[Steve Bentley]

Hey Mark I didn’t know about these guys.
Just watching the video it looks like, after you change the easing with the MTM tool, and then you select they key frames again, the interface doesn’t tell you what the easing on those key frames is set to, only what the last easing you did was.
Is that true?

[Steve Bentley]

Can you upload the actual project? If we’ve solved all those things I’m not sure what’s still not working.

[Steve Bentley]

So many ways – it really depends on the footage you have or the footage you want to shoot.
If you want to now exactly how that one was done, I’m sure that’s in the Cinefx mag – I think the whole issue was on Titanic. But its been a while so I don’t recall the actual method.
Here’s a few ways: (the first one is the most probable as to the way they did it – at Digital Domain I would guess).

1) Shoot your two characters with match frames as close as you can get. Then model or scan the heads of each, Then morph the CG versions of each head so they change from one to the other with lighting and textures and changes in geometry – in that span of time while they are morphing, you can shift your CG camera around so the CG of old rose matches the live action of real old rose exactly. Then it’s just cross fades.

2) You could use reshape in AE to do a morph, but you would have to get the two shots lined up really close for it to work well. I would shoot wider than you need but with a larger frame size than you need, then you could move the elements around to get a good match and the edges wouldn’t show.
I don’t think you need a motion control camera shot here because the frame comes to a standstill during the transition.

3) You could match move a CG versions of Old Rose onto Kate’s face and then do painterly reveal through to the CG Old Rose and then another reveal through to the real Old Rose.

4) If you can get your outgoing shot to match your incoming shot you could just do a long cross fade – but the geometry of things like the eyelid are going to change and it won’t be as seamless as it is here.

Don’t be fooled – a lot changes here, position, rotation, face geometry, wrinkles, eye spacing, skin texture, lighting color and direction, eye texture (which may in fact be CG all the way on either end so that Young Roses irises match Old Roses irises when the actor’s own would not have. The blink is the giveaway.)

[Steve Bentley]

For every animated attribute you can open the motion graphs and change the kind of keyframe you want that to be. (see picture for where the motion graph button is).
You can have bezier handles or broken bezier handles or holds or or really anything you like.
Just make sure to set the graph to Value (vs Speed – the prefs button for the graph area is second from the left at the bottom).
While hitting easy ease etc is really quick, its pretty rare we let animation curves run with these defaults. We are usually tweaking the graph shapes to get the exact curve we want.
Position will have a “speed” curve no matter what because it has at least two values buried in the position attribute. There is a “show separate dimensions” button at the bottom that will split the x,y,z out so you can address each one of these directions in its own curve and use “values” to do so (instead of speed).
Unfortunately there is no such separation available for scale – if you want to do that you can add an expression for each of the x,y,z of the scale attribute and use a slider or another layer to drive these separately.

[Steve Bentley]

Here are two spline wraps so you can see the problem – one has the subdivided cube and one doesn’t.
The geomtry of the undivided cube is not allowed because you have non planar polygons. It will render but it will shred textures.

12347_twosplinewraps.c4d.zip

[Steve Bentley]

Ok lets get the basics right first. Is this the type of thing you are trying to do? Note I have subdivided the cube that is deformed by the spline wrap – this was missing from your file. With out this there isn’t enough geometry to bank or rotate the wrap and you end up with flashing overlapping polygons.
Also, as an option, turn off the spline wrap in this file and turn on the sweep. You may find this easier to work with.

12346_wrapandtracer.c4d.zip

[Steve Bentley]

Would you not have more control by selecting all the keyframes, for say typing, and copying them and sliding them down the time line?

[Steve Bentley]

First, welcome back Adam, we’ve missed you!

Xavier, you can also assign lights to specific objects. That sometimes helps when you are trying to isolate certain areas – something you can’t do in a studio.
You can also limit how far a light travels with its falloff setting. So the light from the theater side of things doesn’t have to penetrate all the way to the background characters (as in the minions set up) and visa versa.

You can also feather the edge of a spot light, so as the character comes forward it passed out of the main circle of light of the spot light that is used for the main “stage” and passes into a new light’s feathered circle, one that is assigned only to the foreground character. The darkness on the character may be less about a shadow from the black box and more about lack of light from the main stage lights.

I don’t know if you have a reflection setting on your material but keep in mind that it will be affected to by your big black box (which you can choose to include or not in a compositing tag).

I don’t know if you are letting GI just handle everything but don’t be afraid to turn GI off and then light the scene as Adam suggested. When you get the lights all set up properly and it looks good, turning GI back on and it will add that much more realism (you often have to tone down the lights though). It’s handy to get good a lighting without relying on GI because things render faster without it.

Finally, if you are using GI, you can build the environment so that the “black box” is built in to the surrounding but invisible environment map. This will affect the light falling on the scene and the reflections too.
Like this:

[Steve Bentley]

Normally I would have said that the back leg (for example) is overlapping the belly a little and you are seeing the line of the belly drop out at that point. But in that gap is another line (the hidden texture perhaps since its thinner?) see image.
S&T can be tricky and can produce popping line segments depending on geometry.
Can you upload your project and we’ll have a look at your settings and the model to see if there is a way around it.

[Steve Bentley]

So you’ve got the spline wrap putting a 6 poly cube around the traced paths of the points on the other cube. That is creating what I would call broken geometry because there aren’t enough facets to really wrap -is that intentional?
With only four tracers (or even 8 or 20) this might be better to animate each object by hand (think how much time you have spent already).
Then you can just keyframe the different textures on each object. You can use the shader effector without a cloner.
I know that’s not so procedural but you always have to take stock of whether its going to be easier to just do it or let the computer do it for you, given that you can’t always get the computer to that last stage of finesse like you could with doing it by hand.

I guess I’m not clear what the tracers are for. The object swept along the trace line will change shape so with the type of texture you are using it will change shape too – is that the goal? I’m sure you have a reason for attacking this way but the fracture, the tracer and the spline wrap seem unneeded and even harmful. What am I missing?
Do you have a picture or a sketch of what this should look like?
I ask that because I think there is a better way and if I could see the look even on a napkin I might be able to suggest it.