OpenSubDiv... what you might be missing.

Become a member of the CGSociety

Connect, Share, and Learn with our Large Growing CG Art Community. It's Free!

Thread Tools Search this Thread Display Modes
  05 May 2013
OpenSubDiv... what you might be missing.

At this year's FMX, I wandered into a a talk by Bil Polson thinking "Subdiv's... know that" but Bill's so future-thinking that I wanted to make sure I wasn't missing anything. And lookie there... I was completely off the mark with the extant of OpenSubDiv technology. Namely, I thought it's a smoothing algorithm and I've got it. Bill's talk opened my eyes, it's a lot more than that.

because most modelers/artists (as opposed to the tech guys) have a similar preconception of the OpenSubDiv, I wrote up a report at my blog. I know I'm sometimes slow to this stuff (jus' give me my wacom pen, eh), but feedback has shown I'm not alone.

we need OpenSubDiv as an industry standard not just so that my meshes get from one app to another in a consistent manner, but so that the underlying efficiencies make sure I can author, move and play with them while enjoying speedy feedback rates.

From my blog:
"advantages from the POV of us artists, because we need to be vocal about getting this implemented across the board… from CAD software to games to film and vfx software. A world of play-nice toolsets and immediate artistic feedback that you can trust to hold up across pipelines is just too necessary. 1) model clean, lean meshes that subdivide according to defined creasing values (or edge-weighting) and send these meshes from one tool to another with consistent results (go to 26:50 in the video above)
2) get rapid feedback of detailed surface displacement not only while sculpting, but while animating and and lighting
3) manage large scenes efficiently via level-of-detail controls that optimize hardware usage. This applies to distance-from-camera controls as well as to localized displacement effects such as Merida’s horse Angus plodding through snow.
4) know that your work is compatible with further feature sets like ptex texturing

SubDivs? You know you want ‘em!"
  05 May 2013
What's the difference between Pixar's OpenSubDiv and Pixar SubD in Modo that's been in it for a few versions now? Or has Modo just adopted the workflow early? They seem to have the same features.

It is certainly a fantastic toolset. Modeling complex SubD surfaces without getting drowned in a mess of edge loops is simply amazing.
  05 May 2013
that's what had me... I was so happy to get subdiv support in modo. Makes me more secure in prepping meshes for use in maya, etc.
So while modo is an early implementer of the format, it doesn't have the whole Subdiv. From what Bill said at the talk, no one rally does yet. It requires fairly deep changes to the way the cpu and gpu are directed. So it's a hardware as much as a software issue.

The larger issue is the heirarchical support - so that you get uniform behavior across apps, but more importantly - fast response while sculpting / animating / shading within modo itself. I don't know how far the implementation goes to be honest, but I'm sure that modo is up there with Maya in pursuing deep implementation.
  05 May 2013
My jaw dropped as I saw a siggraph video last year on youtube. The osd toad character which had millions of polygons was animating in realtime in Maya's VP2 view. A more recent video showcases dynamic tesellation too...

Modo has very powerfull sculpting tools that work effortlessly on multimullion poly geometry today, so I wonder if they've got some sort of OSD tech already incorporated in the latest version...

I was an advocate of having the sculpting and painting power of mudbox within Maya to see my highpoly animated deformations in the view. Now through OSD, this seems to become reality for all apps soon.
"Any intelligent fool can make things bigger, more complex & more violent..." Einstein
  05 May 2013
Quote: What's the difference between Pixar's OpenSubDiv and Pixar SubD in Modo

I can see how this could be misleading...

A few years back Modo worked with Pixar to tweak some of the boundary interpolation and crease rules to match the modes supported by PRman. Somehow this eventually turned into a sticker on the box claiming "Pixar subdivision compatibility"...

AFAIK, Modo uses its own code (written by a very talented developer in Japan if memory serves me right), not OpenSubdiv. Maya is in a similar situation where Autodesks's original algorithms have been tested and adapted to match a subset of the spec found in OpenSubdiv and PRman. I'll add the same caveat as in my previous posts : i can test a few shapes, but a thorough QA validation is mathematically impossible.

Comparatively, while Modo's code is very efficient with topology editing workflows (which currently OpenSubdiv does support well), I don't think it uses the GPU to subdivide or tessellate its geometry (yet).
  05 May 2013
Thread automatically closed

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.
CGTalk Policy/Legalities
Note that as CGTalk Members, you agree to the terms and conditions of using this website.
Thread Closed share thread

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

vB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Society of Digital Artists

Powered by vBulletin
Copyright ©2000 - 2006,
Jelsoft Enterprises Ltd.
Minimize Ads
Forum Jump

All times are GMT. The time now is 02:00 AM.

Powered by vBulletin
Copyright ©2000 - 2018, Jelsoft Enterprises Ltd.