PDA

View Full Version : SIGGRAPH 2010 Trends (Tell us)


RobertoOrtiz
07-25-2010, 11:45 PM
This is a thread for those attending the 2010 conference.
So what graphic threads have you observed in the SIGGRAPH 210 conference?


What technologies have you observed at the lectures?
What tech are companies trying to push?
What emerging technologies have you seen that caught your eye?

Looking forward to your comments and observations.
-Roberto

Dang3rousd
08-01-2010, 03:44 AM
Kind of surprised no one has responded to this yet so I guess I'll start off.

I haven't been to the conference since 2006 so I'm looking at it through a bit of a wider lens than people who go every year and see the change slowly I think. What stood out to me was how heavy mocap was this time around. There seemed to be tons of exhibits that were all trying to sell their new real time mocap tools. It felt like everyone was aiming for a system that not only was cleaner than previous versions but was giving real time visuals to go along with it. Even the emerging technologies was heavy with mocap of varying kinds.

The second thing that I noticed the most was 3D. Everyone is trying to jump into that with both feet. It was in the seminars, the exhibition, the emerging tech's, you couldn't get away from it.

As a modeler I was focused a little more on things related to that and of course Zbrush and Mudbox were both showing off their new goods. Autodesk was really pushing the cross platform use with Mudbox, Maya, Motionbuilder, and the Softimage spin-offs. They even were creating some strong tie-ins with Photoshop and I wouldn't be surprised if in the long run they either created their own version or bought out someone like adobe to bring it all in-house so that you'd never use anything but an Autodesk program.

I guess to sum up, for me, the future as it relates to the cg marketplace is all about mocap and 3D until something newer and shinier comes along.

RobertoOrtiz
08-03-2010, 12:20 AM
OK here is my take on trends;

Stereo 3D is here to stay
We better start familiarizing ourselves with Spherical Harmonics.
GPU computing will soon be the norm
Open formats will gain more traction
And Disney support for SIGGRAPH rocks.

MagnumCort
08-03-2010, 01:05 PM
Agreed - Thanks to Disney for the Geek Bar. A great place to go where it was cool (temperature-wise), quiet, and that had a great wi-fi connection. They gave great talks and I enjoyed their dedicated Disney room (next to the Geek Bar) An excellent sponsor.


Mocap and 3D have been huge the last two years.

It seemed like there was less emphasis on GPUs this year than last, but admittedly my selection of talks was considerably different than last year. That said, NVidia was all over the place. I remember thinking at one point, "sheesh, what happened to ATI?" Granted, ATI had a booth, but it seemed like nVidia was everywhere and teamed up with everyone.

RobertoOrtiz
08-03-2010, 05:38 PM
It seemed like there was less emphasis on GPUs this year than last, but admittedly my selection of talks was considerably different than last year. That said, NVidia was all over the place. I remember thinking at one point, "sheesh, what happened to ATI?" Granted, ATI had a booth, but it seemed like nVidia was everywhere and teamed up with everyone.

Funny I asked a friend of mine who is more into this that I am and he said:

Cuda happened:

http://www.nvidia.com/object/cuda_home_new.html

taffy77
08-03-2010, 07:01 PM
Deep Compositing
Spherical Harmonics
Stereo

ArtisticGore
08-04-2010, 01:56 PM
Yeah I heard that term "Deep Compositing" around in many of the lectures. Not sure if I got the grasp of it. What exactly is deep compositing?

RobertoOrtiz
08-04-2010, 02:10 PM
This might help:
http://www.deepimg.com/

From wikipedia:
Deep image compositing is a recently emerged way of compositing and rendering digital images. In addition to the usual color and opacity channels a notion of depth is created. This allows multiple samples in the depth of the image to make up the final resulting color. This technique is producing high quality results and removes artefacts around edges that could not be dealt with otherwise.

http://en.wikipedia.org/wiki/Deep_image_compositing

noizFACTORY
08-05-2010, 12:40 PM
This might help:
http://www.deepimg.com/

From wikipedia:
Deep image compositing is a recently emerged way of compositing and rendering digital images. In addition to the usual color and opacity channels a notion of depth is created. This allows multiple samples in the depth of the image to make up the final resulting color. This technique is producing high quality results and removes artefacts around edges that could not be dealt with otherwise.

http://en.wikipedia.org/wiki/Deep_image_compositing

Thanks for the links. Very interesting. Here's the paper from that page you linked:
http://www.johannessaam.com/deepImage.pdf

Animal Logic guys seem to be working on this so if anyone from there is reading this then please do shed some light on this technique. The part about storing the deep shadow info from the render camera was a little confusing.
We use the same rendering
approach but store deep data from the render camera, still using
Pixar’s deep shadow map format which records an opacity function
over the distance to the camera for each pixel.
So basically, the way deep shadows are stored from the light's perspective (i.e. the light camera), with this technique, the opacity of each pixel being sampled from the render camera is stored w.r.t/alongwith the distance of that pixel from the camera (i.e the depth)?

CGTalk Moderation
08-05-2010, 12:40 PM
This thread has been automatically closed as it remained inactive for 12 months. If you wish to continue the discussion, please create a new thread in the appropriate forum.