Google Summer of Code: Selected Blender Projects
Ton Roosendaal just announced the students who will participate in this year’s GSoC!
Ton writes:Here are the projects we have in 2014:
- Interactive Quadrilateral Remeshing – Alexander Pinzon. Mentor: Howard Trickey
- Fluid simulation MantaFlow integration – Roman Pogribnyl. Mentor: Nils Thuerey
- Cycles Optimizations – Thomas Dinges. Mentor: Sergey Sharybin
- Viewport FX III – Jason Wikins. Mentor: Antony Riakiotakis
- BGE – Cleanup & Support – Inez Almeida. Mentor: Daniel Stokes
- NURBS Modernization – Jonathan deWerd. Mentor: Sergey Sharybin
- Relative shape keys workflow enhancements – Grigory Revzin. Mentor: Bastien Montagne
Welcome students, and congrats with getting a great summer :)
You’ll get contacted all soon about further steps. Meanwhile, I suggest the mentors to already hook up with a student to have the first bonding going on.For those who wonder – why “only” seven projects: we actually selected 9 very good proposals, but two students also had a proposal accepted at another organization. In both cases we connected about this, and it was clear that the other (smaller) orgs would benefit better from the student’s work than we would.
None of these organizations (nor we) had backup students to fill in. Overall, the quantity and quality of proposals this year was lower, but we also selected a bit more strict than last year (15 total).
Good stuff!Although I have to admit I had my fingers crossed for an implementation of Ptex that would eventually make it into the trunk. And what’s Viewport FX III?
That is what follows and shoud conclude the work done during Viewport FX II and I.
http://wiki.blender.org/index.php/User:Jwilkins/GSoC2013/Viewport_FX_II_Proposal
Yes, Ptex would have been really good.
“Projects we have in 2014:
…
– NURBS Modernization
…
Overall, the quantity and quality of proposals this year was lower”
Poor NURBS, never getting any respect from the cool kids. It’s a shame too.
I’m pretty sure Ton was referring specifically to the proposals that didn’t get accepted. After all, he also said
—
we actually selected 9 very good proposals [from 35 total]…
—
suggesting that he thought mine was decent. There were also a few “stinker” proposals if the mailing list discussion was anything to go by — he may have been referring to those in particular.
In any case, I’ve been hoping for better NURBS support for some time and I’m super excited that I now have the opportunity to take a crack at the issue myself!
Jonathan, I for one am very enthusiastic about any improvements that may come on the NURBS front! It is great to hear that you’re gonna give it a go! I was pretty bummed when nothing ever materialized in the main from the Nurbana code that was donated years ago. ‘Course my dream is to have MoI-like capabilities within Blender, so that may tell you a bit about my overly exalted expectations… :-)
I think BF will never be interested into NURBS cause it’s not interested into CAD or pseudoCAD use of blender. So even if I see this project as very interesting I face with rationality what I think, and I think this modernization will be a start point and at the same time an end point for the next years and the obsolescence of blender on the NURBS side will last forever. Of course I hope I’m wrong, cause I’m a blender supporter, so I am not interested into failure just to say that I was right… hope this is clear.
I work a lot with CAD groups to provide visualizations. I can’t tell you how frustrating it is to perform conversions. Hope you get somewhere with the modernization!
mantaflow seems. it’s about time liquid sim gets updated. Mantaflow is not the particle based fluid sim Nils has posted on youtube?
Some of these are really quite good but I was also crossing my fingers for something that would provide tighter integration between the different editors, like tighter integration between the game engine and blender, tighter integration between the compositor and the rest of blender, tighter integration between the video clip editor, movie clip editor and the motion tracker or something like that but anyways I like the improvement on the game engine, if that one gets finished, I think that will be a really strong improvement on blender.
It doesn’t matter how strong something is, if the foundations are weak then it is weak, I believe the BGE clean up and support although wont add any new features as far as I am aware, it will increase the strength of the foundation of the blender game engine making it a much stronger and efficient platform from which to build upon.
but anyways good luck
Totally. I too think tighter integration and better cleanup is a high priority–the future of Blender will depend on it. Especially if we want to see more new features come to Blender.
If we don’t address Blender’s ad-hoc nature of structure, things will become even more crowded and non-integrated than they already are now, and the UI will reflect the crowded internal structure even further.
Blender sees its issues with its UI because it’s directly the result of its lack of consistent and efficient internal integration. It’s hard to make an efficient UI as its surface when things are scattered all over the place underneath.
It can often prove regressive to one’s workflow rather than progressive, and folk will find other software, now that other solutions are becoming more affordable nowadays.
I myself, while I’d love to use Blender as a one-stop solution, rely on other software for texture-painting, high-poly sculpting, and retopology. For my profession, I use Blender for quick modeling, but I find I often have to turn to another solution for doing a certain job better.
It’s kinda sad if I find it faster to use several software faster than just using Blender, esp. now that we’re seeing some of these areas of texture-painting, sculpting and retopology being addressed in Blender.
With that said, I’m totally excited to see a true remesh solution on the agenda. If I can use an improved Remesh solution, that’d be more one reason to stick with Blender as we wait for internal improvement in the meantime.
Don’t even get started on the remesh improvements lol, the current remesh modifier although as good as it is, is completely useless when a model becomes to complex and is only really useful when doing simple shapes unfortunately, lets just hope that this new remesh algorithm that is getting created for GSC is a lot more powerful, well fingers crossed because if it is then that will improve the sculpting in blender even more.
What’s with existing particle fluids? There was a talk that OpenVDB was needed to be implemented first, to get meshing right.
I would like to know also.
Interactive Quadrilateral Remeshing (aka a quad remesh solution that can produce uniform quads)?!
NURBS Modernization?!
Guys, I overuse this thing, but this truly calls for the button.
o/ http://yeeeeeeeeeeeeeees.com/
Seriously, I’m hoping the Interactive Quadrilateral Remesh (…I’mma’ just call you IQR) is what I’ve assumed it is. More uniform, adjustable remeshing solution, rather comparable to Dynamesh in ZBrush, being added to the current available Remesh options. If so, then we’re seeing a major addition for Blender, folks.
I mean, Dyntopo sculpting workflows would drastically improve. Metaball, Skin Modifier and Boolean mesh operations would work better for quickly forming base models, with an even-quad Remeshing. Remeshing complex models to make them more efficient within Blender.
Seriously, this feature alone would draw thousands of new users wanting a free interactive method for remeshing. And then we’re finally seeing NURBS get some love in Blender–it’s been a long time coming.
I’m assuming that Interactive Quadrilateral Remesh will be similar to automatic retopology in 3D Coat and ZBrush.
Blender’s existing Remesh modifier is comparable to ZBrush’s Dynamesh. They both voxelize a mesh and convert the voxel data back to a mesh using what appears to be naive surface net meshing algorithms. Nonetheless, Dynamesh has several advantages, as expected. Specifically, it is significantly faster, supports boolean operations on the voxel data, and optionally executes post-remeshing commands like “project” and “clay polish”.
I think you missed the biggest advantage of Dynamesh: The results are a predictable, uniform, and cleaner formation of quads.
Yay! Quad Remeshing!
That was one of the missing pieces for making cloth sewing stuff I’ve been working on pleasant to use!
Something simialr to this?
Quadrilateral Meshing with Directionality Control by Packing Square Cells ( https://www.andrew.cmu.edu/user/shimada/research.htm )
Darn, I guess I must’ve upvoted the first time without being logged in. Oh, well. This post makes me so happy, it deserve two upvotes from me.
Considering the quality of the projects, I hope that at least this time all the students will complete the promised work in time. I’m always disappointed when it not happens. If you take a commitment you have to keep it.
interactive remeshing!!! – http://www.inf.ethz.ch/personal/dpanozzo/videos/sketch-based-generation-and-editing-of-quad-meshes.mp4
something like that would kick ass, I can’t wait till it gets into some 3d software
While I’m disappointed to hear that Blender’s current remesher isn’t very useful (I had a few ideas that would’ve relied heavily on such a tool), I am super excited to see what improvements are made. Even more so on the NURBS front. If I’m not mistaken, even the ancient Caligari TrueSpace had better NURBS tools than Blender does today. It just doesn’t make sense that such a gaping hole in Blender’s capabilities gets relatively no attention.
Where can I get latest updates of these projects?