[Steve Bentley]

Hey Steve,
That looks like the element that has the movie texture on it is sitting in the exact same location as another object (surface of the water perhaps?) As the render goes along it constantly deciding what is “above” and below but due to round off errors or just limited decimal places the blocks of texture that are being looked at by the rendered flick in front or behind just a fraction causing banding or pixelation as the rearward object shines through.

[Steve Bentley]

Something else to watch out for:
Any attribute of the camera you haven’t got keyframes for will change permanently if you accidentally move the camera. I’m forever zooming in with the viewport controls on my camera view port and screwing up hours of intricate camera moves. And for me, undo has never worked where it comes to the camera.
So if you have keyframed x and y but not the z position or any of the headings, and you adjust that camera’s veiwport, you are stuck with a new version of what you animated. No amount of pounding on the undo button will help ( I know, I’ve tried!)

[Steve Bentley]

Hi Anika,
Have you been working with the auto key frame option turned on? It sounds like you made some changes and that got recorded as a keyframe and now as you go back and forth through the time line those key frames are changing your settings in the basic menu for that object. Is there a red dot beside the attributes that change? Anything other than a dull gray radio button means a key frame has been set somewhere in the time line.

The auto keyframe button (depending on your version of C4D) will be in the main timeline slider area near the play and next-frame buttons – it looks like two brackets inside a red circle (next to the skeleton key symbol) With that button turned on every change you make records a key frame.

[Steve Bentley]

You can also force a new frame rate in the project window. Select the footage in the project window, then under file, choose interpret footage and then change the frame rate. This is great for putting a 29.97 piece of footage in a 30 fps comp or running a 24fps clip slightly slower but without a 3:2 sequence within a 30 fps comp.
It’s also great for something that was shot at 60fps or rendered at 60fps (where the frame rate flag travels with the footage), and get it to play in its intended slow mo. You just interpret the footage down to 30 fps and it suddenly gets longer in a 30fps comp. No frame duplication, its just taking each frame that was there and mapping it each frame holder available in the time line. It’s also great for the opposite, speeding up a 30 fps clip to twice that (interpret to 60fps and put in 30fps comp)

One other way we use this is when a client wants something slower – if everything in our AE project has been generated within AE (ie keyframes, masks moving etc, but no running footage), instead of changing all the keyframes we’ll just tell AE to render this comp at a higher frame rate and then reimport the footage back in, set it to the original frame rate and voila, clean slower motion with every frame being newly animated.

[Steve Bentley]

Thanks for that. I’ve gotten to the point where we’re going to have to do this in 3D (C4D as well) or go with after market plugins.
I’ll give the revision boys a try. I figure if they can’t solve this no one can. We’ve used their very clever retiming and and tacking software but never their warper.

Thanks

[Steve Bentley]

Luma is still available in “obsolete” section of the effects list

[Steve Bentley]

So three things.
First you are looking at the image in 8 bit – always remember this. The LUT that’s changing the image so you can see it may be faulty or just not expecting the extreme things you are doing to the image.

Second depending on the filters used, you may be using one that is not capable of 32bits and then all your good work at 32 is knocked down to 8 somewhere in the render pipeline and at that point the conversion might fall down and start clipping.

Last, if you are compressing the video so you can see it running, the codec might be doing the clipping for you, or more likely you are trying to send out of AE a 32 bit image into a media holder that is not capable of holding 32 bits, so those subtle colors available in 32bit space get slammed to far ends of the spectrum in the final 8 bit space.

Can you provide any more info on the pipeline?

[Steve Bentley]

I think you’ve already got the answer. But instead of using the wiggle out of the menus, use it as part of an expression and then you can animate the wiggle’s intensity over time.
A particle system would also do this but the ones for AE may give you grief trying to land them on a specific spot.
If the landing area is just an area (like a custom shape or letter) then that could be done with particles run backwards so they are born from the shape in question and then fly along the path of the parent.

[Steve Bentley]

I’m not sure if this what you are looking for but here’s how we solve this – personally I find the clone tool just too tedious for all but the shortest shots. Use the tracking data to move a mask around (or the layer holding the mask) then just put another copy of the video through that mask (using track matte)but shift the copy by however far you want the offset to be to have clean skin show through. These two elements sit on top of the original.

This has the advantage of you being able to keyframe the mask over time so it can better conform to the blemish and it allows you to key frame the copy over time so you can put a better “clone source” through that mask. I find doing things by hand with any of the paint tools can create some terrible “boiling” of the fix.

It also allows you to feather the mask over time for better blending.

The biggest advantage is that you can now color correct the copy and use a different transfer function for how it sits on top of the original. Many times using the colour, hue or other type of transfer function will do a better job of covering the blemish since you aren’t really covering it and then there are no matte lines to deal with.

As for using the expressions, don’t forget about the pick whip. Just add an expression to the layer’s property you want to control with data from another layer, and then drag the pick whip from the destination layer to the data layer’s property, and the expression will be written for you.

[Steve Bentley]

More info after some testing:

It seems that image resolution and ratio vs comp ratio make a difference.
With an HD image reshaped within an HD comp, I’m having perfect results. But if you throw a larger image or an image that does not have the same ratio as the comp, then you get all sorts of different errors.
The closer you get to the comp size and ratio the smaller the errors are.

If the image is larger but the same ratio as the comp, just precomp the image and then do any scaling and reshaping in the smaller outer comp – I’ve had good results with this as well. Unfortunately this does not work if the image is larger and a different ratio.