It was an interesting watch, but it already has me rather concerned. While it looks great for still images and already has a number of nice improvements over the current system, I currently have a fear that it will end up suffering the same problems when it comes to animation.
I just feel things are going the wrong way around, in that animation/simulation/baking/art direction (manual control/adjustments) and interaction between both hair itself and hair with other objects should be worked out first.
There’s no point in having fantastic grooming tools if as soon as you try to move it half the hairs cut through the face or an ear. What does all the randomness matter if as soon as you apply simulation it just all falls straight down. While at the same time, if you have curly or wavy hair, you need it to still hold the shape but also move naturally.
Speaking of natural movement, depending on length/style/type of hair it just doesn’t move like cloth or liquid or particles, it has it own unique way of flowing and even if one does get it perfect via simulation, there are still times where one would then want to make some manual adjustments, which should be as easy as possible to do.
So yeah, I just fear none of that will be considered or tested till after the ‘new’ hair system is in place and then as such, it will just be a best one can easily do with the then new complex hair system that has been put in place.
My takeaway was that the focus on manual/destructive tools came first because the development is in parallel with the next blender movie project.
I’m guessing they needed the improved grooming tools first for developing the characters and the advanced simulation and complex node tools will come after.
O’ yeah I can totally see why its happening like it is and that is no doubt the reason. Just like I said, given that is the reason and the way its happening, I thus have my fears as outlined above.
Even with the particle hair tools we have had for ages now, it was always possible to add and create some really good looking hair on characters, the art gallery proves that. Just pretty much none of them have ever been animated. Lets just hope the advanced simulation tools have at least been considered as part of the process so far and do indeed come after (without too much of a wait).
Not bashing Blender team here. Just being realistic. I would temper my expectations here considerably. Based on public roadmap, current development focus and available resources I’d say there won’t be any animation available for hairs for at least next two years. Unless BI considerably change focus on this area and/or get a huge surge in funding from users and companies to higher more engineers.
Well, I am not sure I’m supposed to post something like this here (as this is like super early), but I want to show that this area is in the dev’s mind even at this early phase of development.
The grooming tools are already well past just defining the starting position and appear to be significantly better then current hair tools. So as I said, all the grooming tools on the planet mean little if it all then falls apart as soon as you try to animate/simulate it. Unless of course ones only interest in hair is for still image.
Hair simulation is an order of magnitude more complex than grooming. I think it’s totally reasonable for the BI to deal with the low hanging fruit first.
Again, as others have said, temper your expectations. Hair simulation is incredibly difficult. It sure would be nice to have, and I’m sure the BI will do something eventually. But don’t hold your breath.
Current hair dynamics can be really heavy computations.
Good trade off between realism and hair grid that can be computed in a reasonable time is not simple to find.
And that is explaining why hair shapekeys were used instead, in Cosmos Laundromat.
New Curves tools are not very different from old system.
Same principle of selection to work on a localized area.
Half of brushes have same purpose.
The other half would have equivalent, in old system, in tools common to edit mode.
Old particle edit mode suffers of a stupid removal of useful pivot point menu.
It has cut tools that are supposed to be assured by geometry nodes in new system.
So, Andy 's demo is smartly showing where there is a gain.
Curves Objects instead of particle systems.
Sphere Falloff for brushes.
New Brushes effects that can be limited to a global selection, instead of transform operators that are requiring multiple smallest localized selections and iterations.
Geometry nodes instead of deformers and child settings.
I hope that we would have edit mode tools to define a specific shape for one curve and propagate it to shapes of neighbours, tools to interpolate between shapes of distant curves.
Currently, we can draw Bezier Curves and convert them to Hair Curves.
We can add interpolated curves. But we can not really act on already existing curves.
A better cursor that indicates what is depth of brush effect would be welcomed, too.
I also agree that lasso select and box select may be as pertinent as brush select.
To me, that’s all that is missing in destructive workflow.
Most interesting stuff and new abilities will be related to geometry nodes.
Demo is not showing node trees.
So, new tools are equivalent to old system but in a lot more intuitive and practical workflow and UI.
We are just missing presets of geometry node trees in master, to perceive a huge difference with old system.
Currently both systems are complementary.
We can create basis with new system, convert it to old particle system to use children settings and hair dynamics.
Based on what I have read in the past, the new hair system is actually a very important part to ultimately replacing the ancient particle module with something modern (as to do so requires the decoupling of the two systems).
I can only assume at this point that once the hair system is usable, the devs. will make a move towards getting the particle nodes up and running for real.
Would it be worth splitting the hair conversation into a new thread? With the new hair system starting to take shape I’d imagine more and more people having input on the subject.
While on the subject, might be worth checking out some of the tools from wig/barbershop -
As far as I understand it, they haven’t implemented everything yet, especially regarding geometry nodes. You can criticize the grooming tools quite easily. On the other hand you may look at them as a useful testing tool to check whether the new hair system is working as expected. Having a solid implementation before starting with something as tricky as simulations makes a lot of sense in my opinion.
Already, with the new hair system, you can easily transfer hair from one model to another with a different topology, and use brushes like the snake hooh, or the slide brush
One could, but I haven’t. As grooming tools for basically static hair they look great and already appear to make getting a good looking result fair easier then the current system.
But what if the initial implementation turns out to be unworkable for simulation and/or manual animation going forward. We even have a recent example of this, Geometry Nodes where one system was introduced, only to discover that going forward it wouldn’t work out so well, so a new system was added and the old way was deprecated and then removed.
So what happens if they do all this work on grooming tools, on integration with Geometry Nodes and whatever else they have planned that makes producing still images of hair easy and look fantastic.
Only to then discover that simulation really only half works and manual control (like the ill-fated shape keys) just becomes unworkable. Are they then going to go back and deprecate this new system and half start all over again or just say its good enough and we end up back to mostly were we currently are.
Now of course I could be wrong, it’s possible that while there is yet a plan/schedule for the whole simulation/animation system for hair to be done, that those working on this new hair system do have that firmly in mind and pretty much know how they are going to do it and know that it will all work.
If that’s the case, I’d love to hear those thoughts, as I’ve already been thinking about what I’d like to see for a type of manual pose control rig one could add to a hair simulation. Using something along the current use of guide hairs that one could place around the ‘volume’ of the hair you want to control, which then has an amount of influence/weight painting to the hairs around it with control points much like brezier curves that one could then pose (move/rotate) somewhat like armature bones.
You are inventing hypothetical situations to worry about.
Just wait.
Set your expectations to zero.
Find workarounds in the mean time.
When they do get around to implementing simulation (which they certainly will) because you left your expectations at zero, anything they do will be a huge improvement. Or you will have gotten so proficient at building workarounds that you won’t even need the new tools.
Do something more productive than pointing out the obvious and pouting that you aren’t the project manager for blender hair development.
Do you seriously believe they forgot to consider simulations?
You can be sure that there are initial problems with simulations, because that happens with every complicated topic. However, Blender has a simulation that can be used as a starting point or reference at least. What Blender doesn’t have yet (and never had) is something to bring hair and geometry nodes together.
The decision what the priorities should be in that regard should without doubt be left to the people who do the actual work. They have a way better understanding of what needs to be considered, what follow up problems might be, … . They have though about integration problems 100x more than anyone else. (And yet, they are still going to run into unforeseen problems)
That was a mistake for sure. They should not have kept the old one as experimental for a while, which would likely have lead to directly pushing the new one as stable. Even though they are still going to make mistakes in the future, for some reason I believe they are able to learn from mistakes…
Well to be frank Most of the Character animation from the user community( as well as the Blender foundation”, has been Pixar inspired “cute stuff”
Where truly realistic Human hair animation can be easily excluded as stylistic choice.
I am an animated filmmaker
great looking still images mean very little to me
I just watched a bit of the old Final fantasy the spirits within ( 2001).
it struck me that 21 years later we can easily surpass ALL of the visuals of that film with Blender +Iclone/Daz and a home PC…except the hair of “Aki Ross”
Mayas Xgen hair can do it
Houdini certainly can.
My most expensive add-on is flip fluids ($79 USD)
followed by Auto rig pro( $40)
I am more than willing to pay for premium add-ons
to enhance Blenders capabilities without waiting for the the Blender foundation.
But sadly it would seem that the indie devs dont see much of a demand for a usable realistic hair/fur animation solution for blender
as the Pixar wannabe productions are being happily churned out without one.
I don’t think that risk is realistic, as the cool thing of this new hair system is that it is just curves. No special weird new object or anything ad hoc. Plain, easy to manipulate curve objects.
It makes sense that they are focusing on the basics first. I am sure that animations will be considered, as it is a topic that Blender cares about and work is being done on it. As usual, the problem is not if it will be done, but when. And for that we shall see.
