Ultimate Texture Project

Hi all,

For the last weeks, I’ve been testing a new texturing system for blender. This is going to be a HUGE project that will require hours, days, weeks and months to get mature enough.

The idea came from using Allegorithm products (with no offense, I personally didn’t like them).
So I decided to make some tests in blender, and for now things are looking very promissory.

The system will include a node based system, and a texture paint system, both connected to each other.
The purpose is to create textures for objects, by either using procedurals, texture painting or both.
Things being planned are:

  • multi layer painting
  • particles and dynamic painting
  • weathering procedurals (nodes and painting)
  • texturing procedurals (nodes and painting)
  • AO, Curvature and other geometry dependent tools
  • Baking
  • and more

Because of the effort this will require from me, I decided to first probe the Blender users, on how should I forward this development.

So I want to invite you all to respond to this poll. And I hope in the next weeks to start posting some results.

For the first option I’ll later ask for donations, so I can develop the addon faster and dedicate most of my time to it.
The second option will make me just post updates, with no rush (as I’ll will need to work more in outside projects).
And the third option is just a check.

My best regards,
Miguel Porces

Thank you!
Cool stuff!

It’s really a bigger proyect for one person. I think. Also in reality actually blender support somethign like multilayers.

But my priority will be

  • Official development in blender trunk (with patreon or quickstarter)
  • Paid
  • Free

Anyway, talk with developers because with blender 2.8 near is better to talk about with him

something i have been working on from quite a long time

easy shaders

Well, if you are doing this, its bound to be good.

Though I agree with DcVertice, Id really want this in C++. Is this a possibility ?
Andreas Esau tried a paint-on-multiple-Layers function for BPainter, but it turned out to be too clunky in Python (?)…
Third question: http://bpainter.artbyndee.de/faq

Did your tests included multilayer painting ?

In general, I would happily support you via patreon or somewhere else for a stable workflow like this.
Also thought about fiddling for a procedural PBR workflow + particles in Blender.
Also not a big fan of Allegorithmic tools, which secretly suggests that you might have nice ideas for improvements, you are really hyping me a bit here.

Yeah, the really good will be have all this native in blender. The paint with multilayer system, specially.

For now, I’m not planning to do much C code, except for fixing bugs that the current api has and that I need for this to work seamlessly.
But this doesn’t mean that C won’t be part of the addon (more exactly OpenGL). Most of it will be taken care by the GPU, and only writing/updating/reading buffers plus the UI will be done in python. Also, only user input is updated to the gpu (when painting), so just mouse position, buttons, and I expect to pass ‘pressure’ and ‘tilt’ in a later phase.

The reason for choosing this way, it’s just because it’s easier to prototype. I make a change, and I can run it right away. In the source code I need to wait for compiling and since this is still in a highly experimental phase, all that time is useless.

About MultiLayers, what I already tested worked ok. Something like: writing at two buffers at the same time took 0.01 seconds for 1024^2 texture size (per mouse input!), in a very unoptimized prototyped… But since I really don’t need to pass the images back and forward from the cpu to the gpu, it’s really quite slow for now…

Of course, my plans for the multilayers are not really for that. I want more. I want intelligent brushes, spray water or paint into objects, make mold grow, put stones side by side, etc.
When all this is working good, then making it into the Blender source will be faster.

2 Likes

Sounds perfect. So, just by the scope and nature of it, blender would already benefit from your fixes and additions you’d have to do in order to get the addon on its feet ?

Fantastic, I would prefer a free version, with you being supported via donations, tbh.

Boy Im getting a bit megalomaniac here. :slight_smile:
Yes, SP lacks modeling functionalities. So its easy to see the potential in here!

Can one support you besides testing and donating ?
i.e.: Getting a nice set of CC0 PBR decals (smudge, grunge, moss…) ready for you ?
Just for convenience.

So you wanna implement in Blender the ‘painting logic’ you find in SD and SP? Wow… I hope the best for that project :slight_smile:

Have a look at AssetGen, for instance.
Allegorithmic has the fastest tool in the market. They have special patent for speedy algos that makes your workflow fast.
Anything python will be simply so incredibly slow. So if you have speed in mind, forget it…

Python is not so bad!)

1 Like

@rbx775, Thanks for the decals offer. For now I won’t need them, as the current state is figuring how to structure core things in order to be able to add features without troubles. Because of the abstraction of this part, I haven’t thought much about having textures, decals, etc. Those will come later, when the underneath is working. Only after that I can implement an interface for incorporating brushes, textures, procedurals, etc.

@kabu, I don’t think Allegorithm products are fast… And about Python, it can be very fast if the code is well designed. I could program something in C and make a stupid algorithm loop eternally before an output, so even in C the problem is the same: Design!

1 Like

I don’t know…as long as it is free…I see a lot of addons going the autodesk route…while I am not against people making some money…I just get tired of looking for an addon and I find one that fits my needs…but then it costs money…it is common that they are not supported longer than a year and this just makes it irritating for me…(this is not all of them ofc) at this point if it is not free I do not touch it period…

as far as the addon itself goes(subject matter) I feel it would be better to start looking toward 2.8 and thinking what painting/texturing tools would be useful there…to start one now for 2.79 would be useless work…2,79 will still be used for at least 2 years for a lot of projects…and if you only plan on supporting it that long you will be fine…after that you will see possibly an exponential decrease in demand.

But all of this is just my opinion.

Great initiative.

Voted for free, not because of being cheap or such, but because i believe in free choice & equal opportunity. Nobody can start jumping over hurdles while not even beginning to walk. Also marketing, advertising, bureaucracy… take a whole lot of extra effort, time, work & skill to really succeed.

On functionality… Do you know about Verve? Would be great to achieve such level within 3D (especially with Blender). Also, from my experience with the tool & forum, i consider Taron (developer) a great spirit, sharp mind & true master of painting code.

About SP&SD…
There’s not much i can write about, since it goes against my beliefs. I prefer light, efficient tools, modern HPC code, free choice (am fan & admirer of the demo scene)… without bloated minds stuck in the past, slaves to profit by hype & propaganda (& all non-generic assets that come along are IMHO bloat & hurdles to creativity - tho i see how it benefits the industry. From what i have observed, lack of artistic creativity & high profits in showbiz, i consider it as a Technology for distraction).
Let artists, creators, developers… profit from their imagination, diversity, entertainment & works. Why i like Blender :slight_smile:

I strongly believe human mind must be confronted with free choice, independence within society and as such can learn the responsibility - to take care for holistic system of happiness & dreams.

You got my support & gratitude.

How will it differ from existing add-ons like Pro-Elements, AssetGen, Armory3d?

This project seems way too complicated for any single person. I like your spirit! :smiley:
If you go the free route, don’t hesitate to ask for help from the community…

Thanks for the comments. I also would like to share this for free. But on the other hand, this will require a lot of time and dedication from my behalf… Something that I cannot have without some monetary help. It’s not that I want to get rich or something (I’ve been sharing great ‘secrets’ with this community for some years, free of charge), but I’ve bills to pay, and I would rather spend my time doing something usefull for the comunity than making works for companies that don’t really care on anyone interests.
As for the 2.8, I’m not worry. I’m aware of its development, and this project will be carefully oriented to 2.8+ versions (actually, I can inform that this won’t work, at least completly, in the 2.79, since I’m already using 2.79.1, and its new python version). Adding to what I already said above, another reason I’m not including this directly into Blender’s source is simply because it takes more time to see it in the trunk, and I don’t want for anyone to need to have a separated compiled branch just to use it. As an Addon, one can simply install it in their current working version, and enable it when needed (even for alpha and beta testing).

I know Verve… it’s a quite interesting toy, but imho that’s all about it. I don’t like its UI, and there’s not much else one can do with it… It would be nice to have that kind of effects in Krita, for example. Anyway, I’m planning to use some ideas quite similar to the ones in Verve in some of the tools, thought they should be more 3dimentional… :slight_smile:

The main purpose of this project started with one problem I constantly face with when I’m writing shaders (I do write OSL shaders and design procedural textures). OSL has some limitations in terms of procedurals textures, mainly because it’s supposed to run at render time. This makes it very difficult, if not impossible, to use technics like Diffusion-Limited Aggregation, Random Walkers and other ‘3d texture space’ dependent algorithms. So I need something that can create textures on the same time it can analyse surface caracteristics, with some interface easy to use but also possible to achieve a plethora of effects.
Pro-Elements, AssetGen and Armory3d are very nice projects, but they still lack the stuff I really need and want to put in this project. The good news is that this project will output image textures that can be used not only in Cycles, but also any render, game engine or software.

@tosky and antoni4040, the algorithms I’m planning to include are more or less defined. I’ve been studying and creating procedural textures for quite some time to be enough comfortable with it. The most problematic part of this project is to design a good base inside Blender that can produce good results, but more particularly an excelent user experience (something most texture softwares lack!)… For those who know me better, know that I favor simplicity above all (clean UIs, simple methods, clear purposes); And that’s my primary goal with this project.

@all, since it seems the ‘free’ version is winning, I created a patreon account for it… Right now it’s brand new, and no information at all about anything, but I’ll populate it as soon as possible. Feel free to take a quick journey there: https://www.patreon.com/secrop :wink:

I hope later to post some of my tests here, but remember that this is still in an embrionary phase… No arms, no legs, just an undeveloped mix of cells. :slight_smile:

Just took a quick glance at the patreon page and it only offered support redeemed by post.
Will there be a monthly option aswell ?

I’m a total noob with this patreon site :slight_smile:
It should allow that option now.

I know Verve… it’s a quite interesting toy, but imho that’s all about it. I don’t like its UI, and there’s not much else one can do with it… It would be nice to have that kind of effects in Krita, for example. Anyway, I’m planning to use some ideas quite similar to the ones in Verve in some of the tools, thought they should be more 3dimentional… https://blenderartists.org/forum/images/smilies/sago/smile.gif

Oh… yup, Verve’s UI is tricky - very stylized & artistic.
But the magic that the tool also provides (along interactive fluid system & canvas influence) is ability to export the height map along with textures/layers as you can build up paint as in the real painting :wink: