Water maybe, but they avoided it for hair by using the classic ‘Blender’ solution and put the hair in a single pony tail. So that would have just been a bone chain with likely some automated movement (maybe via geonodes) and tweak controls for any quick manual corrections.
Is it a classic ‘Blender’ solution or is it really something everyone does, if pothentially in a different form?
Didn’t even Framestore do it?
Think about it: Gravity had a female lead (Sandra Bullock), and allmost exclusively (often fairly long) zero-g shots.
At the same time Bullock had her hair cut fairly short in the movie:
(Image from fxguide.com)
Now I’m not saying short hair on (actual) female space-travelers is unheard of (and personally I imagine it has its advantages from a practical standpoint), e.g. Samantha Christoforetti:

That said, nothing hints at any policy or obligation to cut one’s hair short prior to an actual spaceflight:
Plenty of example pictures of actual long(ish) female hair of actual space-travelers under zero-g conditions from Wikimedia Commons
.









https://upload.wikimedia.org/wikipedia/commons/thumb/0/09/


.
After all, yes, I get you’d rather they jumped head-on into it and developed a perfect, production-ready full-blown dynamic hair-simulation in the course of project Gold. I think you’ve made this much clear already, repeatedly.
And sure, we would all have loved it.
But labeling it a Blender solution, as if everyone else did the full-blown hairsims and only Blender/Blender Studio didn’t, well, I guess not.
edit: OK, stupid forum-software, if you can’t make your own features play well with each other, I’ll put all the images on your server.
greetings, Kologe
No, I labeled it the Blender solution, because for Blender it basically is the solution, repeated time and time again.
Now sure, others may well do much the same thing rather then full-blown hair sims, but that could be due to artistic style, budget or time limits. But with Blender none of that matters, since its just not an option.
Hence it went from full free flowing in early concept art:

To very quickly being just about all platted together in near final model:
Thus removing any need for development work on simulations.
For example,
Now, not sure about you, but I think I’m seeing a pattern.
That very wild painterly style might have been possible if they stopped trying to use a hair sim to get a painted bundle look. I paint, and I don’t paint every hair strand nor do I use Blender’s hair particles to reference - I paint what amounts to jumbles of meta balls and then begin joining and dividing. I think the same solution used for animating the vines and tendrils and Proog’s coat in Elephants Dream could have been combined to the proxy mesh objects for the hair and then creative texturing and filtering could’ve gotten the rest of the way for that specific look there in the shot. This progression feels more like a sculptor/modeler got in their own way trying to be a painter.
Well they tried it in Cosmos Landromat, there is a lot of hair simulation there.
But anyway it’s non-trivial , first because you need to bend the simulation to stylized animation, so it not only needs to feel believable, but the shape of the character needs to be maintained and that should stay consistent when you have extreme changes in posing… Furthermore this also add a step to animation. Instead of having animators in full control of the shot it’s done in several steps and therefore each shots got more back and forth and takes more time to be visible and eventually signed-off.
This might also become a creative choice, as one might prefer a simpler animation of cloth/hair but with a lot of control rather than something automated that might feel too real and not precise enough to fit the particular style of the project.
Therefore, while it’s possible to make hair simulation we probably need more tools on top of that to be able to add creative control over the result. This is probably something that need good design and maturation and be thought a bit outside open movies;
So yeah sure, it’s not an option yet in blender, but it might not be the right time to start looking into that especially when there is a lot of other stuff to look into on each open-movies.
Exactly all that might easily apply to the Blender Studio all the same: Artistic style, budget or time limits.
And probably much more so than to a renounced FX-house like Framestore in the latter two cases.
Sure, I see what you mean by ‘a pattern’, and this much is true. But I think you seem to underestimate the technical difficulties involved here.
I’m anything but an expert in numerical simulation for VFX usecases, but I’m pretty sure hair simulation is simply one of the most tricky thereof.
Things like fluidsims and even cloth simulation are easier to manage, for they have more degrees of freedom in typical scenarios. All voxels or particles in a fluidsim act basically independently from each other and while this isn’t true for cloth, the latter is still typically constrained in an isotropic fashion.
Hair isn’t.
To me this makes it easy to see it’s a technically very challenging problem. Sure it can be solved, but if we’ll see a solution in Blender, I doubt it would be in-house developed during an Open Movie which immediately uses it.
More likely would be an Open Source library falls from the sky and gets integrated or such simulation capabilities get developed outside of a movie production first and then we talk about open, floating long hair in the next Open Movie.
greetings, Kologe
That’s what they did for the first open movie projects and it was clearly not optimal for many reasons.
I hope this can be done within simulation nodes, but with more sophisticated nodes for physics and something like a solver node which handles the complex computations.
That’s very possible, tho then we may end up back to some of the reasons (guess work) as to why the movie got cancelled. In that doing something like the above maybe didn’t fit with the directors vision or to make it all work together would require more development time in areas that at this stage may not even have a clear plan/design worked out.
Yes, my understanding is a lot of that was ‘experimental’ and a total nightmare that would fall apart/crash if you just looked at it the wrong way. As such, much of it never made it to a final stable release of Blender.
Yes, based on theses tests: Cloth Simulation- But With Nodes? - #5 by Xeofrios it seems the first thing we may need is sub-stepping in the sim nodes.
Cloth I’m not so sure about, but fluid, yes, since we already have a very viable solution in Flip Fluids. Not really any sort of equivalent addon for Cloth or Hair at this stage.
That would be nice, for hair and cloth, pretty sure such a library doesn’t exist at this stage.
Which new project will Blender studio work on ?
Maybe a Pet Projects Ep 2…
After all one of the main ideas behind that movie was as a ‘filler’ between main projects and the characters, etc are all done already.
The hair wasn’t simulated in Cosmos Landromat, if I remember correctly, the hair was animated using hair shape keys which was later released in the Gosseberry branch which was also later removed from Blender’s website.
I remember downloading the branch and accidentally deleting it. I never got that branch back.
Hair shape keys could be used to keyframe the hair underwater easily and maybe run a simulation over the animation?
Well I’m pretty sure it’s not the only tool they used, look at these shots :
Hairs/Fur on the two main character was using simulation IRRC , it’s not visible in the teaser but there is a shot where the guy put the collar on the sheep with a lot of interaction with the fur.
You may be correct but I remember reading they used shapekeys to keyframe and fake the fur movements of the sheep and the guy’s hair. That simulations was too time consuming or computer intensive or something like that (this was 8 years ago when hardware was not as good as now) and the results payed off.
I may be wrong though. They added the hair shape keys feature to the gooseberry branch and then for some reason they removed the branch. I can’t remember why. That shapekey hair feature has never returned back to Blender since then for some reason.
Is the project alive again? On Blender Studio’s website there are production logs coming out weekly, still.
On Twitter there is a glimpse of it, and visually it looks like the most beautiful production of the Studio yet. It’s such a pity they shelved it… but did they?
They announced in previous video that what they already did would be shared in upcoming weeks
I see, thank you. I’m glad that what they did is coming out, the artistic production behind it was great
Yeah, I remember this addon too. I think they did.
What were the weak points of this production? Can they work out on features needed for development for 4.3 and 5 release?
It was too complicated and complex. Needed would be a new physics system and some more which isnt coded in two blender versions.
Is project gold still going or when are the assets finished? They have not announced anything else since the we stop it video.




