[Steve Bentley]

Can you post the project? I’ve not seen a layer shrink unless the originating footage changes. I don’t know how a motion blur would make a layer shrink.
Even with a slowly turning wheel you should see a little blur.
Something is up here.

[Steve Bentley]

Thats not so much an effect as some good compositing and planning.
They probably shot a clean plate (no actor), the plate with the actor with the clay overlay mask (the pink hockey mask thing) and then the beauty pass with all the actors and action. They may have run a pass of her in a green screen mask so they can get Adams hands separately.

Then you track the actors face in the beauty pass, add in a the green screen separated hands over top of the real actor and match the move to the the clay reveal pass and do a bit of masking. The multi coloured layers of the peel back are probably layers of clay built into the flesh colored mask so that melt effect was actually done practically as he smushes the clay mask up.

I think I might have eliminated two steps by making the outer layer of clay green and letting the inter colors reveal through thereby making their own mask and keying Adams hands all at the same time, but they may have found in some tests that it was easier to blend in a flesh colored clay mask on to the real skin instead of fighting green spill and a keyer generated mask.

[Steve Bentley]

As for the layers – render your character in C4D at the size you want him with an alpha (you could also scale him down in AE but it’s always sharper to use elements size-as in AE) Then duplicate your characters in Ae and place them in the grid style you want. There are scripts that can help with this or you could use an expression or just do it by hand.
Turn on enable time remapping for your layer before you duplicate it and add the expression too so you don’t have to copy and paste this to every layer (but you can do it after too). The expression will be the same for every layer.
The expression on the time remapping channel should retrieve the current layer’s own position and then reference a common extra layer that contains your pattern like an X or a diagonal line – get the color data from this layer – this turned off precomped layer can also be animated so the pattern changes over time and the characters will shift in time accordingly. You can animate this layer in AE or just bring in some footage and use that. Smooth CG footage should be used you don’t get grain noise as that might make the character layers cha cha back and forth as their expression detects the color differences in grain as it changes in footage.

[Steve Bentley]

You can tie the values of “enable time remapping” to a layer’s lightness channel in an expression for instance. So if you have a diaganol white stripe in the middle of a black field (turned off and just used for reference), you could get the position of the characters instance, ask AE for the color of the pixel at that coordinate and then do some math to figure out which frame that should be based on black being the begining of the footage and white being the end. Make sure to clamp your decimals so you don’t end up with frame 39.3. Enable time remap required whole frames.
You can also use the timewarp effect and tie the value to the source frame (the default in this effect is” speed” so you would have to change that to frame number). The beauty of the Timewarp effect is that it will interpolate the missing frames between frames if the math doesn’t work out to perfect frames.

[Steve Bentley]

It really depends on the shots. You may need a proper 3D tracker like Boujou or FP track. Motion blur can be a nightmare with trackers. Often we’ll overcrank the camera to get a good track and then skipprint the footage and add the motion blur back in.
You can also try crushing the values of the shot for your tracking element – sometimes this will reduce the blur and you can track the blobs that result.

[Steve Bentley]

While not exporting the handle data you can attache an expression to your attribute (pos, rot etc) and then tell AE to turn the expression in to keyframes. From there you can copy the keyframes and paste them into a spread sheet. Then you can scale that data any way you like and keep the easing that came from the handles. Then copy and paste that shifted or scaled data back into AE on an attribute that can take the same type of data.

Not ideal I know but all doable inside AE.

[Steve Bentley]

You should be able to do this with a shader effector (to set the pattern and the time offset) with your character in a grid cloner array. You might also want to set it to instances to keep the animation and geometry down.

But I guess I might be missing something because I would have to ask what’s wrong with the AE solution?
Unless you are interested in the perspective offset (slightly differing views of each character due to distance from lens-center) or want to have different characters in each slot, it seems like a lot of render time for 150 characters when only the showing frame of a single character will change across the face of the screen.
Rendering one character’s entire animation would let you get all the bugs out and the motion bang on and then just duplicate that movie in AE for the effect.
The only downside to the AE method I can see is that if you want to time warp the characters speed, you will be stuck with set frames you have already rendered in Cinema (a way around this is to set a higher frames per second rate in C4D and then you would have more frames that you would need).
Or if you want to move the camera around.
You could also do this in Element 3D right inside AE but I think you would have a better time of it doing 3D character animation inside a true 3D package. (element has a similar mode to the Cloner in C4D but is weak on character/bone hierarchies). Although you could use Element’s cloner on postcards of your single character’s frames to get the same kind of cloner control goodness that you have in Cinema.

[Steve Bentley]

Ah infini-D! That brings back some memories. I actually started with Alias (now Maya) back in the late ’80s but at $87k a seat (which may have included the required SGI computer) it was just untenable. But working within the confines of infiniD after Alias sent me running for EIAS (I was on Mac at the time and it was the only game in town, still I think EI was about $11k a seat).

You should be able to do the blur with built in CCvector blur – or any effect that takes a motion blur data channel.
When I test the new C4d Vector blur I’ll make sure its possible to use a non-RSMB tool with a multipass blur channel as well.

[Steve Bentley]

Without seeing the project its hard to say, but it looks like that node is taking the non uniform scaling from a higher node, or the rotating bit was already non uniformly scaled and once inserted in the hierarchy it is being affected by nodes above that are bringing the rotating part back to its original shape.
When you start linking rotating objects you can often get what’s called gimbal lock (if you rotate 90 degrees in x and z, which way is Y now?) and scaling things can often have unexpected results.
If you unlink the rotating arm what happens? What happens if you unlink it when it’s oddly scaled? What happens if you unscale it when it looks right?

[Steve Bentley]

Check that the you can undo the view manually in the view in question (under View in each view’s menu).
I have a terrible time with the hotkeys not sticking and I’d say every third launch I’m reminding C4D of one or another shortcut which seems to bring its memory back for them all. Its usually a permissions issue with where the prefs file is written.