[Steve Bentley]

But isn’t an extracted depth channel done with extractor already a depth matte? Yes you have to use it as a track matte with another layer but….
With a 16bit comp or higher the slider in the depth matte isn’t any different than crushing the levels of the extractor output. You are just changing the black point and white point of the depth channel to determine where the cut off is. And I think that’s all the depth matte slider does.

[Steve Bentley]

One of the best reasons to use QT Anim Best plus Alpha was the run length encoding. The mostly black (or white) alphas and RGB channel with stuff only existing inside the alpha were a godsend to keep the files small. Is there a reasonably low mb option now that every one is tossing QT overboard?
Are we all a victim of cheap harddrives? No need to innovate when storage is cheap? Its kind of a like the coding arena – code is pretty sloppy these days because the procs are numerous and fast. Who needs elegantly fast algorithms when you can just make it fast through sheer horsepower. Whats next? The return of the boxy-but-good Volvo?

[Steve Bentley]

Re hyperlapse: I don’t think that’s the case here. Hyperlapse tends to add distortion to the images as it tries to line up objects from one image to the other and I don’t see any of that here. But the bigger tell is that because hyperlapse is brought to bear on a number of stills taken at different times (usually farther apart in time than an equivalent running video of a tour) the people move chaotically or even disappear.
A simple radial blur (set to zoom) on sped up footage of a walkthrough would get you there. If you have the choice, just get your crowd to hold still while you walk the camera through the shot – even they aren’t perfectly still the shot will be so chaotic you wont’ notice it. There is a Seal music video that did this sort of thing. (in Paris I think)

[Steve Bentley]

You can also use the Extractor effect in AE (under 3D) to make any channel you want contained in the EXR a depth pass.
But better still is to use the position pass instead of depth!

[Steve Bentley]

Either way you will have to paint the wieghts it seems. I just thought the auto rig would do a better job of it. You can keep the rig you have and just manually paint the weights from where you are now.

[Steve Bentley]

Without seeing your precomp it could be a blend/transfer issue. Some of the blend functions don’t work right when there is no background layer and the continuously rasterize can make it worse.

If the precomposed layer is 3D or the precomp layer is 3D in the outer comp the CR button can have an affect on that was well.

As for the shadow there are number of scripts on aescripts.com that let you do reflections or shadows or you could use videocopilots free VC reflect plug in, or simply flip a duplicate of the precomp or use the transform effect. That way you only have to animate the seal and the “shadow” will follow along.

[Steve Bentley]

I think what’s happening is that the turbulence is to high an amplitude for the mesh level of the object. Even in frame one every poly is at a severe angle compared to its neighbor. If you look at the duck in the example you provided all the polys are pretty small and even and except under the wing all have normals facing outward and the overall distortion is very small. Because of the profile cap on the text and because of the tight inward curves and 90 degree corners it’s not going be long before any deformation (with a small amplitude) is going to have neighboring polys intersect.
What version are you using? I don’t think I’m getting the full picture when I open the file. So I may be using a different one from you.
Is this the bloating thing we were looking at a few weeks ago?
There’s a good web tutorial on softbody deflating (you would use it in reverse) I’ll have a look for and post.

[Steve Bentley]

To be fair I’ve never used the auto rig before. And I may just be glad of that now.
It looks like the weights are screwed up. I’m not sure if that’s because the model doesn’t have enough polys to bend as you would expect and the auto weighting doesn’t speak minecraft, or because of the settings you used during the rigging.

See image. I selected the left shoulder in the weights manager. The weighting is only set for the front faces of some of the geometry. If you move the controllers, the vertexes weighted properly do move (you have to “update” in the animation tab of the bone in question), but it becomes a mess pretty quickly. Some of the hand polys have no weighting assigned whatsoever so that may feel like nothing is happening if you just played with the hands first.
The weights on the legs look correct enough to animate though. So that’s why you are getting success below the belt (there’s a joke in there somewhere)

[Steve Bentley]

Can you animate the graduation or the changing black and white in C4D instead? That way you have full control of the nodes in the gradient and they are animateable. (the gradient can be instant, so a hard area of white moving on a black field is still a gradient, just a really short one). You can twirl down the the attributes of each knot on the gradient color bar to see its position and you can position another color in the next position over to get a hard line instead of a gradient.
You can either animate the plane falloff at the same frame as the gradient keyframe or you can tie the two together with expresso.

If you want to stick with a movie you can click the animation tab in texture in the material and change when the movie starts in the animation and where it ends (this is the two range settings)

[Steve Bentley]

Use the CC power pin if you can – the native corner pin is quite soft.