[Bf-committers] Undergraduate Research

[Ton Roosendaal]

Hi Scott,

The final decisions by Google for accepted projects will get published on 27th of May. Until then we can't say anything about it unfortunately.

In general coding work is always welcome though, especially in the area of fluid simulation. There are not many experts for it though, you might consider to join our irc channel at irc.freenode.net #blendercoders, and seek contact with Daniel Genrich there.

Our todo wiki and open topics for Blender is quite messy in places, I know.

But, apart from simulation code itself, there's a very interesting related area in Blender that's entirely orphaned - Volume rendering for Blender's internal render. It has a distinct section in our wiki todo with links to open issues (bug reports). You could check on that?

Regards,

-Ton-

--------------------------------------------------------
Ton Roosendaal  -  ton at blender.org   -   www.blender.org
Chairman Blender Foundation - Producer Blender Institute
Entrepotdok 57A  -  1018AD Amsterdam  -  The Netherlands



On 20 May, 2013, at 7:15, Scott Aufderheide wrote:

> Hello everyone,
> 
> For this upcoming fall semester, I will be doing undergraduate research for
> a professor who does some work in graphics and visual effects. Even though
> this is for the fall, I was hoping to get a jump start on it during the
> summer, but I have yet to choose a topic to delve into. I want to do
> something with Blender as I have been using it for a couple years and
> really enjoy it, plus I feel that I may be able to do something worthwhile
> as I'll have at least seven months to dedicate to it.
> 
> I am very comfortable with C and C++, I am pretty familiar with python, and
> I have some CUDA experience as well. For a project last fall, I looked at
> implementing different noise functions (Wavelet vs Curl) for Blender's
> smoke simulator to compare their effect on render times, and investigated
> how they could be used to better simulate fire. For this past semester, I
> just finished writing an addon for Blender that simulates swarm
> intelligence using a flocking algorithm (I got the idea for the project
> before I learned about boids particles). The addon takes user input to
> define some swarm properties, and then parallelizes the computation of the
> motion paths for each swarm object on the GPU using CUDA, and then animates
> the objects.
> 
> I tried looking at the projects and todo pages, but I was somewhat confused
> on what I found there. It seemed like many of things mentioned were for
> previous releases, particularly since the phrases "...for 2.5/2.6..." were
> found throughout, and I was wondering if anyone might be able to direct me
> to more recent proposal and project areas.
> 
> I suppose the timing is a little awkward as the GSoC proposals are still
> being reviewed, but a number of those projects seemed like they would be
> very interesting. I was first looking at the modifier section under the
> GSOC ideas since it said,
> 
>    "Modifiers are probably one of the best introductions to coding new
> modeling tools. Modifiers
>    have a well-defined API (DerivedMesh comes in, new DerivedMesh goes
> out), so the student
>    need learn only a small part of Blender's internals before becoming
> productive.",
> 
> and learning Blender's internals is probably what intimidates me the most.
> I was thinking of trying to implement something similar to
> this<http://www.youtube.com/watch?v=u_2qKQpxCRw> as
> a potential modifier project. I know that Blender currently has the
> Fracture addon, but I feel like this is a very limited tool, where
> something similar to Rayfire's Crack would be much more dynamic for
> allowing users to break up objects. Two potential problems with this though
> would be, first: that I have no idea about the legal
> ramifications/restrictions on trying to copy (to a degree) what someone
> else has done, albeit for a different program and we aren't trying to sell
> it. And second: I noticed that in messing around with the fracture addon
> that it slows down or even stops Blender when you try to break an object
> too many times at once. I wasn't sure if this was from the addon trying to
> compute too many things or Blender from trying to process the addon's
> output to the viewport (I was only considering this second option as I
> noticed Blender slowed down some when I tried creating too many swarm
> objects simultaneously in my own addon). For this, I suppose I would have
> to wait and see how the modifier performed and take it from there.
> 
> The other GSOC idea that seemed interesting to me was the thin fluids
> feature for the fluid simulation. My only concern with this is that due to
> my inexperience, I am not sure that I could fully complete it in the time
> I'd have for the semester (again, Blender internal/integration fears). In
> working with the smoke simulator before, I was quite a bit lost in
> everything and kept my small changes to only two files for fear of breaking
> everything when trying to recompile. I like the challenge of this project,
> but I think I would need someone with some more experience who would be
> willing to help guide me and be willing to put up with (probably) some
> pestering and (potentially) noob questions.
> 
> All that being said, I know that someone else was interested in the thin
> fluids feature as their own GSOC project. I think Ton said that the final
> slot allocation due date was the 22nd, so I suppose I would have to wait
> until then to see if thin fluids has been taken.
> 
> Additionally, I am looking into getting my feet wet with shaders or
> textures or rendering. For the first two, I wasn't really sure where to
> start looking for anything so I thought I might ask if anyone is doing work
> or wants work to be done in those areas, and if so, if they might want to
> discuss things further. And for rendering, I know a lot is being done with
> Cycles right now, but I again was hoping that I might be able to talk with
> someone on project ideas just because there is so much there.
> 
> My apologies on the length of this email, but I thought I should explain
> myself a bit. My main hope in reaching out here is that no matter what I
> might work on, I am hoping that I am able to find someone willing to
> discuss problems and implementations in a mentor role for a while, despite
> the fact that I am not part of GSOC.
> 
> Thanks for your time and all the work you guys do.
> 
> Scott Aufderheide
> University of Notre Dame
> Computer Science 2014
> _______________________________________________
> Bf-committers mailing list
> Bf-committers at blender.org
> http://lists.blender.org/mailman/listinfo/bf-committers