What OS? What version of AE precisely?
I just tried it on a wimpy Mac Mini running CC and Mavericks on 8 gb of ram with no problem.
All of Rick's questions plus how many keyframes/ frames are we actually talking about? Inevitably there will be a point when things get slow, but it will depend on the specifics. Most slowdowns relate to actually drawing the keys as they are moved, but that would literalyl mean you have hundreds and hundreds of them...
Thanks for the response. I had updated to the october 2013 version.
There are a lot of keyframes 12900.
What I wonder about, I can move the clip, with keyframes showing and it moves with no problem.
Would that still mean this is a drawing issue?
the mocha ae plugin, is that a factor?
I cut down to about 600 keys and it is workable.
Trying to do a relight and using this new workflow first time. I can of course subdivide clips. Kill redundant keys, etc.
Just trying to learn the parameters and realities of this workflow.
It's probably a combination of things like keyframes evaluating when you move them around in global time vs. being locked to the layers local relative time and the resulting extra updates whe ndrawing the timeline and comp/ layer windows. But 12000+ keyframes is really slightly insane. I think the most I ever had in my 12 years of AE use was somewhere in the 5000... If possible, really try to keep things neat and to only the parts you actually need.
I agree, the thinking being since I really am trying to integrate mocha AE workflow I wanted to know how "real world" the workflow is.
I am relighting a 40 min chunk of footage. Using mocha to quickly (relatively) track and matte the face.
I divided into 3 pieces and the mocha plugin brings in frame by frame for geometry. Hence the huge number.
I thought maybe there was something I was missing in optimizing, but it looks like it boils down to running much smaller chunks. Part of the thought was, if mocha can handle that number, could AE.
Thanks for the input
Yeah, it's 40 whole minutes of video that's killing you. Smaller chunks are good. A rough-cut is called for.
Yes, quite familar with spotting, rough cuts, etc.... I am working with some folks who shot with a handycam (auto magically) and they have students working with footage. Each student can cut as desired. Therefore I'm giving them a rough relight on the whole shoot. Tried taking their roughs and ended up with numerous issues, plus multiple visits to the color times.
This is an experiment to test the workflow
and torture the code and gear.......
40 minutes? Wow that's way outside a normal production workflow. Even if you were Andy Warhol I would suggest cutting it up into manageable chunks and then putting the clips back together. Even if this were a locked off single 40 minute take I'd split it up.
If there are multiple takes in the 40 minutes of footage I'd separate each take, do your tracking, and then put the clips back together, render them to a production codec and distribute the fixed footage to the class. That will take less time than fussing with the tracking each time the shot changes.
If you are teaching folks how to make movies then the first thing you need to teach them is efficient production workflow and I can't imagine any scenario where this one is the most efficient option.
Yesterday I split a 7 second shot for some mocha tracking and roto work into 3 sections and only worked on the 2 second section in the middle where the actor's arm needed to pass in front of the CGI element. I only worked on the arm even though the temptation is to mask the entire actor.
If it were my class I'd show them how to do the relight with a 10 second clip, teach them to rough cut and then apply the relight to only the portion they will be using.
I'm curious why you need to shift the keyframes in time. Tracking implies keeping things lined up.
40 minutes? Wow that's way outside a normal production workflow.
-This isn't a "normal" workflow.
-Even if you were Andy Warhol I would suggest: Leave poor Andy out of it
-cutting it up into manageable chunks: I'm doing that, but due to the nature of the project and in the spirit of Andy.... no wait Spinal Tap, I was turning it up to 11 just to see how far and hard the workflow could be pushed...
"If there are multiple takes in the 40 minutes of footage I'd separate each take, do your tracking, and then put the clips........."
: been doing that for sixteen years, so I understand the concept.... that doesn't serve in this instance.
If you are teaching folks how to make movies: Not teaching anybody, this is a class at a college, taught by a professor and not in the subject of media, production, etc. The production aspect is secondary as far as they are concerned.
teach them an efficient production workflow, If it were my class I'd show them how to do the relight with a 10 second clip, teach them to rough cut...: referencing above, I'm not teaching and they aren't looking to learn that in this course.
I can't imagine any scenario where this one is the most efficient option.
-While I cannot be as descriptive as I would like, due to an NDA, this is the scenario where it is, in fact, the most efficient.
1. Several shoots from a full fledged production crew (us). Great quality (picture and sound) lackluster content (relative to desired result)
2. ENG style shoot, with actual news channel crew: result=usual boilerplate
3. High school kid with one laid back adult, DV handycam and a cheap, small shotgun + wobble pod. In short nothing more than a kid out shooting a school project.
4. The subjects were recorded in different random order, over the summer, by the different crews (where possible)
5. One of the goals: observe the effect of each production size on the content gained from subjects.
6. As I personally suspected, the low key "high school" shoot caught the most honest footage.
7. The course is going to use this footage over several years. He likes the content of the amateur shoot the best and I would agree (in context of what they are working on). So I am correcting gross defects (auto settings, with backlight and traffic in background, pumping exposure, shooting in shade with hot background, etc) In all cases the faces were poorly "lit" The professor is going to have students edit as they see fit, following the protocols of the course work. It will be used over multiple years and courses. The students can, and will (I've already seen some creative cuts) use anything in the footage.
8. So the scenario: clean up the footage enough that it isn't an assault upon the eyes, any portion can be used and the student is not worried about any tech/post concerns. There goal is to simple cut the footage together.
I'm curious why you need to shift the keyframes in time. Tracking implies keeping things lined up. Didn't import that way, maybe due to the insane load I was experimenting with
Thank You for your response,