HappyPolygon

I'd love to see something like this because for the past 20 years there hasn't been any new book about computer graphics that would go beyond the antialiasing algorithm in 2D graphics and phong shading in 3D graphics. Actually the most advanced 3D book I've encountered was about light mapping (I don't want to go that far without first learning about Sub-surface scattering or other geometry/structural methods). Anything else is prehistoric algorithms. All modern structural methods concerning UV mapping and animation are hidden in scientific papers without the well digested contents you would find in a book. Even old-school 90s effects like watercolor and pastel are not present in signal processing/2D graphics books I've encountered, they talk about edge detection and blur and then you've reached the end of the book. I'm still looking for the title of that Maxon book about noises (I hope it's not only in German) and I still don't know the technology behind ZBrush's polygon encoding... To be fair I've seen a lot of modern GC books but they were about real-time graphics that I'm not really fond of (I thing it was a series of 3-5 books called CG Gems or something like this).

Very happy to hear about that Srek. Do you use the old hack of exporting as alembic to make the SceneNode object appear in the Object Manager or am I missing a new way to do that ? I noticed you output to a port while I have that node to communicate to the viewport ... Unless the Tool Hole is a node group connected to other nodes not visible in the screenshot. Oh wait, you made a Capsule Object right ? Will this be the standard way to have a manifestation of the generated geometry ? (so far my use of capsules is limited to making them children of objects and use them as modifiers and when I hear about capsules my mind goes straight to that kind of use)

I think I have mentioned to Srek on CGTalk, as soon as Scene Nodes were first introduced, in the context of a productive criticism, the importance of good, user friendly nodal system that can support Kit Bashing. I'm not sure if I used correctly the term as many artists refer to it as placement of small pre-modelled parts, where I had in mind a mid-sized scale kit bashing between that and the products of https://kitbash3d.com/ . Essentially you have the finalized models or modular parts that when procedurally placed using a constraint-and-condition-oriented system can produce unlimited variety of style-consistent, scalable models for scenes. My example of what the final result of a system like that might look like was a screenshot from the sandbox game called Townscaper. I don't know how much they considered my suggestion due to its MoGraph resemblance. I don't know how feasible my suggestion was as it was too specific as a use case and too game-engine-like of a feature. I don't know even if it is already possible with the R25 version as I struggle with simple forms. As for the texturing... AFAIK, Blender uses procedural texturing on node-generated models. As for the fully node-based Houdini, I'm not sure even that uses nodes to UV unwrap and edit UV maps, maybe Igor knows more about this. I think the main reason for not having many tutorials for really innovative node structures is the same for not having many tutorials on how to code for C4D or other DCCs. I think it's only because of the nature of "experimentation" that these two share and the fact that those involved in these areas are far less than those using the standard modelling options. I've never seen a tutorial about how to program a fully functional calculator with GUI elements. Same applies for nodes. The factors contributing to this are: It takes a lot of time. The final video is going to be very lengthy, and the time required by the author to edit the video for final release is multiple of the video recording time. It needs a good tutor. It takes some serious skills to be able to have a descent tutorial with all the appropriate annotations and guidance from the narrator. It needs preparation. The bigger the project the more complicated it becomes, the more complicated it is the harder it is to replicate in a live manner. For the same reason it didn't got solved in one day, it cannot be deconstructed, analyzed and presented in much less time. It is a new "technology". Sometimes the area of expertise is fairly new and the amount of people involved in it are not that many in order for "tutors" to emerge from them. It's complicated. A tutorial is something that people may refer to as a video or a small series of videos that may not exceed 30 minutes. If a series of videos is comprised of 3 or more 45-60 minutes parts, it is then considered a seminar and can easily fit in the academic realm as a full course. Do you think you could do a full tutorial for people to teach them how you made one of your 2000-lines-of-code plugins ? And if you can, for what kind of expertise-level audience will it be for ? People that just started coding ? People that already familiar with the programming language ? People that are familiar with the C4D SDK ? (rhetoric questions) Let's take a recent "technology" from the IT field. DOCKER. How long did it take for people to start publishing books on how to properly use it ? It took 5 years for it to start appearing in post graduate courses. And to be fair, DOCKER is a quite robust platform. On the other hand Geometry Nodes have been around for two years and Scene Nodes for one. And both are still on development with many additions and changes being introduced quite often. So making in-depth tutorials with them that may not apply in the next application update or soon major release is not productive or profitable. To wrap up, what you ask is not application agnostic, as each company may develop its tools as they seem fit, and as you have already figured out by yourself the algorithms involved to do what you want are not existent. The main reason for that, I think, is the intertwined paradigms (classic and virtual mediums) of how thing should get done. Without the use of a computer the artist first builds the base model, then adds details, then finally paints it. These steps never changed priority. So when the tools became digital the developers kept that sequence and adapted everything to be dependent on that (except rigging, in classic mediums the rig precedes modelling, that's one part where the paradigms get intertwined). Nodes do not operate in a different way from the rest of the app, it's just an automated way to work with the classic tools, so far they seem to be the "macros" of the old apps like Excel. It may offer some low-level control but it is still limited by the internal procedures on how they handle information. If I understand correctly, you want to be able to delete a polygon assigned to a UV map, create a new polygon and assign the now lost UV coordinates to the new one ? A very raw description of an idea that can result in having a UVed human model change from quad topology to hex while maintaining the correct texturing, is that correct ?

Is Chris member of this forum ?

WOW, that OpenVDB technology is wild. I wonder if they are planning on using that only for preview purposes though.. I always rooted for a hybrid renderer that would automatically composite elements of Real Time Graphics and standard Ray Tracing... I've never heard of any renderer of that kind though...

What kind of shimmering is that ? The "Naaaaaaaaaaaaaaaaaaaaants ingonyaaaaaaaaaaaaaaaama-bagiiiiiiiiiiiithi baba !" kind ? sunset.c4dsunset.c4d

I'm glad you've figured it out. Can't wait to see the final project some day.

Train ? The engine is an AI ?

Maybe the Time parameter is not for the effect you want. Have you tried the Delay Effector + a moving Field or a Field with oscillation for a repeated effect ?

I don't see why this isn't possible... all I see in the plugin is a unified UI for strength per applied modifier like the cloner has for the effectors. Easy fixed with Xpresso.

What does U-Render have to do with exporting ? C4D already supports glTF export.

The example in the manual is this: I want the Scaled Blue-Noise and I don't care about the color. Why my setup doesn't work at all ?

Am I the only one who doesn't bother with the new layout because I use the old layout ?

What about the memory ? I've also noticed when C4D hangs there is little CPU consumption. On one hand I'm glad the CPU is low because if it wasn't it would probably never launch the task manager to kill the app. On the other hand it could mean that C4D didn't hang because of the enormous amount of processing load but because it "lost it's way" and caught in a trivial loop.

I don't have the R23 installed so I'm going to step you through it using the R25 I don't have RedShift, I'm just showing you how to break a VDB in two. I did not have any spare VDB files so I made one myself The Cube is set in Subtract mode I selected the Volume Builder Object and go to Export -> VDB I then imported the VDB object and made it a child of an other default Volume Builder I just decided to make the first piece to be the central blob Then I duplicated the "Part 2" And now you have two pieces of the same object You just have to export those two parts as VDBs selecting the Volume Builder, not the Volume Mesher. Re-Import them and color them in Redshift. You can use a Random Field or Shader Field instead of my sphere to achieve more complicated results. As the number of pieces increases the more complicated the procedure gets... Here's the project file and the VDB for you to relink to it cloud.c4d cloud.vdb

You can use the VDB in the C4D under the Volume builder and use any field or geometry to use as a mask and occlude the part you want to isolate. Use that part to partially color your cloud and use the same method to isolate the other half of your cloud.

I'm gonna go on Cerbera mode and say: Welcome to the Core 🙂 please fill your profile info so we can see all your settings and find the answer much faster. You can use the Collision Deformer. You are getting the scene in R25 collision.c4d

Can someone post screenshots from the scene file revealing hidden angles and how the scene transforms in key moments of the demo ? I would do it myself but I don't have Unity and don't know how to operate it... Some wireframe screenshots of the character and environment details, show the outdoor environment (is it just an image?) are also welcomed. I would also like to know if the floor opening hutch effect is a bool.

This demo showcases their new Adaptive Probe Volumes. I guess this is what we see casting shadows on her but I could be wrong. Here's the link with their analysis and some way of getting your hands on the scene file. https://blog.unity.com/news/introducing-enemies-the-latest-evolution-in-high-fidelity-digital-humans-from-unity

You mean project it ? The effectors have the ability to act as deformers on objects outside the scope of MoGraph generators. You just have to set it in the appropriate mode Unless you refer to an AE access on individual clones here's the scene with the cubes un-instanced examplescene 01.c4d

I love this model, can I keep it ?

Something's fishy about this... It's like the whole desk and the character are from an other take and then composited with the environment. The elements that I found to back my statement are: lack of perspective (it's like the camera is only focally zooming to the point of interest that would make the space between the camera and the point of interest small enough to adequately fit in the rolling environment , not that it could really pose any problems) Lighting doesn't seem consistent with the environment (again possible to exclude other objects from being affected by light sources) She really does look like a cut-out from the background (too much feather mask on her hair + the CRT screen effect on the background) I'm not saying that everything is not CG, it just hard for me understand how this whole scene is set in one single Unity project and rendered in one go.

My first try would be to try a spline wrap but that didn't go as planed ... I think I've figured out a way but because I'm a noob at character rigging it didn't work for me either, you might figure it out. Draw a spline at the shape of the model you want to straighten (use one of the orthogonal views) if the model is coiling in all three axis you'll have to use an edge selection and trace the path with the loop selection toot deselecting the extra lines that return back to origin. Then make a spline out of the edge and try to center it in the center of the volume of the model use a Sweep with a circle to the spline you drew scale the circle to make sure it covers all parts of your original model here's the clever part - use the Mesh deformer. Make the deformer a child of the original model and drop the Sweep in the list of the deformer make a copy of the spline you drew and convert it to a Joint Chain Now you are on your own 'cause I couldn't bind the Joint Chain to my Sweep, if you do that you will be able to straighten your model by simply re-positioning the joint chain really easy if you set it to IK. Just press Current State to Object and you have it. maybe the Mesh Deformer part is not necessary and it can work just with bones, I don't know.

That's what I first thought but the description was too complicated for such simple design, and given the rank of OP I didn't think he couldn't figure a solution like that....

You've totally lost me. any references or a simple sketch ?