UPDATE FEB 8 2012: I guess this piece got high up in some common Google query, because I’ve had a lot of vicious comments on it over the years. So let me explain: This article is from 2009. I was nineteen when I wrote it. I also wrote this follow-up piece clarifying what I meant to say, which apparently isn’t as popular on the Googletubes. Anyway, this is a terrible post. I was trying to encourage people to innovate with low-cost animation, but the following is a completely ridiculous way of saying that. Please stop getting angry at me. UPDATE FEB 9 2012: Okay that didn’t work. You know what? Comments are closed. I’m getting tired of seeing this pop up in my inbox all the time.
Perhaps I’m uneducated in the ancient Hollywood art of Unit Production Management, but it’s baffling to me that WALL-E had a budget of $180 million. Yes, it was a gigantic Hollywood production, but consider the fact that all of its visuals were made by pressing buttons and waiting for the images to appear. Well, it was more complicated than that sounds, but that’s basically what they did.
So what exactly cost so much money? Without access to Pixar’s financial records, I’ll take a few guesses. But the short answer is that they’re spending way more money than they need to.
I’d imagine, from my experience, that some of the the biggest expenses in filmmaking are film stock, set construction, stunt coordination, and other tangible materials. Pixar requires none of that. The only materials they really need to buy are computer parts, which are relatively cheap on their own and downright bargains when bought in bulk. Yes, their concept artists need clay and paint and such, but that hardly seems like it would cost more than a couple thousand, even for a gigantic team of the kind of overzealous madmen who spend more than a week in an art school (i.e. me).
So that brings us to visual effects, which is essentially 100% of what you see on the screen. Yes, WALL-E and Happy Feet had live action bits superimposed, but anyone with access to a green screen and a DVX can imitate those. The actual CGI, though, is achieved by making a bunch of 3D models, texturing them, arranging them in a world, moving them around, and then pointing a virtual “camera” at various parts of the world to get the images you need.
It may sound simple (save for the whole “talent” part in terms of modeling, texturing, and animating), but to create the impressive visuals, Pixar needed a very large amount of polygons. Add to that the calculations necessary for lighting, post-processing (focus, motion blur, etc.), and spitting that out into 24 images per second at 4000 pixels wide, and you have something that requires a lot of very expensive computers to eat a lot of expensive electricity for a very long time.
Except it actually is as simple as I made it sound in the paragraph before that. Take a look at this video:
Impressive visuals. If you had never heard of Team Fortress 2, and that video cut off after the first 30 seconds, wouldn’t you imagine that perhaps this was an upcoming Pixar film (albeit more violent than their previous efforts)?
Now, obviously, this is actually a promotional video for a game. Let’s take a look at the game’s graphics…
Obviously not quite as impressive, but you’d be hard pressed to say that the Scout model in the game isn’t the same one in the Pixar-quality promo video. In fact, the line between the two is very blurry — specifically the fact that the promo video has more blur effects based on motion and focus. Well, that’s because it’s absolutely true.
Meet the Scout, rendered in Team Fortress 2, contains a significantly smaller amount of polygons on screen at a time than a Pixar movie like, say, The Incredibles. Can you tell?
On close inspection, you can see that the Team Fortress 2 models are slightly more jagged than Mr. Incredible. However, it’s unlikely that you’d notice this unless you were looking for it in a freeze frame. In addition, Team Fortress 2′s models are actually far less complicated than those from many other video games. Take, for example, BioWare’s Mass Effect:
There, slightly more complex models and the jaggedness is solved, but still significantly less intense than what Pixar does. And most blockbuster video games cost things that are considered “indie budgets” to Hollywood: Gears of War cost $10 million, $5 million less than the pretty much indie-produced film Slumdog Millionaire. And as you can see in that article, many games cost “crazy figures” like $30 million, which is still pocket change to Hollywood. Off-season movies like Taken cost somewhere around $25 million, for god’s sake. Now, if you were to take all the work on Gears of War related to getting the thing to run smoothly on an Xbox 360, and channel all that into putting the camera in the right places, pressing “render,” and letting three laptops run for 10 hours to give you a 2 hour movie, I’d imagine it would cost even less.
So now we’ve thoroughly debunked the myth that CGI is inherently expensive. What’s left? A-list voice talent? Yeah, I can see that being up there. And marketing costs a lot. Indie filmmakers would probably replace those two strategies “five of my friends” and “Twitter,” respectively, but I understand Hollywood’s obsession with such luxuries. That can’t account for all $180 million, though, can it?
Well, in addition to Pixar’s overkill in terms of hardware and technical complexity, they employ a gigantic amount of people, paying them salaries and such. That’s probably necessary because of the technical complexity of their animation.
In theory, a team of four passionate geeks who already have MacBook Pros and copies of Maya, Photoshop, and Final Cut could probably make a Pixar-quality film in their spare time. Perhaps it’s the massive, daunting complexity of Pixar films that discourage budding independent filmmakers from trying to compete using more efficient methods.
Or maybe they’re all developing video games instead.
Comments on this post are closed because nobody read the gigantic bold text at the top, and I’m tired of having to deal with this all the time. If you really must insist on explaining to me just how much of an idiot I am, you can scroll all the way up to the top of the page and click on the Contact tab to find out how to send me hate mail.
Apollo
Im sure its the actors they get to come in cost a boat load to take sound check after sound check. The computers cost more to maintain and aren’t your average desktop. The advertising, and PIXAR likes to send their animators to the most exotic locations to take Hi-Res pictures and get a better hands on grasp for what they are about to animate.
Who knows, maybe they spend it on the most expensive weed and burn it, pumping in through the ventilation.
Apollo
Im sure its the actors they get to come in cost a boat load to take sound check after sound check. The computers cost more to maintain and aren’t your average desktop. The advertising, and PIXAR likes to send their animators to the most exotic locations to take Hi-Res pictures and get a better hands on grasp for what they are about to animate.
Who knows, maybe they spend it on the most expensive weed and burn it, pumping in through the ventilation.
Markus Koivisto
You completely disregard the price of labour, which is a fatal flaw in your judgment. $180 Million or even $30 million sounds like a lot of money, but you need to realize that the lion’s share of all of this goes to pay the salaries of all involved. Consider a team of animators and programmers who each make $5000 a month. A gross understatement, but we’ll let it stand. Well estimate a headcount of fifty people involved. How much does this talent cost? 3 Million in salary costs alone, each and every year.
Now, the team for WALL-E was even bigger, and commanded even higher salaries. A brief look at IMDB[1] gives a cast of 300 characters involved in the making of the movie during a period of 4 years. This gives me a back-of-the-envelope figure of 72 000 000 dollars right there, using the same underinflated salary of 5000/month i used above.
Using game engines for independent movie development is done. There’s a whole genre of these movies out there, known as Machinima. These movies are however limited. You can choose to use the existing characters and settings, but you are thereby limited to said characters or settings and the somewhat limited animation tools provided by the game. You can of course develop new characters and settings, but that takes even more time and effort.
How much time and effort? Recenly, Clear Skies was released as a particularly impressive bit of machinima. It was a macbook and four friends -type of project, that took two years of part-time work to develop for forty minutes of work, and this was with existing character animation simply repurposed for the task. The creator of this piece of work told in an interview that he’d be able to cut down the production time to 8 months in subsequent episodes, but that still leaves us with a daunting 5 minutes/month schedule for a dedicated amateur.[2]
Animation is, simply put, hard and above all time-consuming work. A professional animator can churn out about 30-120 seconds per day, depending on the complexity of the task. There is no easy way around this. Even for a video game using special lip-synch software, the facial animation usually takes on the order of 90-150 seconds per day.[3]
The same math applies to making movies in general. The set pieces etc. cost nowhere near as much as the labour and the talent needed to get things done in short order. Post-production is one of the biggest time- and money sinks in the production chain, and here again it’s due to the vide variety of companies and people involved in the making of the movie that contribute to the budget.
The old adage “cheap, quick, high quality – pick any two” certainly applies here. Getting high quality in a reasonable timeframe takes a hell of a lot of talented effort which doesn’t come cheap.
Actually, I didn’t disregard the price of labor. I mentioned in the last couple of paragraphs that it was probably the reason why the budget was so enormous.
The issue with machinima is that it’s often quite literally done in the game. A video game isn’t exactly friendly to someone who wants complete control over their images. That Clear Skies film you linked to was done in EVE Online, which has a user interface that’s difficult to use even for playing the actual game. It’s not surprising that it was no easy task.
Valve’s Source engine, by contrast, is open to anyone who buys a Source game and downloads the SDK. New animations can be created for all of the in-game characters, using a combination of external 3D programs or Source’s own facial animation system. Or they can create their own models and drop them in. Mod-makers do this on a daily basis; a filmmaker using the same tools could create something rivaling Pixar, as Valve’s promotional videos demonstrate.
Most modern video game engines also have a lot of procedural animation functionality built in, so that can shave off quite a bit of work on all the nuances and variations that occur.
Any independent filmmaker trying to finance their own production is going to find creative ways to imitate Hollywood without spending nearly as much money. It’s something I’m working on right now, and hopefully by mid-April I’ll have a final cut to demonstrate the results. Why limit that to live-action films, though?
Markus Koivisto
As for Valves promotional footage – a 95-second bit doesn’t really compare with a feature film. It is impossible to recreate those 95 seconds of footage with the tools Valve has released. You need a lot of additional work. You can bet that behind that promotional spot was several days of custom animation. And that precisely is my point: you can’t bring that time down if you want to do something custom. 30 seconds of animation is still going to take all day, and you can’t bring that down by an order of magnitude no matter what tools you use. Consider the above for a moment. That spot took anything from one to five days to animate.Extrapolate from there to a feature film, and you can quickly see that there is no real time saved.
Clear Skies was an amalgamation of the Source engine with specific tools for posing and animation, combined with EVE online footage chromakeyed in, which you’d know if you’d looked into it. It really was a split approach, and the Source engine tools you specifically referenced were definitely used. The most time-consuming part was the facial animation, since even though the facial animation was provided by the engine, every phoneme still needed to be extracted and timed.
It becomes painfully obvious to me that you haven’t been involved in any animation work. People always underestimate the duration of tasks they don’t understand. Animation is time-consuming no matter how you slice it.
The reason why machinima can be faster is that you use premade animations to create your content. You are essentially combining canned goods to make your film. As long as you are relying on the work of the character animation of the game company, you are able to make great progress. If you want to do something other than the canned bits of animation, you are SOL and have to start the time-consuming process of animation again.
If you are not using the canned animation work provided by the character animators at the game company, you are not really saving any considerable amount of time. Red vs. Blue for instance, is rather simple to churn out since they did not have to do any custom animation. However, this does show in the end result, which is not quite Pixar-quality to say the least.
The examples you quote (Mass Effect, Valve promo spots etc.) are all the results of much custom animation, and not something that can be replicated easily. Procedural animation is a red herring – things like IK and skeleton animation is of course used in any reasonable animation tool already, and is not something unique to game development. The same tools are available to the animators at Dreamworks or Pixar or what have you.
I suggest you pop over to Machinima.com and take a good look at what kind of work is being produced with the four friends and a macbook -toolkit, and then think about the limitations of the medium for a while.
http://plankhead.com Zacqary Adam Green
You’re correct, I haven’t been involved in any major 3D animation work. I have, however, done fifteen seconds of fairly minimalist two-dimensional animation, and I’m going to estimate it took me an absolutely ludicrous 30 hours over the course of a couple days.
I’ve looked at a wide range of machinima, and I’m well aware of the limitations, and what it is. I apologize for only taking a cursory glance at Clear Skies, I simply assumed it was similar to the rest of the machinima I had seen.
However, I think you were misinterpreting exactly what I’ve been trying to say. I don’t feel that raw machinima, at this stage, can rival Pixar, and I’m not suggesting independent teams try to imitate Pixar to the letter with video game-quality art assets. I’m suggesting drastic innovation and experimentation.
You’re right, it is impossible to recreate the 95 seconds of Valve’s footage with the tools they’ve released. Source Filmmaker is not available to the public yet. Regardless, tools like it aren’t the be-all and end-all of quick, cheap, quality animation. From what they’ve said about it, it is used only for choreography, motion blur, and depth of field. The custom animations still have to be produced in another animation program.
That animation program doesn’t necessarily have to be the time-consuming keyframe programs like the ones used by Pixar. Procedural animation is used in video games, even ones in development by small teams like Wolfire. Obviously, as their more detailed example shows, this can’t account for the entirety of an animation; that’s the exact principle on which professional software like Endorphin works, supplementing keyframe animation with procedural.
But keyframes can be cut back even further with recent motion capture techniques that don’t necessarily have to be expensive. Benjamin Button used a special camera to overlay CGI in real time, but what if a similar technology were used for rotoscoping digital video? Combining a technique which was originally demonstrated by researchers in 1999 with face (or other body part) tracking, it would be possible to capture an actual actor’s performance from a single camera angle, then apply the same animation data to a compatible model.
Perhaps an integrated and easy suite of tools to do all this doesn’t yet exist. But programming things like that have been done by four friends and a MacBook, and they’ve often been released open source.
Markus Koivisto
There is at least one false assumption here that should be cleared up. Animation studios already use a wide variety of tools to speed up the animation process. Don’t for a moment kid yourself and think that mocap and procedural animation isn’t already used in feature-length animations. It is, and extensively, whenever it makes sense to do so. In fact, it is used first in those instances before it ever becomes viable to do with games.
Mocap doesn’t make things any less time-consuming either. Instead of animating, you end up shooting real footage with real actors, motion capturing it, and cleaning up the result. You are just trading one set of problems for another.
What I’m trying to say is that there is all kinds of research that is being done on how to animate quickly. This research is being driven by the needs of the VFX and CGI industry, and they are of course the first to benefit from new developments in the field.
For an indie filmmaker getting into the CGI field, the learning curve is steep. The tools aren’t there yet, as you say, and I would go one step further and say that they will probably never be, since the bar is constantly pushed higher.
Questioning why Pixar chooses to spend the amount of money doesn’t really lead anywhere. The simple answer is that a feature-length production with the level of quality they aim for in any reasonable timeframe will always require a large team, and large teams cost a lot of money. Of course, Pixar is pushing the envelope, so they are doing a lot of R&D. If they wouldn’t stray from the beaten path, they would be able to create more of the same far cheaper, but that’s not what they want to do. They have the resources, and they are turning a good profit, so I would say they could in fact throw MORE money at more challenging problems, not less.
Where does this leave the small team? As with live-action, small team should concentrate on their unique strengths and the power of their vision to turn their limited resources to their advantage. It doesn’t really matter which approach you choose, the important thing is to know it so that you can concentrate on the story instead of the tools.
If the story is compelling, and the basics of good filmmaking are there, the audience will be sucked in and won’t care what you used to create your movie. Learning the latest and greatest technique will leave you chasing a moving goalpost instead of getting up to speed and actually producing your movie.
http://plankhead.com Zacqary Adam Green
I understand where you’re coming from, but you make the false assumption that I’m suggesting anything about “learning” the latest and greatest technique.
One factor which I don’t think I’ve explicitly mentioned is the independent filmmaker’s passionate disbelief in the concept of “impossible.” If what’s already out there is too expensive, too laborious, and too massive, then an indie team with no budget will build their own damn technique from the ground up. It’s been happening since Georges Melies reverse-engineered the Lumiere camera, and it will continue to happen, regardless of medium.
I can’t predict exactly how a small team is going to eventually do something on par with Pixar. I’m merely suggesting what that inevitable technique could be like.
Markus Koivisto
I suppose we can agree on that whatever the technique ends up being, it will involve a metric assload of work :)
http://plankhead.com Zacqary Adam Green
Initially, yes, I’d imagine. I’m a bit overwhelmed trying to figure out how to shoot a movie under blacklight, because the only tips and tricks I can find are for still photography. Probably only a milli-assload compared to figuring out how to imitate Toy Story for under $10,000, though.
http://www.herrschmidt.tv Martin Schmidt
Creative ways to imitate? How is imitation creative? The biggest fault you can make as an independent filmmaker is trying to imitate Hollywood! Why would you do that? Use your creative ways to create your own style, not imitate the faceless Hollywood style. Put your greatest effort into reducing your design, step away from realistic humanoids, how boring are they anyways?! Hell, you can make a movie containing only cubes as characters ( http://www.youtube.com/watch?v=hddYiSixKGQ ). Learn some good storytelling, thats a filmmakers tool.
And by the way, no animator can ouput 90 – 120 seconds per day of full animation! At least no quality animation. Limited animation like Southpark can output this kind of figure, but not the Pixar kind of thing. I don’t think you really need the Pixar animation style to create great films, in fact it might even be hinderly. Just to give you an idea: a Disney or Pixar animator outputs between 3 and 10 seconds of animation per day! More rough animation, like the stuff done for series, takes a day to finish 10 to 20 seconds!
Gosh, it took me 5 years and a couple of short films to figure out that I have to reduce my style even more to make a living as an independent filmmaker.
Kac_gy
I totally agree. I’m sure Hollywood inflates things a bit and then inflates their reports on the real inflated figures as a marketing ploy; “if the film cost this much to make then it must be quality”. We have that same cost/quality association when we buy other goods as well.
As someone who works in the arts I think it’s a bit callous the way that things such as scripts/screenplays, directing talent, designing the look of the movie, etc are all pretty much disregarded.
Pedro Ngalaxi21
Markus am I an Adobe AS3 certified programmer and i’ve made a few stone aged 3d animation my self. But computing evolved it used to take 4 year to make a 3d rendering in the 80′s today is take much less .. all of this time they spend on politcs, script changing and other non thecnical issues.
All those million most are spent in advertising, and ceo’s pockets .. 50 high qualified tech taking 4 years of hard work to make SHERK…give me a break
I’m an animator (brief introduction) just wanted to clarify some things you may not know .
a professional animator cant animate 30 to 120 seconds a day ! and wont . for example in wall-e (which is 80 minutes) there are 72 animators , and the animating process usually takes three months (not sure how it is in pixar but it can be more not less) so each animator gets a little more than 1 minute in the whole movie which is 20 seconds a month ! animating is a very important process in the animation production its importance is like acting in movies . animators must plan their shot choose the best acts for the characters , body mechanics and a lot of more things so when you are watching the animation you believe the characters .
Markus Koivisto
You completely disregard the price of labour, which is a fatal flaw in your judgment. $180 Million or even $30 million sounds like a lot of money, but you need to realize that the lion’s share of all of this goes to pay the salaries of all involved. Consider a team of animators and programmers who each make $5000 a month. A gross understatement, but we’ll let it stand. Well estimate a headcount of fifty people involved. How much does this talent cost? 3 Million in salary costs alone, each and every year.
Now, the team for WALL-E was even bigger, and commanded even higher salaries. A brief look at IMDB[1] gives a cast of 300 characters involved in the making of the movie during a period of 4 years. This gives me a back-of-the-envelope figure of 72 000 000 dollars right there, using the same underinflated salary of 5000/month i used above.
Using game engines for independent movie development is done. There’s a whole genre of these movies out there, known as Machinima. These movies are however limited. You can choose to use the existing characters and settings, but you are thereby limited to said characters or settings and the somewhat limited animation tools provided by the game. You can of course develop new characters and settings, but that takes even more time and effort.
How much time and effort? Recenly, Clear Skies was released as a particularly impressive bit of machinima. It was a macbook and four friends -type of project, that took two years of part-time work to develop for forty minutes of work, and this was with existing character animation simply repurposed for the task. The creator of this piece of work told in an interview that he’d be able to cut down the production time to 8 months in subsequent episodes, but that still leaves us with a daunting 5 minutes/month schedule for a dedicated amateur.[2]
Animation is, simply put, hard and above all time-consuming work. A professional animator can churn out about 30-120 seconds per day, depending on the complexity of the task. There is no easy way around this. Even for a video game using special lip-synch software, the facial animation usually takes on the order of 90-150 seconds per day.[3]
The same math applies to making movies in general. The set pieces etc. cost nowhere near as much as the labour and the talent needed to get things done in short order. Post-production is one of the biggest time- and money sinks in the production chain, and here again it’s due to the vide variety of companies and people involved in the making of the movie that contribute to the budget.
The old adage “cheap, quick, high quality – pick any two” certainly applies here. Getting high quality in a reasonable timeframe takes a hell of a lot of talented effort which doesn’t come cheap.
Actually, I didn’t disregard the price of labor. I mentioned in the last couple of paragraphs that it was probably the reason why the budget was so enormous.
The issue with machinima is that it’s often quite literally done in the game. A video game isn’t exactly friendly to someone who wants complete control over their images. That Clear Skies film you linked to was done in EVE Online, which has a user interface that’s difficult to use even for playing the actual game. It’s not surprising that it was no easy task.
Valve’s Source engine, by contrast, is open to anyone who buys a Source game and downloads the SDK. New animations can be created for all of the in-game characters, using a combination of external 3D programs or Source’s own facial animation system. Or they can create their own models and drop them in. Mod-makers do this on a daily basis; a filmmaker using the same tools could create something rivaling Pixar, as Valve’s promotional videos demonstrate.
Most modern video game engines also have a lot of procedural animation functionality built in, so that can shave off quite a bit of work on all the nuances and variations that occur.
Any independent filmmaker trying to finance their own production is going to find creative ways to imitate Hollywood without spending nearly as much money. It’s something I’m working on right now, and hopefully by mid-April I’ll have a final cut to demonstrate the results. Why limit that to live-action films, though?
Markus Koivisto
As for Valves promotional footage – a 95-second bit doesn’t really compare with a feature film. It is impossible to recreate those 95 seconds of footage with the tools Valve has released. You need a lot of additional work. You can bet that behind that promotional spot was several days of custom animation. And that precisely is my point: you can’t bring that time down if you want to do something custom. 30 seconds of animation is still going to take all day, and you can’t bring that down by an order of magnitude no matter what tools you use. Consider the above for a moment. That spot took anything from one to five days to animate.Extrapolate from there to a feature film, and you can quickly see that there is no real time saved.
Clear Skies was an amalgamation of the Source engine with specific tools for posing and animation, combined with EVE online footage chromakeyed in, which you’d know if you’d looked into it. It really was a split approach, and the Source engine tools you specifically referenced were definitely used. The most time-consuming part was the facial animation, since even though the facial animation was provided by the engine, every phoneme still needed to be extracted and timed.
It becomes painfully obvious to me that you haven’t been involved in any animation work. People always underestimate the duration of tasks they don’t understand. Animation is time-consuming no matter how you slice it.
The reason why machinima can be faster is that you use premade animations to create your content. You are essentially combining canned goods to make your film. As long as you are relying on the work of the character animation of the game company, you are able to make great progress. If you want to do something other than the canned bits of animation, you are SOL and have to start the time-consuming process of animation again.
If you are not using the canned animation work provided by the character animators at the game company, you are not really saving any considerable amount of time. Red vs. Blue for instance, is rather simple to churn out since they did not have to do any custom animation. However, this does show in the end result, which is not quite Pixar-quality to say the least.
The examples you quote (Mass Effect, Valve promo spots etc.) are all the results of much custom animation, and not something that can be replicated easily. Procedural animation is a red herring – things like IK and skeleton animation is of course used in any reasonable animation tool already, and is not something unique to game development. The same tools are available to the animators at Dreamworks or Pixar or what have you.
I suggest you pop over to Machinima.com and take a good look at what kind of work is being produced with the four friends and a macbook -toolkit, and then think about the limitations of the medium for a while.
http://plankhead.com Zacqary Adam Green
You’re correct, I haven’t been involved in any major 3D animation work. I have, however, done fifteen seconds of fairly minimalist two-dimensional animation, and I’m going to estimate it took me an absolutely ludicrous 30 hours over the course of a couple days.
I’ve looked at a wide range of machinima, and I’m well aware of the limitations, and what it is. I apologize for only taking a cursory glance at Clear Skies, I simply assumed it was similar to the rest of the machinima I had seen.
However, I think you were misinterpreting exactly what I’ve been trying to say. I don’t feel that raw machinima, at this stage, can rival Pixar, and I’m not suggesting independent teams try to imitate Pixar to the letter with video game-quality art assets. I’m suggesting drastic innovation and experimentation.
You’re right, it is impossible to recreate the 95 seconds of Valve’s footage with the tools they’ve released. Source Filmmaker is not available to the public yet. Regardless, tools like it aren’t the be-all and end-all of quick, cheap, quality animation. From what they’ve said about it, it is used only for choreography, motion blur, and depth of field. The custom animations still have to be produced in another animation program.
That animation program doesn’t necessarily have to be the time-consuming keyframe programs like the ones used by Pixar. Procedural animation is used in video games, even ones in development by small teams like Wolfire. Obviously, as their more detailed example shows, this can’t account for the entirety of an animation; that’s the exact principle on which professional software like Endorphin works, supplementing keyframe animation with procedural.
But keyframes can be cut back even further with recent motion capture techniques that don’t necessarily have to be expensive. Benjamin Button used a special camera to overlay CGI in real time, but what if a similar technology were used for rotoscoping digital video? Combining a technique which was originally demonstrated by researchers in 1999 with face (or other body part) tracking, it would be possible to capture an actual actor’s performance from a single camera angle, then apply the same animation data to a compatible model.
Perhaps an integrated and easy suite of tools to do all this doesn’t yet exist. But programming things like that have been done by four friends and a MacBook, and they’ve often been released open source.
Markus Koivisto
There is at least one false assumption here that should be cleared up. Animation studios already use a wide variety of tools to speed up the animation process. Don’t for a moment kid yourself and think that mocap and procedural animation isn’t already used in feature-length animations. It is, and extensively, whenever it makes sense to do so. In fact, it is used first in those instances before it ever becomes viable to do with games.
Mocap doesn’t make things any less time-consuming either. Instead of animating, you end up shooting real footage with real actors, motion capturing it, and cleaning up the result. You are just trading one set of problems for another.
What I’m trying to say is that there is all kinds of research that is being done on how to animate quickly. This research is being driven by the needs of the VFX and CGI industry, and they are of course the first to benefit from new developments in the field.
For an indie filmmaker getting into the CGI field, the learning curve is steep. The tools aren’t there yet, as you say, and I would go one step further and say that they will probably never be, since the bar is constantly pushed higher.
Questioning why Pixar chooses to spend the amount of money doesn’t really lead anywhere. The simple answer is that a feature-length production with the level of quality they aim for in any reasonable timeframe will always require a large team, and large teams cost a lot of money. Of course, Pixar is pushing the envelope, so they are doing a lot of R&D. If they wouldn’t stray from the beaten path, they would be able to create more of the same far cheaper, but that’s not what they want to do. They have the resources, and they are turning a good profit, so I would say they could in fact throw MORE money at more challenging problems, not less.
Where does this leave the small team? As with live-action, small team should concentrate on their unique strengths and the power of their vision to turn their limited resources to their advantage. It doesn’t really matter which approach you choose, the important thing is to know it so that you can concentrate on the story instead of the tools.
If the story is compelling, and the basics of good filmmaking are there, the audience will be sucked in and won’t care what you used to create your movie. Learning the latest and greatest technique will leave you chasing a moving goalpost instead of getting up to speed and actually producing your movie.
http://plankhead.com Zacqary Adam Green
I understand where you’re coming from, but you make the false assumption that I’m suggesting anything about “learning” the latest and greatest technique.
One factor which I don’t think I’ve explicitly mentioned is the independent filmmaker’s passionate disbelief in the concept of “impossible.” If what’s already out there is too expensive, too laborious, and too massive, then an indie team with no budget will build their own damn technique from the ground up. It’s been happening since Georges Melies reverse-engineered the Lumiere camera, and it will continue to happen, regardless of medium.
I can’t predict exactly how a small team is going to eventually do something on par with Pixar. I’m merely suggesting what that inevitable technique could be like.
Markus Koivisto
I suppose we can agree on that whatever the technique ends up being, it will involve a metric assload of work :)
http://plankhead.com Zacqary Adam Green
Initially, yes, I’d imagine. I’m a bit overwhelmed trying to figure out how to shoot a movie under blacklight, because the only tips and tricks I can find are for still photography. Probably only a milli-assload compared to figuring out how to imitate Toy Story for under $10,000, though.
http://www.herrschmidt.tv Martin Schmidt
Creative ways to imitate? How is imitation creative? The biggest fault you can make as an independent filmmaker is trying to imitate Hollywood! Why would you do that? Use your creative ways to create your own style, not imitate the faceless Hollywood style. Put your greatest effort into reducing your design, step away from realistic humanoids, how boring are they anyways?! Hell, you can make a movie containing only cubes as characters ( http://www.youtube.com/watch?v=hddYiSixKGQ ). Learn some good storytelling, thats a filmmakers tool.
And by the way, no animator can ouput 90 – 120 seconds per day of full animation! At least no quality animation. Limited animation like Southpark can output this kind of figure, but not the Pixar kind of thing. I don’t think you really need the Pixar animation style to create great films, in fact it might even be hinderly. Just to give you an idea: a Disney or Pixar animator outputs between 3 and 10 seconds of animation per day! More rough animation, like the stuff done for series, takes a day to finish 10 to 20 seconds!
Gosh, it took me 5 years and a couple of short films to figure out that I have to reduce my style even more to make a living as an independent filmmaker.
Brian
One of the software engineers from Walt Disney, which works pretty closely with Pixar, actually came to MIT and gave a seminar somewhat related to this discussion. Mark was correct in assuming that the majority of the money is in salary. They hire artists, who draw up initial character and scenery designs, software engineers, to write programs to control certain aspects like herding in Lion King (simultaneously controlling hundreds of random objects) or wind blowing hair in Bolt (millions of independent hairs that have to move within a certain parameter), and managers, who act as mediums for the two. While old programs can be reused, like the herding or hair program, each character is unique in its facial and body structure and will have unique characteristics. If you look closely many of the random background people will have the same facial structure, with slight changes because it is so difficult to create new characters. Overall, they work with at least 20+ terrabytes of information they have to process. When they begin to reach a certain cap, they start deleting unessential information. They don’t even bother to render scenes because it would take days just to render a scene that might be a few minutes long.
The only way that indie film makers will ever have the ability to try to get on par with these guys would to have access to terrabytes of storage, a ridiculous amount of panel drawing, character, and scenery design, just as much time for coding, and an ungodly amount of processing time. Either you have to be a zealous nut and be willing to waste your life recreating wall-e or you can wait just as much time for the technology to come out.
I’m pretty sure I left out more but I’m getting tired of typing. Sorry if that was kind of long. Just thought I should give my two cents on it.
http://plankhead.com Zacqary Adam Green
Yes, terabytes of data are manipulated for having wind blow millions of different hairs. But why? Bolt is already very stylized, and having to simulate millions of individual hairs seems like overkill. If it were to simulate even just thousands of hairs, the amount of data and calculations would be reduced dramatically, and audiences wouldn’t notice the difference with sufficient post-processing.
Because major studios have the money to do all that, they don’t feel it’s necessary to “cheat” to produce a perfectly fine result for less. The entire reason for this post, though, is because Valve seems to have proved that such cheating is possible and effective, even if that wasn’t their primary intention.
By the way, entering “asdf@asdf.com” in the email field triggers my spam filter, and you also won’t get notified that I’m replying to you. Just letting you know, if you come back to read this.
Brian
One of the software engineers from Walt Disney, which works pretty closely with Pixar, actually came to MIT and gave a seminar somewhat related to this discussion. Mark was correct in assuming that the majority of the money is in salary. They hire artists, who draw up initial character and scenery designs, software engineers, to write programs to control certain aspects like herding in Lion King (simultaneously controlling hundreds of random objects) or wind blowing hair in Bolt (millions of independent hairs that have to move within a certain parameter), and managers, who act as mediums for the two. While old programs can be reused, like the herding or hair program, each character is unique in its facial and body structure and will have unique characteristics. If you look closely many of the random background people will have the same facial structure, with slight changes because it is so difficult to create new characters. Overall, they work with at least 20+ terrabytes of information they have to process. When they begin to reach a certain cap, they start deleting unessential information. They don’t even bother to render scenes because it would take days just to render a scene that might be a few minutes long.
The only way that indie film makers will ever have the ability to try to get on par with these guys would to have access to terrabytes of storage, a ridiculous amount of panel drawing, character, and scenery design, just as much time for coding, and an ungodly amount of processing time. Either you have to be a zealous nut and be willing to waste your life recreating wall-e or you can wait just as much time for the technology to come out.
I’m pretty sure I left out more but I’m getting tired of typing. Sorry if that was kind of long. Just thought I should give my two cents on it.
http://plankhead.com Zacqary Adam Green
Yes, terabytes of data are manipulated for having wind blow millions of different hairs. But why? Bolt is already very stylized, and having to simulate millions of individual hairs seems like overkill. If it were to simulate even just thousands of hairs, the amount of data and calculations would be reduced dramatically, and audiences wouldn’t notice the difference with sufficient post-processing.
Because major studios have the money to do all that, they don’t feel it’s necessary to “cheat” to produce a perfectly fine result for less. The entire reason for this post, though, is because Valve seems to have proved that such cheating is possible and effective, even if that wasn’t their primary intention.
By the way, entering “asdf@asdf.com” in the email field triggers my spam filter, and you also won’t get notified that I’m replying to you. Just letting you know, if you come back to read this.
Joe
Companies like Valve may be able to crank out a 5-minute animation relatively easily, but if you want to scale it up for an 80-minute feature film, they’re going to have some exponential costs associated with it.
Companies like Pixar aren’t merely making movies. They are inventing new technologies as they go along. They didn’t have realistic hair until Toy Story 2. Human skin didn’t have subsurface scattering technology until the Incredibles. They used reflection occlusion heavily in Cars. Developing all these new techniques cost a chunk of money in R&D. Rendering a frame of Wall-E is probably more time consuming than rendering a frame back in the Toy Story days. Every year processors double in power, but the level of complexity of the scene triples.
I should also note that Pixar writes a lot of the software themselves, because the off-the-shelf software that is available to you and me simply won’t work when you scale it up. I can do a batch render on my desktop before I go to sleep and look at the results in the morning, but that software is not going to work when you’re talking about 200 artists and thousands of processors.
Pixar’s render farm (as of The Incredibles) was probably around 2400 cpus . I read somewhere that one of the benchmarks they had consider was not how much rendering power it costs per dollar, but how much power they could cram into a cubic foot of precious render farm real estate. It’s fortunate that Pixar doesn’t have to pay a hefty software license to use its own RenderMan… or else that’ll be another large chunk of change.
The marketing and distribution costs is probably huge too. Think of all the TV airtime, billboards, bus stations, print ads that Disney unleashes prior to the release of a movie. That’s also included in the $180 mil.
The largest chunk I’m guessing, consists of the salaries of the crew. In the third and fourth year of a typical Pixar production cycle the crew swells to about 200 people (I checked the credits for Wall-E). These are the animators, the modelers, the texture artists, cg painters, layout, set dressing, shader writers and lighting artists. You need all of these steps to create a CG movie. Really.
You can farm out some of the work to places like India to cut some costs, but in practice this rarely works out, because you can’t farm out the knowledge and talent to meet the exacting standards that is required. Pixar only accepts the best animators, turning away many. They’re an elite (sometimes cocky) bunch, and deservedly so. If the average American animator doesn’t have the level of training that’s going to make John Lasseter happy, what are the chances of one working from Bangalore? Maybe one day. But for now, not yet.
But anyway, back to your original point.
Do you have to do all that to make a successful animated movie? Probably not. As long as your story is solid, you can have a great feature film be it the South Park Movie or Wall-E. You might be able to do a 3-hour feature a la Don Hertzfeld stick figures, maybe.
Pixar movies cost a lot because they’re a bunch of perfectionists. You may not care if Bolt has a thousand bristlebrush hairs or a million smooth hairs, or whether the characters are smoothly subdivided or are made up of jaggy polygons, or if the animation is snappy and grounded as opposed to loose and floaty. But they do. You might not notice that 1/24th of a second flicker over on the top left corner of the screen, but these guys obsess over little details that seasoned CG artists may not even notice after the tenth viewing.
If you have the money enough to buy perfection, wouldn’t you spend it?
http://plankhead.com Zacqary Adam Green
I guess I didn’t make my intention for writing this very clear, given that most of the comments have been explaining to me exactly how many people are using how many yottaflops to make these movies. That’s my fault, I suppose. I’m well aware of why Pixar spends $180 million on a single movie. What I meant to ask, though, is something more like “Is it really necessary?”
It’s probably better for me to make a follow-up post than to keep making this page approach 1747958261956 pixels in height (I can’t find any plugins for collapsable comments that work with K2), so look for that later tonight. It’ll be less focused on “why are they doing it?” and more on “why aren’t you starving independent artists doing it?”
(EDIT 3/4/09: Okay, make that later today. Been lazy.)
Joe
Companies like Valve may be able to crank out a 5-minute animation relatively easily, but if you want to scale it up for an 80-minute feature film, they’re going to have some exponential costs associated with it.
Companies like Pixar aren’t merely making movies. They are inventing new technologies as they go along. They didn’t have realistic hair until Toy Story 2. Human skin didn’t have subsurface scattering technology until the Incredibles. They used reflection occlusion heavily in Cars. Developing all these new techniques cost a chunk of money in R&D. Rendering a frame of Wall-E is probably more time consuming than rendering a frame back in the Toy Story days. Every year processors double in power, but the level of complexity of the scene triples.
I should also note that Pixar writes a lot of the software themselves, because the off-the-shelf software that is available to you and me simply won’t work when you scale it up. I can do a batch render on my desktop before I go to sleep and look at the results in the morning, but that software is not going to work when you’re talking about 200 artists and thousands of processors.
Pixar’s render farm (as of The Incredibles) was probably around 2400 cpus . I read somewhere that one of the benchmarks they had consider was not how much rendering power it costs per dollar, but how much power they could cram into a cubic foot of precious render farm real estate. It’s fortunate that Pixar doesn’t have to pay a hefty software license to use its own RenderMan… or else that’ll be another large chunk of change.
The marketing and distribution costs is probably huge too. Think of all the TV airtime, billboards, bus stations, print ads that Disney unleashes prior to the release of a movie. That’s also included in the $180 mil.
The largest chunk I’m guessing, consists of the salaries of the crew. In the third and fourth year of a typical Pixar production cycle the crew swells to about 200 people (I checked the credits for Wall-E). These are the animators, the modelers, the texture artists, cg painters, layout, set dressing, shader writers and lighting artists. You need all of these steps to create a CG movie. Really.
You can farm out some of the work to places like India to cut some costs, but in practice this rarely works out, because you can’t farm out the knowledge and talent to meet the exacting standards that is required. Pixar only accepts the best animators, turning away many. They’re an elite (sometimes cocky) bunch, and deservedly so. If the average American animator doesn’t have the level of training that’s going to make John Lasseter happy, what are the chances of one working from Bangalore? Maybe one day. But for now, not yet.
But anyway, back to your original point.
Do you have to do all that to make a successful animated movie? Probably not. As long as your story is solid, you can have a great feature film be it the South Park Movie or Wall-E. You might be able to do a 3-hour feature a la Don Hertzfeld stick figures, maybe.
Pixar movies cost a lot because they’re a bunch of perfectionists. You may not care if Bolt has a thousand bristlebrush hairs or a million smooth hairs, or whether the characters are smoothly subdivided or are made up of jaggy polygons, or if the animation is snappy and grounded as opposed to loose and floaty. But they do. You might not notice that 1/24th of a second flicker over on the top left corner of the screen, but these guys obsess over little details that seasoned CG artists may not even notice after the tenth viewing.
If you have the money enough to buy perfection, wouldn’t you spend it?
http://plankhead.com Zacqary Adam Green
I guess I didn’t make my intention for writing this very clear, given that most of the comments have been explaining to me exactly how many people are using how many yottaflops to make these movies. That’s my fault, I suppose. I’m well aware of why Pixar spends $180 million on a single movie. What I meant to ask, though, is something more like “Is it really necessary?”
It’s probably better for me to make a follow-up post than to keep making this page approach 1747958261956 pixels in height (I can’t find any plugins for collapsable comments that work with K2), so look for that later tonight. It’ll be less focused on “why are they doing it?” and more on “why aren’t you starving independent artists doing it?”
(EDIT 3/4/09: Okay, make that later today. Been lazy.)
http://checkyourhud.com Jenni
That Team Fortress 2 promo is NOT close to Pixar quality. Even with your comparison of stills you can see a huge difference in lighting. Not to mention you’re comparing a full featured film to shorts from a video game… it takes a lot more to animate an entire film than a few seconds of fast paced animation… it’s not really an apple to apples comparison.
http://plankhead.com Zacqary Adam Green
The difference in lighting is caused by the use of hard light in the Incredibles still and soft light in the Meet the Scout still. Repositioning the key light (or, in animation terms, playing with the sliders) would likely bring them closer in aesthetic terms.
Of course, I’d imagine something similar could be achieved with color correction. I’m on my phone at the moment, so I can’t show you an example, but once I get back to Photoshop I’ll see what I can demonstate with the input levels and contrast.
I don’t think you’re correct that it takes much more to animate a feature-length film, at least in terms of technique. Time, definitely, would increase, but for an animated film focusing more on style than tehchnical innovation, video game models with post-processing would definitely work well. Pixar just does stuff because they can.
http://checkyourhud.com Jenni
I disagree that it’s just a matter of the type of lighting. There’s a lot more depth in the incredibles shot–I’d say more light sources and a stronger knowledge of how to make those light sources work together. Anyone can stick a few lights in a scene and call it a day but to create realistic lighting and set a mood takes a lot of know how and finesse.
http://plankhead.com Zacqary Adam Green
In this particular case, you’re correct. I did open up the image in Photoshop to try and get the color balance in the Meet the Scout still closer to that of the Incredibles still. As you can see from the result, the depth issue wasn’t fixed entirely.
That’s not to say more light sources and a stronger knowledge of how they work together is impossible unless you work at Pixar. If Valve had commissioned a trained cinematographer, he or she probably would have clicked the “Add Light” button a couple more times. Personally, I think bouncing a light off of the front-facing side of the crate, then turning down the exposure would make the background deeper. That’s certainly within the capabilities of the software.
Sarah
You are trying to make the point that it is unnecessary to spend that kind of money on an animated movie. Now, I need to ask you…what is the point of spending money on ANY movie? Why should people bother spending money on nicer clothing or cars? Why bother spending money on anything but the bare necessities?
In other words…we do it because we can and we want to!
Sarah
You are trying to make the point that it is unnecessary to spend that kind of money on an animated movie. Now, I need to ask you…what is the point of spending money on ANY movie? Why should people bother spending money on nicer clothing or cars? Why bother spending money on anything but the bare necessities?
In other words…we do it because we can and we want to!
http://checkyourhud.com Jenni
That Team Fortress 2 promo is NOT close to Pixar quality. Even with your comparison of stills you can see a huge difference in lighting. Not to mention you're comparing a full featured film to shorts from a video game… it takes a lot more to animate an entire film than a few seconds of fast paced animation… it's not really an apple to apples comparison.
http://plankhead.com Zacqary Adam Green
The difference in lighting is caused by the use of hard light in the Incredibles still and soft light in the Meet the Scout still. Repositioning the key light (or, in animation terms, playing with the sliders) would likely bring them closer in aesthetic terms.
Of course, I’d imagine something similar could be achieved with color correction. I’m on my phone at the moment, so I can’t show you an example, but once I get back to Photoshop I’ll see what I can demonstate with the input levels and contrast.
I don’t think you’re correct that it takes much more to animate a feature-length film, at least in terms of technique. Time, definitely, would increase, but for an animated film focusing more on style than tehchnical innovation, video game models with post-processing would definitely work well. Pixar just does stuff because they can.
http://checkyourhud.com Jenni
I disagree that it’s just a matter of the type of lighting. There’s a lot more depth in the incredibles shot–I’d say more light sources and a stronger knowledge of how to make those light sources work together. Anyone can stick a few lights in a scene and call it a day but to create realistic lighting and set a mood takes a lot of know how and finesse.
http://plankhead.com Zacqary Adam Green
In this particular case, you’re correct. I did open up the image in Photoshop to try and get the color balance in the Meet the Scout still closer to that of the Incredibles still. As you can see from the result, the depth issue wasn’t fixed entirely.
That’s not to say more light sources and a stronger knowledge of how they work together is impossible unless you work at Pixar. If Valve had commissioned a trained cinematographer, he or she probably would have clicked the “Add Light” button a couple more times. Personally, I think bouncing a light off of the front-facing side of the crate, then turning down the exposure would make the background deeper. That’s certainly within the capabilities of the software.
frank111
actors are expensive… software licences are expensive… see autodesk maya website how much that software cost per computer, to render 3d image using tons of computers, you need alot of licences.. that mean you need a lot amount of money, and that's just one software. thank god now there's open source software that can cut production budget..
http://plankhead.com Zacqary Adam Green
And when open source software is used by a small independent team that happens to be composed of or be friends with people who are pretty good actors, amazing things can happen for no money.
Kev
Zacqary,
Do you believe, if given the opportunity, you could develop a feature film quality (and length) animation? If so, how many people and how much money do you think you could do it for? And how long?
And how much for the same quality but a shorter length – say 5 minutes.
What are your qualifications?
I scanned above but not in intense detail.
Regards,
K
http://plankhead.com Zacqary Adam Green
I’ll give a conservative estimate. Let’s see…
Erica and I spent a combined total of about 600 hours on Episode 2 Your Face is a Saxophone, which is 25 minutes. That’s about one half-hour of work per second of footage.
Obviously, YFIAS isn’t Pixar-quality, but it is (at least in the second episode) certainly TV quality. Feature film-quality would certainly take longer.
The caveat here is that I, personally, don’t aim to create animation that looks exactly like Pixar. As I mentioned in the post, it’s becoming easier and easier to create Pixar-like visuals, but I personally prefer an aesthetic closer to classic Disney or hand-drawn comic art. Fortunately, that style can require significantly less work while still looking wonderful.
I’d estimate that I could get a team of animators to work for three hours per second of feature-quality footage. For a 90-minute film, that would be about 16,000 total hours.
Let’s split that between five people. 3200 hours per person. Now, let’s imagine they work eight hour days (yes, I know it’s industry standard to kill yourself with hours, but studies show that productivity drops when people work more than eight hours and/or die). That’s 400 days of work.
So, with a five-day work week (What? We’re giving them weekends too? That’s so decadent and inefficient!), that would take 80 weeks, or about 18 months.
The average cost of living in the US is about $2000 a month. So five people working 18 months would cost $180,000. This cost could be brought down by lowering the animators’ standards of living or working them longer, but the added stress would hurt productivity, not to mention increase the risk of somebody burning out and derailing the project.
Actors probably wouldn’t add very much to the cost of labor. You can cast the animators, and for the roles they can’t play, there are lots of talented actors willing to work for free or cheap. The cast wouldn’t need to add more than $5000 to the budget. (Or we could be artsy and go for a non-dialogue film, and remove the expense of actors entirely)
Equipment-wise, we’d need to give them each a workstation. Something around $1500 would work fine. So that’s another $7500. Using only free and open-source software wouldn’t add to the expense. (Or the use of pro-pirate-ary software when necessary, but that’s illegal, and nobody ever does that to cut costs, am I right?)
Recording audio isn’t expensive at all anymore. You can get a great-sounding mixer and mic setup for about $1000, and can record in a mildly quiet apartment.
Speaking of which, studio space wouldn’t be an expense at all if the animators all lived together and worked from home. This is a fairly common practice for start-ups.
So, that leaves us with:
Labor: $185,000
Equipment: $8500
Total: $193,500
Round it up to $200,000 for a bit of breathing room.
Now, $200,000 is still quite a lot of money. It’s nowhere near $180 million, though.
For a five-minute film with the same sized team, you could do it in about a week and a half with a labor cost of $2500 (not including actors, but you could easily get that for free if it’s so short). However, with the equipment investment, it’s almost not worth it. (You could save $3000 by having only two animators work for four weeks, but the total still comes to $8000)
Anyway, those are my ballpark figures for doing it comfortably.
kev
Zacqary,
Are you talking about “My Face is a Saxophone” quality – cartoon like animation for this price or the example above (way above) that is similar to the “gamelike” or “claymation” type similar to pixar but not the same?
K
http://plankhead.com Zacqary Adam Green
It would definitely be more technically sophisticated than Your Face is a Saxophone, that’s for sure. Like I said, I’m more of a fan of an aesthetic that looks illustrated — the recent Oscar nominee Chico & Rita piqued my interest quite a bit. It would depend on the subject matter, though; you have to pick the right look for the script.
kev
Gotcha, I think…So, just to be clear, do you think you could do the “pixar-like” TV quality look for the $200 K?
http://plankhead.com Zacqary Adam Green
You’re serious, aren’t you?
To answer your question, yes, my gut instinct is that I could. But maybe we should continue this through email. Zacqary@plankhead.com.
frank111
actors are expensive… software licences are expensive… see autodesk maya website how much that software cost per computer, to render 3d image using tons of computers, you need alot of licences.. that mean you need a lot amount of money, and that's just one software. thank god now there's open source software that can cut production budget..
http://plankhead.com Zacqary Adam Green
And when open source software is used by a small independent team that happens to be composed of or be friends with people who are pretty good actors, amazing things can happen for no money.
http://1centralx.co.cc/color-hair-o46k/category/hair-color-correction-tips hair color correction tips
Check out Big Buck Bunny for example, for an animated short, and we have Valkaama and The Last Drug as real-life Full feature movies in development!
I still think, the hardest thing to find is creativity.
kendwebber
The reason you see big $$$ attached to movies is in the way the finances are structured. They are produced to conform to ZERO PROFIT. Whatever comes in is how much the movie cost to make because everything is nickle and dimed out by using percentages. They do this because IF you have a profit then you pay taxes but if you have ZERO profit then you pay ZERO in taxes. This is why big actors will earn a cut, a percentage of the profits, so that in the end there is no profit left on the books. This is where all your millions go, divided up among the players and actors and the studio heads. The workers get paid chicken feed.
http://plankhead.com Zacqary Adam Green
That makes sense on the level that I wouldn't put it past Hollywood to pull that crap, but didn't The Dark Knight bring in over $1 billion on a $185 million budget? If what you're saying is true, wouldn't that mean a $1 billion gross is a bad thing for the movie studio?
kendwebber
Their budget for producing the movie may have been 185 but that one billion you mention is not GROSS profit because the big players are paid on percentages. All the major movie studios make ZERO profit once the books are figured. They pay the big talent in percents, the lesser talents get chickenfeed. Most of the money goes to studio heads who use it like their own private reserve to fund new movies, reinvestment and maintenance, and as a hedge against bad times, and when bad times come they can last for years sometimes bankrupting the studios themselves.
Anonymous
Pixar are successful because they seek perfection in their scripts as well as their animation. Pixar spend a long time developing their scripts because without a great storyline at this level of entertainment all you are left with are pretty pictures and nothing else. A well developed powerful storyline will pull people in even if the production isn’t cutting edge because its the story not the window dressing that people are interested in.
kendwebber
Their budget for producing the movie may have been 185 but that one billion you mention is not GROSS profit because the big players are paid on percentages. All the major movie studios make ZERO profit once the books are figured. They pay the big talent in percents, the lesser talents get chickenfeed. Most of the money goes to studio heads who use it like their own private reserve to fund new movies, reinvestment and maintenance, and as a hedge against bad times, and when bad times come they can last for years sometimes bankrupting the studios themselves.
stephenconnell
I want to make an animated film based on a story I have written but following this discussion has educated me in the huge task involved in making an animated movie of reasonable standard. I love the story i want to animate any tips on getting going?
http://plankhead.com Zacqary Adam Green
Work with what you have and think creatively. One option is to go with a cartoony art style that's not necessarily realistic but visually striking and interesting. Whatever you do, approach the story in a way you know you're capable of telling it.
But a nice touch, even if you work in cartoony squigglevision, is to simulate depth of field. I do this by digitally blurring different parts of the image which I don't want to be “in focus”; you'll notice I do that a lot in Goliath. Focus is a filmmaking tool just like camera angle and framing, so it's nice to have in animation too.
But the best thing you can do is know yourself, what you're capable of, and how you can make that awesome.
http://plankhead.com Zacqary Adam Green
Work with what you have and think creatively. One option is to go with a cartoony art style that's not necessarily realistic but visually striking and interesting. Whatever you do, approach the story in a way you know you're capable of telling it.
But a nice touch, even if you work in cartoony squigglevision, is to simulate depth of field. I do this by digitally blurring different parts of the image which I don't want to be “in focus”; you'll notice I do that a lot in Goliath. Focus is a filmmaking tool just like camera angle and framing, so it's nice to have in animation too.
But the best thing you can do is know yourself, what you're capable of, and how you can make that awesome.
Bgurrl
Mostly budgets are inflated for marketing. It gets people in the seats. People want to see what 180mill looks like. There are films that are inflated that couldn't possibly cost that much. This is especially the case when movies have limted locations, non-A list talent, payment of crew isn't a whole lot or at least average, but we are to believe some of the movies cost millions to make. Another thing if CG is making sfx so cheap then whey are we still paying 10 or more for tickets?
http://www.usa-basketball-shoes.com usa basketball shoes
In this material life of society, fashion seems to be a habit. Here you can recommend a few mall, I feel good.basketball shoes outlet
VFX wizard
youre an idiot….
http://plankhead.com Zacqary Adam Green
Thank you for your insightful comment, VFX wizard, whose email address is
apparently the non-existent “c@c.com”. While I disagree with your highly
nuanced argument, I admire your verbosity and appreciate your contribution
to this civil and reasonable discourse.
VFX wizard
Well as you may have noted, I put up a more reasonable explanation to the aforementioned quote, but you tore it down. And I don’t like giving people like you my email, its bad for business.
I also laugh at your use of reasonable. Your entire article is unreasonable as you have no real sound facts backing any of your claims. Take a trip trough any vfx company and you will no doubt see a highly technical and innovative process, which requires literally dozens of sets of expertise as well as eyes.
People pushing buttons is a watered down, uninformed ,ad basically stupid statement to make.
IE: youre an idiot.
http://plankhead.com Zacqary Adam Green
I’m sorry, I don’t recall any previous comment from you. Perhaps an error occurred with the comment system.
Awesome_1234
Yes, in theory a team of passionate geeks can create something that resembles a pixar film. IN THEORY.
To do a feature length animation, it requires a lot of planning and preparation before actually animating. In fact, animation comes far down the pipeline. There’s scripting, there’s pre-vis, there’s setting up a pipeline, there’s modeling, there’s recording for lipsync, there’s just so much to do and so much to know!
If a team of geeks where all the members are experts in directing, script writing, animation, modeling, compositing, audio mixing, then yes… IN THEORY they can do it. But it will take a very very long time.
Simon
This post will probably have a plethora of typos. Never the less, some points I’d like to make:
-Pixar develops custom tools for it’s films. There is an army of programmers/mathematicians who are experienced in fluid dynamics, physics, geometric optics and more. Remember Ratatouille’s water scenes? Tools were built to handle Remy and simulate the water. That was all new technology, made all by Pixar, JUST for Ratatouille. No other studio had done it before. The programmers also maintain Marionette, Pixar’s modeling and animation software.
-Valve does develop it’s own tools with it’s game engine, Source. But… They didn’t exactly invent any revolutionary new tools for this promo video as far as I can see. They used Source’s cinematic function, and had some very basic lighting. Nothing fancy.
-Pixar went through the process of building everything from scratch. So environment, models, script, storyboard, animation mock-up, textures, shaders, design, etc. ALL of this was made from scratch. That is a lot of artists… working lots of hours. Earning money.
-All of Valves assets were already built for them during game development. By assets I mean: models, textures, set pieces etc. The things that were actually made new for the scene were script, music, animation, voice.
-Pixar films have so much detail. How much detail was actually put into the TF2 promo video? The difference between the two is overwhelming.
-Valve used 2 or 3 pre-built sets with minimal detail. Valve doesn’t bother with pouring months of work into a set’s detail, as it’s unnecessary. Game systems can only handle so much geometry, and the physicality of the set is way more important than the aesthetic detail in terms of game mechanics. They animated maybe 10 characters at most. Slap a sun in the sky. Boom. Lighting is done. Texturing was already done. Shading was already done.
-Pixar used definitely more than 50 sets built from scratch, which are also filled with an astounding amount of detail. They animate thousands of characters and pieces. They spend hours lighting a set. Shaders are developed from scratch. Texturing all from scratch. Technologies are built from scratch. Everything is from scratch.
I don’t want to be that ass who has to disprove the blogger in every way possible, but there are reasons why Pixar has big budgets.
The articles on that site detail a LOT of the developments that were made specifically for each movie, as you mentioned.
Abgg
have u lost ur nuts ???
http://plankhead.com Zacqary Adam Green
Yes. Of course I have. Has anything about this website not made that abundantly clear?
james braselton
hi there would take only 1 hour too make a 2 hour movie useing solid state flash drives becuase ssd have 10 times read write speeds over hdd thats why i have the 4 gb ssd xbox 360 wii 3ds macbook air 256 gb ssd ipad iphone 4 32 gb flash drives
http://plankhead.com Zacqary Adam Green
That would improve rendering times, but not the actual process of animation.
james braselton
hi there 3d cost more too make
http://plankhead.com Zacqary Adam Green
In theory, not much more. It’s just a matter of rendering out a second “eye”. That does double the rendering time for the whole film, but it needn’t be any more labor-intensive.
A Pixar animators quote is 1 to 2 seconds per week.
Lora.S
I very much dislike this. Written by someone who has no idea what they are talking about.
“The actual CGI, though, is achieved by making a bunch of 3D models, texturing them, arranging them in a world, moving them around, and then pointing a virtual “camera” at various parts of the world to get the images you need.”
Animation is an art form that includes every aspect of art and design. It is not just creating a bunch of stuff and seeing what happens at the end. It’s not simple, unless you want it to look like sh*t. Artists can tell the difference between Pixar and Team Fortress. The characters have no appeal in that team thing, I’m going to stop before I rant too much about the differences and end up screaming at the computer pretending it is you. Just shhhhh. You don’t have a clue unless you open maya and try to build or animate something. Don’t even get me started on setting the keys properly!! There is no way you could create a pixar film on a lap top. Render, render, CRASSHHH.
Ypatel
they are crooking the tax man by claiming exortibant price tag to make animated films
Wow…
ARE YOU AN IDIOT!!!!!!!! Pixar films take about 2 to 3 years to make, they work their asses off. Animation is much harder than you think it is! Pixar has to pay 750 employees a year, roughly that would probably be about 61,125,000 a year. This doesn’t count their bills for electricity, insurance for the employees, benefits, etc. It doesn’t include the cost of acting. THESE ARE TOTALLY DIFFERENT IMAGES, YOU ARE FUCKING BLIND, IT’S A WONDER YOU ARE NOT AN ARTIST. Have you seen the vast amount of imagery there is in a Pixar Film? Especially Wall-E? This also doesn’t cover distribution or marketing since that was in their budget because Disney owns them. They do their own. Distribution per theater costs about $2400. Now multiply that by about12,050 theaters. You are an idiot. You know nothing of film. Did you account for development, pre-production, production, and post? You don’t just hit fucking buttons.
http://plankhead.com Zacqary Adam Green
Lol u mad bro
http://www.facebook.com/genarorg Genaro Rocha
wow you really have no clue, every paragraph above is full of nonsense and unfunded statements.
http://plankhead.com Zacqary Adam Green
For the record, this post is from when I was 19, and I’d have written it differently today. I honestly have no idea how people keep finding it and leaving angry comments, but they greatly amuse me, so I just go with it.
You’re partially right, though. It’s not nonsensical because I know what I was trying to say here: people can, and should, find innovative ways to animate at lower cost. But in hindsight, I can see how this is a pretty weak argument for it.
Maybe I’ll rewrite a coherent version of this someday.
Alex
Wow, it baffles me how completely foolish you are.
Speedymadara
this guys so dumb I am a computer animator and as much money as its sounds the animators don’t see enough of it. I work 14 to 15 hour days on a computer basically have no life in the production process and the voice actors come in do a couple of tracks and leave and make triple what i make you need to alter you subject bud
Why Do Animated Movies Have $180 Million Budgets?
UPDATE FEB 8 2012: I guess this piece got high up in some common Google query, because I’ve had a lot of vicious comments on it over the years. So let me explain: This article is from 2009. I was nineteen when I wrote it. I also wrote this follow-up piece clarifying what I meant to say, which apparently isn’t as popular on the Googletubes. Anyway, this is a terrible post. I was trying to encourage people to innovate with low-cost animation, but the following is a completely ridiculous way of saying that. Please stop getting angry at me.
UPDATE FEB 9 2012: Okay that didn’t work. You know what? Comments are closed. I’m getting tired of seeing this pop up in my inbox all the time.
Perhaps I’m uneducated in the ancient Hollywood art of Unit Production Management, but it’s baffling to me that WALL-E had a budget of $180 million. Yes, it was a gigantic Hollywood production, but consider the fact that all of its visuals were made by pressing buttons and waiting for the images to appear. Well, it was more complicated than that sounds, but that’s basically what they did.
So what exactly cost so much money? Without access to Pixar’s financial records, I’ll take a few guesses. But the short answer is that they’re spending way more money than they need to.
I’d imagine, from my experience, that some of the the biggest expenses in filmmaking are film stock, set construction, stunt coordination, and other tangible materials. Pixar requires none of that. The only materials they really need to buy are computer parts, which are relatively cheap on their own and downright bargains when bought in bulk. Yes, their concept artists need clay and paint and such, but that hardly seems like it would cost more than a couple thousand, even for a gigantic team of the kind of overzealous madmen who spend more than a week in an art school (i.e. me).
So that brings us to visual effects, which is essentially 100% of what you see on the screen. Yes, WALL-E and Happy Feet had live action bits superimposed, but anyone with access to a green screen and a DVX can imitate those. The actual CGI, though, is achieved by making a bunch of 3D models, texturing them, arranging them in a world, moving them around, and then pointing a virtual “camera” at various parts of the world to get the images you need.
It may sound simple (save for the whole “talent” part in terms of modeling, texturing, and animating), but to create the impressive visuals, Pixar needed a very large amount of polygons. Add to that the calculations necessary for lighting, post-processing (focus, motion blur, etc.), and spitting that out into 24 images per second at 4000 pixels wide, and you have something that requires a lot of very expensive computers to eat a lot of expensive electricity for a very long time.
Except it actually is as simple as I made it sound in the paragraph before that. Take a look at this video:
Impressive visuals. If you had never heard of Team Fortress 2, and that video cut off after the first 30 seconds, wouldn’t you imagine that perhaps this was an upcoming Pixar film (albeit more violent than their previous efforts)?
Now, obviously, this is actually a promotional video for a game. Let’s take a look at the game’s graphics…
Obviously not quite as impressive, but you’d be hard pressed to say that the Scout model in the game isn’t the same one in the Pixar-quality promo video. In fact, the line between the two is very blurry — specifically the fact that the promo video has more blur effects based on motion and focus. Well, that’s because it’s absolutely true.
Meet the Scout, rendered in Team Fortress 2, contains a significantly smaller amount of polygons on screen at a time than a Pixar movie like, say, The Incredibles. Can you tell?
On close inspection, you can see that the Team Fortress 2 models are slightly more jagged than Mr. Incredible. However, it’s unlikely that you’d notice this unless you were looking for it in a freeze frame. In addition, Team Fortress 2′s models are actually far less complicated than those from many other video games. Take, for example, BioWare’s Mass Effect:
(Image from Gamespot)
There, slightly more complex models and the jaggedness is solved, but still significantly less intense than what Pixar does. And most blockbuster video games cost things that are considered “indie budgets” to Hollywood: Gears of War cost $10 million, $5 million less than the pretty much indie-produced film Slumdog Millionaire. And as you can see in that article, many games cost “crazy figures” like $30 million, which is still pocket change to Hollywood. Off-season movies like Taken cost somewhere around $25 million, for god’s sake. Now, if you were to take all the work on Gears of War related to getting the thing to run smoothly on an Xbox 360, and channel all that into putting the camera in the right places, pressing “render,” and letting three laptops run for 10 hours to give you a 2 hour movie, I’d imagine it would cost even less.
So now we’ve thoroughly debunked the myth that CGI is inherently expensive. What’s left? A-list voice talent? Yeah, I can see that being up there. And marketing costs a lot. Indie filmmakers would probably replace those two strategies “five of my friends” and “Twitter,” respectively, but I understand Hollywood’s obsession with such luxuries. That can’t account for all $180 million, though, can it?
Well, in addition to Pixar’s overkill in terms of hardware and technical complexity, they employ a gigantic amount of people, paying them salaries and such. That’s probably necessary because of the technical complexity of their animation.
In theory, a team of four passionate geeks who already have MacBook Pros and copies of Maya, Photoshop, and Final Cut could probably make a Pixar-quality film in their spare time. Perhaps it’s the massive, daunting complexity of Pixar films that discourage budding independent filmmakers from trying to compete using more efficient methods.
Or maybe they’re all developing video games instead.
Comments on this post are closed because nobody read the gigantic bold text at the top, and I’m tired of having to deal with this all the time. If you really must insist on explaining to me just how much of an idiot I am, you can scroll all the way up to the top of the page and click on the Contact tab to find out how to send me hate mail.