AuthorTopic: Puppet/Modular Animation: how when and why?  (Read 44488 times)

Offline Conzeit

  • 0100
  • ***
  • Posts: 1448
  • Karma: +3/-0
  • Camus
    • conzeit
    • View Profile

Re: Puppet Animation: how when and why?

Reply #10 on: February 05, 2014, 08:22:36 pm
Crow: oops...I guess you fixed all the videos cause they all work for me now, thanks!.

Guys(NaCL)! This topic is me saying "hey anyone who has experience with puppet animation help me see how I can use it and still do stuff that -I- would like to look at!"...I'm not saying modular/puppet animation is the shit or anything like that! my bones tell me it sucks shit, but I have to be realistic as an animator freelancer and accept it's a good tool! I dont want to become obsolete for being too purist.

@Probo: DUDE that is exactly how I feel about Rumblefish, down to liking just the flourish/followtrough/tweening and hating that stupid grapple move. =O

The clothes thing is interesting, puppet can also do secondary motions! it also reminds me of how they showed damage in Gundam Battle assault by tearing parts off the robots. All these things highlight how puppet is basically applying 3D tech to 2D animation

@Ryu I never expected to see this done well with pixels drawn in the same resolution as they're rotating ( I HATED it whenever they used it in Metroidvanias ). But in Catacomb Kids it works in it's own creepy way, like you said it makes for an unique look.

So far I had only liked the mix of pixel and puppet when the pixels are lo-res and the rotation is in hi res. Gives a vectory quality to pixels kinda like Ridiculous Fishing


@Mathias thanks for the heads up on SPINE! I'll check it out.

What you're saying about the Rayman engine is exactly what makes me feel left out and wishing to use their toys, Vanillaware and Ubisof OBVIOUSLY have awesome inhouse software, but you cant just expect anybody you work for to make you a custom tool that lets you do all that fancy stretch, squash and deformations.
Also, Morph Target/Mesh deformations is what I meant when I said distortion/wraping(if that wasnt clear). I hope Spriter isnt far behind Spine in oficcially adding that feature!

Gosh, the idea of the open modular/puppet animation tools adding deformation and to see it pulled off in pixels (bleed and catacomb kid) makes me think of using it regularly =O what are you guys DOING to me?  :lol:  :P

I love the discussion we're having...talking this kind of stuff out is always fun for me, but I wish someone who's worked with it and got over their purism would chip in which is why I was so glad PxPd was watching over this board (wink, wink, nudge, nudge)...I'm gonna go see if I can fetch someone on twitter  O_O
« Last Edit: February 05, 2014, 08:44:57 pm by Conceit »

Offline PixelPiledriver

  • 0011
  • **
  • Posts: 997
  • Karma: +6/-0
  • Yo!
    • View Profile
    • My Blog

Re: Puppet/Modular Animation: how when and why?

Reply #11 on: February 06, 2014, 06:00:25 am
Great posts.  :y:

Other examples of games that use a mixed process:
Guacamelee -----> loads animations from .flas
CWars ------> pixel art composited with animation tool and exported as data.
Rakuga Kids ------> Mostly drawn frames mixed with deformations.
Paper Mario -----> Vector engine.
King Arthures Gold ------> Minimal rotations but used for some stuff like the archer, trees, etc.
Gunstar Heroes --------> old school example of transformations.

Personally I think this kind of stuff is really cool.
It's a great example of artists and programmers coming together to make something interesting.
Building base tech for games is a lot of fun.
Having an artist add content to an engine and creating new features to accommodate is very rewarding.

but I wish someone who's worked with it and got over their purism would chip in which is why I was so glad PxPd was watching over this board (wink, wink, nudge, nudge)...I'm gonna go see if I can fetch someone on twitter  O_O
I have some experience with this but not really enough to make strong statements about it.
My animation experience so far is a very weird mixed bag of traditional, pixel, vector, and 3D.
I haven't tried a 2D bone tool like Spine or Spriter.
I'm sure there are other much more experienced people that could be more helpful with this topic.
Also I've seen a few tutorial videos around for those programs and they seem solid enough.

I've taken some interest to using flash again after doing the animation in the scrolling thread, but haven't taken the time to do much with it yet.
It has a nice timeline.
Objs can be replaced by other objects so that they can share animations.
Is it good for games that don't run from a swf?
Guacamelee loads animations from fla files.
But not exactlly sure how it works.
I assume they wrote some code stuff that parses the vector data and redraws it in engine.
You could export as a sprite sheet.
But obviously its better to use bone driven animations that export data to be read by the engine for this style.
Better to re-skin.
Easier to tweak.
Faster to see the results directly in game.
Frames could still work for what you want tho.
There's plenty of examples out there.

As far as getting over purism:
This is rather easy for me.
Is puppet better than traditional?
Or the other way around?
I'm very neutral about most things so I won't be able to help answer that.
Traditional animation processes are very old.
They have a lot of appeal and strength but it doesn't change the fact that there are other things out there to give a shot.
Drop about 100-200 hours, realistically you'll need to do a lot more, into something you've never tried before and then make an assessment of it.
Don't just limit yourself because you think "it's bad!" or "it's cheating!" or "it doesn't look the same!".
Even if you come back to another process of animation you will have gained something from that experience that will help inform you with your work.

If you really have interest in using transformations make sure to have some sort of outlet.
Spine looks cool and works with a variety of engines, including Unity which I like a lot.
Pair up with a programmer, or learn the basic tools/interface/code yourself, and make stuff happen.
If a game project is out of your time scope consider setting a smaller goal like a 1 min animated short or something.

Coming from redrawing obsession in pixelart (seriously I avoid copy pasting parts as much as possible)
I copy paste all the time.
Use the select tool constantly to rip things apart and rotate/scale/translate.
Duplicate frames to set up timing.
Then change only small parts of focus and leave other areas the same.
Adding more unique drawings helps for certain effects but it's almost always not the only option.
There really is some kind of social anger/pressure to only do things a certain way.
I encourage you to discard all of that and be active and open minded.
Find things that help you do what you want to do and don't have second thoughts about cheating.
How do you keep the volume of a bouncing ball in traditional animation?
Draw a circle on a long book mark shaped sheet of paper.
Then whenever you need a frame to come back to the original volume of the ball, slide it under your top sheet on the light table, move it to the location desired, and trace it.
Oh what that's cheating?
It's not, it's just a very simple tool made out of paper.

Modular animation is also subject to code. Code can't alter a set-in-stone frame-by-frame animation in the same way.
All these things highlight how puppet is basically applying 3D tech to 2D animation
And for games this is a big deal.
3D animation is driven by blending animation data and outside variables and 2D with bones is really the same thing.
Back in the day a game engine didn't necessarily use transforms to calculate position and movement and rendered pixels directly into a buffer.
Most games these days even if they are 2D use a 3D engine and draw everything onto quads.
All the graphics data goes thru the shader pipeline which requires using transforms.
This creates an engine that is inherently based on transforms, adding animations that are also based on transforms makes them work together well.

On the other hand, while frame data is obviously not as flexible you can still send information to a shader in the form of color, usually for labeling pixels to be modified in some way.
A simple example is the skull girls palette swap map.
They show it in the video mathias posted for a few seconds but don't go into depth about it.
Much more complex ideas can be done with passing color as data that affect more than just palette.
But let's talk about that some other time.  :blind:

an animation of a character screaming while rotating  an arm can have the same charm as swapping between two crappy frames.
It's funny you mention this.
I've been working on a 2 frame animation thread concept but haven't quite completed the thought.
I'm a big fan of limited animation and I think that it can be really strong. -----> Facets Castle of the Winds thread being a current example

and hating that stupid grapple move. =O
But it still counts as a piledriver!  :y:
Rumble Fish looks cool but definitely has some awkward moments.
Consistency in animation is fairly important.
Whatever style or tools you go with try hard to make frames, actions, speeds, holds, etc fit together.
Games tend to repeat animations over and over so strange looking ones will always float to the top.

Also that Rayman editor looks super boss.
In game tweaking is the way to go.

Looked at spine a little.
Will probly try the demo.
The differences between Essential and Professional don't really please me.
But still gonna try it out.

With puppet animation, the shapes that compose a thing don't really change. This kills the squashing and stretching and deformation that are so important to the illusion. Also, it really locks the form into two dimensions. Even with a fixed camera in a game, the frame-by-frame animations can play in the 3rd dimension all they want by rotating, foreshortening, and so on.
Wanted to comment on your post as well NaCl.
I know exactly what you are getting at.
Frame by frame animation is really awesome.
When I look at a game that leans more towards traditional process, like Skull Girls or Vanguard Princess, Rumble fish seems a little ridiculous.

Check out a combo video if you are into that sort of thing.

It also has other things going for it.
The timing is much tighter.
The impacts are better represented.
But there's still something interesting about Rumble Fish's engine, even if some of it is executed oddly -----> I seriously need a gif of that piledriver as my avatar.  :blind:
And the other possibilities that the puppet concept offers games are really cool.

Also this comment was not directed at you.
There really is some kind of social anger/pressure to only do things a certain way.
Just commenting on how artists, including myself, second guess tool choices based on things they randomly hear.
« Last Edit: February 07, 2014, 12:46:37 pm by PixelPiledriver »
And knowing that it is, we seek what it is... ~ Aristotle, Posterior Analytics, Chapter 1

Offline Conzeit

  • 0100
  • ***
  • Posts: 1448
  • Karma: +3/-0
  • Camus
    • conzeit
    • View Profile

Re: Puppet/Modular Animation: how when and why?

Reply #12 on: February 07, 2014, 09:20:23 pm
 :o holy shit PxPd. thank YOU for posting on my topic

loving the guacamelee stuff. How do I miss that? also loved your programmer artist hybrid perspective on things. thanks for crapping on my purist side to help me get over myself :P had a good laugh.

I'll edit this into a better reply later O_O just wanted to say SOMETHING. (unless -someone- replies to this before I get to it =O)

Offline Mathias

  • 0100
  • ***
  • Posts: 1797
  • Karma: +2/-0
  • Goodbye.
    • View Profile

Re: Puppet/Modular Animation: how when and why?

Reply #13 on: February 08, 2014, 03:47:48 am
P-Driva, I'd love to see what you might do with Spine.
I'm almost positive I'll be tearing into it, myself. Just not yet.

I need to check on Spriter again. (good god what an awful website)
Spriter is only $25, but Spine seems like a better product. Much more expensive, yes. But I don't care. I'll pay OUT THE NOSE for anything that helps me with development.

So . . .
Brashmonkey's Spriter  -VS-  Esoteric Software's Spine
Which is better? Am I the only one wondering this?
Both have the same goals. They seem to fill the same niche. Are they redundant?
*Mathias then does some internet snooping, using his powers of analysis and reason.*
Both were kickstarter projects, but get this:
Spriter was fully funded on April 28, 2012. The people of earth coughed up $71,179 for Brashmoney.
"Spriter is still in development and does not yet have all of its Pro features. This Early Adopter Sale price is a discount for those who purchase Spriter Pro before version 1.0. Estimated delivery date of the full version of Spriter Pro is Q4 of 2013." [source] (Ha! Way to keep your site up to date!)
The Kickstarter campaign was 21 months ago.
Spriter is pretty much still in beta.
What's the hold-up, fellas? Bite off more than you can chew, perchance? (I'm not sympathetic towards Kickstarter abuse.)
Now let's look at Spine's history:
Kickstarter campaign funded on Feb 23, 2013
They got $67,569. $3,610 less than Brashmonkey.
It's been 11 months.
And it's lookin' good. Earlier in this thread I mentioned Spine getting "mesh deformation" soon. Well guess what?
The Spine dev's call it FFD - FreeForm Deformation. --- COMPANY BLOG ENTRY --- FORUM POST --- and it's in!
I discovered the origin of Spine; what prompted its development. Check this out. Found on their Kickstarter campaign page:
"BrashMonkey's Spriter began development 15 months ago. We were initially thrilled when Spriter was funded on Kickstarter 9 months ago. We posted our feedback on the Spriter forums when we found it lacked official runtimes, a multiple-timeline dopesheet, tweening curves, and separate keying of scale, rotation, and translation. After a couple months we decided that Spriter may never have the features and workflow we envisioned, and that is when we decided to build Spine."
In my opinion, this paints the Spine dev's as smart, opportunistic, self-driven entrepreneurs.
They saw an opportunity and they seized it. Basically, Spine is intended as a BETTER Spriter.
Check out their pricing structure. Indy guys like us will drop $60-$250USD for Spine. Appears to be well worth it.

Draw your own conclusion. I've certainly drawn mine.

BUT WAIT . . .
A Challenger Approaches!
"The Open Source 2D skeleton animation solution for Flash"
I was once proficient with Flash. Once. Now very rusty. But bones in Flash would rock. That's obvious.
And it outputs game code, like Spriter and Spine.
Aaaaaand it's totally free.
So, that's something.


OH hold on, we got one more here. (wrote this post progressively as I discovered all this junk so bear with me . . .)

"Objecty has been invented to make 2d game development easier! It offers an incredibly easy to use interface for editing & packing textures, animation of 2d sprites, creation of skeletal & tweened animations, building levels and much more!"
The Kickstarter, back in December 2012 failed rather miserably.
Latest news update, on the company site, is from Oct 2011. The project appears inactive.
So forget about this one.


I know this thread isn't about software necessarily, but this post certainly is.
A few years ago, the notion of game dev modular animation tools became a phenomenon.
There were a few earnest attempts.
I'd say, right now, Spine is the clear winner.

Offline Indigo

  • Administrator
  • 0011
  • *
  • Posts: 946
  • Karma: +0/-0
  • Artist, Indie Game Dev
    • DanFessler
    • DanFessler
    • DanFessler
    • DanFessler
    • View Profile
    • Portfolio

Re: Puppet/Modular Animation: how when and why?

Reply #14 on: February 08, 2014, 07:42:52 am
I've had quite a large experience with modular animation tools throughout my career, and have a bit of a business connection to Spriter, so I figured I'd chime in here.

First off, some history:
These techniques of using modular pieces to animate a larger whole are nothing new to gaming and are about as old as sprites themselves; exemplified in arcade cabinet boss fights among many other use cases.  Mostly this technique derived from necessity of the system's limitations (sprite size restrictions, memory limitations, etc) but became a valuable tool to those who could harness it most creatively.

My personal experience with the tools related to this technique was fairly recent in terms of video game history when I joined Gameloft in 2008 in the pre-iphone era of mobile gaming.  We used it fer EVERYTHING.  Although I have no evidence or examples of it, I'm certain that these tools have existed for quite a long time in some form or another.  All that I know for sure is that the technique was used extensively for some time, and inferring based on the complexity of these animations in old school games I would deduce that they probably developed their own internal tools much like Gameloft did.  The impression I got was that these tools were passed on in the form of secret knowledge from company to company as developers shifted around, and at each place they emerged they were guarded as a proprietary advantage over their competition.  This proved to be true when I began work for Glu Mobile in 2009 to discover they had recently developed their own version of the tool.  And it certainly was an advantage!  I was told estimates that cost of development decreased by roughly half when utilizing the new technology.  This meant faster development with more complexity.

More important than the cost and memory savings, in my opinion, is it's benefit on creativity.  Without these tools an artist and engineer would have to work closely together to pull off any artistic flourish that didn't fit in a nice little package such as a flip-book style spritesheet.  If you wanted a fluid motion tween for the menu screen buttons or perhaps wanted to the explosion particles to behave in a specific way, you'd have to convey your vision to an engineer and hope nothing gets lost in translation... and usually it does.  But with a proper tool it decouples the creative aspects from the technical - enabling the artist to be creative within the limitations of the system and without being blocked by engineering.  Just author it the way you want it.  This is the real beauty of it all.  I've seen some amazing things done with these tools that simple never could have been done if you left it to the artists to interface directly with the engineers.

Lastly, there are inherent benefits to using a modular based system in games; Things like hitbox authoring per-frame or per-module, sprite mapping different graphics to the same animations (different heads, clothes, guns, etc), per-frame scripting events, and much much more.  The benefits here are sort of endless and is the major place of innovation with modern versions of this tool; Everything from using bones (early tools didn't have bones or rotations, just simply x-y translation), to mesh deformation.

It was at Glu Mobile in 2009 that I met Michael Parent of Spriter.  We had talked about these tools at length and began planning to make our own commercially available option.  You see, unlike in 3D gaming which saw a standardization of it's format and toolsets, 2D gaming was stuck behind proprietary red tape.  Nobody had access to these fantastic toys - and at the time it was clear the Indie scene would probably embrace such a thing with open arms.  The closest thing to it was flash which was in a sort of walled garden that was great for web-games, but nothing else - not to mention clunky to use and not intended for gaming in the first place.  Unfortunately Michael moved to France and I got busy with work, so he continued the project without me.  I still check in on the project and was later involved in the form of an investor unrelated to the kickstarter.

I'm still cheering them on, but it is indeed a bit discouraging that they've let competition emerge from slow development.  With that said, I still think they've got the advantage here because they're the only ones approaching the problem purely from a 2D perspective.  All other tools I've seen try to map existing paradigms from 3D onto 2D, and not only does that not work too great, but it also limits the potential of the tool by locking you into that way of thinking.  One example, and probably my biggest pet peeve that some of these tools do is locking you into the concept of a rigid skeleton which gives you fluidly tweened puppet animation, but restricts frame redraw and skeleton restructuring.

Why/when you should use it:
Well simply put, frame by frame redraw will *always* be more visually appealing.  There's just no way around it.  If you want something to appear more puppet-like, just draw it that way.  If you want lots of tween frames for smooth animation, just draw them.  Or if you want to go all out and do complex disney-style animation, just go for it.  Want different color palettes on them? Duplicate them and tweak.  It's the option that is the most flexible but also, quite obviously, impractical.  It takes more time of authorship, more memory resources, more ram, etc etc.  The benefits of using modular animation tools are mostly those of practicality.  Depending on the complexity of your project it becomes a no-brainer to utilize a tool such as Spriter.

But I'd argue you should use it for everything, simple and complex alike. Here's why: Using spriter doesn't limit you.  You can still do frame-by-frame redraw in spriter if you wanted to, but now you have access to the other benefits it gives you which may be useful such as hitbox authoring, sound scripting, etc.  By adopting the system for all things, you now have the choice of how you author within the system.  For some elements you may use it a lot, perhaps for UI transitions and animations, and some elements very little - it's up to you, and they'll all play nice with each other in the end.


Efficient - easy on disk and ram
Cheap - for development
Creative - enabling for artists to be creative
flexible - inherent benefits of using modules including sprite mapping, hit boxes, scripting, dynamic effects etc.

Can feel rigid if you're not careful with how you use it
some tools lock you into rigid skeletons much like 3D character models
Draw calls tend to be a lot higher due to drawing so many small pieces - a problem for mobile if you're not batching your draws properly

Sorry for the long-winded response

« Last Edit: February 09, 2014, 08:16:23 am by Indigo »

Offline Probo

  • 0010
  • *
  • Posts: 317
  • Karma: +1/-0
    • View Profile

Re: Puppet/Modular Animation: how when and why?

Reply #15 on: February 09, 2014, 09:45:16 am
some really interesting stuff here. thanks mathias for that research you did, sounds like spine is the way to go to me. I wonder whether using spine to say animate a run, with only flat colours used, then editing the animation in graphics gale to add shading and little details per frame would be an efficient way to get good results. when ive got the cash, i think ill try that!

Offline Atnas

  • Moderator
  • 0100
  • *
  • Posts: 1074
  • Karma: +2/-0
  • very daijōbs
    • paintbread
    • paintbread
    • View Profile

Re: Puppet/Modular Animation: how when and why?

Reply #16 on: February 13, 2014, 03:40:29 am
I've been doing client work in spriter for the past couple days. It's a nice package and I'm surprised by how capable it is for beta. It is very easy to add small flourishes to movements to make up for the fact you are not redrawing every frame. I spent almost all of 2012 working on realistic gun animations in Maya on a 2d plane so I am very familiar and happy with what I'm experiencing so far in spriter, all the technique I learned can carry over.

However I desperately wish it was as polished as Spine looks to be. Spine is too expensive for me atm but I believe when I get into more serious work with puppet animation I will make the purchase and move up.

I will post a thread on spriter next week to show some of the stuff I've done and some things I've learned. There are a lot of AHA and WTF moments coming from traditional animation to puppet, and I think a short introduction to some new concepts could encourage people to make the jump and try something new.

I'm excited about the fact I can spend a few hours making the character art and then pump out a set of animations in minutes. Really gets rid of the aversion I sometimes have to sinking so much time into new movements or fixing bad movements. In frame by frame you need to rework so much when you notice something is off, puppet makes edits and taking crits a lot more worthwhile.

With the proper style adjustments and mindset imo you sacrifice less than you gain by doing puppet. From a utilitarian standpoint. If i was an immortal robot i would do frame by frame of course. for now i will focus on this newer more efficient method so i can buy a shiba inu quicker.

Offline Atnas

  • Moderator
  • 0100
  • *
  • Posts: 1074
  • Karma: +2/-0
  • very daijōbs
    • paintbread
    • paintbread
    • View Profile

Re: Puppet/Modular Animation: how when and why?

Reply #17 on: February 14, 2014, 12:55:53 am

(starts at 30 seconds)

Huh. The new Smash Bros character was introduced with puppet animation. Looks like they did prerenders with special shaders and then moved that around. Why...? Time constraints?

Offline Conzeit

  • 0100
  • ***
  • Posts: 1448
  • Karma: +3/-0
  • Camus
    • conzeit
    • View Profile

Re: Puppet/Modular Animation: how when and why?

Reply #18 on: February 14, 2014, 02:42:26 pm
w00t =O the topic keeps on getting new replies.
that is great to know atnas, indigo. Looking forward to that topic Atnas! as soon as I have a better hardware setup (soon) I'll give it a stab myself.

I'd been hesitant to post this, but this is very much in the spirit of the topic, this is an interview with a the Rauch Brothers, about how they started their StoryCorps shorts with just pure handdrawn animation and moved into puppet

Here's their latest short for StoryCorps
Evidently, this is still mostly frame by frame animation, but the integration of puppet is completely seamless. Just goes to show that if you dont tween the hell out of everything and make smart use of it, you can make puppet be another tool for someone who's mostly doing hand drawn frames.

Here's one last thing which I think is nuts, it's using some kind of morphing effect to make 2D seem 3D
Never tried it, but it seems like it'd take as much work if not more than hand drawing everything (and making 2D seem 3D isnt to my tastes, but you must admit it's impressive).
I I put it here for the sake of completition.

With that and Atna's post of the SmashBros intro for the new char, which has use of perspective and other advanced puppet techniques (Metal Gear Solid Graphic Novel is another example I can think of, there's plenty of examples of this sort of thing) I think we've covered almost the full spectrum of how you can apply puppet, from just aiding hand drawn, to fully animating in puppet, to using perspective tricks and morphing to take 2d beyond what would seem possible

« Last Edit: February 14, 2014, 03:15:11 pm by Conceit »

Offline Mathias

  • 0100
  • ***
  • Posts: 1797
  • Karma: +2/-0
  • Goodbye.
    • View Profile

Re: Puppet/Modular Animation: how when and why?

Reply #19 on: February 17, 2014, 11:16:40 pm
@ Indigo   You just doubled the value of this thread's content.
Thank you so much for posting your experiences. Not "long-winded" at all.
If you have anything to add, please don't hesitate to do so.

@ Probo   Very welcome, sir.

@ Atnas   Yeah, I don't like that Nintendo animated short. Using a sketchy/brushy line style along with modular conflicts visually, IMO. The modular movement gives away the sketchy lines. Looks unnatural.
. . .
I will post a thread on spriter next week to show some of the stuff I've done and some things I've learned.
. . .
*taps foot*!

@ Conceit   I don't really see any modular in "The Road Home". And I didn't find mention of their actual animation tech in that article. I didn't read every word of it either, though.
They're brother animators. That's pretty cool. Wish my bro did graphics stuff . . . but no.

Some simple, but effective Spine anims: