Wow, very .. exciting! love what it’s doing with the image processing and how it goes that easily into the GPU.
-Draise PH: +57 313 811 6821 From: Paul Doyle Sent: Friday, March 13, 2015 14:04:43 To: [email protected] Thought you might enjoy this one: https://vimeo.com/122130309 - image processing with Canvas using our GPU compute capability. "In this video we take a look at using Fabric Canvas for image processing. Since the KL language that underlies Canvas has full GPU support, we experimented with a continuous GPU pipeline for image processing. All nodes in the video (except for the file loaders) are live (not cached). This allows for interactive framerates on HD full float composites. Numbers on the test machine I7-3770k @ 3.5 GHZ, 16GB RAM NVidia K5000 Shot1: CPU: 1.42 FPS GPU: 8.15 FPS Shot2: CPU: 9.4 FPS GPU: 36.0 FPS" On 12 March 2015 at 10:48, Eric Thivierge <[email protected]> wrote: There is no standard particle system / nodes currently but eventually there will be. Nothing stopping you from doing one yourself or someone else doing it. I think TD's will have good employment opportunities in the short and long term with this kind of work. It's not a bad thing. :) Eric T. On Thursday, March 12, 2015 10:46:53 AM, Juhani Karlsson wrote: I would think so. There is not that much in particles going on. I remember seeing some SPH examples. I would also like to see simple standalone for Fabric so there would not be the application specific splice in the middle. Its nice to hear its planned. : ) - J On 12 March 2015 at 16:41, Chris Marshall <[email protected] <mailto:[email protected]>> wrote: From a non-character perspective though, are we still able to do particle type stuff or is Fabric / Canvas not geared up in that way? On 12 March 2015 at 13:14, Eric Thivierge <[email protected] <mailto:[email protected]>> wrote: One thing you kind of see / feel already is that you can start mixing things you normally couldn't in ICE. An example is to read in an alembic and have it as an input into a rig. An example is to load an alembic of your ground and have your leg solvers collide with it. Of note would be that you wouldn't actually have the alembic in your scene with the overhead that Softimage would carry with actual geo in the scene for point and poly editing. You could still draw the ground using the Inline drawing and eventually RTR but from what I've seen it doesn't impact performance hard when you do. You could also write out image sequences of where your character is walking or where meshes are colliding too since they have image libraries available. The work just needs to be done to map those areas to UV space. But it's possible. So many areas are within reach from what seems day one where with ICE we waited so long for so many features we never got. No standard tools for writing out images from ICE. :( Another major thing for me is that you can convert your code into nodes using the provided utility. You code up some methods using KL (at the level of javascript & Python in syntax complexity) and you get the nodes almost for free. One thing I'm looking forward to and that I've already experienced in our current production, is how easy it is to re-use code throughout different systems. I can't go into too much detail but imagine writing a solver that works for rigs in all your DCC's and can work in specialized tools such as crowd tools too with little extra work. :) It's awesome to see things working the same across DCC's and tools. Eric T. -- Chris Marshall Mint Motion Limited 029 20 37 27 57 07730 533 115 www.mintmotion.co.uk <http://www.mintmotion.co.uk> -- -- Juhani Karlsson 3D Artist/TD Talvi Digital Oy Tehtaankatu 27a 00150 Helsinki +358 443443088 [email protected] www.vimeo.com/talvi <http://www.vimeo.com/talvi>
