[Christopher R. green]

You can pretty much disregard the +y thing. I just put that in in case it made any difference in the way the solution evolved. It is, of course, dependent on how the rotation is done.

Thanks again!

cg

[Christopher R. green]

Um…
Perfect?

That is really cool, Flip.

Dan, you must’ve over-thought the problem!

This is going into my “favorite tricks to steal from others” folder!

I’m giddy.

DANG!

nice.

Thank you.

[Christopher R. green]

[Filip Vandueren] “I’m not sure if I understand that “never go negative in Y”, this would only occur if the rotation would be more than +/- 90°, so you have that under control.”

Never go positive in Y, that is, ‘down’.

But now I’m going to see what you came up with …

[Christopher R. green]

Well, you could always try my wacky script that lets you do it in AE, “CompToUI”, available on crgreen.com, at the bottom of the scripts page (“Comp_To_UI_package.zip”).

[Christopher R. green]

I have had a similar problem with the Quad-Core Intel (2x3Ghz) and Apple ProRes files (originally shot on Red camera and exported via Red Cine). I’ve been working at a company that still has a bunch of PPC machines (how last century), and the gamma shift with the new machine(s) is huge.
From the Almost Silver Lining Department:
One way to test to see if you have this problem (even if you can’t compare to other machines) is to import the movie file in question into AE, make a comp for it, open that comp and toggle different bits per channel (bpc) settings by option-clicking the bpc button at the bottom of the project window. When you shift from 8 bit to 16 bit, you should see a major gamma shift if you have this problem. If you do compare with another (‘safe’) machine, you’ll see that the 8-bit is shifted considerably from one machine to the other (check specific pixel values). When you change to 16-bit on the ‘bad’ machine, the values will be much closer to ‘correct’ (16-bit) values, but still off. Of course I am assuming the ‘safe’ machines have the correct values (which I believe is a safe assumption).

[Christopher R. green]

You can also try:
1)Channel>Invert
2)Channel>Channel Combiner (from:Max RGB To:Alpha not inverted)
3)Channel>Remove Color Matting (Background Color:pure black)
4)Channel>Invert

This is a variation on the “unmult without third party plugin” technique seen elsewhere here on the cow. That works for ‘effect on black’ images, this for ‘effect on white’ images.

sorry about the late post.

[Christopher R. green]

[Jimmy Brunger] “I’d love to be able to apply different trackers to diff points on a mask…would that be a script-based command or is it not possible in AE?”

This is exactly why I wrote the Connect_Vertex_to_Point script, which is at the top of my scripts page:

https://www.crgreen.com/aescripts

With my script, you do one vertex (mask point) at a time.

[Christopher R. green]

My aescripts page now treats IE users with special care. It came to my attention (after all this time … tsk tsk tsk) recently that IE users (at least v6 and v7) were seeing something so horribly wrong that it made me all queasy to look at it. Now there is a working, albeit more drab version of the page that will load for the IE club.

enjoy!

cg

[Christopher R. green]

No, the problem exists in QT Pro.

[Christopher R. green]

If it crashes while displaying the splash screen, watch carefully what is being written out as it displays the various components it initializes. See if you can tell what it reads right before it dies. Might be a clue. Could be a bad font, which might involve a little detective work, the start of which may be emptying your fonts folder (this is a reason why I install new fonts into my user font folder so I can empty it to begin diagnosing a problem).
Also, (and if it crashes before the splash screen) check the crash log using Console. This might also give a clue.
That would be about my two cents worth.

good luck!

cg