CinemaAED

Having renal problems is not funny, trust me! Getting up in the morning is an exercise in itself trying to not scream your neighbors out of bed from the pain and then you feel terrible for the first few hours of the day. In fact this morning I felt so nauseous, I almost threw up on the train when I caught a whiff of someone having a beer at 6.30 in the morning (alcoholism is an interesting condition, is it not?; though of course the people affected always say they have no problems with it [just without, obviously] *lol*). Despite all that I had a bit of fun as of course the interwebs are burning with discussions on that Maxon and Adobe thing. Best quote of the day on mograph.net:

User microdot: Sorry to be the curmudgeon, but shouldn’t Adobe fix After Effects before they go breaking C4D?

Todd Kopriva: I think that you may be misunderstanding the announcement. We’ll keep making After Effects, and they’ll keep making Cinema 4D, but we’ll be doing this in a “strategic alliance” in which we work together on ways to make these things work better together. We won’t be breaking Cinema 4D.

Well, at least that much. That should put anyone’s mind at ease that Cinema might end up being Photoshop‘s new 3D panel. Of course that in itself does not mean anything, so let’s have a bit fun exploring what we could possibly expect:

  • There will be some good deals for you. Yepp, I’m positive we all should be prepared to get some sweet deal offered at some point to buy a complementary license of whatever program of the two we don’t have. Start saving the cash! ;-)
  • People have suggested that Adobe should include Cinema 4D in Creative Cloud, but I’m not sure it would be that good an idea. Let’s face it, “serious” 3D is a whole different thing than having a few features to produce some shiny 3D text and introducing a fully fledged 3D app would be a major thing. Adobe would have to support it on many levels from sales to tech support to providing documentation for it on their servers and that alone costs money. If it comes at all, it may take a while. In the meantime, getting a MSA would be the most efficient way of getting your hands on Cinema 4D.

Moving on to the technical side, let me summarize a few things I have sprinkled across various forums. Let me start by saying that strange as it may sound, I rarely ever actually use the 3D export from Cinema 4D, so as great as it may be, my priorities are actually different. If you ask, here’s why:

  • Doing technical visualizations a lot I regularly push around millions of polygons in a scene with complex hierarchies. There’s simply no feasible way to export most of that in a way it can be reasonably handled in After Effects, even if it’s just Null objects. You would also have to bake position data and keyframes, which may add an extra step.
  • For similar reasons, exporting actual geometry to use with Element 3D, AtomKraft or other plug-ins is rarely an option. You would have to spend a lot of time optimizing it just for the export and then still might end up with heavy geometry so even trivial stuff like using a cool looking cog-wheel for some background animation or logo often don’t work. It’s usually easier to do it directly in the 3D program then.
  • Most exports from Cinema 4D are troublesome. The OBJ exporter is rudimentary and I haven’t had much luck with FBX files, either.
  • Hierarchies based on CAD data can get pretty complex and I often find that it just takes too much time to pick out a specific Null. It may sound lame, but often it’s easier to hand animate a 3D layer to put an arrow in your scene that points to a critical part as is often required for instructional videos or promotional films pointing out specific highlights.
  • I find it complicated to assign object buffers and manage render passes. There’s always something you miss and then your overlaps and obscuration don’t look right. For that reason I often even rely on just using the combined “Beauty” pass because it’s easier just to re-render it as a whole than patching up things bit by bit and just the same, the render times on those heavy scenes are intense no matter what. You do not necessarily save rendering time by splitting things up.

Now based on these peeves let me try to describe a simple workflow that would make my day.

  1. I render a sample image, but all the buffers are retained with maximum depth and data in Cinema 4D.
  2. I switch to After Effects. There I fine-tune the appearance, both on the final output and the materials in the context of other effects.
  3. Based on e.g. a normal buffer or a point cloud I can also adjust lights.
  4. I render my scene again in Cinema 4D, now with all the adjustments.
  5. Because I have UVW info, I can now replace textures without re-rendering everything.
  6. Based on some geometry buffer, I can pick out polygons or points, right-click and let After Effects create a Null object or whatever to place some stuff on it.
  7. I can use a combined material and geometry buffer to pick out items in full rendered quality to create additional object or whatever buffers.
  8. I can use a live depth buffer in combination with the geometry to place other effects in the scene and have correct obscuration.
  9. I now render the full scene in all its glory and let the data be substituted in my already existing compositing project.
  10. Aside from some final adjustments, I’m pretty much done.

While this may sound like stating some pretty obvious things and in fact is a quite conservative approach, it would already solve a ton of problems. I could perfectly live with normal image sequences as a result and would not even require so-called deep compositing. Of course that again would add another level, too, but currently it seems you trade long 3D render times for humongous file sizes that cost lots of performance in compositing, so to me it’s only the next step after the initial implementation. Once Cinema and After Effects can talk to one another bidirectionally, also other options may unfold.

  • We could navigate in 3D scenes and restructure them in both programs and the would mutually update. When you rename a layer that represents an object, that object gets renamed as well and of course we would want position and animation data to be updated for objects, cameras and lights as well.
  • We could use 3D models without any import/ export or other data conversions. That should be handled quietly using Melange or another yet to develop shared data format. Plug-ins could simply reference the scene in a pop-up and Cinema 4D would do the hard work and provide e.g. point positions of vertices or particles on the fly.
  • Similarly, a shared method for exchanging baked data like point clouds would be required. People may want to render Thinking Particles using Particular or similar. That for all intents and purposes could be Alembic, but it should not require any user intervention, either.
  • In a reverse workflow to the previous two points, it might be nice to non-destructively use compositions as textures in the 3D program.
  • By using After Effects‘ code base, Cinema 4D could get better access to Photoshop and Illustrator files as well.
  • For all of the previous to work, apparently the programs need to have a shared resource management so they don’t compete for processor and memory usage. May sound scary, but yes, some form of DynamicLink will be necessary. Likewise, they’ll have to have shared caches and access to the same project folders.
  • The render functions in both programs could be unified by creating a shared render settings and render queue architecture. it then wouldn’t matter where you set it up, you get access to more formats and consistency is assured because changes in compositing would ripple through to the 3D scene and vice versa.

Now all of this sounds like the future is very bright, but of course getting all of this to work will take time in both programs. Rather than one team just using existing bits from the other app, this will be a longer process of convergence to unify programming styles, API function calls and all the other geek stuff under the hood. And not to be nitpick-y: Currently Cinema 4D doesn’t have any form of reasonably interactive rendering at all, so Maxon need to do some groundwork first.

Advertisements
%d bloggers like this: