CGTalk > Main Forums > General Discussion
Login register
Thread Closed share thread « Previous Thread | Next Thread »
 
Thread Tools Search this Thread Display Modes
Old 10-10-2012, 05:34 PM   #16
th3ta
Veteran
 
th3ta's Avatar
portfolio
3d Artist
In a dark office, USA
 
Join Date: Dec 2003
Posts: 739
I really have no idea how to do this, but I've been trying to do stuff with Motion vectors in Nuke for awhile now, without any luck. Once you have a way to access that data, there are a lot of really cool things you can do with it.
 
Old 10-10-2012, 05:42 PM   #17
ChrisCandide
New Member
portfolio
Chris Hatfield
Norfolk, US
 
Join Date: Dec 2008
Posts: 21
Wishful, thinking. Somehow we establish a think tank. Still waiting on the paperman siggraph paper to be released.
 
Old 10-10-2012, 11:52 PM   #18
CGIPadawan
Part-Time Blenderite
 
CGIPadawan's Avatar
Giancarlo Ng
Quezon City, Philippines
 
Join Date: Nov 2006
Posts: 6,779
Send a message via Yahoo to CGIPadawan
Quote:
Originally Posted by th3ta
I really have no idea how to do this, but I've been trying to do stuff with Motion vectors in Nuke for awhile now, without any luck. Once you have a way to access that data, there are a lot of really cool things you can do with it.


Forget motion vectors. What I was thinking of was to "cheat" the effect with Morphing and Compositors kinda like this:



Think of it like that line Bruce Lee wrote in his Self-Defence Manual:

"To win a fight quickly, just take aim, and hit the groin."

I'd try this one myself but I'm kinda tied up right now but I think this could be a good starting point to "fake" a similar result.

Forget proprietary "active vectors". The current apps can't do that.
__________________
"Your most creative work is pre-production, once the film is in production, demands on time force you to produce rather than create."
REVERSION
 
Old 10-11-2012, 12:36 AM   #19
ChrisCandide
New Member
portfolio
Chris Hatfield
Norfolk, US
 
Join Date: Dec 2008
Posts: 21
Yea, there has to be a more time productive way though. I know illustrate, a plugin for 3ds max, has away of controlling the line thickness via vertex.

I may have to try, the above method, when my schedule clears up in November.
 
Old 10-11-2012, 12:50 AM   #20
CGIPadawan
Part-Time Blenderite
 
CGIPadawan's Avatar
Giancarlo Ng
Quezon City, Philippines
 
Join Date: Nov 2006
Posts: 6,779
Send a message via Yahoo to CGIPadawan
Quote:
Originally Posted by ChrisCandide
Yea, there has to be a more time productive way though. I know illustrate, a plugin for 3ds max, has away of controlling the line thickness via vertex.

I may have to try, the above method, when my schedule clears up in November.


Yeah.. I honestly think I didn't think this one through enough.. As I don't think "Plate C" can be part of the final composite.. it's just a "cookie-cutter" intended to make sure you get "perfect sihouette" and you have no "shakes" like globs of black lines in the silhouette becoming Morph start/end points.

It's possible though if the line art drawing artist is skilled enough you don't need "Plate C" at all.

So that's going to shave time and steps from it.

Oh and there will be "time trade-offs". For example, the 3D models that will form the base motion and the main color/blurring data don't have to be done to the same "final-final" quality that you'd need if you were doing it as straight CG and you don't necessarily need texture maps, just very nice basic shaders.

The concept is that the line-art drawing will add in all the detail. So technically the time schedule for some of the other stuff in a straight-CG production will be traded for time given to the line-art drawing which will be less frames (say every 48 frames where the Color Render was done straight from 1 to whatever just to make sure you have nice motion blur).

We can think about how to fine-tune it after we have test. :P
__________________
"Your most creative work is pre-production, once the film is in production, demands on time force you to produce rather than create."
REVERSION

Last edited by CGIPadawan : 10-11-2012 at 12:57 AM.
 
Old 10-11-2012, 03:54 PM   #21
DanHibiki
Expert
 
DanHibiki's Avatar
portfolio
- -
Senior 3d Animator
Invivo communications
Richmond Hill, Canada
 
Join Date: Sep 2006
Posts: 1,801
would be really funky if you can mix this with video (extract the motion with some video tracking).

It would definitely make for some awesome Bakshi esque animation.
 
Old 10-12-2012, 12:17 AM   #22
CGIPadawan
Part-Time Blenderite
 
CGIPadawan's Avatar
Giancarlo Ng
Quezon City, Philippines
 
Join Date: Nov 2006
Posts: 6,779
Send a message via Yahoo to CGIPadawan
Quote:
Originally Posted by DanHibiki
would be really funky if you can mix this with video (extract the motion with some video tracking).

It would definitely make for some awesome Bakshi esque animation.


What do you mean? Like "A Scanner Darkly"?
__________________
"Your most creative work is pre-production, once the film is in production, demands on time force you to produce rather than create."
REVERSION
 
Old 10-12-2012, 01:51 AM   #23
th3ta
Veteran
 
th3ta's Avatar
portfolio
3d Artist
In a dark office, USA
 
Join Date: Dec 2003
Posts: 739
How does this compare to what Flat Black Films have been doing with Scanner Darkly and Charles Schwab commercials? It was my understanding that they have a similar process where the strokes are tracked along the video. I guess the big difference is this is using 3d data and not just 2d tracking?
 
Old 10-12-2012, 03:04 AM   #24
ChrisCandide
New Member
portfolio
Chris Hatfield
Norfolk, US
 
Join Date: Dec 2008
Posts: 21
last problem would be finding the best method of interpolating the line work. How we well would flash or After Effects handle. I'm in complete agreement with you CGIPadawan, definitely a trade off interms of workload compared to a straight-CG production.

@th3ta
Quote:
How does this compare to what Flat Black Films have been doing with Scanner Darkly and Charles Schwab commercials? It was my understanding that they have a similar process where the strokes are tracked along the video. I guess the big difference is this is using 3d data and not just 2d tracking?

I think it is a little different. The method CGIPadawan suggested is similar, not in methodology, but overall outcome. It could be considered a permutation of Rotoscoping.
 
Old 10-12-2012, 03:35 AM   #25
ChrisCandide
New Member
portfolio
Chris Hatfield
Norfolk, US
 
Join Date: Dec 2008
Posts: 21
My initial idea was to model the head, in the reference below, and use the method CGIPadawan suggested. The video reference below would be used as a litmus test, to see how well the line work holds. If anyone wants to give it a go, its all yours. Unfortunately, I can't do anything until November. So I'm all talk right now.


http://www.youtube.com/watch?v=sZ7E...eature=youtu.be

[IMG][/IMG]
 
Old 10-12-2012, 04:14 AM   #26
CGIPadawan
Part-Time Blenderite
 
CGIPadawan's Avatar
Giancarlo Ng
Quezon City, Philippines
 
Join Date: Nov 2006
Posts: 6,779
Send a message via Yahoo to CGIPadawan
Quote:
Originally Posted by th3ta
How does this compare to what Flat Black Films have been doing with Scanner Darkly and Charles Schwab commercials? It was my understanding that they have a similar process where the strokes are tracked along the video. I guess the big difference is this is using 3d data and not just 2d tracking?


"A Scanner Darkly" essentially gets rid of the entire original footage and you replace each frame ( as in 1 to 100,000) with hand-drawing. There is a "melting" quality to it and some looseness and quivering. Plus you throw away the entire live action footage.

The Disney method of Paperman.. The principle of it is that you take some aspect of the original footage (Vector, blur, and Diffuse/Color) and then augment it with custom Vector line drawings to produce the final image. They also use auto-morph in-betweening to eliminate any "quiver" or "shake" in the image.

This is why my proposed "hack job" version is simply to composite from select key frames (rather than each and every frame) and use morphing to get in-betweens.
__________________
"Your most creative work is pre-production, once the film is in production, demands on time force you to produce rather than create."
REVERSION
 
Old 10-12-2012, 04:20 AM   #27
CGIPadawan
Part-Time Blenderite
 
CGIPadawan's Avatar
Giancarlo Ng
Quezon City, Philippines
 
Join Date: Nov 2006
Posts: 6,779
Send a message via Yahoo to CGIPadawan
Quote:
Originally Posted by ChrisCandide
last problem would be finding the best method of interpolating the line work. How we well would flash or After Effects handle. I'm in complete agreement with you CGIPadawan, definitely a trade off interms of workload compared to a straight-CG production.


You have to morph.. graphically... between two key frames.. that's what I think has to happen.

Say you have a 5 second sequence... Your 3D app outputs 5 x 24 = 120 frames
The 120 frames have vector blur but no "line art".

You then take one frame from each 24 frame set (5 frames total) and select those for drawing (note the base images include blurs so the artist has to take them into account).

You use an Alpha-Cut-Out "Cookie Cutter" to overlay the line art over the 5 frames you chose.
Then you have to use another software (After Effects?) to morph directly from the 5 frames each to make the 23 frames that should have gone in between them.

You'd have to change the method if the motion is elaborate because you might need more hand-drawn keys for specific poses to transition properly instead of relying all the time on the morphing.

But this method should get you perfect "line art interpolation" with zero shakes. If you want to tweak the motion you just have to pull out more frames from the "mother set" do more line-art.. use the Cookie-Cutter in compositing it... put it in sequence and morph again.

Again.. that's the "hack job" theory.
__________________
"Your most creative work is pre-production, once the film is in production, demands on time force you to produce rather than create."
REVERSION
 
Old 10-12-2012, 01:56 PM   #28
BigPixolin
Banned
 
BigPixolin's Avatar
c none
US
 
Join Date: Jun 2007
Posts: 2,764
I'm confused. I can see the benefits for the people creating the content, but how is the final result that much different from cell shaded 3d we seen a million times?
Is the short online?
 
Old 10-12-2012, 04:53 PM   #29
RobertoOrtiz
[Forum Leader]
 
RobertoOrtiz's Avatar
CGTalk Forum Leader
portfolio
Roberto Ortiz
Illustrator/ Modeler
Washington DC, USA
 
Join Date: May 2002
Posts: 31,806
Send a message via MSN to RobertoOrtiz
Quote:
Originally Posted by BigPixolin
I'm confused. I can see the benefits for the people creating the content, but how is the final result that much different from cell shaded 3d we seen a million times?
Is the short online?

In my opinion, Cell shaded work tends to looks rather stiff, no matter how good the cg animation work is. Master level Traditional animation provides a flexibility to the character that honestly it is very hard to capture in CG.

What they did in paperman is they a way to capture the flexibility of traditional animation, while combining it with CG animation. As the Animation director of the short said at siggraph,
the cell chader of the short is a traditional animator.
__________________
LW FREE MODELS:FOR REAL Home Anatomy Thread
FXWARS
:Daily Sketch Forum:HCR Modeling
This message does not reflect the opinions of the US Government

 
Old 10-12-2012, 05:22 PM   #30
th3ta
Veteran
 
th3ta's Avatar
portfolio
3d Artist
In a dark office, USA
 
Join Date: Dec 2003
Posts: 739
Quote:
Originally Posted by CGIPadawan
"A Scanner Darkly" essentially gets rid of the entire original footage and you replace each frame ( as in 1 to 100,000) with hand-drawing. There is a "melting" quality to it and some looseness and quivering. Plus you throw away the entire live action footage.

The Disney method of Paperman.. The principle of it is that you take some aspect of the original footage (Vector, blur, and Diffuse/Color) and then augment it with custom Vector line drawings to produce the final image. They also use auto-morph in-betweening to eliminate any "quiver" or "shake" in the image.

This is why my proposed "hack job" version is simply to composite from select key frames (rather than each and every frame) and use morphing to get in-betweens.



Yeah I get those differences, I was just speaking specifically to the automated tracking of strokes. That specific tech seems similar to what Flat Black Films have been doing.
 
Thread Closed share thread


Thread Tools Search this Thread
Search this Thread:

Advanced Search
Display Modes

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
CGSociety
Society of Digital Artists
www.cgsociety.org

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

All times are GMT. The time now is 04:13 AM.


Powered by vBulletin
Copyright ©2000 - 2016, Jelsoft Enterprises Ltd.