These look great, and complex! Yes, I belted the entire hood onto the character’s head, since it kept flopping around or slipping off during the simulation. I could probably use a combo of partial belting and maybe a vertex map to make it a little stiffer than the rest of the cloak, and I did try that, but again it was pretty tricky, and kind of not worth it in the end, since I didn’t want the hood flopping onto the character’s face anyway.
A second machine would be handy, and sometimes I think–since I replaced the CPU, GPU and motherboard when I last upgraded–that maybe I should have just built a whole new machine, but alas, maybe next upgrade.
I have the same problem with “Adventure Male”, especially since the character’s motion is faster and more erratic than a simple walk cycle. However, as far as I’m concerned, it’s not just a technical question (i.e. how to do keep the hood on the head) but also a matter of realism. In real life, there would be friction between the head/hair of a person and the material of the hood but, depending on the person’s movement, I would expect the hood to slip off at some time. Since your character (from “Quest of the Key”?) seems somewhat stylized, it is probably easier to get away with a “belted” hood than if the goal is as much realism as possible. Also, while I may notice that the hood is not dynamic, most regular viewers probably would not.
As far as additional machines are concerned, I more or less consider those a necessity for 3D. For more complex simulations and with render times reaching 2 or 3 minutes per frame, the machine doing the simulating/rendering will be tied up for quite some time; and, in my experience, it is not a good idea to use it for other productive work while it’s busy doing that.
Yes, a certain amount of animation is what you can get away with LOL
As for two machines, I can usually render overnight, so it’s not usually a huge problem, tho there are times when I have a longer/more complex sequence that takes much longer. I recently had a render that took 35 hours, so had to stop it to work, and restart it again later, so that was pretty annoying.
Indeed multiple machines is an absolute must
for a major animation projects because you can commit to long render times on your most powerful PC (sometimes days)
and not only have a second machine for setting up the next scene etc. but even if it is just for passive “entertainment” activities
like watching youtube videos or visiting your favorite forums etc while you are rendering.
At one point I had 5 machines running
but since I recently migrated over to 2D animation and comics I took my two older windows laptops out of service and only use My main 3 year old Dell tower for Cartoon Animator and Adobe character animator and MOHO pro.
And My old Imac for mostly internet
and My adobe Suite for traditional 2D comic & Graphic Novel stuff
and an old HP all in one running Linux Mint just …because.
Yes, overnight renders are par for the course. However, I have found that I need to check render progress every few hours because, for some reason I have not been able to figure out, C4D decided to stop rendering (no biggie since I always render image sequences and can just continue rendering, but I don’t like it when the machine is on all night not doing anything productive).
In the old days, when I was using Physical renderer in C4D, I would sometimes have 5 or more machines render one project via Team Render (1 C4D Studio license came with an unlimited number of licenses for render clients). With Redshift that does not work the way it used to with Physical (since each machines needs its own Redshift license).
These days, if I need to distribute a render job to two machines, I do it manually on each machine and don’t use Team Render (e.g. have machine A render frames 1-500 and machine B 501-1000).
For (home) office work, surfing the internet, and so on, I also have a number of additional (old) PCs, notebooks, and mini PCs, so I don’t really need my 3D machines for that (unless, of course, it is for licensing purposes, e.g. when purchasing something from the RL stores).
Computers do seem to proliferate LOL. The only other machine I regularly use is an ancient PC laptop that I use for Optitrack mocap. Still using a very old version of the software, but it does the job. I think the newer version of the Optitrack software has more bells and whistles, but it would be $5k to upgrade!
True.
It has been years since I last sold or otherwise disposed of a PC/laptop no longer in active use (I just keep them around for a rainy day). What is probably my oldest machine dates from 2011 and only gets used a couple of times a year.
I have an old gateway laptop from around that era still running windows 7 and Iclone 5.5.
The display is showing clear signs of imminent failure though, which is why I took it “out of service”
I rendered my entire 90 minute Marvel fan film “Galactus Rising” in Maxon C4D R11.5
on a 2007 Macbook with only 2 gigs of ram.
(it’s now in a landfill in upstate New York)
I hear ya, 6 years is about as long as my first Star Trek film took. The problem is that as the computers and software get faster, you keep bogging them down with cool new effects and capabilities, so you don’t get a lot faster.
Actually, while we’re talking C4D cloth sim, can somebody explain why cloth only seems to work for me if I increase my characters and the cloth item by about 10 times the size in C4D than what I normally render them at. Am I missing a setting or preference somewhere?
Possibly. There is a setting for scale in the Project Settings–>Simulation–>Scene–>Scene Scale. For me the setting is usually 100 cm, because my project is set to cm. (You may want to read up on Scene Scale in the manual.)
While I never had to change that for cloth simulations, I may adjust the Scene Scale to be more in line with the size of the simulated object when simulating small things (like a chain with dynamic links that are relatively small; at the standard Scene Scale the chains would break because the collision detection would not be precise enough).
My machine from 2011 runs Windows 10 Pro (I must have updated Windows at some point; I think it originally came with Windows 7 Ultimate). At the time, it was a pretty fancy rig that was custom built for me with two 6-Core Xeon CPUs for a total of 12 cores/24 threads, two GPUs (GTX 460 with a measly 1 GB of VRAM), 48 GB of RAM, about 15 TB worth of HDD storage, and a 1,500W PSU.
Since there was no GPU rendering at the time and I didn’t game (still don’t), I had not opted for one of the higher-end GTX GPUs available in 2010/11.
Although the percentage will vary from program to program,
apparently the scale does matter quite a bit with most cloth systems.
I know Blender recommends
you scale up your scenes to 5 times the default scale at which Blender launches for cloth sims.
So it makes sense that you are needing a certain scale for optimal results in C4D as well.
Oh, okay. Scale is something I’ve never even thought about in the literal decades I’ve been using C4D, but I guess I’d better start. It never really occurred to me since it’s all virtual anyway, but I guess it kinda makes sense if you’re trying to simulate the real world. Thanks! I’ll look into this.
Well, it makes no sense to me to scale the entire project scene to get the appropriate simulation results; I’d prefer a “scale factor” for the simulation alone because I want to be able to keep all the models, environments, etc. at real-world scale (it also makes it easier to combine assets from various sources). If you scale your entire project scene, you also get different results from your cameras, etc. that may not be desirable.
Also, AFAIK you can have several simulation scenes (via the Simulation Scene Object) with different parameters in one C4D project scene. That way you could simulate garments at one scale and something else at a different scale as long as the simulated objects do not need to interact (if they do, they need to be part of the same “simulation scene”). I have not yet tried using different “simulation scenes” in one project myself, so, at this time, it’s what I understand and not what I have actually tried. Anyway, this is what the C4D manual says:
“The simulation scene object can be found in the main menu under “Simulate”. It has exactly the same settings as the Simulation project settings. So what now? Imagine you have the same scene in two separate simulations. After doing a lot of fine-tuning you achieve the desired result for simulation 1 with 82 Substeps. So far so good: You achieve the desired result for simulation 2 using 112 Substeps. If you had a common default you would not get the overall result you want. The solution: 2 separate Simulation Scene objects. One for each simulation with respective settings.”
Not just with cloth but also with “fluids”: water in a coffee cup behaves differently than in bath tub or the ocean; cigarette smoke differently than the smoke from a large explosion. So, yeah, proper scale is important to get physically realistic results but also so that the simulation can work in the first place (for voxel based simulations, the size/resolution of the voxel grid is also important).
As I understand it, the Simulation Scene Object makes it possible to have the cigarette smoke and the explosion in the same project (at this time, liquid/fluid simulations in C4D are not native (but via plugin such as X-Particles) and therefore not part of the C4D Simulation Scene, but – if I remember correctly – X-Particles offers a similar functionality).
I don’t know how Blender, Maya, etc. handle this scaling issue for simulations, but messing with the overall project scale would not be ideal in my opinion.
Well, it makes no sense to me to scale the entire project scene to get the appropriate simulation results; I’d prefer a “scale factor” for the simulation alone because I want to be able to keep all the models, environments, etc. at real-world scale (it also makes it easier to combine assets from various sources). If you scale your entire project scene, you also get different results from your cameras, etc. that may not be desirable.
Actually I should have wrote: scale up the cloth ,and any collision items, in the scene
by a factor of 5 not the entire scene.
This is a simple matter of putting the items under an “empty”
(Blender’s equivalent to C4D’s “null” item)
and scaling the empty,
After the cloth sim is baked to a cache
Just remove the empty and the baked cloth sim and collision item return to original scale.
Frankly I don’t use Blenders underdeveloped cloth system for animated Characters, As I have Both Maya’s Ncloth
and an older perpetual license of marvelous Designer 10.
Seems like a somewhat strange procedure to me but if it works.
I recently got MD to more easily re-mesh and export purchased MD garments. So far I have not really delved into MD’s simulation capabilities (C4D works for me in that regard); I also don’t know if MD could consider other objects besides the avatar and the garments in the simulation. Even if MD can, I would still need to export the sim as an Alembic for use in C4D…
Yeah that is part of the problem
with doing effects/simulations outside the main render app where you have to interact with the environment.
Moving character animation data between apps is easy.
But it becomes terribly complicated when you
use external solutions for simulation of cloth, fluids dynamics or even ragdoll
IMHO cloth, smoke and pyro, fluid/rigid body sims etc are hardly worth the effort if you cant do it all in the same software/scene.
Its kind of hard to believe that the vestigial poser software was actually doing rigid body cloth and ragdoll with same a python plugin nearly two decades ago.
Agreed.
Except for fluid sims, the rest (cloth, rope, fire & smoke, soft-body, rigid-body) can already be done natively in C4D (and certain fluid effects can be “faked” with the C4D particle system); the C4D hair sim, while native, does not yet play completely nicely with the unified physics system and I do hope that (real) fluid simulation will also be added natively in the not too distant future.
Interestingly, X-Particles can also do most of these sims plus fluids, so as a C4D user you have a choice which tool to use.
Just because software is “old” does not necessarily make it incapable.
The problem is that “old” software is often no longer supported or compatible with more recent OS and hardware. These days, I expect sims to run on GPUs; I kind of doubt that Poser could do that.
