Now that the Blender hair tools are being improved, I think the developers should take a look at it. They’re just describing the system and there’s not too much technical info about it but I think that the way they approach the hair creation, grooming and simulation is something we could benefit from, specially if the Blender’s system is being worked on at the moment.
I think that Blender’s current hair fixes are a (much needed and not-so-simple) bandaid, while a true reimagining of the system like this will require a complete rewrite of the particle and hair systems to make them not so broken before fancy Yeti or WETA-like groom options become available.
Yes, I like it very much too, and Kent Trammell has showed in his tutorials how good can it look when you know what to do and how to do it hehe.
But since a complete rewrite was mentioned several times before I think it’s a good idea to look at other systems and how they do things.
Anyway, it would be very sad if it’s true that they decided to just improve the current system instead of writing a new one.
Indeed, I agree. With the new hair cage system it seems multiple particle systems on a single character will be a lot more easier to control and groom. Kudos to Lukas
Yeah upgrading the old system was a good idea. If a rewrite is indeed needed (one that will likely happen after Particle nodes are in), then it’s great to see they fixed the system enough to make it a solid production tool for the next few years. Because a rewrite isnt likely coming anytime soon.
Has anyone tested the new system at all yet? Are collisions solid and predictable? The system no longer “explodes” in real-world scenarios?
Lukas told me the other day that the particle system is a mess and is extremely limiting from a code standpoint. Most people’s wishes for particles and hair aren’t possible with the current core, but at the same time redoing the whole system is such a gargantuan task that he said it’s basically impossible. The conversation didn’t inspire much hope, unfortunately. Particles are apparently just about as bad as code can get in terms of spaghettification and modularity/ability to update without breaking the rest of Blender. The wait could be longer than people think before a true fix is properly tackled.
I’ve tested the latest builds and the improved editing tools (like curves for clump and noise) and cage mesh to define the hair are really good additions
Collisions are still somewhat random… I don’t know, they just seem to work sometimes.
Anyway thank’s to Lukas for doing a really needed update to the system. I do hope the system gets a rewrite though… actually all the simulation systems should be connected, and (based on some of the comments on this and other threads) that means that a LOT of things would need the complete rewrite from scratch.
The changes he really wants to do to hair can’t be done with the current particle system, period. And the changes that would need to happen under the hood to the basic core of the particle system would require an amount of time that just doesn’t exist for a single coder. If it happens, it will likely have to be a concentrated effort like Depsgraph has been.
Has any of this work already been started? I remember reading about particle nodes and some demonstrations of it years ago.
I have no delusions about the size of the project, I’m just curious if some of this groundwork for a re-write have been laid already, or if all of the work up until now has just been an attempt to patch the old system (particles, not just hair).
In terms of the hair patch, do the devs feel they can at least get reliable collisions and stable sims working before moving on to greener pastures, or is it all a lost cause?
As for the state of the particle/hair tools and what needs to be done, this commit says it best.
This is obviously total madness, this feature is totally unusable whenthe coder is not sitting next to you … But so are the rest of particle settings, the only solution is a future node system.
The whole system in a nutshell looks like it needs to ripped out completely and replaced with a new system that is node-based, but until that happens, we’re stuck with the cruft of the old system.
Lol. I hope they can at least cobble together enough functionality and reliability out of the current system (even if the functionality is limited) before they move on to a full node system.
Oh well, back to using forces.