Elephants Dream Creator Talks to Wikinews

An anonymous reader writes "Three days after the Internet release of the free content 3D short Elephants Dream, Wikinews exchanged e-mails with Ton Roosendaal about the reaction to the film, open source filmmaking, and the changes to Blender that resulted from the production."

Comments

my opinion

[Toba82]

The movie sounded like a feature film - it wasn't. It was a very thought-provoking piece, however. I've got to say the graphics were impressive.

Download it here

[LetterRip]

For those who didn't see it the first link here are the movies again, hope you enjoy it

http://orange.blender.org/

LetterRip

I'll have to watch the anim now!

[Ithika]

Nice interview, although he sidestepped a couple of questions... like the one about closed source sound software. He just seemed to go off at a tangent there.

A review:

[ledow]

Well... yes, the graphics were quite impressive, however the animation looks very clunky at times. Although the static and slow-moving graphics looked fine, the walking motion and some of the fast action looked very bad (I actually checked to see if my player was skipping frames).

The audio wasn't fantastic - a little jingle of music, a few sound effects and Emo has a very strange accent (and, BTW, what is the Colossus of Row-Des, I thought it was Rhodes, as in "roads"... maybe that's just me being on the right side of the pond). There's little emotion or character in his voice, either.

The "plot" is just plain weird but we'll excuse that on the basis that there isn't supposed to be any plot (read into the plot what you like but it's not present so you can say that anything "represents" anything you like... I hereby declare that the plot could be about Emo the technophobe not wanting to use the clunky old tech that his father used, in the same way I use CD's where my dad used vinyl).

By making the plot weird and the animation clunky, they've actually achieved the opposite of what they wanted - they relied on DVD pre-orders and grants to get this off the ground and, now people have seen the result, they won't be getting many of those for their future projects. Plus, when people next say "we want to use Blender to make X", everyone's going to remember this.

I can't see this being something that people will share around to go "wow" at with their friends (unlike that short about the little robot who wakes up in a room on a spaceship (Blue?), anyone remember how much that cost to make?) so very few people are going to realise this even exists. If they do, they are going to be one of the people here just disappointed with what's been produced after they've spent a lot of money on a DVD pre-order.

The arty-farty types will adore this film if for no other reason than nobody else can understand it and it's been called art.

Re:A review:

[Eloquence]

There's more to CG than character animation, and as Ton explains in the inteview, the artists got better during the course of the project. I think judging by the amount of blog buzz the thing has received, it is being shared and copied quite heavily. I see many potential benefits:

• Be taken seriously by studios and animators. Having a tech demo like this out means that people who make decisions about spending money are more likely to take a closer look at Blender as a highly capable free solution for 3D graphics.
  • As a consequence, a studio might even say: "Sure, Blender is cool, but it's lacking features X, Y, and Z. So we'll pay for developing those, it's still cheaper than what we would pay for licensing." Studios are first and foremost about making films, not software, so open source makes strategic sense for them.
• Get young artists interested in using Blender. "That movie had a weird plot, but man, I'd love to do graphics like that." This in turn may lead to increased uptake in academia, as kids want to use their favorite software in university.
• Help people learn basic 3D filmmaking skills. Remember that the DVD contains the 3D models, storyboards, making of, etc.
• Establish working relations with artists, organizations, and so on that can be built upon in future projects.
• Identify key areas where Blender needs work -- this was done during the process, and any new movie project will help to further refine the software.

I don't dare to predict if future movie projects will be successful. I think there's a good chance they will, especially if the basic idea (without spoilers) is published upfront and well-received. I think it would be neat to cooperate with a major webcomics artists on characters and plot. This is a community artform that has already established itself quite well.

Great questions

[nEoN nOoDlE]

I thought the questions asked were really great compared to most animation interviews I've read which usually just appear fanboyish. I'm glad the interviewer asked about the story and character animation which I felt were the films biggest weaknesses. It's too bad that Ton decided to side-step the issue and not admit flaws.

"Yeah, the challenge the artists set themselves - to use quite realistic personages - is also something that easily works against you"

yeah yeah, we all know about the "uncanny valley" (and if you don't, there's a link :]) but that wasn't the problem that Elephant's Dream had. The animation was just bad. It was obvious that most of the people working on it were better at modeling, texturing, and lighting than animation. This is something that's fairly common in CG animation. It's usually broken down into "character animation" and "everything else." Where you'll find lots of great generalists who know about modeling, texturing, lighting, rendering, particles, etc and then you have the animators who don't do the technical stuff as well but can bring the characters to life.

Re:Great questions

[nEoN nOoDlE]

Any animator worth his salt studies real actors, and is a "real actor" in their own right. Acting is a huge part of learning animation, and books have been written on the topic. That said, acting is also a later step in the learning process. Before that, animation teaching is usually started with the 12 principles of animation which are:
    1. Squash and stretch
    2. Anticipation
    3. Staging
    4. Straight Ahead Action and Pose to Pose
    5. Follow Through and Overlapping Action
    6. Slow In and Slow Out
    7. Arcs
    8. Secondary Action
    9. Timing
    10. Exaggeration
    11. Solid Drawing
    12. Appeal

These principles were defined by the original Disney animators in the 40s and 50s, and are widely used today as the base of animation learning. You can check out a great resource about the animation process, written by 2 of Disney's Nine Old Men, here.

Ton in the interview said:
...so you accept a certain level of non-realism easily. (Check the weird walk cycles in The Incredibles for example).

This kinda just goes to show that he's just spouting off a "factoid" he's read or heard about. The "weird" walk cycles in The Incredibles - while maybe not "photorealistic" - are done that way for a purpose, following the principles listed above to make the character and animation more appealing. Animation is often about getting an appealing looking movement than a "realistic" looking movement. It's just that oftentimes, if you make a movement too unrealistic, people who are used to seeing such a movement will notice that there's something wrong with the animation.

Also, whether the character is "realistic" is irrelevent. The 12 principles still apply, but are just toned down. These things weren't pulled out of a hat. They were observed in human movement and exaggerated to make those movements more clear. Nobody could deny that the original Disney animators really observed what they were trying to animate. In fact, during the making of Bambi, Disney brought in a deer carcass so that animators could study the skeleton and muscle systems in a real deer. One of my professors in school who used to work for Disney showed the class the resulting book they made out of that research.

They started anim and voice to late

[Qbertino]

I was at the premiere in Amsterdam and had a chat with some of the creators at one time or the other. Allthough we all grieve a little over the jerky anims one should keep in mind the following:
1) The timeschedule for a project like this was extremly tight. Remember they didn't have *anything* when they started. Not even a basic plot!
2) They had less experience in film project management than a guy that doesn't do blender all day but watches 'making of's'. Bassam (the director) said that he learned a storyboard and animatics are really important but it's important to move on fast from there on. I could've told him that right away. Then again I don't know a tenth of what Bassam knows about Blender.
3) They got stuck in the middle and took the time for experimenting and redoing animatics, which they, sadly, didn't have. Final anim started to late. Even the extra month they added in the end wasn't enough to give them room to breathe.
4) They didn't prerecord. Very big mistake. One guy said "There was so much emotion in the voices we had to redo some anims". Would've you thunk? Actors can act. It's what they do for a living. *ALLWAYS* prerecord unless you've got an acting director who has the skill to railroad the actors into the anim stance. And even then it's still better to prerecord.
5) Blender was extended with features they needed while they where requesting them *without* having a reference to other packages. All these guys are the elite when it comes to blender. IIRC none of them has any notable experience with any other package. Matt likes to toy around with ZBrush but Andy, for instance, is a 100% Blender guy. Watching him Blendering gives you a good reason why. When he's doing a little doodling in a break at the blender conference there's allways a bunch of people crowded around his workstation looking over his shoulder with amazement. It's absolutely fascinating just to watch this guy work. Then again, whith a feature list beforehand the parallel development of Blender would've gone quicker and features would've even been there before they where requested.
6) The jerky anims are paid of with awesome details that you usually don't notice at first viewing. In fact, one could say that the '2nd unit shots' are the actuall piece of art in this. That fits the lack of experience the Orange team had with larger productions. Bassams mechanical characters just plain rock. That's a fact.
7) AFAIK they where rendering in production which took away some time. Usually you outsource that or another dept. does it. I don't think they used renderplanet, which, if not, they should have.
8) All OSS Tools. Thats the single largest obstacle. The OSS tools are impressive, but OSS NLE and Compositing is just plain no match at all for, let's say, Apple Shake or Digital Fusion.
9) The benefits of compositing only became aware at the beginning of the project and key personell didn't have enough playing time to try things out, imho.

All in all I have to say that I am extremely impressed with the results. As for the semi-finished anims: As it is entirely open, there is no one at all stopping us from reanimating the entire move. The strange background of the story offers countless oportunities to extend the original and the fact that the riggs will be published gives pure animators a chance to show off their skills. Everyone can say: If you don't like it, redo it. A true OSS project indeed. Once again the Blender Community has shown true spirit. Ton and Team Orange rul3Z0Rz!.