Finished a spot for a new show on History called Invention USA. Had a short window of time to come up with the concept and go to finish, which I do admit consists mainly of a bunch of things that Cinema 4D does well (and fast), namely stuff falling on a floor and bouncing around. A little lazy perhaps on the concept side, but it was one of those situations where I just needed to come up with something fast, in the midst of doing 12 other things, and see how it flies. Client liked the concept so I went with it.

I gotta say though, when I'm in a tight bind, C4D always comes through for me. Setting up the dynamics required minimal effort. It just works. And the rendering was fast as well. I started off meaning to use GI and also tested the scene with the new R13 physical renderer, but because time was an issue (needed to go straight to finish in a very short amount of time) I went with the standard renderer. The still above was rendered with the new physical renderer, however the final shot (not shown here) didn't really look much different in the end and rendered in a fraction of the time. Hope to eventually post the spot. Some style frames also included below.

I am currently getting ~8 min per frame rendering at SD widescreen. Only rendered a short sequence (130 frames) so I looped it a few times.

SSS was turned OFF as it was taking ages for the irradiance thing to cache (this was with optimized SSS settings).

I have 1 bounce for self illumination, AO on, motion blur and DOF on.

At the moment it looks like I still have some more optimization options to explore. Render times feel too heavy at the moment for long-form medical animation (usually between 4 to 10 minutes of HD animation) but maybe Ok for short shots for spots and such. We’ll see.

EDIT 12/01/2011

Continuing my experimentation. Removed sub-poly displacement and put SSS back in. Reduced settings significantly. see below)

Got much faster render times (~2min/frame). Still needs refining. Lighting is all HDRI now so could use a bit more directional light. Turned on reflections as well instead of using an HDRI environment map for fake reflections. Also added a second layer of phospholipids (now its a proper bilayer) and threw a couple of proteins into the dynamics group.

What started out as a simple test to explain camera-driven shader effects in Cinema 4D to a friend in need, turned into, well I don’t know, THIS I guess. Staring at that Optical Flare in the hazy distance makes me ponder the meaning of existence. Then I just click on a cat video and I’m all better.

But I digress. I’m using the gradient shader in “3D spherical mode” set to camera space applied to cloned spheres. As objects come closer to the camera, they glow. Objects in the distance are dark. A nice way to fake fog in your scenes (and a lot of other things depending on which channels you use it in)

The radius value (below) determines the length of the  gradient effect (left to right, left being the camera which is at the center of the “virtual” sphere). I copied this shader setting into my diffuse channel to darken the objects in the distance.

I’ve done some post work on the clip above which kind of makes the camera falloff luminance harder to spot, but hopefully it makes sense. I got a little carried away.

I came upon this technique maybe a decade ago as described by CG/VFX artist Richard Morris (aka Jackals Forge of Gallerie Abominate fame) on his  website documenting his cg production process for the science documentary BodyStory 2 for BBC. Make sure you check out his work. Although his website is over a decade old, the work still looks great and the information can still very useful to those in the biomedical visualization field. The relevant information can be found in the “Dynamic Shading” section of the article.

Jackal’s Forge BodyStory 2 Tech. Click here for website/article

I was inspired by Maxon’s marketing image for R13 of the grapes by Marijn Raeven, which I’m sure many of you have already seen. Didn’t want to waste time modeling anything, so I did the next best (i.e. easiest) thing, and cloned a bunch of dynamic spheres and applied SSS. 13 minutes later I had this image.

Over the next few months I’ll see if I can reign in the times and quality to see if its feasible to use physical rendering in animation production, but in the meantime I think I will be sticking with R12.

Here is the first image I created. Its a model of a trichinomas (yucchy parasite). I chose it as I thought it would be a good object to test the SSS due to its general blobbyness. Its a starting point. Still a ways to go…

I always wanted to try to see if I could pull off a decent looking human eye render, and recently had the chance for an animation project which gives an overview of a particular disease of the cornea—the clear outer layer of the eye. I started with a Daz3d model and a high res map that I purchased for it. The eye needed to blink, so I exported 2 states of the eye from Poser; one opened, one closed, then imported both models using the plugin Riptide, which is useful for importing and exporting .obj files to and from Cinema 4D. I used the morph tag to keyframe the blinks, by linking the 2 states of the model into the tag.

I found I needed to re-build the eye from scratch in order to get the level of detail I needed for the closeups in the animation. I was able to take the high res maps of the eye that came with the maps I purchased, and edit them in photoshop to get more detail and a displacement map as well as an AO map out of them. I also needed to animate the disease state of the eye (not shown here) gradually developing on the cornea. Once the eye ball was modeled, I was able to rig the pupil and iris to dilate, again using the morph tag.

The model I was working with had a really low res solution to the eye lashes, just a few overlapping planes with an alpha-channel lash map on it that broke up when you got too close. I fixed this by modeling a lash and cloning it along an extracted path from the edge of the eyelids via the mograph cloners. I used the random and step effectors to make the lashes look less regular and scale down across the eye.

I also used the chanlum shader and vreel skin in the luminance shader of the eye to get some sub-surface translucency.

Although there are some issues, like the skin map being a bit too low res at super close distances, and not being able to control the reflections on the cornea as much as I’d like, I was quite happy with the results.

A few years ago I got to work on an episodic TV project called Factory Floor. You may have already seen my other posts about it earlier on in my blog. Yeah, I know, I’m milking it. It was not a particularly successful series (it only lasted one season) but looking back it was really a great experience overall, and I learned a lot about what its like to be involved in a TV series production. There were ups and downs, funny stories, sleepless nights, stress, and ultimately, a feeling of relieved satisfaction bordering on triumph when we delivered the last approved graphic call after a roughly nine-month period of working. I still marvel at how I and my then business partner R. Scott Purcell, our producer Bob Larkin, (and the occasional 1 or 2 extra freelancers) were able to turn around 100 graphics calls while jumping through all the usual hoops of a production schedule.

The image above is the beauty pass render for a piece of machinery called a pencil extruder. It squeezes out bits of lead that will be inserted into the pencil. Below is what the final call looked like after compositing (and thats CAULK GUN by the way. Get your head out of the gutter.;)

One of the things I really liked about this project was that it required me to QUICKLY build and rig dozens of mechanical objects, mostly factory machinery, but sometimes the odd item like cheese curds, toilet paper or potato chips. I’d get a script, and usually a rough cut of a given episode, and from the low res quicktime footage, as well as google picture reference, I’d need to knock out various models in a short amount of time. I’d usually spend no more than 5 or 6 hours a night researching, building a model, then get started on roughing out the animation timed to the corresponding piece of the rough cut. Each night I’d shoot to have a shot ready to go. After a few weeks, (and then months), the process was like second nature. Research, build, animate.

The style of the graphics was to resemble a sort of moving 3D blueprint. We developed a workflow right from the beginning in which we had a master Cinema 4D file which included all the shaders and lighting we would need in every scene.

The rendering and compositing workflow would go like this:

CINEMA 4D
1. Beauty Pass. We would apply realistic shaders and textures to our models with a decent lighting setup and render this pass usually with ambient occlusion if rendering time allowed. We had a set of metallic shaders we would reuse again and again for the various pieces of machinery.

2. X-RAY PASS. We also developed a set of x-ray shaders using a variety of colored fresnel (angle of incidence) gradients loaded into the alpha channel of the shader. We would take our master project of the given scene and then apply these shaders and render that pass.

(below , the X-ray pass for the lead extruder call, a fire extinguisher, and a taser X-ray  right below that)

3. LINE ART PASS. We heavily utilized the Cinema 4D module Sketch and Toon to generate a bluish line-art pass on white. Although the S&T render method can give a big hit to your render times, we found that setting S&T quality to low and turning off AA sped up rendering times considerably and still looked great. Sometimes we would create custom overshoot and perspective lines and render these in yet another line art pass.

4. BUFFERS. Each animation would require us to hold and highlight a specific part of the object or machinery. When this occurred we would need to glow each part. To help with this we rendered object buffers for each part of the model. If a piece of the model was obscured in some way, like for instance, a battery or valve inside the main object we would render those pieces out separately in their own scene file.

COMPOSITING
Scott developed a a template in After Effects from which we would create each and every shot from. It included a blueprint overlay, background textures, labels and leader lines as well as a layered workflow to style the renders into that blueprint look which would need to mix with the textured background. Scott used the linear light layer pass which yielded the perfect look. We would also use feathered masks to accentuate certain areas, and bring out more of the beauty pass, line pass or X-ray pass as needed.

1. We would typically drop the beauty pass in first then apply the layer effects to mix it with the background texture.
2. Then we would bring in the x-ray pass, then mask out areas as needed, and fade it back considerably.
3. We would drop the line pass in and mix it a bit, set to multiply.

Below is an example of a typical AE Comp.

In the end the final look was nice, and completely in line with what was appropriate for the show, as well as being exactly what the producers wanted. I’ve recently gone back through the files, remembering how much time went into building them all. I thought Id share some of the pre-processed beauty passes and showcase them a bit, as much of the detail gets lost in those original blueprint renders. These are by no means examples of rendering excellence. No global illumination or linear workflow, just some simple down and dirty 3-point lighting and metallic shaders with a touch of AO.

I may continue to add more images to this post, so check back. I have tons of them.
Thanks for looking!

I think the GI looking text on the reflective cyc might be a bit overdone these days– so easy to do.

For reference…

Never finished it completely–the boots were meant to be snowboard boots (Its Danté, the Snowboarding  Demon from Hell!), textures were never  refined and sub surface scattering’s a bit blown out. There are some anatomical problems too. Blah blah blah…

But thought I’d share anyway–just to show that it’s not always microbes and slime for me.

Took it a little further….

Special thanks to C4D-ers Dominic Faraway and Brett Morris for their feedback and assistance.

You can grab it here:

MICROFLOATIES

This zip file contains the Microfloaties C4D project file as well as a library file for use with your Content Browser in Cinema. If you wish to  keep Microfloaties handy in your Content Browser, just drag the microfloaties.lib4d file into your Browser folder found within Maxon>Preferences>Library folders.

Here is a tutorial to help get you started.

Enjoy!

EDIT: JULY 11th

To all who have been asking about adding functionality to allow the change of materials and cloner objects… Yes, I do intend to look into  adding these functions to a future version of Microfloaties. For now, it is what it is