Procedural Content Creation

Bevel shaders are nice for rendering highlights onto hard edges but you can spot it easily specifically also when the bevel is bigger.
thats why I prefer real geometry bevels rather

The thing is that a lot of people need this for videogames and hardsurfaces, not only to make renders.

Not really following, I thought the everything-nodes was the big thing they were working on now, animation-nodes, particle-nodes…?

No, it´s been postponed. It´s too big of a project, and not enough people to work on it for now. So BF decided to focus on getting viewport + workflow improvements done first, and then revisit “everything nodes”. Probably not until next year, would be my guess.

Beeing the pessimist that i am, i don’t expect this “everything nodes” in Blender for at least 2 years.
I would suggest to use the time to learn some Houdini, since its NonCommercial version is freely available.
Importing geo out of it is not a problem, which makes it imho an ideal companion application for Blender.

Houdini is a really simple program… in few days you can learn all the idea of the program. And I don’t think that blender foundation would make a everything node, it’s more easy to make a fork. But make some part nodes…

Oh, absolutely. When I say next year I mean that´s the earliest they would start working on it, and then probably at least a year to a workable state.

I think that only one year of work is a really optimistic plan.

Notice the word “workable”, not finished :wink:

Yeah, and it’s optimistic xD

Lol, that’s the most optimistic approach to Houdini I ever heard… I’d say a more realistic proposal is, in a few days most people have realized they won’t get very far using the shelf tools and a month later they’re on Khan Academy doing a refresh on linear algebra… :wink:

And I don’t think Houdini is complicated either, but there’s a lot in Houdini, you could spend a year playing with the fluid micro solvers, alone… :wink:

Woah, I truly thought that was the next big thing coming in Blender, I thought the only reason they hadn’t implemented the particle nodes was due to focusing on the all-nodes…

Now I’m truly surprised… :eek:

Houdini is a really big and complex program, but the base of the program, the concept is really simple and easy to learn. You see houdini and all have coherence, all works in the same way,… obviusly make things with houdini is hard.

I agree 100% but it’s really hard to understand that Houdini is simple and that threshold is the biggest problem of all. There is an idea that Houdini is difficult, and it acts like a barrier for people trying to get into it - it becomes a self fulfilling prophecy, of sorts. But when you get past that threshold, you’ll realize it’s the most consistent 3D app of all.

That being said, for me the idea of everything nodes in Blender, or animation nodes, particle nodes, if they were to be implemented separately, has nothing to do with making Blender work as Houdini, it’s all about the genius of node based, procedural workflows. Be it in 3D or compositing, nodes based workflows have so many advantages, from being non destructive to helpful as visual aids, problem solving to the overview you have over the setups. They potentially also offer that low level setup that is so friggin nice to have, when you want/need it.

Here’s a spider web setup I did in Houdini and there’s really no other app that can pull a setup like this off…

And here’s the node tree for it. This is the power of node based setups. :slight_smile:

I disagree 100%.
There is no self-fulfilling prophecy.
Its a real problem if you are an artist with no knowledge and interest in programming/scripting/math.
Once you exhausted the usability of the shelf tools, to get further you need this knowledge.
Ergo you have to learn it, which costs lots of energy and time.
There is a real strong learning curve involved which IS a barrier for many people.

Hehe, now I disagree fully. All 3D apps are “hard” and I dont believe for a second it’s easier to get into Blender or Modo than Houdini. And if you spend the same amount of time learning, studying l Houdini, Maya or C4D, I’m 100% sure you would be able to do more on Houdini after 3 months than Maya or C4D.

Now, I can of course be wrong, perhaps my personal experience makes me biased, but I really don’t believe that’s the case. :slight_smile:

I do agree with the points that a procedural approach to modeling is powerful, but when it comes to having that in the general workflow, there has to be a balance between ease of use and flexibility (with the system preferably being built with the intention of working alongside a traditional toolset and workflow).

Cases in point can be seen with some of the Houdini videos, I saw a video where a crystal was modeled in Houdini and how it took over 20 minutes and a lot of node tweaking to get it done (and it wasn’t even a cluster of crystals or a cave filled with them). Houdini’s node approach can indeed be of great help for large, highly detailed objects and heavy amounts of small details in others (but it’s still a turnoff for artists if it also means a much longer period of time to get forms that traditional modeling tools can do quickly, and chances are your noodle will go to absolute insanity once you bring creatures into the mix).

In my mind, the right approach for Blender would be turning the modifier stack into something that works in a node-based fashion, with the core of it based on a special input/output socket type for mesh information (much like how things work in Cycles).

I did a little test a while ago

It’s about using smart/efficient tools and creating yourself a smart/efficient workflow. Nowadays (when the client isn’t really sure about their own taste & wishes) it takes me more time to search the library/web for specific models than it takes to prepare different iterations of procedural models - 3d bashing.
Evolution is the mix of the fittest :stuck_out_tongue:

Here’s one of the oldie-goldies
// for those interested in virtual architecture… long, repetitive & tedious tasks are best done when procedurally automated with options to set parameters as you go. Are precise and prevent mistakes which come after loosing concentration.

Hi everyone!
I know this is an old topic, but I recently started to develop a SideFX Houdini inspired node based procedural modeling system. It has features like realtime mesh preview, input parameters and contains the frequently used modifiers and operators (edit/object modes) in Blender.

The project is available on Github and you can track its progress on Trello (links provided in the thread)