The Kubrick

I’d like to introduce you all to The Kubrick.  It’s the new gaming system I’m building, named in honor of Kubrick’s 2001 : A Space Odyssey.  I don’t like the way a lot of gaming computers look, an aesthetic appealing to teenage boys without any design sensibilities, so I’m going to house all of this inside the massive 800D from Corsair.


You can see where the name comes from.

Actually, a lot of my components are coming from Corsair.  Having been out of the PC scene for quite a while, I’ve been doing a lot of research and the consensus tends to put Corsair at the top of every ‘Best of’ list, so I’m not worried.

An ASUS Rampage IV Extreme motherboard forms the foundation of the system.  This will allow me to do some fiddling with overclocking fairly easily.  I don’t want to do a full water setup, but I will be using a Corsair H100 to keep the Intel 3960x Sandy Bridge hexacore processor that will be mounted in there cool.

I’m going to start off with 32GB of Dominator GT RAM, in a 4 x 8GB configuration so that I can always add another 32GB without hassle.  I’m going to add the Airflow Pro memory coolers as well, just to make sure all that RAM is being used to it’s best capacity.

Graphics wise, I’m going to try and get my hands on 2 EVGA GTX590 Classifieds, setup in SLI.  Because the cards are dual GPU, it’s like having four cards in there.  I don’t want to play any game on anything less than ‘extreme’ settings for a while.  I’ll also be able to run the system in stereo, should I choose to.

For storage, I’m going to use a Force Series 3 120GB SSD as the system disk, running Windows 7 64bit.  I’ll also install some games to it, anything I plan on playing a lot, like Star Wars: The Old Republic.  Main storage will be on a 2TB Caviar Black from Western Digital.  I’m probably going to mount these in the hot swap bays, but I’m not sure.

Powering all this will be the AX1200 power supply, also from Corsair.  One really nice feature of this for me is the dynamic voltage handling, so I can build it where I am in England, and have it all work right away back in the States.

I might add or swap out some of the fans in the case to increase airflow, but they’ll be simple black fans, not any of the LED nonsense aftermarket fans feature.  I want a nice clean look to The Kubrick, I even chose my optical drive based on it’s clean appearance.  I’m also considering replacing the windowed panel of the 800D with the solid panel available from Corsair’s website, but I won’t be sure until I see how ridiculous the internals look.  Even then, I’m going to wait until I return Stateside permanently.


The Andy Circus

I managed to stir up a tempest in a teapot yesterday on Twitter over Andy Serkis and Tintin, and the role of a performer in the whole ‘performance capture’ thing.

Now, before continuing, a disclaimer.  I’ve worked with Andy Serkis before, when I helped run the virtual capture system on the Tintin prototype.  I both professionally respect and personally like Andy.

I bet a lot of you, given my advocacy of VFX workers and their place in the industry as a whole, are expecting me to say - “However, the push to get Andy Serkis an Oscar nomination is both misguided and ignores the contribution’s of the dozens of animators who helped translate his performance to the the screen”.

Well, you’d be wrong.

I know, I really should be protecting my claim to the role of VFX’s most bitter and jaded asshole, but I think, in the case of performance capture, the principal driver in the performance is the actor.  To deny that is to deny the process of acting itself, the skill and craft of building a character.

If we’re going to talk about acting, then lets talk about acting.  A lot of people think acting is what you see on screen.  Not so.  That is merely the external manifestation of the internal process that truly is acting.  Actors are always asking - either themselves or the director - “What’s my motivation?”.  They don’t just care about what they have to do or say, they want to understand why they are going to do whatever or say whatever they’re going to in a scene.  But they don’t just think about it in terms of a shot, or a scene.  They think about it in terms of the entire arc of their character through a film.  There are actors who inhabit their character at all times - 24/7 - to prepare for a role, such as Daniel Day Lewis.  Others stay in character the entire time they’re on set, or in costume.  And some just go, take their marks, say their lines, and go home.  But those in the last category generally don’t take home Oscars.

Actors who inhabit the role to any extent often have little flaws or character that only they know.  They grip their hands in a certain way because they decided their knuckles were hit by nuns as a kid.  Or they grew up poor and never had sunglasses so they have a permanent squint.  Or if you’re Brad Pitt, you’re always eating in a scene.

It’s these things which add nuance and life to a character.  That make them seem more than a collection of actions to deliver plot points to the viewer.

Counterpoint that to animation, where shots are doled out based on the animators skill to technically execute the shot.  This guy is good at lip-sync, that girl is great with dancing, the new guy gets to do background characters, etc.  Even if there is an animation director who is thinking about the characters journey, he needs to communicate it to others.  And he’s receiving direction from the director, who expects to get exactly what he needs, and not have a discussion about it with the team.  It’s character by committee, not by an individual performer.  I feel this is one of the reasons Frank Oz as Yoda works so much better than a bunch of very talented animators as Yoda.  And yes, I think Frank Oz should have been nominated for Empire Strikes Back.

This is not to say that a character’s performance is entirely the creation of the actor.  The most obvious example of this is the stunt double.  This is such an accepted part of making movies that it’s a big deal when an actor does his own stunts.  Also of note - there is no “Best Stunt” catagory.  There may also be a body double, often used by actresses for nude or sex scenes.  The first time we see the eponymous “Pretty Woman” reaching for her alarm clock after a slow tracking shot over one hell of a bare midriff, her head is covered.  But due to the cinematic language, we make the connection and assume it’s Julia Roberts, when in fact it’s just some faceless actress with a rockin’ midriff.  This is also accepted as part of what we do in making movies.  People double if an actor is uncomfortable with an action, it’s too dangerous, or if it’s a skill they don’t have.  And despite some people’s recent attempts to cash in on the job they were hired to do, part of the deal is you are also paid to shut up about it.

The most subtle way to viewers, but most dramatic way to filmmakers, to alter a performance is through editing.  Shots are pulled from multiple takes, shots are flopped, shots are speed-ramped.  Sometimes a closeup from an entirely different scene is thrown in.  All these things affect how a character is perceived by the audience, and decides what actions are seen and words are heard by the audience, and all by a third party.  This has been happening since Life of an American Fireman, yet we’ve never questioned that the performance of a character belongs to the actor.

Things change once we enter the digital world of motion capture, of course.  We can manipulate the data to create exactly the performance we want.  We can now do tight closeups on our character during stunts.  We can mix and match performances, even within a single shot.  Animators are cleaning up data, interpreting an actors facial performance onto a rig, sometimes creating performances out of whole cloth.  None of which changes the fact that the performance originated from the actor.

Cleaning up the data is just a necessary part of the failings of technology.  We don’t even blink when a paint/roto person is asked to remove flyaway hairs, or a fleck of lint on a character.  This is essentially the same thing- noise in the acquired performance.

Interpreting the actors performance to a rig gets a little wonky, especially if a director decides to change a performance - something I actually don’t think they should be allowed to do.  If an actor has enough juice - your Tom Cruises and Tom Hankses - they don’t really have to listen to the director if they don’t want to.  If Tom Hanks insisted on wearing a beer hat for the entirety of Castaway 2: Castawayier, I’m sure the studio would concede the point.  In the computer of course, we can just delete the hat.

We’re getting to the point of being able to manipulate 2D data in much the same way.  In Star Wars, George Lucas would often take performances of actors from different takes and digitally composite them into a single shot.  There’s even one shot - during the scene when Anakin is talking to the Emperor at the Opera of Glowing Water - where two separate takes of Hayden Christiansen are combined in a single shot through the use of a morph.  I’ve seen eyebrows raised, lips quivered, and smirks neutralized, all in post, and all without the consent or knowledge of an actor.  Yet we never question it’s still their performance.

So, if we really look at what a performance is - the physical manifestations of a character - the question becomes not what the medium that stores that performance is - film, digital video or computer bytes, it’s what the method used to create the performance.  Was it the internal decisions of an individual actor, enhanced through the use of stunt doubles, editors and other cinematic tricks?  Or was it the fabrication of multiple individuals creating separate moments?  If you look at it like that, then it’s clear that an actor doing motion capture obviously deserves recognition as much as any other.


Side note to VFX people - we WANT actors to feel we are on their side when it comes to performances - they are very protective of them.  And performance capture heavy movies create a LOT of jobs in our industry.  The more actors feel comfortable with the medium, the more work we’ll have.


Fixing it in Post

I got into it last night with Steven Kaplan, who represents I.A.T.S.E. Local 839 - a.k.a. The Animation Guild a.k.a TAG. It centered around TAG’s attempt to unionize Sony Imageworks in 2003, when I was there. We had to agree to disagree in the end, but what we both agreed on is VFX is in trouble, especially in terms of labor.

In 2003, when the vote happened, I forget what the actual vote tally was, but it was in the area of 10 ‘Yes’ votes to 500 ‘No’ votes. I was actually embarrassed for them. Now, to people who know me and my views on labor rights for VFX artists, they might find it surprising that I was one of those that voted ‘no’. They might find it even more surprising that I would still vote ‘no’ today.

VFX no more belongs with feature animation than it does with art directors, makeup, or acting. Just because people use the same tools to do their job, doesn’t mean they do the same job. Multiple unions have protected the same people doing the same job for years. SAG and AFTRA have coexisted side by side, though soon they may merge. Perhaps a better example would be SAG and Actors Equity. One represents the needs of film actors, the other actors on the stage. Many actors belong to both. No conflict.

VFX needs something new. VFX needs something of their own. Before we can talk about the future, let’s take a look at how we got to where we are.

Many people think the modern VFX industry began with Star Wars and Industrial Light & Magic in 1975. This is not entirely accurate. When George Lucas went to make Star Wars, he discovered that 20th Century Fox had shut down their visual effects department. Up until then, studios had their own in house divisions to handle VFX. Same with sound, wardrobe, make-up, etc. But with the break up of the studio system during the 70’s, coupled with the type of movies being made at the time, having a visual effects department wasn’t necessary. So George Lucas formed his own group, which would become ILM All this is covered amazingly well in “Industrial Light & Magic: The Art of Special Effects” by Thomas G. Smith and “The Making of Star Wars” by J.W. Rinzler. [CORRECTED] At the same time as Star Wars was being worked on, another group was doing Close Encounters of the Third Kind for Steven Spielberg. I always think of Encounters as an ILM movie, but it wasn’t. It was a group of people working directly for the production.


ILM as we know it today - sort of, but we’ll get to that soon - formed in 1978 to do The Empire Strikes Back. But the real key moment was their second project as the new ILM - Raiders of the Lost Ark. Both Star Wars and Empire were done for 20th Century Fox under the auspices of Lucasfilm. Raiders was still a Lucasfilm production, but for Paramount. They made a few more films for Paramount, before MGM and Universal and suddenly the VFX Vendor System was how we worked. This system is the fundamental flaw to visual effects as a business, though it did not become apparent for many years.

For a while, ILM was pretty much the only game in town. Even if their town was 400 miles north of Hollywood in Marin County. Eventually other companies would form - Boss Film, Dream Quest, et cetera, and the bidding process was born. When a movie required visual effects, the production company - now independent of the studio system - would take their script, storyboards, and other design elements to the visual effects companies.

Knowing the scope of the work and what was expected, companies could give a reasonable estimation of the costs. This system worked fine for years. No one was getting obscenely rich, but people put food on the table, roofs over their heads, and could even put a little towards stuff to do in their free time.

This system came under attack beginning in the early 90’s with the advent of computer generated imagery. Before people jump down my throat about Star Trek 2, TRON, or Young Sherlock Holmes - I’m talking about when Hollywood noticed, with Terminator 2 and Jurassic Park. CGI allowed for things never possible before in film. But that wasn’t what was revolutionary. It was the freedom it allowed filmmakers to improvise on set. With the camera, with the action, with everything. It took a few years for everyone to realize and fully leverage this capability. Don’t like the sky? Change it later. Boom mike in shot? We can paint that out for you. Have no idea what the creature is going to look like? No problem, we don’t need to decide for another 6 months. ‘Fix it in post’ became the rallying cry of lazy filmmakers everywhere. Rather than having a definite plan of what was going to happen, or even a finished script, productions learned to just wing it. How can you effectively bid on what the fire breathing dinosaur will be doing in the third act when it wasn’t in the movie when you signed the contract?

Right around the same time as directors, producers and studios were becoming truly aware of what digital VFX were capable of, a fundamental shift happened, thanks to Microsoft of all people. At the time, CG was split into two distinct toolsets. The high-end software like TDI, Prisms, Alias and Softimage cost tens of thousands of dollars and only ran on high end Silicon Graphics ( SGI ) workstations running UNIX. At the other end of the spectrum was ‘prosumer’ grade software like 3D Studio ( eventually Max ), Imagine, and Topas which ran on DOS. You also had the Video Toaster on Amiga which ran Lightwave. Then Microsoft purchased Softimage with the sole intention of porting it to it’s WindowsNT operating system. I remember in 1996 a bunch of my classmates at SCAD took out loans or had their parents buy them the new ‘low cost’ SGI ‘O2’ - for only $9,000. In a deal worked out by the school, this included PowerAnimator from Alias, which SGI had purchased in response to Microsoft’s Softimage purchase. At the same time I was running 3DSMax at home on my $1,200 PC. I knew where this was headed.

No - Really!

When Softimage was released on NT, it was great. It ran fast. It worked the same as the IRIX version. It even looked like the IRIX version. This was a boon to everyone who wanted to not spend tens of thousands of dollars on SGI hardware. Soon, Alias followed with Maya, first only running on SGI’s custom WindowsNT box. Then, on any hardware. It still cost a pretty penny. In fact, when I won a free copy of Maya in 1999, I was terrified of the tax implications. But those prices soon came down.

Within a blink of an eye, you could have Softimage and Maya running on a system for the price of a decent used car, rather than a new house. Suddenly, lots of smaller and midsize shops appeared on the scene. The cost to startup a facility was now so minor, anyone could do it, and did.

If you need to understand how drastic a shift this was, you just need to look at the headquarters SGI built for themselves in the late 90’s. It’s still there. Only now it’s the global headquarters for Google. They went from building a massive, start of the art facility in the middle of Silicon Valley to not existing within the space of a few years.

Paid for by fake dinosaurs

This affected the price of VFX, of course. Price for anything is determined by a simple equation: ( Materials + Labor ) * profitMargin. Materials can be steel, leather, corn - whatever. For CG, materials are things like software, hardware, electricity, and the building you put it ( and your labor ) in. Facilities like ILM, BOSS and the like had banks of SGI’s and hundreds of licenses of Alias, Softimage, and Prisms. They had teams of custom software engineers and extravagant benefits like ‘health care’ and ‘air conditioning’. They were competing against startups that were thrown together in someones spare bedroom with five guys from work. Sure, they couldn’t do all 800 shots in the movie ( remember when 800 shots was a big show? ) But they could do that one sequence when the hell-pigs attack the kid in the pool! And for half the price ILM was asking for! We started to see shows split amongst multiple facilities. First just two, maybe three. Today, it’s not uncommon for a movie to have nine or ten different vendors. Often working on different elements in the same shots. This gave rise to the production VFX supervisor. One individual who’s job was to oversee all the different VFX vendors and maintain one common vision.

During this decade long explosion in the demand for VFX came a demand for talented VFX labor. Not just animators and modelers, but technicians, programers, and all the other skills that make up a modern VFX facility. And at the beginning, there weren’t that many people that had the combination of artistic skill and technical comprehension to work in VFX. And technical talent was in demand from all the ‘.com’ companies springing up. Salaries skyrocketed. Talented entry level VFX artists could expect anywhere from $50,000 - $75,000 to start. Rockstars - and there definitely were some who thought they were - could demand well over $200,000 before even overtime kicked in and get it.

Magazines like Time and Newsweek noticed, and touted ‘Computer Artist’ as one of the top ten careers young people could go into. Studios and governments also noticed. Studios, seeing the rapid rise in the costs to produce films with VFX audiences were ‘demanding’, looked for any way to lower those costs. Politicians, always eager to claim they created high paying jobs for their constituencies, began offering tax incentives for doing VFX in their states, provinces or countries.

And so the race to the bottom was on.

Facilities began to go under. Warner Digital. Dream Quest ( which had been purchased by Disney and renamed The Secret Lab ). BOSS Film. Large facilities advantage over the midsize facilities was being able to offer VFX never seen before. But the pressure was on from government subsidized overseas facilities like Weta Digital and the London houses like Double Negative, Framestore and The Moving Picture Company who could offer the same groundbreaking VFX at a lower cost due to that wonderful government lobbied for by the studios. ( For full disclosure, I currently work for Double Negative in London. Further, my writing represents only my own personal views ).

American facilities responded the only way they could. People who had spent a decade plus at certain facilities got laid off. Satellite facilities were opened in New Mexico, India, Vancouver and Singapore to take advantage of lower cost labor and government subsidies.

Small and mid-sized facilities closed. The Orphanage. CafeFX. The Asylum. The bread and butter work that sustained facilities like this between hero sequences was now handled by the offshore branches of the major facilities.

And across the board, at facilities large and small, pressure was on with schedules. Ever since ILM pulled off a miracle with ‘War of the Worlds’ - completing the post-production in only twelve weeks, every show is expected to be done on that schedule. So VFXers are routinely expected to work 60, 70, 80+ hours a week. Often, illegally, without compensation for the overtime worked.

So here we are, in 2011. With more VFX than ever, and the state of the industry as dire as ever. VFX driven movies like Avatar, Transformers and Harry Potter regularly add millions to the bottom lines of the studios, yet many VFX artists starting off today struggle to pay off student loans, pay rent in the markets VFX work exists in - London, Los Angeles and the Bay Area are some of the highest cost of living areas in the world. As the VFX industry has become global and driven by cost, artists have had to become nomadic. Los Angeles today, Toronto tomorrow. London next year, stopping off in Vancouver for a month in between. And while that may seem glamourous or even attractive if you are 22 years old, it seems less so when you must say goodbye to dear friends for the fifth time, your spouse wishes to have a career of their own, and your kids want to go to the same school for more than one semester.

American artists fear a union will push what work remains overseas even faster. Europeans live in the same state of denial I saw amongst LA artists five years ago, thinking the work will never go elsewhere. And Weta just gets the government of New Zealand to change the law whenever they fear they might lose work.

We need change. The problem is that any change that makes the lives of VFX workers better makes movies cost more, or less profitable. And that’s something the studios definitely are not in favor of. So when an organization like the Visual Effects Society ( VES ) says ‘the studios are not the enemy’ they are flat out wrong. But because the VES is headed by facility owners, managers and producers who have a vested interest in maintaining a good relationship with those same studios, do you think they wiould ever say that? The other problem is, it’s not the vendors, for the most part, treating artists poorly. It’s the clients ( aka production companies / studios ) putting pressure on the vendors with insane schedules and change orders, all with the promise of future work that may or may not materialize. I don’t need protection from my employer. My employer needs protection from the client.

This post isn’t just complaining. I have ideas as to how we need to proceed forward. Long term, we need to rethink how VFX is done. A model where artists work directly for a production company - paid directly by them, would be ideal. Then, and only then, would they see the true cost of their behavior. But that can’t and won’t happen overnight.

Artists are afraid to speak up because they’ve been made afraid that they are replaceable with the guy who just took some courses at a skillmill like Gnomon or fxPhd. So those that are irreplaceable need to step up. And it’s remarkably easy in concept to do.

A union of VFX supervisors should be formed as a part of IATSE or the Teamsters or, given especially IATSE’s foot dragging and lackluster efforts so far, our own organization. Gain the support of the other creative guilds - SAG, WGA, DGA and AFTRA. Sign a Collective Bargaining Agreement with the AMPTP so that any production done by one of their members has to use a union VFX supervisor. Union supervisors will only work with vendors who abide by a Code of Labor Conduct that starts with the simple demand that artists should be paid for the work they do. Artists need to stop working for free. Any hour worked should be paid, and overtime should be compensated for appropriately. Anywhere in the world, regardless of local laws, as long as it doesn’t violate local laws. Ideally with a worldwide minimum salary for positions.

This won’t solve every issue. THere’s still the issue of illegal trade subsidies. And healthcare. And pensions. And credit placement.

But it’s a start.

Stand together. Or fall apart.

One can only hope we have this kind of solidarity someday.



Nearly five years ago, I began a small project designed to allow me to take animation from Autodesk's Maya and bring it into Luxology's Modo.  At the time, Modo had just released version 201 which included a beautiful GI renderer, but did not support animation- it didn't even have a timeline.

I used some voodoo with blendshapes ( morphmaps in Modo terminology ) to get animation into Modo, even with motion blur after an undocumented feature was revealed to me by Modo's chief architect.  Stuart was so impressed with my efforts he gave me an enourmous hug at Modo's party during Siggraph '06 in Boston.

Admittedly, he was quite drunk at the time.

Some people were using Modomotion, but it was limited in it's usefullness.  I knew this, and I had done it more as to settle a bet with Luxology president Brad Peebler than anything else.

With Modo 301, Luxology added a timeline and support for .MDD files.  This made the very concept of Modomotion more viable.  I rewrote Modomotion to output .mdd files, and released it - again, still for free.

Immediately upon release, the complaining began.  It didn't work the way certain people wanted.  It wasn't fast enough.  I didn't provide enough documentation.

I never intended Modomotion to be a 'product'.  It was something I played around with because I like animating in Maya and like rendering in Modo.  I released it for free online because I could.  I wrote the tool to work the way I like to work.  In fact, my private version of the tool automatically creates my preferred directory structure and outputs files the way I like, but I conceded a little bit to other users.

I wrote Modomotion as a script instead of as a plugin because I wanted it to be availiable to as many people as possible.  I don't think most users at home are running 64Bit Linux flavors of Maya, which is the only way it would be released as a plugin.  To maintain a plugin across multiple versions of maya, across multiple platforms, in both 32bit and 64bit requires way more time and resources than I was willing to devote to a free project.

I decided to rewrite Modomotion in python, simply to learn more about Python in Maya, something that wasn't around before.  During Avatar, I was so engrossed in Motionbuilder, Maya advanced without me paying attention.  So I needed an excuse to do a Maya Python project, and Modomotion 3 was it.

Last year I released the rewrite and an accompanying video showing how it worked.  While Modomotion3 has had well over 2000 downloads, the video has been viewed 135 times.  Obviously, there were people out there downloading it and not watching the video.

The complaining continued.  On message boards, on Twitter, and occasionally someone would locate my website or email address and complain there.  I pretty much ignored it.  As I said, I created Modomotion as basically a hobby and for my own use, and distributed it because, hey, why not.

Then I got three emails within three hours from one individual, a student at a university on the West Coast of the US.  This is important once you take into account timezones.  He was trying to complete an important project, and had done no stress testing.  All his animation was done in a manor which wouldn't work with Modomotion.  He was now trying to export it out so he could render, and obviously it wasn't working.

The first email was asking for help.  It was slightly brusque, but that's fine.  I'm aware that if you send somebody an email you don't know, briefness is often a virtue.

The second email was slightly rude mixed with panic and pleading.

The final email was rude and full of insults, belligerence and even legal threats ( gotta love Americans ).

I got all these emails at the same time when I woke up in the morning.  I hit delete. And I decided I was done with it.  I announced I was pulling Modomotion.  I gave people a week to grab it if they wanted, and now it's gone.

It will not be coming back.

To quote an axiom recited by my friend Wes - 'Give people an inch, and they expect a mile.'

If you need a replacement for Modomotion, I would suggest PointOven .  It costs $99.  With that, you get support and can make all the requests you want.


48 Hours

Out Of Time ( Director's Edit ) from David Stripinis on Vimeo.


I'm also releasing the script under Creative Commons.  Feel free to make your own version, with accreditation to me.

The Script 


The script was written in approximately 90 minutes, so be gentle.