PDA

View Full Version : GENOMA and the 12 principles of CharacterAnimation (VIDEO) How to do this?



robertoortiz
12-28-2014, 08:09 AM
http://vimeo.com/93206523 ( cool short video on the 12 principles of Character Animation )
What are the 12 principles?
The 12 basic principles of animation were developed by the 'old men' of Walt Disney Studios, amongst them Frank Thomas and Ollie Johnston, during the 1930s. Of course they weren't old men at the time, but young men who were at the forefront of exciting discoveries that were contributing to the development of a new art form. These principles came as a result of reflection about their practice and through Disney's desire to use animation to express character and personality.


One of the coolest developments in recent memory for Lw is the ongoing development of GENOMA.
Of course this is work in progress and it still growing.

I think one are that could assist this growth is a more seamless implementation of Disney's 12 principles of Character Animation.
Yes you can do them in Lightwave now, but you have to jump through hoops to get them to wo.

What are the 12 pinciples?
The 12 basic principles of animation were developed by the 'old men' of Walt Disney Studios, amongst them Frank Thomas and Ollie Johnston, during the 1930s. Of course they weren't old men at the time, but young men who were at the forefront of exciting discoveries that were contributing to the development of a new art form. These principles came as a result of reflection about their practice and through Disney's desire to use animation to express character and personality.

So what do you guys think? Lets think outside the box.
I am looking forward to your comments and suggestions.

-R

RebelHill
12-28-2014, 08:34 AM
you can do them in Lightwave now, but you have to jump through hoops to get them

No you don't... at least not any more so than you have to in any other package (be it 2D or 3D).

The principles of animation are a collection of artistic and technical guidelines for how to go about creating animation, and the performance of the craft... that's not really something you can somehow "build in" to a tool in any way... it's up to the animator to apply them. Or do you expect that there can be some sort of "make striking/charismatic pose" button??

robertoortiz
12-28-2014, 08:52 AM
How make striking/charismatic pose" button??
You cant, but you can do things like,
The ability to save poses .
Automatic symmetry tools on rigs


to be fair to LW I do have a beef with the development of animation tools in 3d because I think there are things that could be done more seamless.
the way animation in 3d has evolved it seem (at least on the of the shelf side, proprietary animation studio software is another story) it seem to have evolved from principles form the early 90s.
what I am suggestion it to use the computer more as an assistant than just another tool.

And yes I feel strongly that you can develop/streamline software around core principles. it is like something to aim for in development.

And so we are on the same page here are the 12 principles

.1 Squash and stretch
1.2 Anticipation
1.3 Staging
1.4 Straight Ahead Action and Pose to Pose
1.5 Follow Through and Overlapping Action
1.6 Slow In and Slow Out
1.7 Arc
1.8 Secondary Action
1.9 Timing
1.10 Exaggeration
1.11 Solid drawing
1.12 Appeal (OK ill give you appeal)

http://en.wikipedia.org/wiki/12_basic_principles_of_animation

ernpchan
12-28-2014, 09:01 AM
If the software can do all these principles on its own, then a bunch of us are gonna be out of work soon, haha.

Some of that stuff is already possible, we have dynamics for secondary animation. Timing and slow ins and outs is graph editor manipulation. messiah has a neat breakdown function that let's you skew the timing of a pose.

For me, making it easier to pose a character with fast and as few controls as possible would be nice. Rigs often have so many controls it's overwhelming to have to deal with so many nulls, sliders, switches,etc just to make one pose.

robertoortiz
12-28-2014, 09:09 AM
For me, making it easier to pose a character with fast and as few controls as possible would be nice. Rigs often have so many controls it's overwhelming to have to deal with so many nulls, sliders, switches,etc just to make one pose.
Agreed. It is amazing how some of the new studio high end studio animation pipelines are embracing simplicity in their workflow.(Dream works comes to mind)

RebelHill
12-28-2014, 09:19 AM
(OK ill give you appeal)

Really?? What about exageration, staging, anticipation, slow in and out, straight ahead and pose to pose, follow through, overlap... and even ALL the rest?? How EXACTLY can you provide for these things in some magical way that you don't already have plenty options for or which are somehow quicker or easier to achieve than those already available??

Pose to pose... set keys to stepped... done.
Follow through and overlap... You can't "automate" that... you have to key it. Should an overlap/follow through happen over one frame? 2? 50? Should it be a small motion, medium, large? Should there be a hang at the extreme, or a rapid return?? How on earth can you somehow "provide" for an INFINITE number of such options other than by just letting the animator put things where they need to be at a given time (keys and graphs)??

It just plain nonsense... there is no way you can somehow "build in" the animation principles like this... they're artistic principles for animators to follow... NOT something you can just provide for with software.


The ability to save poses . Automatic symmetry tools on rigs

Thats it?? Well then, here you go... Done!

https://www.youtube.com/watch?v=7LxtV07BEl0&list=PLTds3QePYrWGUd7fMuNAivIsORvpnKDqX&index=28


messiah has a neat breakdown function that let's you skew the timing of a pose.

Yep... Done that too... https://www.youtube.com/watch?v=YCQrwJvO7oM


Rigs often have so many controls it's overwhelming to have to deal with so many nulls, sliders, switches,etc just to make one pose.

And that one... https://www.youtube.com/watch?v=7x83r9J0OI4&t=30m55s


It is amazing how some of the new studio high end studio animation pipelines are embracing simplicity in their workflow.(Dream works comes to mind)

Umm... seriously?? Have you ever actually seen any of the behind the scenes stuff from the like of dreakworks, pixar, et al... Its not uncommon for a single character to have OVER 1000 individual controls and animatable parameters available to the animator... Just because they're not all scattereed around right there in the viewport in some unholy visual mess doesnt mean you've created something that's simpler to operate.

brent3d
12-28-2014, 10:24 AM
How make striking/charismatic pose" button??
You cant, but you can do things like,
The ability to save poses .
Automatic symmetry tools on rigs


to be fair to LW I do have a beef with the development of animation tools in 3d because I think there are things that could be done more seamless.
the way animation in 3d has evolved it seem (at least on the of the shelf side, proprietary animation studio software is another story) it seem to have evolved from principles form the early 90s.
what I am suggestion it to use the computer more as an assistant than just another tool.

And yes I feel strongly that you can develop/streamline software around core principles. it is like something to aim for in development.

And so we are on the same page here are the 12 principles

.1 Squash and stretch
1.2 Anticipation
1.3 Staging
1.4 Straight Ahead Action and Pose to Pose
1.5 Follow Through and Overlapping Action
1.6 Slow In and Slow Out
1.7 Arc
1.8 Secondary Action
1.9 Timing
1.10 Exaggeration
1.11 Solid drawing
1.12 Appeal (OK ill give you appeal)

http://en.wikipedia.org/wiki/12_basic_principles_of_animation


LIghtwave Principles of Animation Tutorials (http://www.twinvillecreative.com/#!tutorial-preview/c1ysd)
or here YouTube Training Vids (https://www.youtube.com/user/alleyne3d)
Please see my vids regarding Squash and Stretch, Anticipation, and Follow Through/ Overlap at the links above. LW is definitely capable of doing the principles of animation, but there aren't many examples online which is why I made my videos, I believe that's mostly because LW isn't in schools. The classroom is where the knowledge and performance of the principles of animation are used as the guideline to what 2D or 3D animation is, and whatever the software the school uses has to show clear examples of that. So if your saying "hey everybody let's get some good examples and tutorials of the 12 Principles up online to help our users and to show how easy it is to animate in LW" then I'm totally down with that. Good examples and documentation is all that's needed at this point, the rest is skill, passion, and dedication.

Anticipation example:

https://www.youtube.com/watch?v=NillI-PzCHQ

Surrealist.
12-28-2014, 11:16 AM
The trend is away from zillions of controls and to more of an artist friendly interface. Controls are best hidden. I love this idea.

Pixar's presto:

http://www.cartoonbrew.com/tech/watch-a-rare-demo-of-pixars-animation-system-presto-98099.html

Animat in Blender:

http://blenderartists.org/forum/showthread.php?350402-ANIMAT-Widgetless-On-Model-Animation-in-Blender-UPDATE-1-13-Oct-2014

https://www.youtube.com/watch?v=OI81pu4KzgU

RebelHill
12-28-2014, 11:38 AM
Thats not really moving away from having all the controls though... Its just kinda "proxy picking"... The controls are still there, they're just either hidden, or represented with geo rather than nulls... Its a piece of cake to unclutter your character in LW like this.


https://www.youtube.com/watch?v=Lpasdl_x8ZM

Surrealist.
12-28-2014, 12:42 PM
Yes something like that. Where the controls are hidden. I like the Blender one, as it actually hides the control even when you have it selected, so it is like dragging the mesh around (which could be a draw back but there might be an option to turn them on when needed), but I love the idea of just always looking at the mesh and not having stuff in the way. It is distracting - to me. I always hide the controls to see my animation back (as of course most animators do). But I like the idea of it not being there at all even when I am animating.

For instance I love how the Pixar version works on the face. It is just click and drag on the mesh.

This, in my opinion would make it easier and faster to pose and so on.

RebelHill
12-28-2014, 01:13 PM
I like the Blender one, as it actually hides the control even when you have it selected, so it is like dragging the mesh around

yah... also works fine in LW...


https://www.youtube.com/watch?v=44F9kv3gRAk


I love how the Pixar version works on the face. It is just click and drag on the mesh.

Also totally doable... All you gotta do is chop your mesh up into bits, and use these proxy pieces as your control null items, pick, operate, etc. At the end of the day... however it appears visually, what's hidden what isn't, that's one thing. But what it doesnt represent is any kind of simplification of control in any way, as some seem to think. You've still got the exact same controls, same number, operating the same way... its just their "item shape" which is different.


This, in my opinion would make it easier and faster to pose and so on.

I fail to see how... again... the actual controls themselves are no different, just their shape/appearance you still gotta do the exact same amount of work manipulating them. If anything... Id presume that for all the nice visual of it, it might actually serve to slow you down whilst working, as if you're relying on being able to accurately pick "invisible" controls, then I can guarantee there's gonna be plenty of times where you "miss" and get the wrong one (click slightly the wrong point in space, wrong face, whatever), causing you to have to hunt around more for the right one.

hrgiger
12-28-2014, 01:27 PM
I really wouldn't want to see much of an attempt to somehow automate in some way, most of the principles of animation into a rigging/animation system as it just ends up taking control away from the animator. Rigs should be fast, allow for my specific posing needs, and get out of the way for the rest of it.

jeric_synergy
12-28-2014, 02:06 PM
Thats it?? Well then, here you go... Done!
Ha, I do so enjoy a Rebel Hill tirade, even when it's aimed at me. I can just HEAR his voice! :D

And, I totally agree: with the possible exception of Secondary Animation (via dynamics) the Principles are not IMO automatable. For instance, "Timing"? Humans struggle with that. This will all be automated when we have Turing-compliant AI.


RE: controls: you guys would know more than I, but is some sort of "animation LEVEL" approach practical? Like you have the gross controls, and you drill down to very finicky stuff? IOW, you'd never see the full set at any one time, or rarely.

ernpchan
12-28-2014, 02:10 PM
I really wouldn't want to see much of an attempt to somehow automate in some way, most of the principles of animation into a rigging/animation system as it just ends up taking control away from the animator. Rigs should be fast, allow for my specific posing needs, and get out of the way for the rest of it.

You also run the risk of a bunch of animation looking the same because that's how the software wants you to animate versus the animator doing what they want.

cagey5
12-28-2014, 02:42 PM
:agree:

stiff paper
12-28-2014, 04:58 PM
Hah! Thanks for bothering RH.

I wrote a reply earlier today, read it back and then deleted it. I still feel as though I mustn't quite be understanding what the original post is going on about, because if it's saying what it seems to be saying then it manages to be both completely meaningless and staggeringly ... umm... silly at the same time. Maybe somebody will explain it to me.

Animating is animating. Also, arrange the controls how you like. Who cares? They're just controls. Having different controls won't make anybody a better character animator. Nice? Yeah, sure. But if you want to do animated-movie quality animation you're going to have a quintillion controls. You just are.

On a more useful note, LW could use a rationalized and better set of deformers to help with CA. That's something that would, you know, really help. (And not just in an imaginary way in somebody's head.) Although I'm sure that the LW3DG is already aware of that.

Actually, now I've said that, I can think of a list of useful things. Quite a long list, really. Maybe somebody should start a thread, it must be five years since the last "LW CA needs this" thread.

robertoortiz
12-28-2014, 08:04 PM
It seems he paradim for high end production tools in 3d is shifting more toward simplicity.

Here is a pair of videos from both the top end 3d animation houses in the world Dreamworks and Pixar.

Both are developed new in house advanced tools that simplify the animation process (hell Dreamworks uses a tablet!)


Dreamworks : https://www.youtube.com/watch?v=iepGUcvsWB0#t=46 (Apollo)
Pixar: http://vimeo.com/90687696 (Presto)

Source:
http://www.cartoonbrew.com/tech/watch-a-rare-demo-of-pixars-animation-system-presto-98099.html
http://www.toonzone.net/2014/07/dreamworks-animation-debuts-new-animation-software-dragon-sequel/

Ernest
12-28-2014, 09:44 PM
A lot of the principles are physically based. They are based on things we do (or that happen to objects) because of momentum and friction. The inertia on one part of the body reacting to muscle movement in another part, our skeletal structure trying to maintain a subconscious upright position when we move, the force we need to overcome in order to start or stop moving, etc.

As we become able to add more physics to the rig and the ground, some of those things will start happening "automatically" and the animator will mostly have to define the movements that the character actually intends to make or that defy physics, and set envelopes for the physical constants.

ernpchan
12-28-2014, 09:50 PM
It seems he paradim for high end production tools in 3d is shifting more toward simplicity.



Those are tools to simplify the animation process. Results of years of animator feedback and evaluating production pipelines. I think it's more an evolution of technology and learning from what works and what does not work.

They have nothing to do with the principles of what makes good animation.

robertoortiz
12-28-2014, 10:00 PM
A lot of the principles are physically based. They are based on things we do (or that happen to objects) because of momentum and friction. The inertia on one part of the body reacting to muscle movement in another part, our skeletal structure trying to maintain a subconscious upright position when we move, the force we need to overcome in order to start or stop moving, etc.

As we become able to add more physics to the rig and the ground, some of those things will start happening "automatically" and the animator will mostly have to define the movements that the character actually intends to make or that defy physics, and set envelopes for the physical constants.
One of my dreams with Genoma future is to have physically based rigs that are influenced by bullet dynamics.

robertoortiz
12-28-2014, 10:05 PM
Those are tools to simplify the animation process. Results of years of animator feedback and evaluating production pipelines. I think it's more an evolution of technology and learning from what works and what does not work.
O
They have nothing to do with the principles of what makes good animation.good points, but the reason I feel that we should look again into these prices Copley for optimizing genome is that they predate our current work flow concepts , developed mostly in the 90s by decades. Lacking having a cal arts trained animator giving input for new workflows, the seem like great touchstones for new ideas.

Ryan Roye
12-28-2014, 10:16 PM
One of my dreams with Genoma future is to have physically based rigs that are influenced by bullet dynamics.

This is already possible thanks to the inclusion of python in Genoma2. Someone of course has to make a python script that can read, process, and set up Bullet constraints based on what it sees in the Genoma rig, but know that the infrastructure is already in place to facilitate dynamics-based rigging workflows.

ernpchan
12-28-2014, 10:48 PM
Sure. But are we saying that the way Pixar and DreamWorks animate should be the standard? I'd be more interested in the sdk and technology in LightWave be open and capable enough for the user to create this workflow...if they so desire. RH and chazriker have certainly demonstrated that with scripting knowledge they can create very capable tools. But even then, they're creating tools they personally feel are an appropriate/superior workflow. (that statement is not intended to be a diss of their tools)

A studio's proprietary pipeline is unique and always evolving. And they vary widely from place to place.

If we were to always emulate with what's "standard" we'd all be using Maya.

chikega
12-28-2014, 10:57 PM
And then there's Tom Roth's Course on Character Animation (in LW):

https://www.youtube.com/watch?v=Ed1i1UjPKqw

jeric_synergy
12-28-2014, 11:12 PM
And then there's Tom Roth's Course on Character Animation (in LW):

https://www.youtube.com/watch?v=Ed1i1UjPKqw
Dang! :) Thanks for that link-- somehow I didn't see that one.

And I remind everybody that Liberty3d is having a sale until, IIRC, 1/1/2015, although for safety order earlier (that bit me once, and Kat was kind enough to extend the sale one day).

Just go look! :thumbsup:

Julez4001
12-29-2014, 12:24 AM
Wouldn't Rebel Hill's Animation Toolbox be more "12 principles" in terms of helping the animator animate.

I can't remember what video is was but he had a clip that had a "mini animate dashboard" for jumping to certain rig features (squatch and stretch and the like).
Course it didn't do it for you but t gave you flexibility right there in the rig.
https://www.youtube.com/watch?v=o0OBLtD6LjM

I would like a way to way to quickly turn off rig visually and
a pose sheet for the character.





---- stuff like this ------

TA Poseworks are a collection of LightWave LScripts that let you Cut/Copy/Paste keyframe data between Item Selections, with the option of INVERTING that Data over user-specified axes, or MIRRORING that data on a Left/Right symmetrical Rigging over the X=0 center-line.

Pose Copy -
Copies the selected Item(s') current Pose (all channels for all selected Items to preserve Pose integrity) data into a copy-buffer in your current LW Config Directory.

Pose Cut -
Stores the selected Item(s') current Pose (all channels for all selected Items to preserve Pose integrity) data into a copy-buffer in your current LW Config Directory, then deletes the current Key for the selected Item(s).

Pose Re-Select -
Attempts to re-select the Items in the copy-buffer, in order, based on the names of the Item(s) stored in the copy-buffer WRT the Item names of those currently in the Scene.

Pose Paste -
Pastes the data for all channels, for (all) selected Item(s), from the copy-buffer onto the current frame. Different Items may be selected for Pasting than those in the copy-buffer, but the same number of Items need to be selected, however you may paste a single Key to multiple Items. (If there is a mis-match between the number of Items selected and the number of Items stored in the copy-buffer, "Pose Paste" can attempt to re-select the Items in the copy-buffer, in order, based on Item name.)

Mirror-Paste -
Pastes the data for all channels, for (all) selected Item(s), from the copy-buffer onto the current frame, with the option of either INVERTING the data over user-specified axes, or MIRRORING the data over X=0 based on parsing user-definable PREFIXes for Left and Right. Mirror-Pasting is dependent upon name-parsing, and so will attempt to auto-select and act upon the mirror of Items referenced by Left/Right prefixes stored in the copy-buffer. Thus, having the same Item names (and their Left/Right mirrors) in the Scene as in the copy-buffer is vital, however, you may Inverse-paste a single Key to multiple Items. (If there is a mis-match between the number of Items selected and the number of Items stored in the copy-buffer, "Mirror-Paste" will attempt to re-select the Items in the copy-buffer, in order, based on Item Name.)

Undo Pose-Paste -
Will attempt to undo/redo the action of a Pose-/Mirror- Paste based on the names of the Items in the undo-buffer.

New to Version 1.0... Now uses Windows TEMP directory, to make it easier to copy/paste animations between Separately running Instances of LightWave using different Content Directories!

Julez4001
12-29-2014, 12:37 AM
https://www.youtube.com/watch?v=_na2oKmdoYQ

RebelHill
12-29-2014, 03:00 AM
It seems he paradim for high end production tools in 3d is shifting more toward simplicity.

Here is a pair of videos from both the top end 3d animation houses in the world Dreamworks and Pixar.

Both are developed new in house advanced tools that simplify the animation process (hell Dreamworks uses a tablet!)

If that's what you think then I fear you've completely misunderstood what's being presented in those videos. They don't demonstrate the slightest amount of shift towards simplicity... they demonstrate a shift towards IMMEDIACY, which is a very different thing.

In both, the big ticket item is the amount of geo/textures/deformers/etc that can be present in a scene at one time whilst retaining realtime feedback for the animators to use... that's basically just raw number crunching. No bad thing sure... but it only provides for a more immediate workflow rather than a simpler one.

Did you fail to notice all the talk in the dreamworks one about their push to get subtlety out of character performance... did you not spot that bit where the animator is doing a detailed pose on a characters eyelid and has like 20 odd controls scattered over that area? The push isn't to somehow make animation simpler... its to provide more control for the animator without the expense of having to wait for all that stuff to compute. As for the tablet thing... well wow... You do realise that it's not some mystical interface where you can just grab and push a characters parts or face around right?? You do realise that what in fact is happening is that there is a face rig present with a large (probably very large) number of controls available which are merely "hidden" and his pen "clicks" simply "auto pick" the closest controller to the click point?? (not to mention that said face rig will be a very complex best setup in great detail by some TD).

Go back and look at the vids I posted before where the controls for the character are hidden... then take that setup, get out your tablet, and boom... there it is, LW working exactly the same way... Just an awful lot slower.

What none of this however demonstrates is some tool "built around the principles of animation"... which is, quite frankly, as ridiculous a notion as a cement mixer built around the principles of architecture.


I would like a way to way to quickly turn off rig visually

Yep... RHiggit does that too.

tcoursey
12-29-2014, 07:53 AM
All I want is a little of that GPU action in the Pixar video... http://vimeo.com/90687696 sprinkled around in LW. They've proven they can do it with Chronosculpt.

chikega
12-29-2014, 09:29 AM
... there it is, LW working exactly the same way... Just an awful lot slower.

If the LWG could make LW as snappy as messiah, then we'd be in very good shape.

robertoortiz
12-29-2014, 09:56 AM
Look I get the difference between simplicity and immediacy.

But why can we have both?

You can frontload the complexity of the character in the rigging stage, BUT..and here is the kicker,
you should have the ability to reuse these controls in other rigs.
Just like in Genoma. This is why I like it a lot.

And yes the DreamWork animators have access to a ton of controls, but they are organized In a tiered fashion
(screenshot of Apollo).
It looks ALOT like After Effects.
http://www.itsartmag.com/features/itsart/wp-content/uploads/2014/07/making-of-How-to-train-your-dragon-2-b.jpg


You can see the animators controlling the characters and deforming them using a tablet.
http://www.bloomberg.com/video/how-dreamworks-draws-dragon-2-faster-better-07fyOznmRG~UpXVQ68bsjg.html

And Iget them impression that you might feel that I am talking about doing automatic animations, far from it.
I am talking about developing better tools (inspired by the principles I have some ideas) and organizing them to improve LW for animation.

For more on Apollo

http://www.itsartmag.com/features/how-to-train-your-dragon-2-discovering-apollo/

ernpchan
12-29-2014, 10:05 AM
I am talking about developing better tools (inspired by the principles I have some ideas) and organizing them to improve LW for animation.



What are your ideas?

It still feels like you're saying better animation will be the result of better technology...and that's true to a point. Garbage animation can still be made with awesome technology. Speed and ease of use certainly help. But if someone doesn't have a good sense of awesome timing and acting, all the tech in the world won't help them.

Maybe some ideas by you will help steer thus thread into the right brain storming direction.

jeric_synergy
12-29-2014, 10:18 AM
Analogously, editing: our tools are orders of magnitude faster, and there's STILL crap editing.

RebelHill
12-29-2014, 10:27 AM
Look I get the difference between simplicity and immediacy. But why cant we have both?

You can... BUT simplicity REMOVES control away from the animator. What you CANNOT have is maximum control with minimum number of controls. Flat impossible.


You can frontload the complexity of the character in the rigging stage, BUT..and here is the kicker,you should have the ability to reuse these controls in other rigs.

Thats?!?! Your kicker?!?! Are you kidding me?

Well holy hell... why didnt i think of that... Oh wait... I did...

http://rebelhill.net/html/rhiggit.html

Reuse poses, motions, all sorts between any character and any other... irrespective of differences between the rigs. Why don't you try WATCHING some of this stuff... there're your inspired next gen rigging and animation tools for LW.


I am talking about developing better tools (inspired by the principles I have some ideas) and organizing them to improve LW for animation.

Then how about you actually try stating those ideas for once and describe how such things could actually work (in actual practical detail I mean... not some pie in the sky vagary)

Though still... I think perhaps you dont really understand WHAT the principles of animation really are... They're merely a guide book for how one performs animation itself... much like you might lay down principles for how to approach dance choreography, or acting method... You cant just make some tool that "inserts" them somehow as if by magic, and as for saying a tool is INSPIRED by them... well... tbh, you're just talking guff, mate unless that tool contains some kind of specific, practical implementation... which again, means giving control to the tool, and not the practitioner.

robertoortiz
12-29-2014, 10:31 AM
Then how about you actually try stating those ideas for once and describe how such things could actually work (in actual practical detail I mean... not some pie in the sky vagary)

.
I guess it is my style.
When we organize projects here at work we go from BROAD IDEAS and GOALS to specifics.
That allows for us to think outside the box instead of regurgitating the same old over and over.
And hell my ideas might be silly, I will be the first one to admit it, but at least I am trying to start a conversation.
And this is not a fight my friend.

We are just talking ideas to improve a tool. It is just ideas nothing wrong with them.




It still feels like you're saying better animation will be the result of better technology...and that's true to a point. Garbage animation can still be made with awesome technology. .

but the thing is that at least content is being made.
Great content.
Softimage had great animation tools and voula
you got a ton of great content.
The same with Maya.


Anyway before Genoma Lw was seen as stuck with is calcified animation workflows.


Ok getting back on track. Some concrete ideas.

Give me a few guys.

ernpchan
12-29-2014, 10:39 AM
I guess it is my style.
When we organize projects here at work we go from BROAD IDEAS and GOALS to specifics.
That allows for us to think outside the box instead of regurgitating the same old over and over.


From a tool design or programming point of view, this is too nebulous. When I talk with our programmers they want an actual layout of expected process and results. I can't just say "make my animation process easier". They'll look at me weird and then say "how?".

RebelHill
12-29-2014, 10:40 AM
at least I am trying to start a conversation.

You start a conversation by having a point... not by throwing out some nebulous nonsense as if you were teacher and tell everyone else to "discuss".

Do that, and you're basically that guy at the office who gos round trying to rev everyone up for a banging night out on a friday... but doesnt offer any suggestions of where we should all go, never volunteers to be designated driver and is mysteriously absent when it comes to the question of who's gonna get the first round in... Just throws it out there and expects everyone else to make it happen.

We have a word for that guy where I come from.


before Genoma Lw was seen as stuck with is calcified animation workflows.

The animation workflows are exactly the same both before genoma and after it... it hasnt changed a thing in that respect.


From a tool design or programming point of view, this is too nebulous. When I talk with our programmers they want an actual layout of expected process and results. I can't just say "make my animation process easier". They'll look at me weird and the say "how?".

Q for total f'ing A

robertoortiz
12-29-2014, 10:44 AM
The animation workflows are exactly the same both before genoma and after it... it hasnt changed a thing in that respect.



Q for total f'ing A

Yes they are, and I have stated that they need to be addressed. But the rigging has been improved.

Again this is not a contest or a debate. I just want to talk ideas.
And here is where I do something you hate, dropping my credentials.

I am programmer with a bachelors in Computer Science, with 15 years of gathering requirements and coding.
So I GET IT.
But I also have a art education so appreciate the idea of seeing things from another angle.
There is nothing wrong with dissention, jebus.
But the idea for this thread, by it SUBJECT is to discuss IDEAS.
Jezz

ernpchan
12-29-2014, 10:49 AM
Yes, rigging process certainly got improved. How one animates is pretty much the same. I don't think it's that different in other packages too. Pose, keyframe, pose, keyframe, repeat.

It's probably easier to see huge improvements in the rigging process since it's largely a repetitive technical process. A lot harder to automate creative skill. Now one can help that process by eliminating a lot of thinking and making things fast so one can move at the speed of creativity.

vncnt
12-29-2014, 10:50 AM
From a tool design or programming point of view, this is too nebulous. When I talk with our programmers they want an actual layout of expected process and results. I can't just say "make my animation process easier". They'll look at me weird and then say "how?".

Programmers are not exactly software designers. You are talking to the wrong person.

RebelHill
12-29-2014, 10:53 AM
And here is where I do something you hate, dropping my credentials.

I dont hate u mentioning your qualification at all... but here's an idea... why not try USING that knowledge and experience to actually explain a method by which something can be done...

I'll post this AGAIN...

http://rebelhill.net/html/rhiggit.html

Cos out of the 2 of us, mate... I seem to be the only one who's DOING SOMETHING to actually address these shortcomings and MAKE these improved tools a reality rather than spewing out wishy washy nonsense.

robertoortiz
12-29-2014, 10:53 AM
Yes, rigging process certainly got improved. How one animates is pretty much the same. I don't think it's that different in other packages too. Pose, keyframe, pose, keyframe, repeat.

It's probably easier to see huge improvements in the rigging process since it's largely a repetitive technical process. A lot harder to automate creative skill. Now one can help that process by eliminating a lot of thinking and making things fast so one can move at the speed of creativity.
Yes agreed. Pose, keyframe, pose, keyframe, repeat.

The KEY question is how to optimize that.
This means honestly a reorganization of the interface around these concepts.
for example lets look at Dreamworks Apollo.
http://www.itsartmag.com/features/itsart/wp-content/uploads/2014/07/making-of-How-to-train-your-dragon-2-b.jpg
How clean the interface looks.

ernpchan
12-29-2014, 10:53 AM
No keep throwing stuff out there Roberto. It may seem like arrows are being thrown at you and for me that's certainly not my intent. Text doesn't always communicate proper tone. Some of the threads on here have certainly helped me develop scripts as I see a need or workflow that I don't see or use.

vncnt
12-29-2014, 10:54 AM
Yes they are, and I have stated that they need to be addressed. But the rigging has been improved.

Again this is not a contest or a debate. I just want to talk ideas.
And here is where I do something you hate, dropping my credentials.

I am programmer with a bachelors in Computer Science, with 15 years of gathering requirements and coding.
So I GET IT.
But I also have a art education so appreciate the idea of seeing things from another angle.
There is nothing wrong with dissention, jebus.
But the idea for this thread, by it SUBJECT is to discuss IDEAS.
Jezz
You mean things like HoldPose, TimeStretching certain bodyparts, recycling poses and parts of animations?

ernpchan
12-29-2014, 10:55 AM
Yes agreed. Pose, keyframe, pose, keyframe, repeat.

The KEY question is how to optimize that.
This means honestly a reorganization of the interface around these concepts.
for example lets look at Dreamworks Apollo.
http://www.itsartmag.com/features/itsart/wp-content/uploads/2014/07/making-of-How-to-train-your-dragon-2-b.jpg
How clean the interface looks.

Well that's a UI issue. LightWave's non-dockable panels is a whole new thread itself, lol. Papou has something that helps (haven't used it myself) but it gets close to replicating workspace presets.

RebelHill
12-29-2014, 11:01 AM
Yes agreed. Pose, keyframe, pose, keyframe, repeat.

The KEY question is how to optimize that.
This means honestly a reorganization of the interface around these concepts.
for example lets look at Dreamworks Apollo.
http://www.itsartmag.com/features/itsart/wp-content/uploads/2014/07/making-of-How-to-train-your-dragon-2-b.jpg
How clean the interface looks.

Well... How DO you optimise that??

Lets see... autokey on... I drag something, wow... a key is created for me, I didnt even have to push a button. Can THAT be optimised??

As for posing... AGAIN... the animator NEEDS as many controls as they need to achieve a specific pose... so unless you're automating stuff HOW can you possibly hope to optimise that?? Its like the old "one null to drive finger rotation" thing... sure, its great, and fast... but guess what, it leaves you with a control system where certain poses CANNOT be attained any more. By optimising an action, you've lost control... you've automated away from the animator.

This clean interface you keep whining on about is just that... CLEAN... its no more "optimal" from a viewpoint of how long it takes to actually do the work, or how many controls, graphs,keys,etc an animator has to work and pick through. And as Ive already pointed out... more than once... if you want to declutter the LW interface in the same way, it is TOTALLY doable, QED (and has been since like LWV6).

Dexter2999
12-29-2014, 11:02 AM
Analogously, editing: our tools are orders of magnitude faster, and there's STILL crap editing.

My response to this has been, "Having editing software doesn't make you an editor any more than owning MS Word makes you a writer."

On a different note, simplicity vs immediacy, I think it is actually a bit of both I see in the Dreamworks video. The process was slow because the user is bogged down by the steps required to move through the process. By making a much more complex program they appear to be able to make it less labor intensive on the user side by turning "click, type, enter, click, move" into "click, move". So, that to me seems both simpler and more immediate. I couldn't tell from the video because there is an edit there, those 20 extra markers Rebel points out, if those were hidden/revealed through a keyboard shortcut or a pen click or if the user had to go to a menu. But the idea of having that ability with a pen click is exciting.

It is a glaring contrast to issues I have had in the past with LW such as camera and render resolutions set in entirely different windows, hypervoxels set up in one window but ACTIVATED in an entirely different window. Completely counter intuitive, although I'm not going to argue that the end user ALWAYS wants one setting to be the other, I will argue that these kinds of things should be tabs in the same window.

I'll also throw out there that I personally have never found it useful to keep the scene editor and the graph editor open at the same time as they take up so much screen real estate and you don't really use them at the same time. You use one or the other in combination but not simultaneously, so why shouldn't they be tabbed as well? Instead of stacked windows? Or made like browser windows where you can tear off a tab and make it another window? For those who feel they do need both open.

I do have a question for RebelHill, I have seen both lighting and rigging tutorials where there are double set ups. Set up A controls one aspect while Set up B controls another. Is this best practice? Or is this the developed practice from working around limitations in software?

RebelHill
12-29-2014, 11:05 AM
I do have a question for RebelHill, I have seen both lighting and rigging tutorials where there are double set ups. Set up A controls one aspect while Set up B controls another. Is this best practice? Or is this the developed practice from working around limitations in software?

Err... not sure exactly what you mean by double setups?? Usually things are "separated" to either give specific control over different aspects or modes of something without interference, or to create an overall workflow efficiency. If you've a specific example, I could likely explain better.

Dexter2999
12-29-2014, 11:26 AM
Okay as an example.

Lighting double set up. One for general settings and one specifically for specularity. Because specularity is set in surface properties not light. But in practice if you have a model with multiple surfaces going through each one to adjust for spec. is tedious and a time suck. So separate light for spec. is put in just for this purpose.

But if the software allowed the spec. to be adjusted inside the light control wouldn't this negate the need for a second light?

This is what I mean by, is this best practice or is it a practice dictated by software limitations?

No offence to you sir, sincerely, but I find that many people who are extremely good at what they do, take for granted the work arounds and limitations sometimes as just part of the process that has become ingrained upon them. Sometimes it takes a set of fresh eyes of the inexperienced to say "this could be easier". (I will also admit freely that the inexperienced are likely blind to the full scope of a process and WHY things are the way they are is because there is more power and flexibility built in but not always being utilized.)

RebelHill
12-29-2014, 11:32 AM
Well... specularity IS a surface property, its not just "set" there.

As for doing a second spec only light to work around this... imo... kinda stupid. Its basically like saying "I hate having to adjust all these different surface colours, so I'll just use a coloured light"... great... but now everythings the same colour... useless. Ofc... you can use this spec only light to ramp up/down specular for ALL surfaces, effectively keeping their relative differences... but it's better practice not to get yourself into that position in the first place. If someone's suggesting that as some kind of "workflow/workaround"... it's probably because they've not learned properly how to make decent surfaces.

There is one exception, and that's in the case of radiosity/GI... which doesnt return specular hits. Thus, when using GI, you may need to add spec only lights to "fake" the GI specular illumination. This ofc is still a cheat for reflection (which is all spec is anyway) and it's still much better practice (though obv more time consuming render wise) to setup surface reflectivity properly if using GI.

RebelHill
12-29-2014, 11:43 AM
I am trying to start a conversation... I am talking about developing better tools and organizing them to improve LW for animation.

Also... you know...

I tried to begin such a conversation (somewhere else) with the folk at the helm of LW about 3 years ago now... Lots of points about how could things perhaps be laid out, and tools designed to attempt to best optimise an animators workflow, whilst also delivering them maximum control, how should the needs of the practice of animation influence the design, and deliverables of the tools created for that purpose, etc, etc...

Wanna know what the result was??

Nothing... sweet FA. No clue or understanding about what I was trying to point out, or the issues around such.

The simple fact of the matter, mate, is that there ISNT anyone on the inside of LW who REALLY understands the needs of character animation, nor who has the vision of how tools could or should work more holistically together. So you can start as many conversations as you like about it all, but trust me...

You're wasting your time.

Dexter2999
12-29-2014, 11:48 AM
Not sure I can agree with your outlook on the second light for controlling spec. I mean there is a reason it is broken out into its own render pass.
Because aesthetically it can be undesirable or distracting from a composition, accurate or not. Or perhaps you want it intentional? You want something to glimmer or sparkle? Is it really better to need to envelope every texture in a treasure trove rather than to be able to have independent control of a specularity setting?

And this is lighting I'm talking about but rigging has its peculiarities as well. Like when you have to put in an extra bone just so you don't get gimbal lock. To you that is a "no-brainer", to a newbie that is a "WTF?" moment. Your experience tells you how to avoid the pitfall before you even get to it, but a new person says "this needs to be fixed".

RebelHill
12-29-2014, 12:02 PM
In which case I submit that such issues to do with "tweaking" spec in that manner are better done in compositing than in 3D... something which I believe you'll find most compositors would agree with... give them the passes they need for the control they want, leave them with as little baked in as needed.

As for the gimbal lock issue... its not something which CAN be fixed easily.

There are only 3 real methods for representing 3D rotations mathematically/numerically (which obv is how they must be represented in a computer).

1. Euler angles
2. Quaternions
3. Matrices

Euler angles are MUCH more natural for animators to work with, they give cleaner, more easily interpreted graphs, and thus are THE de facto standard in all 3D animation. They do, however, suffer from gimbal lock, and eulerian flipping... this CANNOT just be "fixed"... its an intrinsic consequence of the math used to create the representation. Any fixing must be allowed for by the rigger or technical director.

Quats CAN be animated, but are FAR less natural to work with, and you don't have the same number of options for how you can tween them. No locks or flips, but VERY counter intuitive.

Matrices.... ahahahahaha!!! Good luck basing an animation system on those. lololol!!

If you can come up with some other method of representing 3D rotations, which solve all these issues... you can collect your fields medal.

jeric_synergy
12-29-2014, 12:50 PM
I am a well-known math ignoramus, but shouldn't UI representation of rotation insulate the user from the numerical complexities of how a rotation is numerically treated? Or do you mean specifically the interpolation between any two rotations will be non-intuitive (aka ugly) in many situations no matter which numeric representation is used?

RebelHill
12-29-2014, 01:08 PM
A UI can only do that up to a point... unless you're going to have a system where rotations are presented to the user as "left a bit... right a bit, right a bit more"... Then you HAVE to give the users some kind of "marker/measure" which represents the actual rotation in play (and if you can think up some measure that's represented without the use of numbers, well done you).

As such, if you wish to present the user with a clear definition of the exact rotation... then that definition must be computed... which brings us back to the problem.

At the end of the day... it makes utterly no difference how good or bad you are at math... can you even begin to imagine a system that represents rotations accurately and measurably that doesnt require numbers or values being presented to the user?

robertoortiz
12-29-2014, 03:24 PM
Ok I invite you guys to take out the sharpie...
Getting back to the concept of the thread I have created some templates in Illustrator based on CORE animation exercises (that do come fomr the old school of 2d animation).

The idea that I have is to imagine quick and easy ways to do this exercises in 3d.

Imagine that you had total control of what you have on screen, what would it be?
What would you need to make this as easy as possible?
I just printed a whole bunch of these to make with me home and brainstorm a bit.
I invite you guys to do the same.

jeric_synergy
12-29-2014, 03:30 PM
A UI can only do that up to a point... unless you're going to have a system where rotations are presented to the user as "left a bit... right a bit, right a bit more"... Then you HAVE to give the users some kind of "marker/measure" which represents the actual rotation in play (and if you can think up some measure that's represented without the use of numbers, well done you).

As such, if you wish to present the user with a clear definition of the exact rotation... then that definition must be computed... which brings us back to the problem.
At a guess, I'd probably go with some sort of gizmo* that would represent the rotations graphically for the item in question (rings and wedges), but since I don't have a dog in this fight, I'll just bow out. But it's a very interesting discussion.


*it'd probably get very busy, visually.

ernpchan
12-29-2014, 03:33 PM
How is what you're posting not just a series of keyframes?

I think one thing that gets taken for granted is the experience that a really talented animator has from hours and hours of animating (well). A really skilled char animator can blitz through these exercises. For example, one thing that 3d animators struggle with is timing. But if you reference how timing is handled in 2d, a person with a stop watch sits over x-sheets and times it all out. Over time they can blitz through pages of x-sheets noting how fast something moves, how many frames of anticipation, etc. After awhile they don't even need a stopwatch. There's no sliding of keyframes in a graph editor, experimenting and playblasts. It's all internal from years of experience and knowing how to do what they do well.

I still feel like I'm missing what needs to be done to make your idea of animation efficiency as fast as possible.

RebelHill
12-29-2014, 03:35 PM
graphically for the item in question (rings and wedges)

Thats only a visual representation... and if you're gonna have wedges (as in slices of a pie)... then those ARE numbers, even if they're not "displayed" as such... thus... the problem remains.

RebelHill
12-29-2014, 03:38 PM
How is what you're posting not just a series of keyframes?

How is any animation not?? Right??

I mean bouncing ball... seriously.

Lets see... you need to place the ball at different positions in space... thats a move tool... you need to do it at different points in time... thats a timeline/timeslider... And you need a way to tween the arcs with ease in/out thats quicker than having to manually place every single inbetween keyframe using aforementioned move tool... that's a bezier spline.

Guess what... short of having a "bouncing ball button"... there is no simpler way than what exists.

ernpchan
12-29-2014, 03:44 PM
Guess what... short of having a "bouncing ball button"... there is no simpler way than what exists.

Well we did have "eagle in a barrel" at one point. So I guess this button is possible...:D

There is dynamics...but I find dynamics animation very very different from character animation.

robertoortiz
12-29-2014, 03:58 PM
How is any animation not?? Right??

I mean bouncing ball... seriously.

Lets see... you need to place the ball at different positions in space... thats a move tool... you need to do it at different points in time... thats a timeline/timeslider... And you need a way to tween the arcs with ease in/out thats quicker than having to manually place every single inbetween keyframe using aforementioned move tool... that's a bezier spline.

Guess what... short of having a "bouncing ball button"... there is no simpler way than what exists.


Ok how about this.
The ability to draw a spline in 3d space and assign what key frames the points belong to. The kind of vector line drawing that can be done with easy with a 2d vector app like Illustrator and it IMPOSSIBLE to do in LW layout right now.
Imagine beign able to control using that spline the smoothness of the motion.

Hell I can top that.
Imagine beign able to select the motion line n beign able to select the number of frames its has. Or being able to assign its motion attributes to any of the parameters of the ball (x,y,z ,r,p,h).
HELL
You could use that type of control spline to do LATTICE like deformations with multiple splines.





It is the simple things that should be done first WELL to build a strong foundation

ernpchan
12-29-2014, 04:12 PM
Ok how about this.
The ability to draw a spline in 3d space and assign what key frames the points belong to. The kind of vector line drawing that can be done with easy with a 2d vector app like Illustrator and it IMPOSSIBLE to do in LW layout right now.
Imagine beign able to control using that spline the smoothness of the motion.

Hell I can top that.
Imagine beign able to select the motion line n beign able to select the number of frames its has. Or being able to assign its motion attributes to any of the parameters of the ball (x,y,z ,r,p,h).
HELL
You could use that type of control spline to do LATTICE like deformations with multiple splines.


Well in animation the points represent space and time...not just space. I do kinda get what you're talking about...but I still see that as keyframe manipulation, both in time and space.r To change the number of frames of a motion line, just change the start and end frames in the GE, not sure what time you'd be saving that's so radically different from what exists.

ToonBoom Harmony has something similar to what you're describing where a keyframe has a value in space and time and then you can add additional points to the spline to change the shape that don't actually exist as points in time. So you can't adjust the timing there...just the shape of the curve.

Now being able to adjust the shape of the spline with handles would certainly be nice. You can'd to that in Layout so you're stuck futzing with TCB, bezier handles, etc in the GE.

RebelHill
12-29-2014, 04:23 PM
The ability to draw a spline in 3d space and assign what key frames the points belong to.

Yeah, ok... pen tool... draw a path, boom... its an animation path. nice idea... but... what happens when my cack handed attempt to draw a path has been imperfect the first time? Do I have to start again, and again until I eventually pull a homer and knock out a perfect path drawing, at each stage having to reassign what keys belong elsewhere on the path... Or will there be some method of "refining" this path after Ive sketched it in??


The kind of vector line drawing that can be done with easy with a 2d vector app like Illustrator and it IMPOSSIBLE to do in LW layout right now.

Oh... so you mean like... a bezier??

Nope, never seen them in layout.


Imagine beign able to control using that spline the smoothness of the motion... Imagine beign able to select the motion line n beign able to select the number of frames its has. Or being able to assign its motion attributes to any of the parameters of the ball (x,y,z ,r,p,h)

So... like you do in the graph editor then?

jeric_synergy
12-29-2014, 05:47 PM
At least we're getting to concrete suggestions.

Re roberto's example and RH's reposte: I'm not so sure about using drawing per se as a UI method, but for altering keyframes, how about if we could [alt/ctrl/shft/whatever]+drag on a motion path keyframe, and it would stay where it is in space, but slide up and down the timeline? (I expect users would want the main time cursor to follow with, although that could be optional.)

That is: currently we can (optionally) see motion paths and IIRC keyframes are indicated. Would the ability to slide a keyframe temporally, not spatially (which of course we already have) within the main window UI be useful to animators?

One immediate drawback is that purely rotational motion paths wouldn't be accessible, you'd have to use the timeline just like now.

An advantage would be you could see the inbetween tick marks/frame markers move in response to adjusting the keyframe time.

To address purely rotational or scale-centric motion: As I'm thinking about this I'm wondering if perhaps a sort of HUD mini-graph editor appearing around the keyframe would be helpful-- for me, a lot of UI suckiness involves searching around the screen, and often between the GE and the main window, pulling my attention around instead of letting me focus in one spot, added to mouse targeting in the same action. Imagine a 3-sided HUD reticule that appears around the motion path keyframe with a pared down representation of the GE curves-- three sided to show Pos/Rot/Scl.....

Meh, all my ideas wind up making the UI really cluttered. :bangwall: --But I do like the idea of not having to look over at the GE constantly.

ernpchan
12-29-2014, 05:49 PM
Would the ability to slide a keyframe temporally, not spatially (which of course we already have) within the main window UI be useful to animators?



Yes, that's what you can do in ToonBoom. I'd definitely find that useful. Changing the timing of a keyframe yet maintaining the shape of the path that is.

RebelHill
12-29-2014, 06:10 PM
I say no (or rather... whats the point?)... because if Im sliding the timing of a keyframe in the viewport... where's the "scale" against which I can see exactly where in time Im moving it to?? To judge that, I have to be looking at the timeline anyway... so may as well just slide it on the time line. It is after all, something that's always there, just like the viewport, so it's not like Im saved from having to go into some panel or other.

Plus... what if Im trying to shift time for several keys (at different points on the motion path) or for multiple items (with different paths scattered at different positions around the viewport)... is that somehow easier to select than having them all grouped nicely together on a single timeline/dopesheet?? I really don't think so.

To me it makes about as much sense as saying "how about changing the positions of items via the timeline rather than the viewport".

Julez4001
12-29-2014, 07:07 PM
How about a more re-fine dope sheet editor?
I find Lightwave version not as easily intuitive, I constantly have to watch William Vaugh youtube video.
(There was a third party version a while back alled KeyTrak that was nice.)
I did enjoy Messiah Studio's dopesheet editor as it easily swapped with Graph Editor.

I had a technique where I would run two Lightwave, or two Messiah or one of each, where one had the character rig and the other was the head of the character fully face rig/morph with a camera.
I would save the mdd and it would update the character rig that had the face added as a separate object.
I could have a high-res meshand multiple rigs to animate sections of the face in one scene that would not interfere or slow down the main character rig.
I was taking the old morph gizmo workflow up another level.

It was all about making the scene lite as a feather..

jeric_synergy
12-29-2014, 07:24 PM
... where's the "scale" against which I can see exactly where in time Im moving it to??
The scale would be the other tick-marks in the visible motion path.

I'm assuming keyframes would not move 'thru' each other.

RebelHill
12-29-2014, 07:28 PM
So then... If I want it on frame 65... Id have to already know what frame number the previous keyframe in the path was at, and then count all the little spacer ticks to find my desired spot... yep?

Im not sure how thats an improvement over what exists now.


How about a more re-fine dope sheet editor?

Yeah... I tend to think that this line of approach is far more productive and capable of leading somewhere than these quite futile attempts to reinvent the wheel... Take those tools and workflows which are already present and simply try to find ways to improve them, make them more interactive and make it faster to get things done.

jeric_synergy
12-29-2014, 08:46 PM
So then... If I want it on frame 65... Id have to already know what frame number the previous keyframe in the path was at, and then count all the little spacer ticks to find my desired spot... yep?
Not really, the idea is more to getting the process to be more, gahd help me, "intuitive", or at least more visible. I'm pretty sure that one would want the timeline cursor and the keyframe to become sync'd, so that as you scrubbed the keyframe along possible places the user would see the other bits of the animation s/he was trying to relate to.

So, in my little fantasy, the user sees that a keyframe is a little early.

Alt+drags (we'll go with alt here) the keyframe tick mark in the main UI to the (screen) right,
(the timeline cursor jumps to the keyframe frame),
and the keyframe is advanced along the timeline (i.e. +time),
with the timeline cursor syncing to the keyframe.
The motion path display is constantly updated to reflect the change, so users can see the change in the curve.

The main advantage, and maybe it's not much, but... is that the user's focus stays in the main window, it isn't dragged away to the GE or even the dopetrack.

The British cycling team won some huge awards by concentrating on very small things, but very very many small things. Maybe allowing the users to keep their attention in one place in the UI would be similar: another small 1% gain, added to many others.

ernpchan
12-29-2014, 10:41 PM
This is ToonBoom's motion points tool that let you change the shape of the motion path without adding a keyframe.

https://www.toonboom.com/resources/tips-and-tricks/use-motion-points-modify-your-animation-path

lightscape
12-29-2014, 10:52 PM
The rigging tools in lightwave are ok. I think its the way we interact with the clunky lightwave gui that can be improved. The scene editor, dopesheeet, grapheditor, on screen controls, etc. It is more clunky in lightwave than other appz because the gui is a bit dated.

You can find only a handful of animators using lightwave that can be compared to Pixar and Disney animators.
I count only Tim Albee and Tom Roth. Even 11second club animators kick lw animators butt and some of them are students.
We should ask Tim and Tom how to make animation in lightwave better since they have the experience and skills.
Tafa is a product of trying to improve character lipsync workflow and it shows how fun it can be done and how fluid it is compared to morphmixer sliders.

Ernest
12-30-2014, 01:09 AM
This is already possible thanks to the inclusion of python in Genoma2.

Not really. Not with the current evaluation speeds in LW. The whole key to this thread is speed. Rigging has become very competitive in LW, but the rigs have to be kept simple. Otherwise, by the time we have four characters in the scene, it would be too slow to interact with. If the speed increased to Presto-like levels, then all the additional power could be added to the rigs. Especially facial rigs. And then the same system looks suddenly so much more artist friendly.



Ok how about this.
The ability to draw a spline in 3d space and assign what key frames the points belong to.

And then drag and drop poses and motion mixer clips right on the spline and slide and stretch them along the spline?
The ironic thing is that this would go quite opposite from traditional animation. The planning would go "this character would start the jump when he turns this corner where the overturned bench is, and would land right on this puddle", instead of "would start jump at 3:00:00 and land at 3:02:20".



So... like you do in the graph editor then?

I see what you're trying to say there, but the graph editor has the comparative "disadvantage", compared to what he's proposing, that it is completely abstracted from the scene. I think his idea refers to having limited graph editor abilities represented visually within the scene and editable within that visual context. Animating in the scene, just like we pose in the scene, instead of moving out into the timeline, the dope sheet or the graph editor to animate. It is actually interesting if it has good filtering. If the filtering is less than great, the scene would probably look like a timelapse of Mexico City at night.


Which brings me to the point I really wanted to make. Filtering.

The one thing that appealed to me when looking at Apollo (other than the speed, of course) was how easy to find were the controls for the specific character that the animator was working on.

There might have been dozens of characters in that scene, but we only saw the one character they were actively animating (which can easily be done in LW) and the animation controls for that one character.

Imagine if, in Modeler, all the layers of all the objects we have open would get jumbled together alphabetically in the layers panel and the layers widget. It would be a nightmare.

In layout, we should be able to select a character in some way and only the animation controls for that character should remain visible (not only in the the viewports but also in the menus and dropdowns). All the bones, goals, targets, constraints, morphs, sliders, proxies, etc. of all the other characters should completely disappear from every list and dropdown everywhere until we reactivate their respective characters. This would require LW to understand the concept of character so it would understand all the objects and nulls (and lights, sometimes) that make a character.

vncnt
12-30-2014, 03:29 AM
I like the idea of using the layer selector gui element from Modeler inside Layout to control visibility of character hierarchies. I'm also a fan of Markers and Regions. They should cooperate with many tools.

RebelHill
12-30-2014, 03:48 AM
the idea is more to getting the process to be more, gahd help me, "intuitive", or at least more visible.the user's focus stays in the main window, it isn't dragged away to the GE or even the dopetrack.

Ok... so that's fine when you're eyeballing it... but when you want precision (which is a LOT of the time)... no dice, right? You're still also only thinking in terms on one key, for one item, on one path... How are you gonna select multiple keys, or keys for multiple items (which may be at different positions in space) if you're having to click drag on single keys for single items on single paths. Working on one key at a time doesnt sound that great to me.

Also... your attention, at present, ISNT constantly dragged away to some other window to do this... you just grab the key on the timeline, and drag. There's the viewport right in front of you, you can see the changes in realtime, whilst, even, playing back your animation... so u see the change not just to the path... but to the motion, interactively... I do this all the time, and it's fast, easy, intuitive and immediate.


I think his idea refers to having limited graph editor abilities represented visually within the scene and editable within that visual context. Animating in the scene, just like we pose in the scene, instead of moving out into the timeline, the dope sheet or the graph editor to animate. It is actually interesting if it has good filtering. If the filtering is less than great, the scene would probably look like a timelapse of Mexico City at night.

So like maya's viewport path tool then? Sure... handy to have.. I'll take it. But it's hardly some revelation or great leap forward, and as noted... bit of a mess if you're trying to do multiple items paths at once.


Not with the current evaluation speeds in LW. The whole key to this thread is speed. Rigging has become very competitive in LW, but the rigs have to be kept simple. Otherwise, by the time we have four characters in the scene, it would be too slow to interact with.

No complaints from me about improving speeds... I'm all for that... Though tbh, rig evaluation speed just in itself isnt bad, at least if you build efficient rigs. Go see my RHiggit promo again... one rig running at over 200fps... 5 in a scene at over 50fps... Simple characters, no geo/deforms... that's pretty fast (actually about 60% faster than a comparable genoma rig). Now obv... you add geo and deforms on there... yup, big ol slowdown... So there's PLENTY of scope for speed improvements.


Which brings me to the point I really wanted to make. Filtering... All the bones, goals, targets, constraints, morphs, sliders, proxies, etc. of all the other characters should completely disappear from every list and dropdown everywhere until we reactivate their respective characters. This would require LW to understand the concept of character so it would understand all the objects and nulls (and lights, sometimes) that make a character.

Where again, there is a great opportunity for a speed increase... Having a hard time getting LW to crunch the raw numbers to calculate all those rig controls or deforms... Well how about if LW was "character aware"... or even "chain aware"... and only kept active rigs for characters, or even single limbs that were currently being called for evaluation?? There I am posing one guy, and all the others in the scene are transparently deactivated, until I pick a part that belongs to one of them. Ofc... as soon as you hit play... you'd be back to having everything acitve, so that's not the be all and end all of solutions either.

However... as you mention understanding the concept of character, and all the bones, nulls, etc associated with a given character.. that is EXACTLY one of the things RHiggit does... Go see the animation toolbox vids. That toolbox is fully rig aware, thus, you can put it in rig selection mode, and key all controls in a rig at once without having to do any multiple selections... or just a "part" of a rig (arm, leg, spine)... or child parts (select the hand, hit key... you key all fingers, but not the hand itself). And this same awareness of whats to be operated on trickles through to ALL the other tools... pose and motion copy and paste, curve filtration, etc, etc.

Would I like to see this concept extended further in LW itself... sure... but I doubt its gonna happen anytime soon (or even not so soon)... So for the time being, Im trying to bring as much of this sort of thinking forward as possible within the constraints which exist.


Even 11second club animators kick lw animators butt and some of them are students.

That is simply down to the fact that there just arent very many skilled character animators working with LW... but don't worry... Im working on that too ;)

vncnt
12-30-2014, 04:52 AM
Having a hard time getting LW to crunch the raw numbers to calculate all those rig controls or deforms... Well how about if LW was "character aware"... or even "chain aware"... and only kept active rigs for characters, or even single limbs that were currently being called for evaluation?? There I am posing one guy, and all the others in the scene are transparently deactivated, until I pick a part that belongs to one of them. Ofc... as soon as you hit play... you'd be back to having everything acitve, so that's not the be all and end all of solutions either.
When LW is having a hard time, smart solutions should be added to keep things organized and tidy so the CPU will only crunch raw numbers for features that are needed at that moment.

The animator should focus more on the goal: expressing emotions using puppets.
Anything else is noise.

RebelHill
12-30-2014, 05:21 AM
When LW is having a hard time, smart solutions should be added to keep things organized and tidy so the CPU will only crunch raw numbers for features that are needed at that moment.

Sure... thats why I added a rig on/off button.


The animator should focus more on the goal: expressing emotions using puppets.
Anything else is noise.

Quite true... however I tend not to like things that steer towards this idea that there can be some kind of "pure" artistic environment... No matter what kind of task you're undertaking, artistic or otherwise... You always have to deal with and be aware of some kind of "infrastructure" issues. Even someone working on hand drawn animation has to make the effort to keep their pencil sharp and their flip chart in order.

jasonwestmas
12-30-2014, 07:10 AM
A pure artistic environment is the imagination, which ime is very fluid without any containers to maintain the fluid. Nothing stays planted in it for very long making things hard to control. :) Kind of like how pure gold is far too soft to make into jewelry.

robertoortiz
12-30-2014, 09:29 AM
Ok guys I had a nice metro ride back home and it game me some basic ideas.

Some ideas:

INDEPENDENT MOTION PATHS
What if allow for the independent creation of motion paths. Allow them to become independent entities on their own
that can be assigned to either NONE or multiple objects.
When they are assigned they overlay the values for that object on the attribute they have been assigned to.

Also this approach allows for other cool tricks


ILLUSTRATOR LIKE SPLINES on VIEWPORTS
And what if we created those splines on the VIEWPORTS using controls similar to the ones found ADOBE ILLUSTRATOR.
And what is we simplified this using a 2 click analogy to keep it simple.
CLICK
Add a point, ENTER THE KEYFRAME (Default CURRENT FRAME, for additional point no less than PREVIOUS point)
CLICK
Select the TYPE OF POINT:

TCB SPLINE
HERMITE SPLINE
BEZIER SPLINE
LINEAR
STEPPED



HAVE LATTICE LIKE CONTROLS OVER THE SPLINE.
Imagine being able to tweak the spline with handles that surround it.


Here are some diagrams I polished up from my notes.


126349

126348

126347

robertoortiz
12-30-2014, 09:35 AM
In layout, we should be able to select a character in some way and only the animation controls for that character should remain visible (not only in the the viewports but also in the menus and dropdowns). All the bones, goals, targets, constraints, morphs, sliders, proxies, etc. of all the other characters should completely disappear from every list and dropdown everywhere until we reactivate their respective characters. This would require LW to understand the concept of character so it would understand all the objects and nulls (and lights, sometimes) that make a character.

This is a BIG BIG problem in LW. Filtering of information. I agree with your point.

RebelHill
12-30-2014, 09:56 AM
INDEPENDENT MOTION PATHS[/B]
What if allow for the independent creation of motion paths. Allow them to become independent entities on their own
that can be assigned to either NONE or multiple objects.
When they are assigned they overlay the values for that object on the attribute they have been assigned to.

How is this any different in a MEANINGFUL sense to having a single null with its recorded motion path? In both instances its just a scene level item... Also, doing it this way adds nothing in the way of attaching that path to things as you suggest, as copying and pasting motions or key sets between items... which effectively is copy/reassigning a motion path... is already present, easy, and takes only a couple of clicks.

How is this anything but a rearrangement of the deckchairs to that end?

Also... fine for motion paths of things, but an awful lot of animation doesnt require spatial motion paths... it requires rotational animation. On an average character rig I only have 3 items moving through space... the root, and 2 foot IK nulls, everything else just rotates on point in its parent space. Sure, you could go assigning curves youve made to those rotation channels... but that's gonna be MASSIVELY uninuitive... rotation curves aren't something you can easily "draw", in the same way you can "projectile" paths.


ILLUSTRATOR LIKE SPLINES on VIEWPORTS
And what if we created those splines on the VIEWPORTS using controls similar to the ones found ADOBE ILLUSTRATOR.

As previously mentioned... thats basically mayas viewport path tool. Sure, I'll take it, its nice and sure is handy... But it wont remove the need to do channel specific edits on curves for which the graph editor is better suited, and once again... is only of use for things moving around in space... not for rotating items. So its still a limited use tool at best.


And what is we simplified this using a 2 click analogy to keep it simple.
CLICK
Add a point, ENTER THE KEYFRAME (Default CURRENT FRAME, for additional point no less than PREVIOUS point)

So... I'm either having to "guess" where I want those keys to fall in time... or, Im still having to move my time slider along to create key points at the current time... BUT... this is for an "individual" path with no item motion on it initially.

Is it not more productive to simply do this as we currently do, either on a null, or the target item itself where you can immediately see the resultant motion of your new keys, adjusting them as you go if desired, or choosing to leave them before coming back to perform the retiming pass??

As the first point again... How is scrubbing through the timeline to just place key points in the viewport on some arbitrary path any more useful or meaningfully different than simply keyframing an item as we currently do??



HAVE LATTICE LIKE CONTROLS OVER THE SPLINE.
Imagine being able to tweak the spline with handles that surround it.

Meh... OK I guess for quick n dirty adjustments... but again, its only going to come in useful for such broad tweaks like that... it wont remove the need to do detailed precision work for the vast majority of the time. Also... note the CPs on the lattice in your example, which marry up with key positions measured from the lattice local X and Y axis... That means if I want to use that to drag a key position around, I have to use 2 controls to do it, rather than the one we have presently... whats the point??

All in all... hardly some great reimagining of animation workflows.

chikega
12-30-2014, 10:01 AM
The number one issue for newcomers trying to rig in Lightwave is the concept of 'resting' bones. In messiah, for example, the "Setup" tab is where you would layout the bones and do all the rigging, but tab over to the "Animate" tab to animate without the need of remembering to "rest" the bones over and over and over again.

robertoortiz
12-30-2014, 10:08 AM
How is this any different in a MEANINGFUL sense to having a single null with its recorded motion path? In both instances its just a scene level item... Also, doing it this way adds nothing in the way of attaching that path to things as you suggest, as copying and pasting motions or key sets between items... which effectively is copy/reassigning a motion path... is already present, easy, and takes only a couple of clicks. .

The couple of clicks is what is the deal breaker for me.
Also the fact that in order to tweak the motion with precision I have to jump through hoops.

And it feels more intuitive to treat motion (via the path) as a entity.


And entity that can be modified with PRECISION ON SCREEN.

the idea is to decrease
CLICKING and SEEKING.
that is why the DreamWorks Apollo video impressed me.
The idea of the lees clicking is that in theory you can animate with a tablet.


The Spline path is ease to click and maintain.
Click a point on the SPLINE path and you can change
the attributes for the TYPE OF POINT, KEYFRAME, the LENGHT of the Spline, and apply LATTICE deformations to it.
All in the VIEWPORT.

BTW No amount of training is going to change that the way LW animated is CLUTHERED, and frankly counter intuitive.
To tell a NEW user that you have to do a ton of click to Animate a BALL with PRECISION is to kiss that user goodbye.

ernpchan
12-30-2014, 10:35 AM
BTW No amount of training is going to change that the way LW animated is CLUTHERED, and frankly counter intuitive.

LW is very click-centric. It would be nice of LW auto-selected. I can see how that would be nice. Also the fact that LW isn't always good at selecting what you want in the viewport doesn't help. This is more of a UI thing than say an animation system issue.

How is animating in LW counter-intuitive? How is what LW does that drastically different than say Maya or XSI? They have the same workflow do they not? I haven't seen a DW or Pixar animator in action but I imagine it's the same. Pose, scrub, pose, scrub. Their system looks to be about showing everything (hair, cloth, etc) in realtime with a reduced interface. Also bear in mind, that while those systems are nice there's a ton of training involved.

I do like the idea of being able to draw a motion path in the viewport. Not sure what I think of the idea of assigning frames to a control point, to me that's the same as scrubbing and creating a key.

The lattice on a motion curve is a nice idea. I'd prefer to have lattices over the graph curve. XSI has that.



To tell a NEW user that you have to do a ton of click to Animate a BALL with PRECISION is to kiss that user goodbye.

Maybe I'm old school. I don't see how it's that labor intensive to animate a ball with the current key, move, key, move, tweak in graph editor system. It kinda sounds like you're looking for animation presets which is a different topic. Presets are nice, but then everything has the danger of looking the same.

RebelHill
12-30-2014, 10:47 AM
The couple of clicks is what is the deal breaker for me.

Honestly... have you actually thought through your own idea here??

Select the spline... theres a click.
Select the target item... there's a click.
Hit some "assign" button... there's a click.

You're saving NOTHING.

Can you please explain the EXACT set of steps required to perform this action which could possibly require fewer than 2 clicks??


Also the fact that in order to tweak the motion with precision I have to jump through hoops.

No you don't... what hoops?? You're talking nosense.

If you want to tweak timing, you can do it in the track or dopesheet fantastically fast and easily. Wanna change the key positions along the path in viewport... use the path tool that we already have... Grab a key point, drag it around... How can your method POSSIBLY be any easier than that? What's it gonna do... read my mind and move the positions without me having to perform a single click or action??


And it feels more intuitive to treat motion (via the path) as a entity.

And its IDENTICAL to having a null with a path... youve got 1 item, which is basically nothing more than a motion path. You've not erected a new spire, you've just repainted an existing one.


And entity that can be modified with PRECISION ON SCREEN.

Path tool!!!


The idea of the lees clicking is that in theory you can animate with a tablet.

Have you actually paid attention to any of my previous posts?? You can do that in LW TODAY!!


BTW No amount of training is going to change that the way LW animated is CLUTHERED, and frankly counter intuitive.
To tell a NEW user that you have to do a ton of click to Animate a BALL with PRECISION is to kiss that user goodbye.

You don't have to do such a ton of clicking though... or at least... I can see NOTHING in your suggestion that allows for doing any less... the only possible exception being the "lattice path" idea... which is a broad stroke tool, which removes precision control. Can you possibly tell me how you get precision placement of EVERY keyframe without somehow cliking, or selecting that keyframe??


How is animating in LW counter-intuitive? How is what LW does that drastically different than say Maya or XSI? They have the same workflow do they not? I haven't seen a DW or Pixar animator in action but I imagine it's the same. Pose, scrub, pose, scrub. Their system looks to be about showing everything (hair, cloth, etc) in realtime with a reduced interface.

Honestly I dont think Roberto actually understands what he's seeing in those videos... One thing is happening and I think he imagines something else is going on... though he seems unable to articulate exactly what that is other than by saying "simplicity", which doesnt really mean anything here without some explanation behind it.


I don't see how it's that labor intensive to animate a ball with the current key, move, key, move, tweak in graph editor system. It kinda sounds like you're looking for animation presets which is a different topic. Presets are nice, but then everything has the danger of looking the same.

Nor I.

robertoortiz
12-30-2014, 11:03 AM
LW is very click-centric. It would be nice of LW auto-selected. I can see how that would be nice. Also the fact that LW isn't always good at selecting what you want in the viewport doesn't help. This is more of a UI thing than say an animation system issue.

How is animating in LW counter-intuitive? How is what LW does that drastically different than say Maya or XSI? They have the same workflow do they not? I haven't seen a DW or Pixar animator in action but I imagine it's the same. Pose, scrub, pose, scrub. Their system looks to be about showing everything (hair, cloth, etc) in realtime with a reduced interface. Also bear in mind, that while those systems are nice there's a ton of training involved.

Good points.

For me the idea is intuitive is open for interpretation.
for me an intuitive systems allows you to see attributes of something when clicked.
POINT CLICK- give me info about this
RIGHT CLIK-Show me ADVANCE attributes



I do like the idea of being able to draw a motion path in the viewport. Not sure what I think of the idea of assigning frames to a control point, to me that's the same as scrubbing and creating a key.

The lattice on a motion curve is a nice idea. I'd prefer to have lattices over the graph curve. XSI has that.
.
the idea of assigning frames it that it links it to the timeline WITHOUT having a navigate in the timeli



Maybe I'm old school. I don't see how it's that labor intensive to animate a ball with the current key, move, key, move, tweak in graph editor system. It kinda sounds like you're looking for animation presets which is a different topic. Presets are nice, but then everything has the danger of looking the same.
ill give you that one

ernpchan
12-30-2014, 11:12 AM
POINT CLICK- give me info about this
RIGHT CLIK-Show me ADVANCE attributes

What's your definition of "info about this" and "ADVANCED(D) attributes"?

Is this the same as Maya's channel box?

I would like to see more attributes/channels with items in LW, we're limited to Pos/Rot/Sca and even then you can only see one group at a time. 8~



the idea of assigning frames it that it links it to the timeline WITHOUT having a navigate in the timeli


Ok, still seems like a minimal amount of savings though. I'm also not seeing how this would work for more complex animation, like posing a character. This workflow seems very specific to plotting a motion path and therefore limited in it's scope. Therefore may not be worth the resources to implement compared to the practicality.

robertoortiz
12-30-2014, 11:18 AM
Is this the same as Maya's channel box?
.
I guess yes.





Ok, still seems like a minimal amount of savings though. I'm also not seeing how this would work for more complex animation, like posing a character. This workflow seems very specific to plotting a motion path and therefore limited in it's scope. Therefore may not be worth the resources to implement compared to the practicality.
You are right again.
And that is why I am going ONE exercise at a time to refine the ideas.
The original idea of the thread is to try ideas.

ernpchan
12-30-2014, 11:26 AM
I guess yes.


I'd like more info a/la Maya's channel box. But the info in the Channel Box are additional channels tied to that object for animation control purposes. So they're not necessarily channels always associated with an object. I suppose in theory, you could have one object with all the channels of a character tied to that one object and then just animate by sliding the values. Not a way I'd recommend to work though, haha.

But let's say you start putting the keyframe type (linear, tcb) and those associated values (tension 1) in this expanded channel area, you're just pulling the info from the GE up into the UI. So you sorta eliminate the need to go into the GE, but you're cluttering the main UI. I guess another option would be UI hotboxes if you wanted to get access to the data in a keyframe. It sounds cool and fancy, just not sure how practical. I belive IKBooster does this via a right+click mechanism.

Dexter2999
12-30-2014, 11:31 AM
Which brings me to the point I really wanted to make. Filtering.

The one thing that appealed to me when looking at Apollo (other than the speed, of course) was how easy to find were the controls for the specific character that the animator was working on.

There might have been dozens of characters in that scene, but we only saw the one character they were actively animating (which can easily be done in LW) and the animation controls for that one character.

Imagine if, in Modeler, all the layers of all the objects we have open would get jumbled together alphabetically in the layers panel and the layers widget. It would be a nightmare.

In layout, we should be able to select a character in some way and only the animation controls for that character should remain visible (not only in the the viewports but also in the menus and dropdowns). All the bones, goals, targets, constraints, morphs, sliders, proxies, etc. of all the other characters should completely disappear from every list and dropdown everywhere until we reactivate their respective characters. This would require LW to understand the concept of character so it would understand all the objects and nulls (and lights, sometimes) that make a character.

Having a "Solo" button somewhere so that panels such as the scene editor and graph editor follow/filter by selection seems like a great option to me.

I have also thought I would like to have the Graph Editor "follow" what I am working on. Keeping the active curves visible while filtering out channels that I am not trying to actively control. The issue being here that hitting the space bar in the Layout window switches between Position, Rotation, Scale and hitting the space bar with the Graph Editor active switches between Move, Add, Stretch, etc.. referring not to the layout window but rather the mode you are working on points on the motion curves.

To eliminate "clicking" in one aspect will create it in another.

And again, all of this is a difference in workflow between a novice and a pro. As a novice many people do everything with the mouse. Then as they get more advanced they can use the mouse and keyboard shortcuts. Similarly functions not used by a novice might seem like logical sacrifice in workflow to change, while pro's are gnashing their teeth and screaming at the screen "YOU'RE ALL IDIOTS!" Because they realize all of the extra work that would be created by slanting the workflow towards novice users.

I still always fall back on the preference isn't in making changes for what we think is the best but rather to offer options. Because what is best for some will not be best for all.

robertoortiz
12-30-2014, 11:35 AM
Having a "Solo" button somewhere so that panels such as the scene editor and graph editor follow/filter by selection seems like a great option to me.

I have also thought I would like to have the Graph Editor "follow" what I am working on. Keeping the active curves visible while filtering out channels that I am not trying to actively control. The issue being here that hitting the space bar in the Layout window switches between Position, Rotation, Scale and hitting the space bar with the Graph Editor active switches between Move, Add, Stretch, etc.. referring not to the layout window but rather the mode you are working on points on the motion curves.


great post.


And again, all of this is a difference in workflow between a novice and a pro. As a novice many people do everything with the mouse. Then as they get more advanced they can use the mouse and keyboard shortcuts. Similarly functions not used by a novice might seem like logical sacrifice in workflow to change, while pro's are gnashing their teeth and screaming at the screen "YOU'RE ALL IDIOTS!" Because they realize all of the extra work that would be created by slanting the workflow towards novice users.

I still always fall back on the preference isn't in making changes for what we think is the best but rather to offer options. Because what is best for some will not be best for all.



Good points.
A happy medium can be reached.
I guess one thing we can agree is that LW can stand for some de cluttering.

vncnt
12-30-2014, 12:01 PM
Having a "Solo" button somewhere so that panels such as the scene editor and graph editor follow/filter by selection seems like a great option to me.
Legato.
And it has a "Mute" button too.

I have also thought I would like to have the Graph Editor "follow" what I am working on. Keeping the active curves visible while filtering out channels that I am not trying to actively control. The issue being here that hitting the space bar in the Layout window switches between Position, Rotation, Scale and hitting the space bar with the Graph Editor active switches between Move, Add, Stretch, etc.. referring not to the layout window but rather the mode you are working on points on the motion curves.
Again: Legato.
By default it will display all "current tool"-related channels that are not locked.
But you can also display position/rotation/scale/morph/mc channels that are related to the current item.

Oh, and it can transfer poses and animations, defined by markers and regions, from both external LWS files and internal markerregions too.
And organize your schematic viewport.
And store/recall up to four render presets.
And reverse/timestretch/amplify/clone/move animations too. For entire hierarchies or subhierarchies.

Legato beta 1 rev O12 is expected to be published soon.

robertoortiz
12-30-2014, 03:47 PM
Thanks for sharing that. But one of the thing I was hoping is to come up ides to suggest to expand and improve Genoma beyond the rigging of models.
Having said that Legato looks great.

RebelHill
12-30-2014, 04:35 PM
I was hoping is to come up ides to suggest to expand and improve Genoma beyond the rigging of models.

Hate to have to say it again... but...

Done!

Its called RHiggit... fully integrated rigging and animation tools that talk to and understand one another.

jeric_synergy
12-30-2014, 05:15 PM
Roberto, it seems to me you are MULTIPLYING clicks. Why not have the UI do a bit more, without clicking at all.

For some inspiration, look here: http://www.dontclick.it/

I'm not a big fan of the MAYA channel box, to me it's like something an engineer would design-- really, that's how I feel about the entire MAYA UI-- the rivets show, and they like it that way.

Just throwing this out there, along the lines of easing access to the animation curves: How about a highlight system that shows you what item the UI thinks you're currently hovering over and when you hold down a 'sticky-key', a HUD-style version of the GE appears in the main UI window. You can do whatever you need to the curves in this HUD, while the viewport 'behind' shows you the results of your alterations. Another press of the sticky-key and the HUD goes away.

Features would include: excluding items from auto-selection; specifying curve sets to be displayed; presets for same; proxy-picking if an item is constrained by another; complete color and opacity control; selection following thru the usual methods without dismissing the HUD;more....

(I may be misusing the term 'sticky-key'-- in this usage I mean it does NOT have to be continuously held down to activate the feature.)

++++++++++++
Also, Roberto, unless you are fully up to speed on RHRigget, you might want to slow down a bit and see what's in that package.

RebelHill
12-30-2014, 06:31 PM
I'm not a big fan of the MAYA channel box, to me it's like something an engineer would design-- really, that's how I feel about the entire MAYA UI-- the rivets show, and they like it that way.

And yet it is THE software of choice for the vast majority of animators... must be doing something right. And that something is the exposure of control, rather than the obfuscation. Animators (artists in general in fact) need control... absolute control, not silly tools that try to do things for them. THAT... is generally what engineers have a bad habit of doing, making stuff that does stuff for you, because... well, that's what engineering is about... lifting some of the workload.

Why can't I change the keyboard colour on my iphone... because we already made it the best colour for you.


How about a highlight system that shows you what item the UI thinks you're currently hovering over and when you hold down a 'sticky-key', a HUD-style version of the GE appears in the main UI window. You can do whatever you need to the curves in this HUD, while the viewport 'behind' shows you the results of your alterations. Another press of the sticky-key and the HUD goes away.

See, now that's more like it... all the tools, all the controls... accessible with ease and immediacy. Ofc, you can argue that its better just to have docked panels for a ui, so that you can always have the GE open, or have it as a fly out, but so long as that option is still there for those who want it, there's nothing wrong with having it pop up elsewhere too.

(oh... and I think you meant HOT key.)

jeric_synergy
12-30-2014, 06:48 PM
The term "sticky-keys" has a distinct meaning in the UI design toolbox, but as it happens none of the applications I use use them, so I'm a bit vague on the specific definition.

It's either:

A mode or feature that is called up and dismissed by a given key, or
A mode that is only active while a key is held down.

If it's the first, I don't know why they would term it 'sticky'. If it's the latter, what exactly is sticking?

Very likely I've got it all wrong.

Oh, in the second definition, Photoshop's "spacebar=move screen" thing, and 3dCoat's "ctrl=Smooth Tool no-matter-what" qualify. For the HUD idea, the first definition would be preferable, I believe, because you might want to be tweaking curves for quite a while.

jasonwestmas
12-30-2014, 06:59 PM
A sticky key is a button/function that has to keep in contact with your finger (like it is sticking to your finger) in order to toggle it on. Release the key and the function is toggled off. It's convenient in the way that you don't have to remember to hit the key again to turn it off later, making workflow faster for functions that you use often but don't need to be turned on for lengthy periods of time. Like snapping to different types of components.

RebelHill
12-30-2014, 07:17 PM
I thought a sticky key was the one which you pushed once to turn on, and again to turn off... thus the key stuck as if held down... whilst a key which you actually have to hold whilst performing an operation is a modifier key.

jeric_synergy
12-30-2014, 07:28 PM
The Photoshop usage ("position image in viewport") is very useful and I use it constantly. The closest thing in LW, I think, is the alt key behaviour of the viewports, or the ctrl key behaviour of various tools (constrain to axis).

I don't think that's the behaviour we'd want for the GE-HUD, for the reason stated above.

hrgiger
12-30-2014, 09:01 PM
I've asked for sticky keys for some time in LW, and sticky keys I've always thought as keys you hold down to enable a temporary function and then returned to whatever the former function was before the key was pressed. So if you're in move mode, hold down the Y key, rotate your object, release the Y key and you're automatically back in move mode.

jeric_synergy
12-30-2014, 09:50 PM
OK. --I think it's bad nomenclature, as 'sticky' doesn't seem to apply, but whatevs-- as a UI method, It Is Good. :)

And good news! You already got sticky keys in the form of the alt key!

Julez4001
12-31-2014, 12:43 AM
I think the problem with comparing APOLLO is that (I am pretty sure) none of us have used the system, at least in a way we can confirm that its less clicking and intuitive for an animator. What I took from the videos was that all the gotchas with heavy scene is that Apollo could burn through multiple Ik systems, cloth sims, hair sims, particle sims and dynamics with minimal CPU usage compared to other system where they would just choke.

Robertoortiz, have you used RHiggit?

ernpchan
12-31-2014, 12:47 AM
I think the problem with comparing APOLLO is that (I am pretty sure) none of us have used the system, at least in a way we can confirm that its less clicking and intuitive for an animator. What I took from the videos was that all the gotchas with heavy scene is that Apollo could burn through multiple Ik systems, cloth sims, hair sims, particle sims and dynamics with minimal CPU usage compared to other system where they would just choke.



Ditto

lightscape
12-31-2014, 02:28 AM
That is simply down to the fact that there just arent very many skilled character animators working with LW... but don't worry... Im working on that too ;)

Looking forward to that but I think lscripting is beyond what a native solution can bring.
How would someone outside the lightwave dev fix the undo issue in lightwave in general, ge, se, etc?
Also make it so that we don't have to click then move, rotate, scale a selection. Something like ikboost with click manipulate would be faster.
Without thirdparty plugins, I think there's no way to tell lightwave that this is a character with its own rig and controls, this is a car with its own rig and controls.
It would be great if clicking an entity in layout would bring up its related customizable gui. Contextual and organized tagging for entities. The way to set this up is even more important.
Don't laugh but I would look at Iclone for consumer level products to see how setting up entities for puppeteering and UX interaction. Characterizing and setting up entities in iclone is well thought out. Its not simply a character or item picker. Its deep intergration.
Maybe genoma will automatically do this someday so multiple rigs in layout can be tagged easily for the user in a nicely presented user friendly gui. I think that's one nice aspect those animation demo roberto posted and the sheer power of the software engine.

RebelHill
12-31-2014, 05:03 AM
Without thirdparty plugins, I think there's no way to tell lightwave that this is a character with its own rig and controls, this is a car with its own rig and controls.
It would be great if clicking an entity in layout would bring up its related customizable gui. Contextual and organized tagging for entities... Maybe genoma will automatically do this someday so multiple rigs in layout can be tagged easily for the user in a nicely presented user friendly gui.

No... other than a plugin which "knows" a rig and its control layout, there is no way to "let LW know"... When I say that RHiggit has rigging and animation tools that understand one another, it is ofc only within the context of the RHiggit tools themselves (namely the body builder and animation toolbox).

The problem that you have ofc with creating such a system... a characterisation system... is that things MUST conform to a certain standard. Thats how iClone, motionbuilder, etc do it. You can't just take any old arbitrary rig and somehow magically "tag" it as a character. Is the rig you've given it using IK or FK arms, or switchable and how has it been designed hierarchically? Are the legs using pole vectors or local space twist? What does the foot system contain, are there only a couple of controls, or many, how is the hierarchy of those controls designed to function, how are the controls laid out and designed to co and counter rotate with/against one another, is there even an actual foot "system" at all?

And so this is where the fundamental idea behind the RHiggit system comes from... The fact that you need a design underpinning character rigs as a whole. Easy to do for a single type of character, such as humanoid bipeds (ala motionbuilder... there is essentialy only one control rig, ad all characters are conformed to it)... But to do that for a wide range of possibilities is trickier, and requires a great deal of planning. And it is this which causes RHiggit (under the hood/behind the curtain) to be a VERY complex beast indeed. Parts are designed in such a way that they conform to a general principle, and not just one specific design per part... but multiple. RHiggit understands different setups on spines... FK, IK, spline, Handles... it's able to understand different hierarchy orders, and adapt itself to these varying possibilities. Parts can "talk" to neighbouring parts in a hierarchy and conform themselves logically (like where a neck can see a head attached to itself and will switch its bank control into the head system... (meaning no counter animation of head and neck that needs to be maintained or wrestled by the animator).

This is what allows you to be able to simply switch parts to different styles of controls with just a click... https://www.youtube.com/watch?v=Yp8NJVypTY8 and the underlying "character system" is able to handle these different control methods.
It's what allows you to actually DESIGN the ordering of controls to achieve different styles of control for a single method that best fits the way you want a character to work... https://www.youtube.com/watch?v=sjA8VPWTvvc and the system can adapt itself to those changes and maintain its own internal consistency.

And because no matter how you've stitched things together it always keeps track of the character as a "whole"... it's what allows any character rig to be "compared" to any other without the user needing to go through some awkward or long winded "mapping" procedure, and why, consequently, you can take poses and motions from any one character, and apply them faithfully to any other... https://www.youtube.com/watch?v=7LxtV07BEl0&t=08m26s despite differences in the number of controls, how those controls and parts are laid out or structured hierarchically, or differences in the character anatomy.

And yet all of this, despite the complexity of the actual rigs themselves in places, maintains efficiency... Genoma 1 rigs are pretty dog slow... Genoma 2 much faster... but still... a comparable RHiggit rig (same basic anatomy and control systems in place) is up to 60% faster in scene playback than the Gen2 rig.

I don't see LW nor genoma delivering this kind of flexibility or "internal connectivity" any time soon... if at all, because there's no central "design ethic" underpinning those tools. Genoma is pretty much just a way to "re-pose" any single preset, fixed rig into different body shapes. There's no kind of attempt to maintain some kind of "parity" between different character designs. No underlying, unified, way in which the rig control systems are pieced together or desiged, and thus, no sign of it all being able to fit together into some kind of "character system".

Too much clever... not enough smart.

robertoortiz
12-31-2014, 07:51 AM
Ok guys, sorry i have not been able to participate more, but I got a newborn at home.
I m going on vacation unitl the 7th, but I will bring with me the diagrams to see if I can come up with some ideas. Keep the fire warm until then.

-R

chikega
01-02-2015, 02:30 PM
I'm still plodding through both the Rhiggit 2 tutorial videos (50 of them and counting) as well as the paid Rigging 2 tutorials (to learn how to rig from scratch). These are essential to learning how rigging works in LW as well knowing it's weaknesses. RebelHill knows the dark underpinnings of LW rigging for sure.

jeric_synergy
01-02-2015, 04:25 PM
He has forgotten more than I will ever know, and what's he's forgotten was not important.

Davewriter
01-02-2015, 08:27 PM
He has forgotten more than I will ever know, and what's he's forgotten was not important.

When I first thought of "learning" LW I just thought that rigging was something you just tossed in. A big part of me hoped for that as well, being a lazy sot and just wanting to make my creation. Since I have seen the light and realize how important it is and certainly want to tip my hat to and wish I could shoot a pint through to you RH.
Who knows... some day I may even learn it :)
Still owe you that pint!

jeric_synergy
01-03-2015, 03:26 AM
.. and wish I could shoot a pint through to you RH.
Well, really you can. Just purchase one of the versions of RHrigget or one of his other products (tutorials and whatnot).

VonBon
01-04-2015, 07:49 PM
Too much clever... not enough smart. :ohmy:

If newtek wants to improve animation in LW, then they will need to
change what ever is needed under the hood so that we can have
a better GUI. Not just the look, but how items are handled inside
the engine. Docked windows want cut it, we need ways to group,
organize and tag items within large scenes, while also consolidating
motion tools so we're not all over the place.

Davewriter
01-04-2015, 08:03 PM
Well, really you can. Just purchase one of the versions of RHrigget or one of his other products (tutorials and whatnot).

Then I feel better. At that rate I just picked up the evenings bar bill.

01-05-2015, 04:14 AM
They are called "Principles" for a reason.
They are the underpinning of the philosophy by which they wanted all their (Disney) animations to have as it was/is a known formula for success.
Go look at the cute cat in the Happy New Year thread of Dmind for an example of this as applied through LW.

It really isn't that Genoma can or cannot do this (12 Principles) as that is not its purpose. The purpose of Genoma is to put together a rig and for YOU, the animator, to bring forth the 12 Principles to your art.


The thread title really needs to change. The question I see being of worth to ask here is: How can we use Genoma to express the 12 Principles of animation?


A myriad of answers follows this question on the 'net, just waiting to be extrapolated into LW-ese; RebelHills new tutes might hit the spot; posing an animation with the intended principle as a backdrop to be followed might get you there.

The diagrams would probably make good practice pieces, at best.

jeric_synergy
01-05-2015, 12:00 PM
The thread title really needs to change. The question I see being of worth to ask here is: How can we use Genoma to express the 12 Principles of animation?
Well said! :thumbsup:

Dexter2999
01-05-2015, 01:53 PM
Well, my comments before weren't targeted at Genoma specifically per se. And if we are going to be that specific about it, of the 12 principles, Genoma directly pertains to Straight Ahead Action and Pose to Pose, which I equate to FK/IK.


But lets break down those others real quick.

Appeal
Solid Drawing
Design and modeling/texturing issue?

Exageration
Squash and Stretch
These can and/or do pertain to deformations. And that is what bones and joint do. But it extends beyond that to deformation cages/lattices also. We don't really have that natively. (I do enjoy the videos from 3rdPowers and what they have devised.)


No, I think most of the relevant points of the "12" have to do with the Graph Editor rather than the rigging system.
Anticipation, Follow through, Arc, Slow In and Out, these things have to do with the principle of "S" curves in animation and that is more directly addressed in the Graph Editor.

Staging
Layout in general

Timing
Really isn't about the software at all. It is about the artist.

Secondary Action
I don't know that this fits firmly in an aspect of the software. It is animation so I suppose putting it in Genoma is as good a place as any. But layering action and performances has more to it than that.

This is just my assesment. Any differing opinions as to how these things get classified?

The point of this, is that if we are going to get very specific about how Genoma can be improved we need to be examining what it applies to. And I will also pose the question, " is the needed improvement in one specific area or are should we be looking at how two areas like posing a rig and working in the graph editor work in tandem?"

Dexter2999
01-05-2015, 06:35 PM
This is in response to the "Sharpie Challenge". I don't have the experience to speculate effectively on what should be open. So, I took a screen shot from the Dreamworks Apollo video and broke it down and highlighted sections that I "think" is what is being reflected in the screenshot.

I think it breaks down to the following components;

Left is
Stage
Playback Control, Audio Scrub, Timeline
(What I think is) Scene editor type strip for selected object

Center is
Selected Object
Active Control for Selected Object
List of Channels for Active Control showing spreadsheet values

Right is
Graph Editor Showing relevant channels for Active Control

126422

Any corrections? Comments?

creacon
01-06-2015, 07:09 AM
Yes, rigging process certainly got improved. How one animates is pretty much the same. I don't think it's that different in other packages too. Pose, keyframe, pose, keyframe, repeat.


There's one thing missing in your list and that's "undo". As long as this problem isn't solved in LW our animators won't even look at it for CA.
Another important part is some kind of deformation stack (it already exists as the displacement plugins list but the bone deformation isn't part of it). I wonder how much time it would take a developer that has access to the LW source to implement this. The skeleton wouldn't be part of the object anymore (as is the case in other packages), but it would open up tons of possibilities.

creacon

bobakabob
01-06-2015, 03:20 PM
There's one thing missing in your list and that's "undo". As long as this problem isn't solved in LW our animators won't even look at it for CA.
Another important part is some kind of deformation stack (it already exists as the displacement plugins list but the bone deformation isn't part of it). I wonder how much time it would take a developer that has access to the LW source to implement this. The skeleton wouldn't be part of the object anymore (as is the case in other packages), but it would open up tons of possibilities.

creacon

Can you say more about your LW "Undo" frustrations? LW has limitations but I sometimes feel this is overblown. Of course it's not as luxurious or expansive as Maya where virtually everything "undoes". However, if your IK rig components are set up well with controllers, Undo in LW works OK, no? Also FK moves can be undone no problem. You can also undo changes in the Graph editor, albeit only once (so yes this definitely could be better!). Saving LW scenes incrementally is good practice and this doesn't take up disc space. Lately I've been animating in Maya (using Advanced Skeleton) and Lightwave (using RHiggit and Genoma) and enjoy working in both apps - rendering in Lightwave - without major headaches.

Dexter2999
01-06-2015, 09:42 PM
Also, FWIW, I did some clicking around. That immediacy that RH refers to comes a premium cost. The Pixar presentation specifies a Quadro K6000, which run about $6K. And the Dreamworks guys are using 24 Core Hp Z820 machines, which start at about $6K and go up from there. (To be fair, some of the Dreamworks guys are only using Z800's)

Either way, these guys are using gear worth more than my car, to get that high end performance.

jwiede
01-06-2015, 10:33 PM
Can you say more about your LW "Undo" frustrations? LW has limitations but I sometimes feel this is overblown. Of course it's not as luxurious or expansive as Maya where virtually everything "undoes". However, if your IK rig components are set up well with controllers, Undo in LW works OK, no? Also FK moves can be undone no problem. You can also undo changes in the Graph editor, albeit only once (so yes this definitely could be better!).

Yes and no. Part of the problem with LW is that the "undo" in Layout is "local" as opposed to part of a "unified undo set". So yes, GE has a single undo, but it also only triggers in GE. If your sequence of actions is something like "move element", "tweak graph", "move element", etc. you've created an ordering dependency, and unless you remember where that "GE undo" falls among the "general undo" ops, you have a problem. If you do undos in the wrong order in LW, you often will wind up in a new (not previously occurred) state rather the prior state you expected -- that's a serious limitation compared to other apps where there's typically a singular undo stream containing all "undoable" events.

And yes, incremental saving is an alternative, but it too has issues -- as an alternative, it is neither as convenient nor as reliable in LW as undo is in other apps, because of the object-vs-scene split. Just incrementally saving the scene does not always capture all changes, and having to continually incrementally save both scene and objects is nowhere near as convenient a situation as dealing with undo. Also, again, you have a dependency issue as well, because there's no automatic coordination between incremental object save revisions and incremental scene save revisions in native LW. Yes, you can find scripts which offer versions of that functionality, but even there you're typically relying on filenames as the sole representation of the state contained within the files (and that's not very adequate if you're moving back more than a few saves' worth in time).

There is nothing, IMO, "overblown" about the undo-related complaints w.r.t. LW. There is basic functionality absent from Layout that is ubiquitous in virtually every other 3D pkg (and nigh-all other apps), and worse, that's been the case for a embarassing length of time. Even setting aside the relative competitive aspect (though that too is a strong indicator), the sheer number of customers who mention undo as an issue amply demonstrates that it is a major problem.

m.d.
01-06-2015, 10:44 PM
to complicate the undo system further....
When doing some character animation the other day in LW 2015...i accidentally had my 3dconexxion mouse set to 'item' mode and not 'viewport'. I grabbed a bone bumped the mouse and sent the bone to the other side of the screen.....

But....because of the way LW treats the 3d connexion moves, undo wouldn't undo the move, and scrubbing the timeline wouldn't pop it back either. The only way to restore the position was restart LW (thankfully had scene incremental saving)

We need a true undo that works in every context, from surface editor to animation.

creacon
01-07-2015, 02:47 AM
Try this:

Autokey is Off
Move something
create a key
Move again
create another key
undo
undo
undo
undo

First think about what you would expect to happen, then try it.
Node editor has full undo
Surface editor: no undo at all
Graph editor: 1 undo


Can you say more about your LW "Undo" frustrations? LW has limitations but I sometimes feel this is overblown.

So, no it's not overblown, it's a disaster. ;-)

lightscape
01-07-2015, 03:56 AM
The undo. I wrote in bold letters in my previous post about this long standing issue in lw. Its not fixable by third party.

Ernest
01-07-2015, 11:52 AM
Also, FWIW, I did some clicking around. That immediacy that RH refers to comes a premium cost. The Pixar presentation specifies a Quadro K6000, which run about $6K. And the Dreamworks guys are using 24 Core Hp Z820 machines, which start at about $6K and go up from there. (To be fair, some of the Dreamworks guys are only using Z800's)

Either way, these guys are using gear worth more than my car, to get that high end performance.

You can find a 24 thread CPU in ebay for less than $1500 and a compatible motherboard for around 500. Most of the rest can be reused from an old PC. Still a lot but more plausible. But I'm sitting on 32 cores and I don't have that immediacy.


How one animates is pretty much the same. I don't think it's that different in other packages too. Pose, keyframe, pose, keyframe, repeat.

But what does the word pose mean?

Does it mean opening a dropdown at the bottom of the screen and scrolling down a list of 400 items until you find Chuck's left arm's IK goal and moving that, or does it mean clicking on the graphic of your rig, like in Maestro, or from a curated list of controls like in Rhiggit?

Does it mean rotating your bones into the desired pose and letting the mesh take the corresponding shape as best it can, or does it mean meticulously sculpting the shape as seen on the camera view so the silhouette will have the exact curve from the reference character sheet?

In which case, does it mean moving dozens of little helper bones, or using dynamics to drag points, or swooshing a brush, or reshaping with a 3D lattice, or reshaping with the camera's 2D lattice?

And how many pose, keyframe cycles do you do? Ten because your morphs are only linear or two because you can adjust the path like with smartmorph.

jeric_synergy
01-07-2015, 07:20 PM
And how many pose, keyframe cycles do you do? Ten because your morphs are only linear or two because you can adjust the path like with smartmorph.
??? Is it best practice to use morphs in CYCLES?

Cuz it doesn't sound very efficient.

Ernest
01-07-2015, 09:53 PM
??? Is it best practice to use morphs in CYCLES?

Cuz it doesn't sound very efficient.
The premise I was replying to was that animating is always a series of pose-keyframe cycles. Pose, key, pose, key, pose, key... That's the cycle. And it's true in a literal sense, even if not in a practical one. The "cycles" in morphing, in that context, then, are the number of times that you pose and key to go from one morph shape to the other. I wasn't referring to real cycles, but at the number of times you have to repeat the pair of actions "pose and key". Because setting a morph target is posing. So, with only linear morphs, you have to pose-key-pose-key-pose-key-pose-key many times to make sure the morphing action follows an ideal action curve. With smorph.p, it may be possible to have just one pose-key at the start and one at the end and define a satisfactory action curve just from those two poses.

01-08-2015, 04:13 AM
Smartmorph... Brilliant.

jeric_synergy
01-08-2015, 09:48 AM
Ah, thanks for the clarification Ernest: "work cycle" not "animation cycle".