my.cebas.com | Contact
 
| Home | News | Products | Online Store | Community | Service & Support | Downloads | Company | Login

 Testimonial

thinkingParticles‘ rise to the challenge in “The Witcher 2”

 

thinkingParticles works its magic in the EPIC intro film of "The Witcher 2"!

 

 

 

 

 Åukasz Sobisz' works as a Lead Simulation TD at  Platige Image, Poland.

 His professional background is a mixture of  programming and technical skills  including cloth, fluid  and particle simulations.

 

 

 

cebas: It seems that The Witcher 2 was an important project for you and Platige Image. Can you tell us how you were involved?

Platige Image was involved in all previous cinematic trailers and intros for The Witcher games, but this one appeared to be very special - not only for its original and a little unconventional storyline - but also tremendous amount of work to be done. Including all kinds of simulations.

 

                                                       © CD Projekt RED | Platige Image | Namco Bandai | WB Games

 

cebas: Which shots/images/sequences in particular were you involved in?

Our fx team was involved in all computationally heavy effects and simulations including cloth, particles and of course RBD destruction.

 

cebas: What cebas software did you use and why?

It was thinkingParticles from the very beginning. Although it was the first project to rely on tP so extensively, it was quite obvious for me that no other software (especially within 3ds max environment) can guarantee stable rigid body sims to this extent while providing flexible and procedural approach at the same time. We've used tP to simulate rigid bodies in all demolition shots, but it was also the main tool for all kinds of particle effects including debris, dust, spells or blood.

 

 

cebas: How did you use our software to achieve the effects?

When the layout part was done, we knew it's more important than ever to keep our work as close to the animatic as possible. To minimize sim iterations, we've used layered dynamic sets and caching tools in tP to operate on smaller chunks of data at a time. At first, the primary setup structure was created. Particle groups were designed to reflect the logical structure of the ship and stood untouched through most of the shots. That allowed us to move between the shots without too much pain and exchange data smoothly.


cebas: What features in particular helped you achieve your goal and how?

Main credits go to the volumeBreaker and SC nodes; I hardly remember any failure of these two. vB gave us the control to break everything procedurally and minimize or avoid any pre-fragmentation. It's great to have this kind of freedom when you iterate through the shots. Shape Collision is incredibly stable and fast even with complex stacking and lots of contacts.

 


cebas: What was the most difficult aspect of this project and how did you solve it?

In terms of demolition sims, we've had some hard times getting into the material management; mostly because of the huge amount of objects carrying complex shaders and the fact that nearly everything had to be considered breakable. All these had to get pushed through common dynamic sets and come out with proper material IDs assigned. With few types of objects it's easy to track errors, but having hundreds of different materials in the sim was quite a challenge. We ended up with a script, which modified all the source materials, so they were consistent in terms of IDs. That way, whatever was added to the system - the result was correct and rendered properly.


cebas: What was a step-by-step breakdown of a typical shot/image/sequence?

For demolition sims, we've tried to use a similar approach for every shot. Every simulation layer was calculated with great relevance to the previz version and tweaked until the desired look was achieved. Consequently, we've been adding successive dynamic sets - usually going from the deck and collapsing mast through shattered sails and ropes on top of that, followed by debris. Every piece of input objects was covered with ice geometry, so each simulation layer was usually followed by a corresponding ice layer. The setup was also handling Fumefx emission. To deal with changing topology of fragments we decided to create another layer of particles to use them as emitters on cracked surfaces.

  

 

cebas: How did the cebas tools perform for you and how was the experience of working with them like?

thinkingParticles is very stable and predictable. We've had some minor crashes, but it's not a surprise with such heavy scenes. Generally, it was a whole new way of thinking about procedural particle setup.

 

 

cebas: How did it compare to other software you may have used?

I think that tP, with its deep procedurals, outperforms other particle solutions for max. However, I find pFlow still quite useful for simple tasks. On the other hand, the Blackbox feature - among others - encourages combining and reusing your solutions, so thinkingParticles' field of usage is growing quite rapidly in various tasks.


cebas: How did cebas software integrate into your production pipeline? How straightforward was it?

tP integrates very well because it works in a familiar max environment. No problems here.

  


cebas: What was the most fun or rewarding part of this project for you?

I think that the most precious thing was to watch the team evolving and exploring new areas of possibilities. It’s amazing how much you can learn, when the tool lets you experiment and at the same time gives a solid structure to rely on.


cebas: What do you wish cebas software did that it's not currently doing for you?

As far as I know, the most important things are on their way already :) However, there are some smaller issues that could make life much easier.

I miss predefined vector and matrix math nodes. It would be nice to see more nodes using multithreading (pSearch, PPassAB) and the whole system evolving in this area. Robust particle collisions with deforming meshes would be great too. Last but not least: geometry lookup node, so we could directly find nearest point on the mesh without using workarounds.

 

cebas: What new projects can we expect from you in the future (if you're able to tell us).

There is a lot happening right now at Platige. Unfortunately I'm not allowed to talk about any new projects at the moment.

 

 

We can't wait to learn about your upcoming projects!

Thank you Åukasz for your time, we absolutely love your work :)

 

Check out the final product of these behind-the-scenes footage HERE ...Simply.Flawless.

You can check out Platige Image's website to find out more about their accomplishments so far :)

Don't forget to check back soon for more exciting one-on-one interviews to come!

 

©  2013 cebas Visual Technology Inc.  |  Follow:  Twitter   YouTube   LinkedIn   Facebook   RSS  |  Privacy Policy   |  Terms & Conditions  | Terms of Sale Sitemap