sine wave instead of a random noise wave. I dunno.

Here's a example case. The vase's Y translation is on a Rand() expression. RandomTest.prj Pick any specific time like 1:00 and scrub to it, out of it, back to it, before it, back to it.... the vase will always land in the same spot at 1:00 The Rand() function generates a random sequence but the sequence is repeatable no matter how many times the animation is played. That is as it should be and the example behaves that way.

Here's my mystery so far... If I set a Null's Y value to be equal to Sin(ChorTime()) it floats up and down in a sine wave manner as expected If i set another object's Y value to be equal to AtTime(1,the Null's Y value), the object takes the Null's Y value at 1 second and stays put there, as expected, even as the Null continues its sine wave motion. If I set the Null's Y value to be equal to Rand() it jumps up and down within a 0 to 1 range, as expected. If I scrub through 1 second the particular value it has at exactly 1 second is always the same, as expected. However... the object, which is set to use the Null's Y value at 1 second, doesn't stay put anymore. It is jumping up and down as if it were constantly tracking the Null's Y value instead of just taking the Y value at 1 second.

I'm trying to figure out a way to assign a value to a bone that it had at some specified time in the past. That ought to be possible without creating a circularity.

Anyone know where the full list of expression definitions is?

Here's AtTime.... I think these will indeed be helpful. I forgot we don't have actual variables. Lemme think about this a bit more but i have no doubt that what you want can be done.

Instead using the Rand every frame, set up counter that steps through x frames before the Rand is used again.

Welcome to the A:M forum, Arturo! I don't recall the specifics of them but I recall Holmes Bryant doing some face related tutorials... Weighting a Face Face Pose Basics Control Points Weights If you have questions, come back and ask!

Adam, if you have time someday it would be great to make a post that explains the Mac-specific stuff you've found out, for the benefit of other mac users. Here's what I get on PC... clip3512NetrenderNodes250.mov

Is it possible it really takes a 0 to 1 value?

Have you tried it with integers? You might try a progressive test where R G and B are all constants, Dose that survive? then try R getting a random number and see if that works. then add G getting random...

A slo-mo Thanks-giving! I bet your family enjoyed seeing themselves in animation form.