Hi all

Thanks to Jahka who told me that he suspected the current harmonic force to act in a similar way to my spring implementation, I have performed some researches and he was indeed right.  Just that the harmonic force doesn’t consider a rest length parameter. So I have merged both implementation into one, where the harmonic force now uses the spring force code with rest length and reuses the previous Harmonic velocity damping.

Now everything is ok: if the user sets the rest length of the new harmonic to zero he will get the old harmonic behavior, while if he set its value to any other’s one he will get the new spring force behavior. The new patch is at http://www.pasteall.org/9882/diff

Now it requires even less changes  plus it inherits its gizmo

Cheers

The Spring force has being merged in the Harmonic force, improving it

Hi all!

Lately, Stephen and I were busy trying to implement spring forces for the Fluid Particles. we had made many test and something was always wrong, leading to some stall in this project but sometimes you just have to let your brain does the homework in a background process, take some rest, enjoy your life and suddenly it will pop you with a correct solution

Yesterday I was trying to reduce the problem to its simplest form: adding spring forces between two particles, outside the fluid solver, I was too focused on the fluid solver that the solution was always in front of me and I couldn’t see it , that simple switch of point of view lead me to the solution: implement it as a force and not as impulses like suggested
the main paper, suddenly everything started to flow and as a nice side effect I implemented a new particle force type: Spring force, not just for fluid particles but for everything!

cheers

Searching to add spring forces for the fluid particles I ended up with a nice patch that adds also spring force as a new general force type that could be used like any other blender forces: for objects, particles, etc. Needless to say the importance and wide use cases for spring forces in 3D, previously this functionality was available in Blender only in softbodies,cloth and constraints but with this patch this could be extended to particle systems for many VFX’s and also to objects, the possibilities are endless It’s interesting to note how the particle system integrate features from softbodies and fluids It’s very easy to implement and does not break any previous code, it only will need a new UI slider for spring rest length but I’m not good at UI coding, for you it should be easy to add

http://www.pasteall.org/9806/diff

the size of the spring force gizmo changes according to its strength

I just sent the patch to Jahka for review and possible commit

Hope you will like it and I plan to upload soon some videos about it

Cheers to all  Farsthary

Hi all
the other day i have received the happy news that I have won a relevant prize in the National Computer forun with my thesys Volumetrics in Blender!!!!!

I’m very happy for it
Cheers Farsthary

Hi all

Thanks to Mcreamsurfer’s feedback I was able to correct the bug that prevented the particle emittor with the verlet integrator from being correctly animated, now it is solved. I have sent the patch to Jahka , just be patient because he is currently busy.

I’ve also made some tests about the nice particle system feature of fading in time for cached simulations: It allows to see the flowlines

Cheers

Hi all

The Verlet integrator patch was just commited by Jahka , so it is now in trunk
thanks to him who kindly review and integrate my patches.
I’m very happy for that

Cheers Farsthary

Hi all

Currently the particle system support 3 integrators types (in english integrators are what actually advance particles in time)
and for many users that´s the kind of options that fall under the I-always-will-use-the-default-value category,
and while for many set ups different integrators provide very similar results, in others they could make the difference.

The simplest integrator, Euler, is very fast but not very accurate, internally Blender performs only one cycle of some calculations.

The second integrator, Midpoint , is sligthly slower because it performs 2 cycles but is more accurate than Euler and is the default integrator.

And the third , RK4 (Runge-Kutta) , as you may guess is the slowest and more accurate of them, and performs 4 cycles of calculations in Blender

The good news here is that I have found another integrator type that is as fast as the Euler integrator (the fastest) since it requires only one cycle of calculations and is more accurate than the Midpoint even, the error is of fourth order: The Verlet integrator  http://en.wikipedia.org/wiki/Verlet_integration#Basic_Verlet
have several advantages over the others and is very stable and fast, and gives realistic results.
with many particles interacting each others the Verlet integrator clearly shows the speed boost even over the Euler integrator.

the hand made patch is at http://www.pasteall.org/9364/diff

I have implemented it in Blender seamlessly and I suggest it as the default integrator, the previous integrators suffer from an increase of energy in time, note how the amplitude of each bounce is biger than the others in the previos integrators while in the Verlet is uniform as should behave in a conservative enviroment (no friction,damping,etc).

It is very easy to implement in Blender and is a small step toward improvement, hope you like it
Cheers Farsthary

Hi all

Since SPH particles are , well, just particles, they can seemlessly be integrated with the whole dynamic system, providing an incredible range of particle FX’s, here for example I show you some simple yet beautiful cases:

the previous videos show SPH particles rendered with pointdensity textures (volumetrics)

And this one shows an interesting effect with instanced particles and some force fields with no gravity

Hope you like them
Cheers Farsthary

Wow!

I was happy that this season we didn’t have any hurricane over Cuba, but I guess this was too much to ask
This weekend we had a little hurricane: IDA, that reached the 2nd cathegory (2 of 5)
nothing compared to last year season where we had 3 hurricanes of the 4th and 5th cathegory that put on the knees my already devastated country. And as a curiosity IDA has become an extratropical hurricane, an invernal one.
Nothing really happens this time, I was just locked in my home the whole weekend, and luckily without electricity blackouts so I could continue my projects

Hi all

An open source developer should not have any problem in dropping a project for joining another if the second one promises better features, as working in two seperate projects will only make things more difficult and slow and at the end necessarily one of the two solutions has to be dropped. The good thing about being open source is that this is not a race among devs, this is a collaborative work
So I didn’t have to think twice to make the switch once I had contacted Stephen and reviewed his implementation of Particle Based viscoelastic fluid simulation (that is also an SPH method) based on several reasons:

It’s much more stable
It’s more artist friendly
It avoids duplicate efforts
It speeds up the development

This is not the first time something similar happens to me , the first was Volumetrics, then Photon mapping (which is currently on hold until some physically correct material system gets added or some future decisions are being made) and now SPH particles so I have experienced that

We are currently working togeither (also with the help of Jahka) to deliver very soon a testing build.

Cheers Farsthary

Hi, this is -again- a friend of Farsthary!

As Farsthary made last week a “little” excursion in the beautiful particle world of Lord Jahka (sorry if this sounds cheesy…), I thought you would be interested by some pieces of information:

- Jahka has kindly accepted to review his patch, so stay tuned to Jahka’s blog

- Stephen Whitehorn aka chickencoop (an early supporter of Farsthary) recently started his blog about other amazing particle VFXs! Kudos to him, we wish him a lot of success!

All progresses relative to the particles system are managed by Jahka, who directly adds them to the trunk. This means that if Jahka speaks about it on his blog, you can find the new tools a few days later in the fresh 2.5 builds from Graphicall (thank again to them).

Regards

p.s: sorry about the troubles with the new wordpress layout (problem under firefox for the comments, strange formating..), but this is one of the rare thing that Farsthary directly manages, so I don’t want to “steal” that to him by doing it without his authorization… so please wait a few days so that he corrects this himself..  Currently, if you want to let or read  a comment, you have to go to the previous post, then click “next”..

Hi all

Wow! every time I start a new project I get some comments of concern from the blender community about myself spreading my code time across them . I have started many projects since I got involved in Blender development:
Volumetrics, a smoke simulator, GI pathtracing (unbiased), GI photonmapping (biased), Bloxel (Blender voxel system), a small patch for using object groups as a larger manager and now SPH particles, among them currently only Volumetrics have gotten into trunk because Matt Ebb got involved, rewrote everything and put his skills on it (thanks to him). Luckily Daniel Genrich has also integrated the smoke simulator of Niels Thuerey independently.

I have stated many times that my main approach to Blender is as a CATALYST: Due to my internet issues (low bandwith/speed: believe me I don’t know what web videos are :( , working near isolated, no access to IRC chats , live talks, etc. ), I could not directly commit to the blender sources or have my own branches. Instead, I have to rely on core Blender devs (Like Matt has kindly done with volumetrics) to review and commit my code. Sadly one of the hardest stuffs for a programmer is reviewing other’s code; that’s perhaps why in the Open source world there’re so many redundant applications (millions of sound players,text editors,3d apps, etc, etc) many of them of low quality/unfinished simply because it’s often easier to write everything from scratch than to dive into existing projects (you have to learn its arcanes) and focus development on them (like blender have succesfully done attracting 3d devs). What is the best way of learning a foreign language? it’s not to pay a teacher or watch lots of courses, no, the best way is to get thrown in that country where that language is native and you must learn it or die : I bet in a matter of months you will speak very fluently that language . The best way for me to master Blender is to tackle projects in many fields of Blender that will force me to go to different areas of its source code. This will allow me to build a global idea of Blender inner working, that’s something you simply won’t get reading the whole source code even if you have the lifetime for that, I could focus on one project at a time but if I had done that I’m sure I would have quitted blender dev long time ago: is simply boring. I’m not a coding machine, in order to move on I need to feel JOY about what I’m doing. And what motivates me the most? Encouraging comments, positive reviews/feedback, someone using/needing what I’m doing, support , resources and a place to render my creativity .

With every new project I tackle I learn new skills that are very difficult to get when working alone. I try as hard as I can to deliver a good PROTOTYPE in order to attract the attention of a core dev; I can not say: -”!Hey, integrate this please for tomorrow!”.

My projects are unfinished prototypes? non standard? delivererd too early? I’m not hiding that, actually I’m proud of that, for me they have lasted an eternity due to the large amount of new knowledge I had to gather for doing them (more than in college), but I’m still a newbie in Blender development! and in graphics technologies and in open source! I started blender programming two years ago (April 2008): for some peoples it can be slow, but this is my rate of learning skills nearly alone, and also in order to make the life or the reviewer easier the code should remain like that, a functional prototype. In order to master one field of knowledge a person have to spend an average of 10 years, I WISH to became a Blender pro and if life allows me I’m willing to pay that price in time, because many things that a blender user/artist forgot is that a developer DOES pay a price for it (as well as an artist do, of course).

I don’t have another blender dev in miles surrounding me, sadly, so I can not walk to someone else place and say: -”Hey! how do you add this button? How do I access this data from that?”, and lots of other minimal questions. Sure, there’s the bf-commiter list, but I can not flood it and bother a BUSY blender dev with them (my only way of communication), I simply have to do my homework and learn it the hard way: try and error, and this takes time.

I hope to have some days power over the Blender sources, creating my own branches, reviewing others patches and adding new functionalities without constant assistance of other core devs, in other words becoming myself a core dev, but I simply will have to EARN that, I will be the first one to forbid myself to apply for that until I consider myself skilled enough to tackle TO THE END whatever project I choose.

But I’m not worried, I trust in myself and know that I simply need time and plenty of MOTIVATION and JOY about what I’m doing. Not long ago, before I started the photon mapping stuff (which happen at the same time as Matt and who has allowed me to reuse part of his code) I was enjoying deep creativity but right after a motivational crisis that I have only shared with my closest friends. Indeed, while Matt was polishing Volumetrics, I saw in the horizon its completion with less and less for me to contribute in. That was good for Blender but bad for me because I have a DEEP CREATIVITY NECESSITY that I need to keep constantly fed otherwise I risk emotional collapse. (go ahead, call a doctor, I’m crazy )

So you know now my secret, I really NEED to create stuffs, as soon as what’s left for finishing a project is the “plumber job” of integrating it into Blender I start to get demotivated: not saying I will not finish it, it’s just that I need to jump elsewhere in order to gather new strengths and skills that will allow me to comeback and ease that pain later. One of the best things that have happen in my life was using blender as a dev, because it will keep my demon busy for the comming years . Since my childhood I have never gotten enough by watching TV, reading, playing. I
used to write books, draw, paint (on canvas), sculpt, make chemical experiments, animate with the corner of my books (sorry, no computer back then), and many, many small things that I sometimes miss

I think my main contribution is not really these patches, but the HOPE they bring: I’m not just poking in a forum saying: -”I wish blender have implemented an integrated antivirus because it will be helpfull for….”
NO, I’m showing HOW blender will LOOK LIKE and BEHAVE with that feature, I’m saying : “look at this, this is very DOABLE and might be used as a REFERENCE or simply a HINT of where to start or don’t start ! “the first thing an open source dev does before start a project is to search for prevously existent attempts, and the more they find, the more probable he/she success will be (aside from his/her skills of course) so by now, when a dev will google for volumetrics, pathtracing, photonmapping, smoke and SPH in Blender,  Google will not answer only forums entries .
Sure, many of those projects can eventually be made without me as nobody is a God here, but I’m sure volumetrics would have been pushed further in the future if I had not started “True Volumetrics…”, and also my first smoke
simulations rendered in Blender have touched many peoples hearts and contributed a bit on making Daniel’s smoke to come true. GI has also shaken some minds about Render Internal development much more than years of asking for it in forums, the Pathtracer project contribute a bit to revive the interest in Luxrender, and what have happened with Luxrender for me? It’s an awesome project, but so complete at the point I planned to join it that it simply killed my motivation (Second deeper emotional crisis). For joining it I have to do only buggs correction jobs, that is OK and very preciated and needed but simply don’t take my creativity factor in that equation   I will come back to it when I get enough skills to tackle a big project in addition to bugfixing.

And last but not least, something that many people haven’t realized: Every submodule of Blender is ALSO an open source project and is not only the full responsability of the main dev of that area but of every interested dev: the
more devs working on a module, the better it will get, and I have always been open in my projects: I welcome every dev who want to take it and improve it. Help me, show me, teach me and I will do my part
Don’t worry, I don’t have a deadline and while those projects take time before completion, I keep making the community dreaming but not an ethereal dream, a very real one that will eventually come true: I’m working hard to become a pro.

Cheers  Farsthary

TADAAAAN! Smoothed Particle Hydrodynamics (SPH) for Blender particle system
(True Fluid Particles) !

Hi everyone

Long time without posting huh? well , I have prepared a surprise , I have been making recently some research in the area of SPH simulators (don´t blame me, I found this while I was gathering resources for Bloxel and since its development is shorter than Bloxel I decided not keep the community waiting ) and came up with the first draft of SPH physiscs for the powerful Blender Particle System from Janne Karhu (aka Jahka).

For the ones who think that SPH means something related to the stars I have to tell that in some way you are right: SPH was firstly developed in 1977 for simulating astrophysical processes . The fundamentals of the SPH theory are the interpolation theory, the units in the SPH method represent a region in space and not a real particle, thus field variables (pressure,viscosity,velocity,color, etc) are found by averaging, or smoothing the field variables over a region of interest. SPH is often described as a rugged, robust technique that could easily be extended to N dimensions while Eulerian approaches (Grid based methods) suffer from the curse of dimensionality (computation grows exponentially with dimensions).

Pros:

• Arbitrary scene set ups , no need to define a domain.
• Intuitive and artist friendly.
• Possibility to handle arbitrary complex interactions like hetero-fluids simulations, and unexpected deformations.
• Robust, adaptable and extensible.
• Good results even with a low number of particles.
• Represents a natural extension to the current Blender particle system.
• Excellent for some crazy fluid like particle FX´s .
• Represents one of the most versatile and flexible fluid simulator type.
• Does not depend on boundary conditions.

Cons:

Achieving the same level of accuracy as a grid based simulator like El’beem requires a high number of particles.

Todo

• Improving its speed and stability (particle KD-tree?)
• Jahka’s review and possible integration in his system?

(reviewing requires time, so don’t forget to support him and to mention this patch if you want to motivate him )

This new method is not a replacement for the Blender El´beem fluid simulator, it’s just a complement and a natural extention to the current particle system, like many other packages have implemented (Softimage, Houdini, Realflow, etc.)

I plan to release the patch for next week for further development and review, also the GUI is not finished and some design considerations should be made, this patch is nearly ready to go and it will not require any traumatic change in Blender so will be safe to commit.

Hope you like it

Farsthary

Hi all!

Though are not quite news, I thougth it worth post for it.
Matt Ebb (that is now a full time Blender dev at BF) have recently added full shadow support for volumetrics, at my opinion that was the final milestone needed to make it production ready, so currently volumetrics are an entire suite for handling participating media for artist
Of course, with time it will evolve and get better and better.

Cheers to all Farsthary

Hi, this is a friend of Farsthary.

Although the volumetric renderer is now maintained/optimized only by the great Matt Ebb (aka Broken) with some assistance from Alfredo de Greef (aka Eeshlo), I thought it would be interesting for the supporters of the previous blog “True Volumetrics in Blender” to have some news from time to time, so…
* If you follow Matt’s blog, you will notice that he has implemented the interaction between external shadows and volumetrics. Hence, for any volumetric-related stuff, follow his blog by now instead of this one.

* Nick  (aka NRK) has kindly recorded and shared some volumetric tutorials with Blender 2.5 (development version, official beta version to come in one month, grab a build here)

Volumetric + Nodes

Point density

thanks to him!

P.S: If you follow Daniel’s Blog, you can only be amazed by his additions: smoke and fire (in the Blender 2.5 development build) can now interact with force fields! Well, he has done in a few months an incredible work, which will save a lot of time to the Durian’s devs for other cool features!

You can also notice another fund-raiser. Honestly, he is a modest guy, so even if the previous fundraiser for buying his new computer (except the graphic card: he currently uses the integrated and low-end one) went up to 200%, it was below 500$ (see “Hall of Fame”) so he did not buy very expensive hardware. So, this new fundraiser is for him to buy a powerful graphic card compliant with the opencl standard.

What does it mean? Well, opencl is a new programming language which allows to make algorithms run on various GPU (recent nvidia and ati graphic cards). For the smoke and fire simulator, this means that he could develop with a new graphic card a way quicker simulator thank to the new graphic card computing power. (new graphic cards can outperform CPU for heavily parallel calculations like those involved in the smoke/fire simulator –> realtime previzualisation?). So every blender user who has a recent graphic card and want a ultra-fast simulator has a potential answer…

By the way, Farsthary will also investigate opencl programming, I just sent him recent drivers + programming documentation.