Sunday, August 26, 2007

Want to Make an Open Movie?

One recurring theme on the Blenderartists forum is the quest to have a group of online volunteers produce an "open" movie. Brimming with enthusiasm after getting their hands on the means of production, many newcomers to the CG scene decide that the only thing standing between them and an Oscar-winning short, or even a feature movie, is time and ability. The next assumption is that ability will come quickly and that the time problem can be overcome with the help of a bunch of other modellers and animators who are presumably watching the forums waiting for just such an opportunity.

In my time on the forum, I've seen countless such proposals (and probably missed quite a lot too) and there is a general consistency to them. In most cases there is no script - and often no concept at all. The typical pitch goes something like "I think we should make an open movie. It will be brilliant and set the world on fire. Everyone can contribute and we can all vote on everything to keep it really open. All models used in the movie will be made available to everyone. I don't have a script so we can all work that out too. I haven't animated anything before but I'm a fast learner and I know that if we all join together, we can do it."

Then, the starry-eyed Spielberg wannabe awaits the flood of over-excited members literally throwing their services at him (or her) all wishing they'd thought of this idea themselves. It's at this point that reality usually hits home, like a bus hits an errant pedestrian. Invariably, the innocent request for participation is met with a barrage of advice to "put up or shut up" and the discussion descends into a long lesson about how movies don't just happen and how it would be a complete waste for anyone to spend one minute of their valuable time on this project - if it can even be called a project at this point. Other respondents offer finished models for a project that doesn't even have a solid script yet. In some cases, the project may have enough legs to last a year or two but ultimately fails without one second of finished animation being produced and sometimes leaving the most committed contributors, and the originator, almost distraught at the thought their effort was all for nothing. It can be a nerve-wracking experience.

REALITY (and hindsight) suggests that there is little chance of completing a successful movie project where anyone and everyone is encouraged to contribute to and vote on everything from the plot to the final render. It would be like expecting all the world's religions to develop a single spiritual model.


So, do open movie projects stand any chance of success and if not, are they of no value whatsoever? As a past-participant in one project, let me say that I learned a lot about modelling and animating and that the motivation of contributing meant that learning was fast-tracked. I wouldn't trade that experience for anything and am very grateful for the opportunity to participate. But I also learned something about how these projects work - and why they ultimately fail (or, at best, drag on with no end in sight).

Reality (and hindsight) suggests that there is little chance of completing a successful movie project where anyone and everyone is encouraged to contribute to and vote on everything from the plot to the final render. It would be like expecting all the world's religions to develop a single spiritual model. It's a feel-good philosophy and it just isn't going to happen, no matter how well-meaning the dozens of contributors might be. Even very short animated movie projects struggle as the decision-making process becomes increasingly overwhelmed from the wide variety of ideas and opinions thrown at them. Give people more than two choices to vote on and, like political elections, you'll be lucky to get a genuine majority decision. What you will get is a lot of discussion and some disappointed members as the largest minority vote presumably wins .

However, what really seems to kill the projects is not just the naive attitude of open-ness but a complete disregard for standard movie-making processes. Models start to appear before a script is even partially written and animation begins before there's a story-board. In fact, huge, complex, poly-heavy sets can be assembled before there's any visual indication of what sets are actually required to tell the story - if there ever is a finished story - or any idea how the final product is to be rendered.

Numerous contributors will spend countless evenings making model after model without so much as a style guide or even a solid commitment to an era, so there's little chance of cohesion unless the director is willing to ultimately exclude some of these contributions and risk upsetting someone who believed the project was open to all styles. And that's if there is a director!!! Surprisingly, this crucial job seems all-too-absent in some of these, almost anarcho-syndicalist, projects.

SOFTWARE is modular and users experience only those parts they need to at any given time but movies are stories and stories are linear and the user must experience every part from start to end if they are to make any sense of it.


Without fail, open-movie proponents point to successful open-source software projects as an example of the viability of the open approach but they miss a fundamental difference between software and story-telling. Software is modular and users experience only those parts they need to at any given time but movies are stories and stories are linear and the user must experience every part from start to end if they are to make any sense of it.

Take Blender 3D animation software as an example. There's no doubt about the increasing success of this product and the collaboration that drives its regular improvement. But Blender isn't a movie, it's a tool. I have used Blender for three years and there are substantial parts of it, like the game engine and a variety of physics-based functions, that I've never even looked at. People have spent hours, weeks or maybe even years developing these aspects of the application and I've no doubt many other people are glad they did - but I can happily use Blender completely oblivious to those parts that don't interest me for now. The same can't be said of a movie. If someone wants to insert a character, building, tree, joke or an entire scene in a movie then when I watch, I have to endure those parts no matter how irrelevant or distracting they might be to the storyline. Imagine a movie where every contributor's "good idea" is included as a result of pursuing a genuinely all-inclusive ideology. The end result would be disastrously fractured (and overly-long) and far from embracing everyone, it would ultimately alienate them.

Online collaborations also fall victim to the tyranny of distance. These projects move very slowly since some of the crew are always asleep or at work or school while others are awaiting or making important decisions. Even simple questions can wait days for answers as key people miss each other on the forums or chat lines - or get sick or go on holidays without informing anyone. Remember, no one's getting paid to do this - it's all voluntary and no one can be forced to make themselves available (after all, you can't fire them for failing to comply).

AND then, when everything is finally done and all the team members look back over the years they've spent contributing all their time for absolutely no financial reward, they give the whole lot away to anyone who wants it. Utopia! Nice idea but it's just not going to happen.

Now, a movie short has some hope for completion but for a feature-length movie collaboration, expect to spend 6-12 months getting a "finished" script together. Spend 6-12 months making concept art. Spend 6-12 months story-boarding the script. Spend 6-12 months getting models made to match the story-board and concept art. Spend 6-12 months getting a rough voice track down. Spend 12-24 months making a 2D story-board animatic to match the voice track. Spend 6-12 months getting voice actors to lay down the finished voice track. Spend 12-24 months animating (this might be a conservative estimate so allow 48 months). Spend 12 months polishing the animation with lighting and effects. Spend 6-12 months writing and recording a musical score and recording an effects track. Spend 6-12 months compositing.

That's at least seven years and while there will be a little overlap in some of these tasks I suspect many of my time-frames could be considerably underestimated if approval of all stages is to be collaborative too. Remember, unlike normal feature films where people are employed to produce the movie, no one is working on your project full-time so hours become days and days become weeks. Tell your prospective contributors at the start that you'll expect them to contribute non-stop for at least seven years and see how many useful volunteers you get. Let's face it, given the difficulties encountered by numerous defunct short-movie collaborations, a feature-length online collaboration is extremely unlikely to even come close to fruition without some serious expertise and management on-board.


Another area where these collaborative projects fail is that many contributors are guided by software features rather than the story. So a typical project starts out with incredibly detailed, high-poly scenery, ray-tracing, fluid simulations, soft-body and cloth simulations, explosions, fire and other intensive halo-based effects, reflections, refractions, ambient occlusion, sub-surface scattering, fur, hair, physics engines and so on and so on. The list will grow over the years as more features are added to the software. With this in mind, the project team needs to allow a few years just to render the finished frames.

And then, when everything is finally done and all the team members look back over the years they've spent contributing all their time for absolutely no financial reward, they give the whole lot away to anyone who wants it. Utopia! Nice idea but it's just not going to happen.


Having said all that, I think there is a possibility one of these projects will see results. Since it is said that every good blog should have a list (and who am I to ignore such wisdom?), here's my list of things to consider in order to stand some chance of open-movie success. This assumes the typical situation in which the project leader and contributors DO NOT have solid, professional experience in making movies:

  1. STORY IS EVERYTHING: Movie-making is story-telling. Without a story, the movie is pointless unless it is being made to serve some other purpose such as developing or advertising software capabilities. And don't expect amateur modellers and animators to necessarily be good script writers. If you need scripting advice, consider seeking guidance from writers.

  2. THINK SMALL! Animation takes a long time. Depending on complexity (see "KISS" below) and experience, a one or two minute short will be enough to test most online collaborations. Keith Lango wrote in his Fool's Errand series about the problems faced by the independent animator and Keith's list of concerns is equally valid for open-movie collaborations. If you can take a one-minute short to completion using a collaborative model you will have succeeded where many before you have failed and you'll be far better placed to move onto something bigger.

  3. HAVE A STYLE GUIDE: Once you have a rough script and some firm idea of where you're heading, you can start on some concept art so prospective team members will know if this project is their "cup of tea" (some people like fantasy, others like hyper-reality while others like toons). Better to start with the right people than to try and retro-fit a team as the project progresses. Ultimately, style flows through to more than just the models and sets - it involves decisions on lighting, using fluids, fur, cloth and other features that need to fit together to deliver a cohesive finished product.

  4. BE A DIRECTOR: (or appoint one). While the opinions of employees at film companies might be valued, the director makes the final decisions and this must be the case with an open movie project too, if it is to succeed. The director has authority to decide what does and does not make it into the script, the sets and the movie. He or she should probably take advice from a select group of project members (see below) but it just isn't realistic to make every decision a democratic one for the whole group.

    Don't underestimate the importance or workload of a director. Keep the team informed with regular updates , be prepared to say "no" and don't go on holidays just as things get difficult. If you don't think you can wield supreme executive power or stick with a project over the long-term, then maybe you should contribute to someone else's project instead of trying to run your own.

  5. APPOINT TEAM LEADERS to handle some specific areas. Team leaders should have demonstrated experience in their area but they still work for the director. Team leaders are the first level of control in ensuring contributors are following the plan as laid out in the script, storyboards, art-direction and animatics.

  6. KISS: (Keep it simple, stupid) Again I'll refer to to Keith Lango's advice regarding the urge to mimic Pixar in your own animations. Unless you have a big budget or free access to a render farm, you might want to forget RAM hogging features like ray-tracing and ambient occlusion. At the very least, be clever enough to limit the use of these features to a few scenes where they're necessary. You may even be able to tell your story without complex scenery. Look at everything in your arsenal to limit your modelling, animation and rendering times. It's much better to have a simple movie that looks good than a complex movie where corners had to be cut just to get it finished. It's really a question of balance. If you keep things simple then a 30-minute short wouldn't be out of the question for a well-managed project but if you want or need visual complexity then trade this off against the length of the movie.

  7. FOLLOW THE RULES: DVD movie releases from major animation companies like Pixar, Blue Sky and Dreamworks are awash with examples of the movie-making process yet most wannabe movie-makers seem all-too ignorant of the basics. Either that or they simply don't believe they need to consider them. You can't expect to make a movie efficiently if you start modelling before the script is written and the concept art and story-boards completed - it is the first sign of a pending train wreck.

    If you're going to do it, then why not do it right? Have a story, have a design concept, have story boards and an animatic, with a rough audio track. Then, when you know what you actually need for the movie, call for models, record a finished soundtrack THEN move on to animation and effects. The director should ensure no one is doing things before they are required. Even if someone volunteers to make a model, politely ask them to wait until requirements and specifications are published.

    You'll also want to understand accepted techniques (you can ignore them by choice, but you still should know them). If you don't know a pan from a dolly, learn. If you don't know a cut from a fade, learn. If you don't know what "the fourth wall" and "crossing the line" are, learn.

  8. MANAGE YOUR ASSETS: The story-boards and animatic should give enough information to allow the director to determine exactly what assets are required for each scene. If a building is to be seen in the very far distance, there's little point using a highly-detailed model containing thousands of polygons. Either use a low-poly model with an image texture indicating architectural details or better still, make the building part of a matte painting (you can render a single frame of a complex model and use this as a matte painting).

    If a shot involves an exterior view of a building, don't use a model with a fully-dressed interior that won't be seen. Every polygon affects render time. Take a tip from live-action movie sets where many buildings are little more than facades. Also, background shadows can be baked to save on render time and scenes can be split into separate render layers or even rendered separately and composited in post-production to save time.

    Another thing to consider is where to host all this activity and how you'll get it rendered. You'll need good forum and database facilities with fast and easy access. As the project grows, members will be handling huge amounts of data as they upload and download assets. You will definitely want to take the project out of the very public eye of a general forum community if only to maintain some level of sanity. Regular reports can still be made to other forums in order to maintain public interest and possibly attract new members but for the most part you'll hopefully be too engrossed in production to be dealing with constant questions from passers-by. And make sure the host site is secure and stable (idiot hackers apparently get their jollies by demolishing anything vaguely positive) and that you won't lose everything at the whim of the server's owners (this happened more than once on the project I contributed to). If the site requires subscription, make sure to keep it current. It's embarrassing to have project members see some weird generic splash screen where their favourite open movie site used to be. For rendering a high-poly movie, you'll need a decent render farm that supports your file-type.

  9. KEEP IT LEGAL: If a project stands a good chance of success it would be wise to ensure all members are truly committed to the open-source ideology. It would be very embarrassing, and possibly costly, if a contributor decides that their contribution had strings attached and demands compensation for their effort. Spell out the nature of contributions, using one of the many Creative Commons licence-types or GPL, if appropriate, then get the release forms signed and returned by everyone.

    Keep tabs of who worked on what and if you don't have a release (or some form of consent) for something, don't use that asset. You could make it clear that every model submitted must be accompanied by a text file releasing that model under the chosen licence-type or it won't be used. (NOTE: I'm not qualified to offer legal advice)

  10. HAVE FUN: That's why you're doing it in the first place. Never lose sight of that. Maybe this should have been point number one. Oh well, if you want it that way, consider this a count-down list and your wish is granted ;)


To anyone looking to contribute to one of these projects I suggest that you should do so if the project itself interests you. If the style and story interest you and there are areas where you feel you can assist - and if that assistance is wanted - then go for it but go in with your eyes open. If the project fails some of the above tests then it may ultimately fall victim to the ever-present open-movie Grim Reaper. But as I said at the start of this article, the commitment of contributing can help to fast-track your learning process. Just don't set yourself up for heartbreak by assuming that all other members are as committed to the project as you are in your early days of contributing. Be prepared for the likelihood that your contribution will ultimately not see the light of day in a finished movie and if this happens be happy to write the whole exercise off as a valuable learning experience.


The term "open movie" (and open-source or open-content movie) is relatively new and has come to have a variety of meanings. It can refer to a movie produced using only (or almost entirely) open-source software, but the film itself may have restricted rights applied to it. Open movie can also mean a movie produced by a public collaboration of people who choose to contribute to it - ie, it is "open" to anyone to participate - but it may involve proprietary software in parts of the production. In possibly its broadest definition, and this is largely the situation I'm discussing in this article, an open movie is one which is open to public participation AND which uses only open-source software AND which is released to the public domain without restriction AND where all the source files used to produce the movie are ultimately released for free to anyone who wishes to use them for any other purpose.

Monday, August 20, 2007

A good helping of Ratatouille

Well, I have now seen Ratatouille, two weeks before it officially opens here in Oz. Pixar have done it again. The film is a visual feast and contains all the elements of hope, drama, despair and comedy that typify Pixar story-telling. In an audience that was made up of around 80% children - many of them very young - it was good to be able to watch the movie undistracted by bored kids running around - as happened with Cars. This was a little surprising as Ratatouille doesn't seem to be aimed at young kids, it really is a film that many adults will appreciate. But the young-uns clearly connected better with the furry critters in Brad Bird's ratty tale than they did with the slick, shiny machines in John Lassetter's Route 66 adventure (which I quite enjoy, by the way). And the adults had a ball, laughing loudly at all the right moments.

I won't critique the film as I'm a hopeless watcher. Neither movies nor novels ever hold my uninterrupted attention and with modern animations I'm easily distracted by the CG feast. As an oil painter too, I was easily distracted in Ratatouille by the wonderful, glowing light that permeates every scene. But I knew from all the preview material that this would be the case. One area where I was disappointed was the in-your-face moral lesson that underpinned the story. In fact, overall I feel the moralising overwhelmed the humour and I really don't need life-lessons from a cartoon.

One thing that did amuse me was the much-feared food critic Anton Ego. His funereal appearance reminds me of ex-Australian Prime Minister, Paul Keating - or at least of the caricatures that entertained newspaper readers during the time of his political career.

Potential spoiler here... > I couldn't help but wonder if Ego's ultimate summary of the value of food critics might also have been intended as a bit of a swipe at the many film critics who had this rodent-filled film dead and buried before the last frames were even rendered?

Now I'm looking forward to the DVD release. Hopefully Disney-Pixar will come good with the extras for this one and not short-change us as they did with Cars, which came with very few "accessories". I think most, Pixar fans will expect to be given a good helping of side servings with their Ratatouille. If this DVD is as skimpy as the Cars DVD was then I'll be waiting until it hits the discount shelves - the extras have become a large part of what makes Pixar DVDs so special and are one big reason why I own so many Pixar titles on DVD despite already having perfectly good VHS copies. The apparent decision to release a fully-accessorised version of Cars, 12 months after the initial DVD release, leaves me feeling a little cynical that someone's trying to double-dip by enticing fans to "upgrade" to the new, improved release.

Thanks again to the organisers of the Telethon previews.

Wednesday, August 8, 2007

A Taste of Ratatouille

Movie-goers in Western Australia may have an opportunity to get an early look at Pixar's latest release "Ratatouille". Well, early compared to the rest of Australia but still late compared to many countries. Anyway, that little grumble aside, the really good news is that Telethon is holding preview screenings of Ratatouille at selected cinemas on 19th August, a good two to three weeks earlier than the official opening date of somewhere from 30th of August onwards.

If you haven't done so already, find your nearest screening and hope you can still get a ticket. At $8.50 each it's a win-win-win situation. Early screening, cheap tickets and you'll be supporting Telethon!!!

Big thanks to everyone who has anything to do with making this happen. I'll be there!