you are here [x]: Scarlet Star Studios > the Scarlet Letters > the advantages and drawbacks of cg
<< before sock mending party 2008
after >> print gocco basics at the iprc

January 12, 2008

the advantages and drawbacks of cg

by sven at 12:01 am

During the past decade, we've seen an enormous amount of computer animation in feature films.

Among the stopmoes at, I hear a lot of frustration about this. We've lost ground. Seeing stopmo in feature films is rare now. Many of the artisans who used to have the skills for doing feature-worthy effects haven't been able to find enough work, and have had to move on. Hard-won craft knowledge is dying off. There are regular rants about the "suits" in Hollywood not understanding what can be accomplished with stopmo, and not putting money behind it.

My strongest love and loyalty is to stopmo. Yet, part of my mind wants to get beyond the rants and understand in depth what's behind the rise of CG. Call it playing the "devil's advocate" -- or just really wanting to know "why?"


Where in feature films do we see computer animation?

It seems to me that there are two areas where CG has come to dominate: monster films, and "cartoon" animation.

For three generations, if you wanted to have full-body monsters on screen, you needed to use stopmo. Willis O'Brien was the progenitor of this heritage, animating the 1933 King Kong. He was followed by his protege Ray Harryhausen, who single-handedly produced the effects for such classics as "The Beast From 20,000 Fathoms," "Jason And The Argonauts," and "Clash of the Titans." Ray's mantle as master of the art form was passed on to Phil Tippett, working within the company Industrial Light and Magic. His style came to be known as "Hyper-Harryhausen" -- more photo-realistic monsters, but with the addition of motorized "go-motion" to create blurs (e.g. for the taun-taun in "The Empire Strikes Back," and the dragon in "Dragon Slayer").

"Jurassic Park," (1993) effectively ended the tradition. A few more films used stopmo for monster effects after it came out... But Jurassic Park's success began the swift transition over to CG.

The other area where we see CG films dominating now is in what I'll call -- for lack of a better word -- "cartoon films." Films like "Ice Age," "Finding Nemo," "Over the Hedge," and "Monster House" previously would have been done using hand-drawn cel animation. The landmark film that began the shift to CG was "Toy Story"... The death-knell signaling that cel animation had seriously lost ground: when Disney dissolved its cel animation division.

I hasten to point out that "cartoon" feature films are a genre that stopmo has NEVER dominated in the USA. Stopmo feature films have always been rarities. If I'm not mistaken, the first feature-length clay-animation stopmo film was Will Vinton's "The Adventures of Mark Twain" in 1984. The first feature-length puppetfilm in the USA followed in 1993: "The Nightmare Before Christmas."

Since then, we've seen further feature-length puppetfilms and claymation -- but coming from an astonishingly limited number of artists/producers. "James and the Giant Peach" and "Coraline" come from Henry Selick, and "Corpse Bride" comes from Tim Burton -- the same two individuals who produced "The Nightmare Before Christmas." Will Vinton studios has dissolved and reemerged as Laika, which is employing Selick to produce "Coraline." Meanwhile, it seems to me that the spiritual successor to the Vinton Claymation tradition has manifested in the UK in the form of Aardman studios -- whom we have to thank for "Wallace and Gromit."

Thus, when we look at "cartoonish" CG films and wish that Hollywood producers were making them in stopmo instead, we would do well to recall that this is not a genre that we have lost ground -- rather, we are only beginning to make our first inroads.



What are the advantages of using computer animation to create monsters and other similar effects? Here's the list I come up with:

1) Editable.
Whereas a stopmo creature delivers a performance which you cannot go back and tweak, computer animation can be tweaked and edited endlessly. Given the financial risk associated with producing a feature film, this is a huge plus.

2) Crowd shots.
Ray Harryhausen pushed the limits of what can be done with stopmo by having Jason fight seven skeletons simultaneously. Stopmo is best suited to dealing with just a few puppets at one time. With CG, on the other hand, you can produce whole armies of creatures.

3) Less dependence on master animators.
When you have to get a shot right on the first try, you hire a master animator to accomplish it. When you have the freedom to edit, you have a much broader pool of talent to hire from.

4) You only need to build one model of a character.
When "The Nightmare Before Christmas" was being filmed, there had to be a dozen or so copies of Jack Skellington, so different animators could be working at the same time on different sets. When your "puppet" is a data file, you can give duplicate copies to many animators, with no further expenditure of time or money.

5) Simplification of materials.
When you do computer animation, everything is made out of pixels. When you're building stopmo puppets, there are many materials to procure: foam latex, silicone rubber, paints, steel and silver solder... And each of these different materials requires specialized tools: brushes, ovens, spray booths, lathes... This is not to say that CG doesn't require specialized softwares and specialized skillsets (modeling, rigging, animating) -- but things are simplified nonetheless.

6) Less physical storage space needed for props.
After you've made a film, what do you do with all the puppets and sets you constructed? Some might be auctioned off -- but a lot will go into cold storage, which consumes space (which also costs money). Keeping data over the long term is highly problematic -- but in the short-term, it's a huge space-saver.

7) Models don't degrade over time.
Puppets get dirty or break and need to be cleaned or replaced. After a few years, foam latex starts to rot. CG models remain immaculate during the course of filming. (In the long run, "bit rot" is a problem, though.)

8) Lighting.
One of the more tricky parts of compositing a stopmo monster puppet with live action footage is getting the lighting to match -- it's a delicate art. With CG, you can use virtual lights on your subject... Another instance where being able to edit CG sequences inside the computer makes the filmmaker's life easier.

9) Shadows can be accomplished without miniature sets.
In stopmo, if you want a puppet's shadow to fall on something, you can't do that with a green screen -- you have to build a miniature set. With CG, you can create transparent shadows that get composited in over live action footage. You may need to model planes that the shadows fall on, but this is usually a fairly simple matter.

10) Dark shots.
When you shoot against a blue screen or green screen, you need a fairly high level of illumination in order to make sure that the backdrop is a uniform color. You probably can't shoot a monster that's supposed to be in a darkened room and just green screen it into your shot -- you'll probably have to build a miniature version of the set. With CG, shadowed creatures are easy to composit.

11) Key-framing.
Both cel animation and computer animation allow you to pose key-frames and then make "inbetweens" to connect them. This isn't an option in stopmo -- which is one of its special challenges. You can plan an animated sequence in stopmo by shooting a "pop-through" -- but you don't ever get to use those actual photos in the final.

12) Complicated sequences can be animated piecemeal.
In CG, you can animate in passes... First just animating the monster's spine, then going back and doing its limbs, then finessing its claws, and finally working on the facial expressions. In stopmo, you have to pay attention to all of these things all at once.

13) Reusable sequences.
After you've animated a monster in CG, you can re-use that performance many times, looking at it from different camera angles. (For instance, if you have an army of monsters fighting.) With stopmo, because you only get one camera angle, you are almost never able to reuse a performance.

14) Algorithm-based motion.
With CG creatures, certain motions can be accomplished through programming rather than key-framing. A millipede's legs or a robot's walk, for instance, are good candidates for this technique. Obviously, it's not an option for stopmo.

15) Easily combined with other effects.
Because CG already exists in the computer environment, it's relatively easy to combine it with other special effects -- e.g. fire, smoke, water, explosions. Fire and water are notoriously difficult to accomplish with stopmo... It's often easier to use other means to create such illusions -- but then you're left with the difficulty of how to fuse them with the original stopmo performance. (Imagine getting a dragon to realistically breathe fire, for instance.)


What about the inherent drawbacks of using CG? My list is much shorter:

1) Less textural.
Most of the textures we see on CG monsters are simulated. The scales on a dragon generally aren't modeled in virtual 3D space -- they're painted onto flat polygons. See, the more polygons you have, the longer it takes to render out frames. To an extent these painted-on scales can be programmed to catch light and shadow -- but this strategy produces less convincing results than when you do actual 3D modeling. Stopmo, on the other hand, uses real textures -- which have an inherently "real" look to to them.

2) Unrealistic lighting.
CG lighting often has a flatness to it. In real life, shadows are often stark and whites are blown out -- but you hardly ever see this in computer animation. To an extent, it's the result of aesthetic choices. In a "good" image, you don't choose to have blown out whites -- but that's not necessarily the most realistic choice. Rendering lighting conditions such as "radiosity" (ambient, reflected light) and sub-dermal glow (flesh's subtle translucency) require a lot of computation... In terms of the time it takes to create these effects, they're very expensive. Stopmo, on the other hand, by using real light, bypasses many of these problems.

3) "Floaty" animation.
From its earliest days, computer animation has fought against its tendency to look "floaty" -- as if things are moving around without being impacted by gravity or friction. Much has been done to improve this tendency -- and yet, its roots are inherent in allowing the computer to create inbetweens for you. Stopmo is often accused of being "herky-jerky"... But in reality, living animals move with some jerkiness. To an extent, what has been perceived as a flaw of stopmo adds to its feeling of "life."

4) Not hands-on.
There's a computer screen between you and the thing that you're trying to animate. To me at least, it's easier to relate to how a thing is supposed to move when I can actually touch it.


When I spell out all the advantages of using CG for monsters, it seems like a really staggering list to me. I don't find it surprising at all that CG would become the first tool of choice for a filmmaker when confronted with a special effects challenge... And I can understand why, over time, there would be an impulse to just do all of your effects work with CG and forget about the other options.

On the labor-supply end of things, I can also see why CG has become so successful. There's a uniformity of software -- which makes it easier to train potential employees. Learning the art of stopmo has largely remained a master-apprentice process (or perhaps even more often, a matter of being self-taught)... Learning how to use a piece of software like Maya, on the other hand, is easily accomplished in a classroom context. Hollywood needs an army of interchangeable CG modelers, riggers, and animators -- so the institutional schooling system responds by offering relevant majors to students.


Stopmoes should remember that computer animation is not just displacing our own work -- artists who make hand-drawn cel animation are also in jeopardy.

It seems that computer-created cartoons are evolving in two main directions: ones created using 3D modeling software such as Lightwave and Maya -- and ones created using 2D vector-based software, such as Flash. The 3D cartoon look is exemplified by films such as "Cars," "Ratatouille," "Happy Feet," "Veggie Tales," and "Barnyard." The Flash cartoon look is exemplified by TV shows like "Powerpuff Girls" and "Samarai Jack." For this essay, I'll limit discussion to 3D productions.


So: What are the advantages of making something like "Mickey Mouse" or "The Secret of NIMH" using computers rather than pencils and paint?

1) Characters are guaranteed to stay "on model."
When you're hand-drawing characters, it takes a lot of skill to keep mass and shape looking correct. With CG, this is a non-issue.

2) Elimination of the inbetweener's job.
Instead of having to pay people to draw the inbetween pictures, all you need is someone who'll do the key poses (in theory).

3) Ease of editing.
When you want to make a minor edit to a sequence, instead of having to re-draw it you can simply push your digital puppet a bit more this way or that -- and the computer will take care of the rest of the fixes for you.

4) Easier to rotate geometrical shapes.
In hand-drawn animation, it's much easier to rotate objects that are round and squishy... Hard-edged rectangular objects are difficult to rotate accurately. Not so for a computer.

5) Lighting effects are easier.
Want a shadow? Want to change the color palette of a scene from high noon to midnight? No problem.

6) Savings on film stock, cels, paint.
Computers aren't cheap -- but (theoretically) they represent a one-time expense. In place of materials costs you have... Electricity bills.


What about the disadvantages of using CG for cartoon films?

1) Fewer cheats.
You can't just imply a location impressionistically with a few lines and swaths of color -- everything you want on screen has to be modeled.

2) Less life in the inbetweens.
A lot of the exciting character of animation happens in the inbetween poses. If you leave that work to the computer, the product is going to be more boring at a very fundamental level.

3) Less squash and stretch.
Yes, to an extent you can squash and stretch computer models... But if you go too far, the rigging (digital armature) will break. You can rig special models for special effects -- but it takes a conscious effort to create the extremes that a pencil can draw with complete ease.

4) Fewer lively "off model" poses.
There's a school of thought (championed most loudly by John Kricfalusi) that focuses on creating truly unique poses and expressions for animated characters. These are, almost by definition, "off model." It's an approach that is contrary to what computer animation does best: uniformity.

5) Absence of line quality.
A huge amount of expressiveness is conveyed through the hand-drawn lines that an animator makes. These don't exist for CG characters.


When the problem is how to create a monster that interacts with live-action actors, CG and stopmo offer two different solutions -- but there is a common criteria for judgement: do the results look photo-realistic? All other considerations aside, CG will usually win out because it is able to provide images that are on the whole more complicated and better integrated into live-action sequences.

When we compare CG and hand-drawn cartoons, however, the products don't look remotely the same. The shared goal? To tell a story that can't be told with live-action. [I'm tempted to say "a story with talking animals," since that is a frequent commonality -- but it wouldn't include a film like "The Incredibles."]

CG cartoons and hand-drawn cartoons ought to be able to co-exist as two separate and unique forms of animation... And yet, how can we ignore Disney dissolving its cel animation branch?

It seems to me that while CG has not delivered the deathblow to hand-drawn cartoons that it's dealt to stopmo monsters, displacement and domination are apparent. Personally, I would say it's largely due to the entertainment industry's aspirations to be... Well, industrial.

The same art school students who are being trained to do modeling, rigging, and animating using Lightwave and Maya for special effects -- they're easily repurposed for CG cartoon films. Companies like Disney were essentially factories to begin with -- but with the standardization that computers (and computer training) provides, the working parts of the entertainment machine (i.e. animators) become even more interchangable.

For an entertainment corporation, the only purpose that "artistry" has is to win Oscars, which act as a form of advertising for the product. So long as the product is "good enough," selling enough units to turn a profit, artistry is expendable...

At least so long as brand recognition doesn't suffer. If different companies' products don't look different from each other -- then there's a reason to start bringing artistry back into the mix!


The phenomenal success of CG as an animation technique is also its Achilles' Heel. Because all of the big entertainment companies have rushed to embrace it, the old techniques have essentially become new again.

There is room for stopmo to be revived for monster films -- and not just as retro pastiche. However, it can never again be the default. From now on, it has to be used as a conscious choice. A name-recognition director very much has it in their power to go this route.

An example: Wes Anderson. Anderson is known for his unique vision; people go to see his films in part if not largely because he's the author. In "The Life Aquatic with Steve Zissou," he bucked the trend and used stopmo instead of CG for all his underwater creatures. The most important of them all -- the Jaguar Shark -- is essentially a monster. This isn't your typical (read "cliched") monster film -- but it's an example of stopmo being used for realistic monster effects, nonetheless.

With regards to cartoon films, again, because the big studios have all rushed to embrace CG (for fear of being left behind?) there's now a void ripe for a daring entertainment company to exploit.

Enter Laika. See, Pixar's doing CG, Disney's doing CG... If Laika starts pumping out stopmo features like "Coraline," it doesn't have any competition! If it does well with this film, it has a chance to quickly establish market dominance.

So how will the other big players in the USA respond? Well, rumor has it that Disney's agreed to do a stopmo remake of Tim Burton's "Frankenweenie." [There's Tim Burton again, one of three-or-so individuals who's currently getting free reign to make stopmo films when he wants.] It seems to me that Disney has seen their blunder... And rather than letting Laika get the upper hand, it's going to counter with its own stopmo product.

Laika and Disney both putting out stopmo products? Remember, there's only been a handful of feature-length stopmo films made in the USA -- EVER. This is an unprecedented scenario... Which, optimistically, could lead to a new boom for stopmo.


I think the year that "Wallace and Gromit" and "Corpse Bride" were both up for Oscars represents a turning point. "Coraline" will build upon that momentum... And if we're very lucky, we might be looking forward to a decade or more of a stopmo film coming out ever year or two.

This would be an excellent thing because -- (and this is a surprise ending to the essay that I didn't see coming) -- what I think we desperately need is more big name directors who have an affinity for the stopmo artform... Because right now, all we've got is Tim Burton, Henry Selick, Nick Park, and Wes Anderson.

If the big studios commit to producing stopmo product, then some new blood might have a chance to rise to the top ranks... And once the directors have had a chance to taste the process of making stopmo films, how can they not wind up pushing to do even more of them?

posted by sven | January 12, 2008 12:01 AM | categories: stopmo, writing