[Steve Bentley]

It seems to be there on the perspective shot but not on the overhead shot. (the pegs that are real are blurred but your text is not)
The shortening of the layer may be due to the footage from premiere being a different FPS than the AE comp.
Check in the project footage and hit interpret footage under the file menu to see what fps AE thinks it is. It should match your comp fps or the footage will shrink or grow in time.
AE doesn’t really know about FPS. It works in discreet frames so its best if you can match one frame of footage to one frame of comp time.
Wanna post the project and I’ll have a look? Footage not required.

[Steve Bentley]

I’m not sure what you mean by “Speed of comp”.
If you have an expression going you can always add a slider (expression controls/slider) and then tie that value to your expression. Then you can keyframe that value and change it over time.

[Steve Bentley]

I wasn’t aware you could so that (bring a premiere project into AE with animation keyframes intact – but again I don’t use premiere)
It wouldn’t surprise me that rotation didn’t come in because so few of adobe’s package play well together.
But then again if position and scale come in then rotation should too.
In Ae you can hit the attribute to light up all the keyframes and then hit copy. Then you can paste that data into a spread sheet or text doc. Perhaps you can do the same for premiere and then paste that data onto the rotation attribute in AE? Just make an initial keyframe in rotation in AE and paste the premiere rotation date on top of it.

[Steve Bentley]

The problem is that a bass hit is instantaeous. So the flare will jump to the end of the path then retract to the beginning. Is that what you are after?

[Steve Bentley]

Good point Mark.
Tom, just keep in mind that once you convert each layer to 3D and use a camera and start pushing your layers into the Z, when you scale them up to get them to the size you want (and not to simulate a camera move) that percentage of scale doesn’t mean much anymore. So a 1000 pixel image pushed to 5000 units in the Z will be very very small. Scale it up to fill the frame (even though its way back there in the distance) and your scale factor might be 500%. That doesn’t mean you have blown up the image to 500% and that the quality will be horrible. It just means its 5 times bigger than it looked back there on the horizon.
As long as the image in real life is larger than it appears when the camera is as close to it as its going to get then you will be below 100% of real world scaling and the image will remain sharp.
Keep in mind too that when you have a moving camera you should have motion blur turned on and that blur can put a bandaid on many overly scaled/soft elements. A good rule of thumb for scaling is nothing over 130%. While it will be softer at this size you can get pretty close to the original sharpness with filters. Beyond 130% and you have to go with something like Genuine Fractals to get the kind of blow up you want.
I mention this because in projects like this sometimes original photos etc (like in Mark’s doc) will only ever be so big (1990’s video was only 640×480 remember: 1/6 of today’s standard video). So if you do need a still made bigger there are packages like GF that can blow it up to about 600 percent before it starts to degrade and before you put it into AE.

[Steve Bentley]

Correct. AE measures the values between one keyframe and the next and then does the math to figure out how much motion blur to apply. No keyframes come from premiere (that I’m aware of – but I’m not a premiere guy). What comes out of premiere is just a movie. You might be able to export keyframe data in an edl but I doubt it. Do the animation in AE (where it should be done anyway for best control) and apply the motion blur there and then bring that rendered element back in to Premiere (with alpha if needed)

You could apply a smart motion blur effect like Real Smart Motion blur which can look at the pixels and figure out the blur from a movie that has no motion blur on it but the effect has its limitations (and it’s not cheap) – if something comes in from an edge the effect doesn’t know what to do with those new pixels because as far as it’s concerned they haven’t started moving yet, they just appeared. Overlapping objects at different speeds can also cause trouble. Its not perfect but it is impressive given what it’s got to accomplish. It’s literally tracking every pixel. A doff of the propeller beanie to the guys who invented that!

[Steve Bentley]

I’m not sure what you are trying to do. The loop out expression just repeats the motion set by the keyframes in the position parameter once the last keyframe has been handled.
If you attach the motion of the lens flare to the output of soundkeys (I assume that’s what you are after) that flare is going bounce all over the place like crazy.
Are you trying to have the flare travel along a path so that the soundkey data controls its position along that path?

[Steve Bentley]

You can also use the transform effect and take advantage of the skew feature. Trying to do this in AE’s “3d” world will not get you there because there is no perspective here. The example you sent is isometric so all “perspective” lines are parallel.

[Steve Bentley]

I’m not sure what you are asking. Do you want to slow down the standard formula that comes with that deformer? Just write a new formula, that’s the whole point, or in a pinch just do a division at the end of the formula:
This is what comes in mine when I apply it: Sin((u+t)*2.0*PI)*0.2. So you can just do this: Sin((u+t)*2.0*PI)*0.2/2 to cut the effect in half (or just get rid of the *2 in the middle of the equation which amounts to the same thing). And that “divide by” can be anything you like except zero.

[Steve Bentley]

I’d go with the multifile motion vector trick in AE. Its all we really use now. There are so many things that motion blur doesn’t work with in C4D even with the physical camera: things through transparent objects dont blur, things reflected in other objects dont blur, certain shadows, certain particle or displacement effects etc. It seems we always run into something so we’ve just adopted this pipeline.
With this piple line you can also have other objects carry the blur data and then in AE you can apply that data to objects that didn’t blur.