The Making Of: Blade Runner

The Making Of Blade Runner

How do you sum up Blade Runner’s influence on videogames without something along the lines of ‘Let there be light’? Bringing together four masters of speculative science fiction – author Philip K Dick, illustrator Syd Mead, filmmaker Ridley Scott and special effects pioneer Douglas Trumbull – this bleak film noir unknowingly landscaped vast regions of today’s virtual worlds.

Its vision of a near-future Los Angeles, a society pushed to the limits of architecture, industry, technology and population, remains a touchstone for designers whatever their trade. And as 2019 approaches looking nothing like it, it’s not relief that we feel but regret; we’ve got its problems of isolation, privatisation and the demise of the natural world, yet we’ve failed to invent the flying cars, sky-piercing towers and offworld colonies. No wonder our artists choose, with apparently endless enthusiasm, to build such futures for themselves.

A commercial slow burner, Blade Runner’s imagination stretched far beyond the reaches of the Hollywood system, its bickering creators and the confines of the silver screen. It was an anarchic, accidental, colossal undertaking with backstage drama to rival any of its 116 minutes. Or 113, 114, 117 minutes – however many its different versions, each considered definitive by someone or other, happen to take up. “I wouldn’t say it’s an ensemble piece,” suggests Louis Castle, co-founder of the now defunct Westwood Studios, “I’d just call it ‘screwed up’.”

He should know. When the Blade Runner Partnership, having gone through EA, Sierra and Mediagenic (aka Activision), approached Virgin with plans for an official game, the complexities of the licence had yet to be revealed by Paul M Sammon’s essential document, Future Noir: The Making Of Blade Runner. Scott’s movie just swallowed money, exhausting investors and secondary investors before ending up with The Ladd Company, run by media mogul Sir Run Run Shaw, and Tandem Productions’ Bud Yorkin and Jerry Perenchio. True to form, it bled even them dry, all the way down to favours and promises.

“Ridley had to sign lots of deals with actors for residual rights,” explains Castle of his favourite movie. “The entanglements were very deep in the sense that different people had different deals. And worse yet, the production ended on a sour note and many of the records were lost. So no one knows who might come out of the woodwork claiming ownership of the film, or even pieces of it. We were prohibited from using any footage or audio because we’d never know whose rights we might be trampling on.”

Yet this dire situation is what, in a roundabout way, landed Westwood the job. Before any of it emerged, Castle’s pitch already described a very rare kind of tie-in that would “add value rather than borrow it.” You’d play a character within the Blade Runner world while the movie’s story unfolded. You’d be part of it and able to affect lots of things, but nothing already captured on film. “You’d be behind the camera, amid things left unexplained, wallowing in the mythos,” says Castle.

Westwood knew that Yorkin’s partnership was on a strict timeline. It didn’t know why – that this was the latest of many negotiations with publishers and studios – but had to deliver its pitch in little over a week.

“We put together a demo of the first few minutes of the film, from the point where the Spinner flies in and lands,” recalls Castle. “We crunched all weekend and I went in there with our pitch. I went down the list: you’ll be a character who won’t know if they’re Replicant or human, and the game is centred on the fact that you’re never quite sure. As you play through, you should always feel you’re successful no matter which route you choose, so it never sends you back to start over. At the end, you should be able to finish with very different expectations of what and who you are next to someone else who’s played the game.”

The clincher: “’That’s our story – now let’s show you what it looks like.’ And we played the video, which used a compression technology we’d made for the original Command & Conquer games that offered full-screen, high-resolution [640×480 pixels at the time], full colour graphics. Bud said: ‘I was going to ask what you’d use from the film; now I need to ask why you’d bother.’” Every scene in the game, it was confirmed, would be generated from scratch in 3D.

It wasn’t until months later, when the game was actually announced, that Westwood learned just how many studios had been approached to make it. Some even thought they were still very much in the running. “A bunch of folks called and said: ‘I was working on that!’ I didn’t even know we were competing,” insists Castle. “But I guess it came down to the unique approach. In my discussion of what is and isn’t appropriate for a property, I’d made a lot about how just running around shooting Replicants would be very unsatisfying for those who like the Blade Runner ethos. I think that’s what convinced them, together with the quality of the visuals at such short notice.”

Unable to use footage or even the beloved Vangelis soundtrack, Westwood could at least use people from the movie, though it pretty much ruled out two of them from the get-go. “I guess Ridley had a pretty bad falling out; we weren’t able to get hold of him or Harrison Ford,” says Castle. “I don’t know if you know Harrison Ford’s position on the games industry, but he’s of the mind that it’s taking liberties with licences, so he’s firmly against it in many ways. He’s disenchanted. So I don’t know if we failed to get hold of him or just couldn’t get past that bias. But virtually everyone else on the film we did get to work with.”

Contributions from Syd Mead added to a wealth of concepts for LA 2019. (When asked for materials for this feature, Mead’s agents insisted his file on the game was slim, his work on it short.) But despite “executing it faithfully,” the team soon realised that something wasn’t right. The onscreen Blade Runner, after all, is quite some departure from Mead’s perfectly conceived drawings. The new version was faithful, all right, but to the wrong source material.

“So, I went and found the guy in charge of doing the sets for the film – not the designer but the engineer,” says Castle. “We hired him as a consultant and said: ‘Look, here are the concept drawings of the set for, let’s say Deckard’s office or home, and here’s what’s in the film. How did you get from there to there?’

“He said: ‘Well, we looked at Syd Mead’s stuff and said we’d love to do it, but we don’t have a million dollars to build each set.

So we just went to the scrapyard, these props rooms, and grabbed anything similar and just bolted it on, spraypainted it and whatever we had to do to get it as close as we could.’ With that revelation, I went back to our 3D artists and said: ‘Look, you have access to these 3D libraries with all this stuff you can use, and you’re no longer permitted to make anything from scratch. You can only cull things and modify them; you can cut it, repaint it and scale it, but you can’t rebuild.’ Using that same discipline, we got a look in the game that felt very, very close.”

Extraordinarily so, in fact, for a game released while 3D hardware was still finding its feet in the PC market. Just a small handful of sets were used to make the movie, recycled and positioned to look like many more. The game had to build 100, many open to player exploration, “with richness and density in the millions of polygons. Graphics cards at the time could manage, at best, a couple of hundred polygons per frame. Even the highest-end graphics card would have had a terrible time reproducing one character, much less these rich environments with all their lighting effects.”

So Westwood built a new technology based on one of gaming’s timeless curios, voxels. Used to make units that could scale and rotate in Command & Conquer: Tiberian Sun, Blade Runner married them to another bleeding-edge invention. In the mid-’90s, a game that attempted 1,000 motion-captured sequences would be rightly considered a dangerous folly; Blade Runner features 20,000.

Each of its scenes is essentially a miniature movie, and sees visual and geometric data compressed and uncompressed on a frame-by-frame basis. Opening the door to a primitive yet effective form of dynamic lighting, this lets the game’s voxel objects – the characters and flying Spinner cars – add light to an otherwise prerendered scene. It’s a form of deferred rendering, effectively, introduced years before games like Killzone 2 popularised it.

“But where the technology let us down was in the time and energy needed to process the motion-capture,” says Castle. “We were capturing at 60 frames per second, and the characters in voxel space took up so much that we had to go through them, removing those frames by hand using keyframing.” The original voxel objects weighed in at around seven megabytes in size while hardware at the time could manage only about 100K. Automatic optimisation would leave them unbearably blocky, leaving Westwood to do it by hand. Naturally, the studio didn’t have time for 20,000 sequences, which is why some of the game’s objects look plain bizarre. Castle sighs: “It just took too long.”

Was dedicated hardware ever considered? “People ask me that,” he replies. “’Why didn’t you just use 3D?’ Well, that’s easy to say now we have full rendering pipelines and shaders, but back then you just didn’t. You had a few vertex-shaded polygons and the texel mapping just wasn’t accurate on a lot of the cards. Nowadays I’d do it all in 3D. But I gotta tell you, the time’s coming when voxels and raytracing will come back into the fold because the rendering pipelines are so complex that we’re flipping back to computational load, a pixel-by-pixel solution. All technology ebbs and flows.”

Which is more than can be said for game design methodology. Beyond something like a Police Quest or the odd RPG/point-and-click hybrid – Microprose’s Bloodnet, perhaps, or Esprit Software’s BAT – few other games play anything like Blade Runner. Hopping between on-site and menu-driven detective work, making occasional use of the movie’s Voight-Kampff and ESPER machines and even rarer use of the gun, it knows the darkest secret of adventure gaming – that if you want to be really interesting, sometimes you have to be really quite boring.

“Blade Runner creates a very specific emotional experience – not one of horror or action-based fear but one of terror,” says Castle. “There’s actually very little action in that film, but when it happens it’s violent, explosive and deadly. I wanted to make a game where the uncertainty of what’s going to happen makes you quiver with anticipation every time you click the mouse. Things could go horribly wrong but most of the time they don’t; most of the time it’s simple.”

Indeed, much of your time as hero Ray McCoy is spent in point-and-click limbo, visiting and revisiting people and places, wondering which overlooked action will open the next chapter. Most of your shots are fired in a training room as part of a routine that includes trading clues with a police mainframe, travelling in lifts, feeding a dog, travelling in more lifts and occasionally getting some sleep. “You had to be doing something that was distracting you from the physical, so when the violence came you were shocked,” says Castle, before describing how he’d apply the same principle to Scott’s other sci-fi classic, Alien: “Most of the time you’d be dealing with the problems of the Nostromo, how to fix it and who the android is. But when the alien appeared, by God you’d have to fight like a devil.”

Who the androids are is, of course, the crux of a Blade Runner story, especially when looking in the mirror. Westwood didn’t want to spend four CDs’ worth of time and money on a game you only played once, so the problem became one of recycling uncertainty. How do you solve the investigation, expose the fugitives and read the script without ripping the game’s heart out? Simple, thought Castle – you make a ‘story simulator’ instead, a script that even the writers can’t predict. “You’d think I’d have learned by now,” he laughs, “not to be seduced by these easy-to-express, difficult-to-execute ideas. A living, breathing city that adapts to what you do? That’s very easily said.”

In early prototypes, all of the Replicant identities were randomly assigned, but this proved baffling for playtesters. Have you ever retired a human by mistake? Yes, they answered, after bungling the investigation at the first opportunity. Making the first and last suspects – meathead chef Zuben and the majestic, psychotic Clovis – always be Replicants helped ground the game’s story, but still couldn’t save a valiant QA department from a reported 2,500 playthroughs. The game’s script, penned by Bud Yorkin’s son David and lead designer David Leary, had swollen from 80 to 500 pages.

Castle recalls how Yorkin’s first draft, an attempted videogame script, just didn’t work. Instead, he was asked to write a movie script for a hypothetical Blade Runner sequel. This gave Westwood a cast of characters so in keeping with the movie’s bitterness that they’re now considered canon by fans. There’s animal counterfeiter Runciter, troubled teenager and potential Replicant Lucy, hotshot Blade Runner Crystal Steele, and a new band of escaped Replicants, some looking for revenge, others simply looking for a life.

“David [Yorkin] might have one line in his dialogue that read: ‘The brownstones looked particularly out of place because they were lost in the 1950s.’ So we’d research all that stuff and build a set based on that one observation,” recalls Castle. And where the locations were too close to warrant cutscene transitions, more were built to connect them. “We did a lot of location stuff in LA, on the sets where they’d done things like the Bradbury building. We included a lot more of that building than was seen in the actual film.”

Blade Runner had, fittingly, become a colossal undertaking of its own. Castle admits: “It took longer than expected and cost more than we wanted, and there were poignant moments where we said: ‘OK, can we reduce the scope?’ But I’d say that sheer tenacity and a desire to do it right got us through those things. Before we even started we had beautiful work, and as the game came together looking just like it, people were so enamoured that we could have just kept going and gotten any amount of money. Time was our enemy; we just couldn’t get to everything we wanted.”

It was a familiar outcome sadly never followed by a ‘final’ or even ‘director’s’ cut. A ‘designer’s’ cut, yes, but that was just a menu option to trim the game’s dialogue. There were “deep discussions” about a sequel, and the game’s success – it outsold 1997’s The Curse Of Monkey Island by three to one, claims Castle, shifting over a million copies – made it seem inevitable. “But even with that kind of volume, the mere fact it was four CDs made it a very expensive game. And the deal we had with the Blade Runner Partnership meant it was not terribly profitable. It didn’t do as well as you might think.”

Looking simply at gross sales and ignoring the costs of making, much less publishing such an epic, the Blade Runner Partnership wanted an ever bigger slice of its successor. In the end, Castle himself pulled the plug. “It was untenable, too great a risk. We were under the impression they’d get someone else to do it, but I gather they couldn’t find anyone to go along with the terms they thought were fair.” Want proof? Ask Gearbox.

Westwood, headed up by Louis Castle (front, centre), was liquidated by EA in 2003. Some staff moved to EA LA, while others formed Petroglyph and Jet Set Games.

In 2009, the Borderlands creator described how it was offered the licence, calculated it would cost $35m to develop and that it would never make the money back. It would be a game, agrees Castle, to “challenge anybody’s sense of a reasonable business rubric. You’re not going to succeed unless you do it extraordinarily well, and that won’t happen without an extraordinarily good team with an extraordinary amount of money. And then you have a partnership that demands a giant chunk because it’s not their job to take risks. Put that in a recipe and bake it and you don’t end up with cake.”

Tragically, and despite the resurrections of games like Beneath A Steel Sky and The Secret Of Monkey Island, there won’t even be a remake. To restore almost a terabyte of assets, whether for new prerendered backdrops or full realtime 3D, would itself cost tens of millions of dollars. What’s worse, you’d have to find them to start with, which might well require a time machine. Stored on magnetic platters, most of Blade Runner vanished when Westwood was liquidated by EA in 2003. It’s lost, as a mad Dutchman might say, like tears in rain.

This article was originally published in E222, November 2010.