[Steve Bentley]

Thanks Mark,
I haven’t found it yet – the description of all the RG suites isn’t very robust and with everything being a “Suite” now there’s not much differentiation. I’ll have to download a bunch of user manuals just to see what all the tools are within each package. It might be buried in Image Lounge (I didn’t know they were still making that; what’s next, Kai’s Power Tools?).????

[Steve Bentley]

It always feel great when Adobe confirms it’s not just you.

Here’s another reason to consider the “jump”. We just took delivery of a new monster machine: 256 GB of Ram, 32 cores each @ 4ghz and an unspeakable expensive graphics card. The PC was cheaper than the graphics card, but a Mac (as close as we could get to that config but still falling short) was going to be 3 times the price. Maybe if they cut back on the titanium in the case? While this isn’t a usual example, the cost difference does hold to 2-3 times higher for the Mac equiv across the product line. My $3000 tricked out PC laptop compared spec-wise to a $6000 Macbook Pro.

The only things I miss from the Mac days is the way the mouse moved, and the file structure control (the way the file structure appears to you on a PC changes depending on whether you are inside an app, outside an app, across the network, looking at it from a windows tool, from a shell, saving, or opening a file. I guess I shouldn’t expect too much from a company that can’t even keep the right click menu consistent throughout the day.) In some ways Mac wanting control over the entire user experience was beneficial, but in other ways, very limiting.

[Steve Bentley]

Most of my stuff is on PC’s so the ProRes pipeline gets messy in that world, but yes I assumed that it was still supported.
My understanding of the whole mess is that QT is not being developed anymore and as such can’t be made compatible with the higher bit-depth-color environment inside AE. Which I totally understand, but it’s the way Adobe implemented it without warning that has us all fuming. What are people to do with legacy projects?
In reality its probably more the Adobe/Apple pssing match.

Its kind of like when we all had to bite the bullet and embrace 64 bit architecture on the hardware side and OSX appeared on the scene – the main reason I jumped ship from the mac platform (sniff, OS9 I miss you!). It was necessary pain to move forward with faster architecture. Rip the band-aid off and keep going!

But given all of that and the odd way FCP developed and then exploded, I’m wondering if there is something that isn’t totally 16 or 32 bit clean about the ProRes codec, or, if an older codec is installed on your system which shows up as ProRes but isn’t the latest version that others might be using who we’ve not hear complaining.
Usually these black issues are color space, or monitor gamma, or color management issues. (what are you working in with AE BTW?) It can also be a YUV vs RGB issue.
Just for a test try rendering in PNG in C4D and then work in 8 bit with ProRes in AE and see if the problem shows up. If not then it’s the higher bit depths that are not translating well.

Does it have to be ProRes? What about BlackMagic?

[Steve Bentley]

It looks like the culprit is the blue channel. This is the problem with compressed footage being used as a source for the key, the compression adds noise to the noise already there from the camera.
But the green channel is quite clean (this is why we use green screen whenever possible).
So here’s a file without the usual standard keyer that produces a better result and keeps the shadows.
Have a look and see if that does it for you. You will need to do a little work on the highlights and reflections, but these should be there so that you can pump color from the new background through them as though they are reflecting the environment.

12577_greenkey.aep.zip

[Steve Bentley]

You can apply materials at any time. The problem might be that because the model seems to separate parts, there will be a texture for each part. This may produce seams at the texture or model break points.
It all depends on whether you want to combine those parts into one model (then you could use body paint to stitch all the textures together into one big UV mapped texture), or you want to have separate limbs and then you can have separate textures.
The only caveat to all this is you may need to apply the textures at the home or T position. When a mesh is warped it needs to warp the texture with it, so often the texture will need to be applied at the default position time, then, when the character moves the texture will “stick” to the mesh.

[Steve Bentley]

Humpty was a politician? (Now that I’ve said that, I have a memory of Carroll depicting Humpty in the shape of Churchill or some other parliamentarian)

I haven’t tried this but once you get the traces from the particles can’t you feed those into the sweep object to be the path? Or, because it’s one emitter for the particles, does the sweep see the traces as one giant path made of all the traces? (hmm what would that look like?)

[Steve Bentley]

Hey Jim, since turbulence only works on the particles, could you not then “trace” those disturbed particles to get the sweepable paths required and wouldn’t those paths now be turbulated (is that a word)?

[Steve Bentley]

For expresso you can pump in a random number into the link list to pick the object from the list, and then you can examine the output of the list on the other side of the node. You can do a compare on that output (either via the name of the object or it’s position in the list) to send that object output to whatever “silo” goes with that object.

It’s really just a bunch of “if” statements. You can do that either with a bunch of compare nodes or it might be cleaner to use a coffee script where you can use the “switch” command to send a single input (of the output of the link list) to different outputs based on some parameter. It’s like you are sending models (or in your case words) through a pipeline that shifts to deliver them to the right spot.

[Steve Bentley]

I think you have discovered the problem. The bone structure and binding you are describing is expecting a single mesh instead of a different part for each limb or fashion accessory.
The problem with a disconnected model like this is that when you rotate the shoulder socket, the “fabric” of the chest won’t change, the arm will rotate as though it’s a barbie doll with a seam at the shoulder. If the character is meant to be an assembled doll, that’s find, but if it’s supposed to look normal, you will end up with gaps where the boots don’t meet the leg etc.

[Steve Bentley]

If you put them in a cloner, you can just enable the random selection under the “clones” parameter in the Object tab of the cloner.
That will spit out all the objects in the cloner in a random order. Then use a Random Effector to position your elements in space randomly. Or better yet: a Shader Effector – it’s a random effector on stereoids that lets you use a texture (usually with an animated noise) to control the random positions of the objects emitted.

And it may have just been in the writing, but you don’t actually apply the random effector by putting the cloner “in” a random effector. You have to assign (or drag and drop) the random effector to the Effector tab of the Cloner object (or just choose a new effector while the cloner is selected and it will be assigned automatically.)

You could also do this with expresso and make a link list of the objects and have C4D pick the items from that list to generate from an emitter in a random order by injecting a random number into the Index port of the Link List node.