II:  The Astro-droids of 1977

The first droids

This section documents aspects of the construction of the original R2-D2 droids, which were designed and built between 1975 and 1976. These are often referred to as 1977 robots, since that’s the year that Star Wars/ANH came out.

In other words this chapter focuses solely on the key ANH droids, which were mostly built by Peteric Engineering, with mechanical and electrical work by John Stears’ team at EMI Elstree and some subcontracted work by C&L Developments. The eight fibreglass droids built later for the Empire Strikes Back are discussed in a separate chapter.


When I started writing (the Star Wars screenplay), I found the most intriguing thing was to take two robots and make them into human beings, and make them the most interesting characters in terms of the comedy element. So I wanted to make the film around them, using them as a framework for the film. And I knew when I did that that I would be getting myself into a lot of trouble.

– George Lucas, 1977.


Figuring out the specifics of our friend R2-D2 is difficult. The ANH/1977 robots were famously inconsistent in details, reflecting the limited budgets and time available for their construction. The ones built for ESB were more consistent among themselves, but introduced minor variations from their predecessors. The prequels, sequels, and TV shows introduced more changes. And so there are tons of continuity errors in the movies; often within the same scene. Which droid is “correct” is really up to you!

Kenny R2

For shots where R2 was meant to react and express emotion, the mechanical effects of the 1970s simply couldn't cut it. A human had to be involved to give the robot a genuine sense of character and personality, but the short squat R2 design meant few people could possibly fit.

Children weren’t an option. Not only would a child not be strong enough to handle a solid metal barrel, but working hours would be considerably limited, especially in places like the sweltering North African desert! An extensive search for small actors was made, and eventually Kenny Baker was discovered, apparently by Star Wars production supervisor Robert Watts.

Kenny Baker

When he was cast as R2-D2, 42 year old Kenny Baker had been working for years as a variety performer on the British stage, usually in partnership with Jack Purvis. Billing themselves as the Mini-Tones, the duo had developed a musical comedy act, and after years of touring (they met in 1960) they were hoping to break out nationally via the TV talent show Opportunity Knocks. The Star Wars team had to convince him of the glamour of moviemaking, and promise to employ Purvis so he wouldn't be left in the lurch, to get Baker to sign on.

Kenny Baker as R2-D2. From documentary footage filmed by Peter Shillingford.

In an interview with Cinefantastique published in 1978, Baker said, “...we were convinced stardom was just around the corner. We didnt want to commit ourselves to a film for six months when we could anticipate a lot of lucrative bookings. I turned the film down four times for that reason. Les Dilley, John Barrys assistant, persuaded me to do it. It was obviously a decision he never regretted!

The Mini-Tones: Kenny Baker (left) and Jack Purvis (right). Jack went on to play the chief Jawa in Star Wars, an Ugnaught in Empire, and Teebo the Ewok in Jedi. The duo also starred in Terry Gilliam's Time Bandits in 1981.

At 3'8" or 1.12m tall, Baker was perfect for the task, and R2-D2 was specifically built around his body size. His engaging stage presence, and previous experience playing animals and fantasy creatures in costume, meant he was able to convey R2’s emotions through effective non-verbal movements. He was also strong, and able to handle the heavy R2 suit. One of the problems with casting the role was finding a small person who could cope with the rigours of being locked inside a metal can that was meant to walk!

Kenny Baker, on location in Tunisia. From documentary footage filmed by Peter Shillingford.

Building the Kenny R2 Units

Two of the metal-bodied droids were designed to be operated by Kenny Baker. The photo below shows the interior of an ANH “Kenny droid.

The seat gave Baker the ability to rest his legs between takes, and the harness helped him support the weight of the droid when he walked. Weld lines and spots are easily seen, as well as the circular race that allowed the dome to rotate. A pair of vertical bars at the front gave Baker something to hold onto. He complained about the sharp metal edges inside the droid, and you can really see why!

Note also that the rotating neck ring at the top of the barrel had neatly spaced holes, drilled out to save a miniscule bit of weight. At least one of the Kenny droids is believed to have been done this way.

Working on a Kenny droid at Elstree.

The beige foot, ankle, and battery box on the right were test models produced in fibreglass. They were essentially prototypes to measure fit, and were not the final design. Note also the skirt component that the guy on the left is picking up. This contains two triangular wedge-shaped components, which were an early design for concealing Kenny Baker’s legs. They were replaced in the final droids by the flexible dryer hose.

It’s not easy to see in this photo, but a fun bit of trivia is that the diagonal trough-shaped thing in the foreground is actually the mould used to cast the rubber dianoga tentacle seen in the trash compactor scene.

In the shot below Kenny Baker prepares for a take during the filming of his first scene - the droid auction shot in Tunisia. (not sure why everybody calls it an auction when it was simply a sale...) Note how his legs are angled awkwardly into the feet, and flexible ribbed hose is used to conceal his ankles.

The wires for powering the lights were surprisingly messy. You'd think they would have wound all the wires together into flexible tubing or something to keep them all organized, but apparently not. Baker has commented in interviews how awkward the wires were.

We had all the wires everywhere. And the head was on a track of ball bearings (note: the ESB ones apparently were, but the ANH ones weren't). I used to grab whatever I could grab a hold of, and turn it physically - that way and that way. Not very far, because if I went too far the wires would go around my neck!

– Kenny Baker, 2004

Baker said in a 1978 interview that there were “two switches, one for lights and one for the motors to power the (animated) lights. There was so much buzzing and whirring going on that when Lucas yelled ‘Cut!’ I went on until someone hit me on the head ... to stop me.”

Here's another view of the ankle-concealing grey hoses which resemble the air ducts used on household clothes dryers. These are present on most two-legged “Kenny” R2s. At least one Star Wars Kenny droid seems to have had cream-coloured leather sleeves, and some second-generation Empire Strikes Back models had white leather sleeves for filming in bright coloured rooms such as the medical frigate.

The base of the Kenny droid's feet usually had metal brackets with cylindrical rollers on them. These served to keep Baker's feet from sliding back, but still allowed him to shuffle forwards. According to John Stears in the April 1978 issue of Interface Age, the cylinders contained non-return roller mechanisms that prevented them from rolling back. The rearmost roller looks larger and has a side piece perhaps the ratchet that Baker referred to in an interview. Fans sometimes call these “rollerskate feet.”

I had about 3-4 inches of movement in each heavy leg so the solution was to rock it, which is what I did, and it gave the robot a lot more animation.

– Kenny Baker, Cinefantastique, 1978

Occasionally R2 would shuffle about in two-legged mode, but generally the Kenny Baker droids were reserved for reaction closeups and other emotion-based performances. However, in one brief scene in Return of the Jedi (just after the Ewoks cut him free) R2 is shown rolling forward in two-legged mode. This is kind of an exception, however.

Finally, the Kenny R2s also had springs or shock absorbers built into the legs. This is plainly evident in scenes where R2 is sort of hopping about – you can see the legs compressing and extending at the ankles. It also means that the Kenny R2 was a bit taller when empty than if he were inside. (note: the images below are from a brief shot, in the rebel base, that was flipped horizontally in the final film. This was done intentionally to change the direction of the scene, but of course means that R2 is a mirror image of himself.)

This brings up another point, incidentally. Some people think that Bakers arms extended down into R2-D2s legs. They didn't. There simply wouldnt have been enough space to fit his arms. You can see the deep recesses at the shoulder pivots, for one thing.

Most of the droids built for ANH were never seen again in any film. But this one's an exception. That is indeed an original hero Kenny R2 from the first film, being used to film the Phantom Menace twenty years later. You can see the projecting neck ring pins and the ANH-era foot groove spacing. However, it also features some changes - most notably the legs have been reversed to match the ESB/prequel button sides.

Baker got R2 to move around by shifting his body weight and moving his feet. As noted earlier, there were also metal handles on the interior of the body barrel to make it easier for him to grab and move.

Kenny Baker. 1934-2016.

The radio-controlledRC R2”

For shots where R2 had to roll around on wheels, a single three-legged mechanized droid was built in 1975-76. This unit contained radio-controlled (RC) motors for movement and steering, and third-leg lowering capabilities.

Sadly, things never quite worked out the way everyone wanted. For reasons of time and budget the prop builders werent able to get the RC droid to function properly for the first movie. It couldnt steer reliably, it lacked a head-rotation mechanism, its middle foot failed to lock in place during the droid auction sequence, the non-removable batteries ran down and were hard to charge, and the RC R2 generally just resulted in a lot of heartache for the filmmakers and builders.

Everything was a prototype on the first movie. Like, “Gee, we're going to build this — we have no money, but have to try to make this work .” But nothing really worked.

– George Lucas, Starlog Magazine Issue 48, July 1981

This behind-the-scenes shot, a still from production footage, shows one of the prop technicians, on his knees and disguised unconvincingly in a Jawa's robes, desperately trying to get the three-legged R2 unit to tilt back, lock its legs, and extend its third foot. He eventually gets it to work by shoving it.

The technical problems must have been supremely aggravating for all concerned, but to be fair to the original team, they were being asked to do complex work that nobody had ever done before, and on a very tight budget. Making things worse, finances were heavily delayed by 20th Century Fox, the studio, not wanting to commit funds.

Things started to go wrong. The Artoo unit was not working right, the batteries would run down too early, and they were hard to replace. The kick-down leg didn’t work, and the head didn’t turn right.

– Gary Kurtz

Shooting in Africa with robots, in the sand, is not anyone’s idea of a good time.

— Mark Hamill

The all-metal construction was also extremely heavy, putting considerable strain on the motor mechanism. When it came time to make The Empire Strikes Back, a different company was hired to produce lightweight fibreglass shell R2s, and the metal droids were retired. Kind of a shame, as the originals looked terrific. Especially unfinished – the raw metal has the elegance of an Airstream trailer or an unpainted 1950s Cessna. In fact, years later an all-aluminium droid was used very effectively as a background robot in Rogue One.

The photo above shows the original RC R2 undergoing tests in England prior to painting. They're testing it on a simulated sandy surface, but it wasn't the most accurate test possible, as the sand they used is very shallow. In deep sand or gravel it's impossible for small wheels to gain any traction. In fact, in most of the movie plywood sheets were laid down outdoors to give the robot something smooth to drive on. Any time the droid is outside the studio, and you can't see his feet, he's on plywood. Maybe they should have gone for a man on crutches approach, after all!

There was no way that R2-D2 could walk in the sandy desert scenes as (the surfaces) shift all the time when weight is placed on them. Les (Dilley) got round this by laying boards into the sand so R2-D2 could run along them and, filmed from the right angle, you couldn’t see them. Les spent a lot of his time hiding these grip-boards in every location where R2-D2 was needed to walk along. The desert and salt flats are not ever flat, and are often soft or rocky ground, and R2-D2 was just not equipped to move on anything other than a flat surface.

– Roger Christian, Cinema Alchemist

Filming in Tunisia. This is the scene moments before the chief Jawa (Jack Purvis) leaps up and zaps our robotic hero.

This next shot is from the opening of the film. For some reason the three-legged R2 seen trundling down the Rebel Blockade Runner's corridor has extremely obvious fastener holes on its body. This scene, though appearing at the start of the movie, was actually one of the last to be filmed in the UK before Fox shut down principal photography be for budgetary reasons. And poor old R2 is looking pretty beat up by this point!

The windshield wiper Motors

People have wondered for years what kind of electric motors were used to drive the RC R2. Descriptions from John Stears were typically pretty vague – mentions of “traction motors” and the like. I’ve even heard them described as being motors from an old UK “invalid carriage” (strange three-wheeled mini cars made in the UK from the 50s and 60s) or from powered wheelchairs. But the motors we see in the photos didn’t look like anything that could work in either type of vehicle.

However, I noticed that windshield wiper motors looked pretty close to the photos, with a sort of offset integrated gearbox. And so I and others spent some time trawling the web, but never found anything quite right. Eventually in mid 2022 droid researcher Robert Jackson of OpenR2 hit the jackpot. He discovered a British product that matches photos of the motors perfectly – the Lucas (no relation to George) 17W windshield wiper motor. This was a relatively large permanent-magnet motor designed for use on vans and similar vehicles. The unusual triangular arrangement of bolts on the mounting bracket corresponds exactly to the Stears foot drive drawing.

So it appears that another droid mystery has been solved!

One of the original R2 windshield wiper motors, in situ.

Shoulder catastrophes

In addition to the unreliable middle foot mechanism and lack of head rotation capabilities, another serious technical problem with the three legged robot involved body transitions. It was designed to change shoulder angles when it switched from two to three leg mode, but the shoulder pivots weren't able to handle the weight of the droid! As a result, in three-leg mode the robot's feet spread further and further apart, the legs would splay badly, and eventually R2 would suddenly veer off course and swerve the wrong way.

C&L’s Neil Anderson has said that there were specific plans to beef up the shoulder joints and strengthen the crossbar system, but the rush to get the robots packed off to Tunisia meant that many things got dropped, including this alteration. Which is a shame - the droids may have worked far more reliably if the techs had just another week or two to work out the final details.

The truth is that the robots didn't work at all. ...during the whole shooting of the picture I couldn't get R2 to go more than three feet without running into something. So you'll notice in the film that he moves very little. [...] You get the impression that he moves through the whole movie, but he doesn't. 

– George Lucas, Starlog Magazine Issue 48, July 1981

The outtake above, from the Death Star shoot, shows how badly things could go wrong. But even footage in the final film reveals this problem. Check out this making-of screenshot from the droid auction.

Yep! There's our hunched-over technician from the photo earlier in the page, shuffling sheepishly on his knees towards the multi-armed Treadwell robot. He's trying to look inconspicuous in a Jawa costume, but it doesn't fit him and his bare forearms are visible. Oh, the humiliation!

And theres our hapless tech in the finished film - the picture above is an actual screenshot. Even the Special Edition didn’t paint him out. However, the audience is so focused on R2-D2 in this scene, plus his robes are basically sandcrawler-coloured, so most people miss this amusing detail. Anyway. The splaying leg problem is definitely visible here as well.

The same steering problem happened in the scene where R2 lowers his third leg and heads towards the camera after receiving the Death Star plans from Princess Leia. You can tell he's actually heading towards the projecting bit at the bottom of the wall, and outtakes show him crashing into it. The final movie footage simply cuts away just before that happens. Much of the original film was saved in the edit, and this is a classical example.

This problem necessitated the installation of metal angle brackets, joining the skirt to the front edges of the battery packs. In fact the R2 manuals that John Stears wrote up specifically state, “In the three legged configuration the leg ties must be in position for stable performance, in particular: when executing tight turns.”

Help me, ugly angle bracket. You're my only hope! A closeup view of the three-legged R2.

This is how the foot brackets were used. The first shot below shows a two legged R2 with the third foot just visible below the skirt. Then you can see the third leg has deployed on its release mechanism and the sprung rear legs have tilted back, leaving the feet at an odd angle. Next you can see the droid start to move forward, but its legs start to splay badly. At this point the scene cuts away to a different shot, and the camera is stopped.

Finally the scene cuts back to the robot with angle iron brackets installed, bringing the legs back together and letting our droid move forward in a relatively straight line.

Done with wires

The Death Star shot above is one of the few scenes where the radio-controlled R2 is seen to change direction, albeit crudely. (the auction lineup is another) And the general inability of the three-legged robots to steer reliably lead to all kinds of other production delays. In the end, a host of tricks had to be used just to get shots in the can. We know from interviews with the principals that fishing line (monofilament fibre) or wire was used.

Worried about time, Gary (Kurtz, producer) asked Les (Dilley, art director) if he had any monofilament, and Les went rushing round the back of the (Tatooine) set completely flummoxed, having no idea what monofilament was. Les asked everyone, but no one knew. Someone suggested it might be fishing line, so Les went back and revealed our secret R2-D2 unit, already prepared to be pulled along with the fishing line. “Yes, said Gary, that's what I meant.”

– Roger Christian, Cinema Alchemist

Maxi and Co had dressed me up as usual. Then they abandoned me, brushing out their footprints as they backed away to the very distant camera. At my side, Props had placed an empty Artoo shell on skis, attached to a long piano wire which disappeared into the haze, towards the crew.

– Anthony Daniels.

Daniels is describing, incidentally, the scene in which 3PO and R2 are seen walking away from the escape pod in the Tatooine dunes. If you've ever wondered how they got the droid to roll across loose sand, that's your explanation! They used the fibreglass R5 droid body in this shot. The 4K scan really shows how R2's feet are pushing their way through the sand rather than gliding over it.

The only footage I’ve seen of someone actually hauling a droid along on a string is this photo from behind-the-scenes footage shot in Tunisia. It appears to be a rehearsal of the “these aren't the droids you're looking for scene, since the white and green droid (played by the RC R2) is missing its dome. John Stears, wearing the RC R2’s control unit around his neck, can also be seen in other footage of the same scene.

You can’t really tell in this still image, but in the video you can clearly see there’s a person crouched down by the wall, pulling the yellow droid forward by yanking on fishing line, hand over hand. I've circled the person's hands in this picture.

The comedy shot of C-3P0 and R2-D2 crossing the hallway of the Tantive IV during the gunfight is another shot that appears to have used wire. This behind the scenes photo reveals a faint line, not parallel to the gaps between the floor tiles, leading away from the droid's front foot. This line is not visible behind the foot, suggesting that it is in fact a wire were seeing. Note also that this isn't the actual three-legged R2 again here, but the same droid as above (framed charge bay door; round skirt; long ankle) with an R2 dome.

Two droids roll by a minute later, in another scene aboard the Blockade Runner. This seems odd, as only one motorized three-legged RC R2 is believed to have been built for the original film. The shot involves a pan that shows the end of the corridor, and cuts just before the frontmost R2 crashes into the wall. How was this brief but memorable scene filmed?

Conjecture: the droid in front was the motorized RC R2s body, repainted grey and black and equipped with a fibreglass dome. The droid directly behind it was simply attached to the first by a short length of wire or fishing line, and pulled along. Note that the second droid has a round skirt and no under-shoulder details, so might be the Identities body, thus named because it has spent years touring the globe with the Star Wars exhibition of the same name. Behind-the-scenes photos show that this cream/green droid had a button-back left leg and no RC electronics inside its body barrel.

However, this is just a theory and there's no real proof. They might have drilled tiny holes in the wall of the set near the floor, and pulled both droids along by wires, but the lead trooper crosses over from right to left at the end of the shot, and would have tripped over the wire, so that seems unlikely. And of course no wires or lines can be seen, and repainting the RC R2 seems a somewhat extreme thing to do. So who knows?

Incidentally, what do you think this scene represents? When I first saw it as a kid I assumed that the two droids were simply being taken captive or arrested or whatever by the troopers. I still think that today. But I've heard people assuming that the droids were Imperial droids marching the humans, along with the troopers, to their doom.

Double droids playing one was another trick used on set. In this view of a Mos Eisley street, for example, two droids can clearly be seen. The first was hauled in a straight line from left to right, and then passed out of frame. Close examination of the photo below reveals a screwhead or something on the bottom of the middle foot, where the pulling wire was probably attached. Another version of this photo that I've seen has a faintly visible wire running from the left-hand droid's foot.

Another fun couple of bits of trivia about this set. Although it was meant to be a city street outdoors, it was an elaborate set on an Elstree sound stage. The staircase to the left was actually built in forced perspective, which means it had increasingly narrow steps. The blue sky was a painted backdrop. To the right is the doorway to Docking Bay 94. Notice the stylized 94 carved into the wall!

The second droid, probably the RC R2, then rolled up into frame and moved further down the street. Filmmakers sometimes call this trick a “Texas switch,” though I don’t know why. Unfortunately this illusion never quite worked for timing reasons – if you watch the movie youll notice that the location of the two droids doesn't actually make any sense given their speed! There wasn't enough time for R2 to swing around unless he had suddenly accelerated. Incidentally, the Adywan fan edit of the film solves the problem by showing two droids rolling down the street.

Another point is that the left-hand droid is the hybrid aluminium/fibreglass droid that had probably just played a conehead R4 droid on a different Mos Eisley street. The recesses in the door above the coin slots, the round skirt, and the extended ankles, indicate that it's not the RC R2. It also has odd grey-painted panels on the front, most likely to cover up other colours that had been there before.

Finally, wire or line was also used to tug the motorized droid into a new trajectory as it rolled along. This was probably done for the Death Star sequence where C-3PO fools a stormtrooper. I don't have any photos of this being done during the filming of the OT, but the shot below shows a technician yanking on a piece of wire or fishing line during filming of Episode I. He was attempting to get R2 to change direction and not crash into the wall, with limited success.

Screenshot from Bad Droid Karma, a mini documentary about the production of Episode I.

Could R2 turn his head in 1976?

Well, seems obvious, doesn’t it? Clearly he could - you see it in the film. However, it’s usually the two-legged R2 that you see. Kenny Baker is inside, grabbing the inside of the dome and spinning it around. At least by a certain amount only - the dome wires would threaten to garrotte him if he turned it too far. However, the three-legger is a different story altogether.

Artoo cost us at least a week just because his head wouldn't work. It just wasn't well designed and it was not that well executed in the end. I said I want those heads to work, to turn, and we gave them a lot of time and put on more people during production toward the end when we were really using them - and they never worked.

— George Lucas, in Rinzler

Photos of the interior of the radio-controlled three-leg R2, taken around the time of construction or production, exist. However, no dome-rotation motor mechanism is apparent. The photo below clearly shows piles of servomotors and stuff inside the head, but there's nothing that could turn the head. So, in ANH:

“R2-D2 didn't really function that well. He could run along in the three-legged position, but he couldn't turn his head at the same time.”

– production supervisor Robert Watts , 2004

That said, there are three shots in the actual 1977 film which seem to show a three-legged R2 turning his head. Impossible, you say? Why no - it seems a little creative trickery was required.

Aboard the Blockade Runner

Immediately after C-3PO says “they’re headed in this direction!” there's a closeup on R2’s dome. If you're watching the original cinematic release of the film, or the Despecialized Edition, you'll also notice a slight but noticeable jump or lurch between the moment where his head returns to looking forward and before he starts rolling off. This lurch implies that a number of frames of film were edited out in the middle of the scene. The inadvertent jump cut was digitally fixed by the Special Edition.

You can tell this wasnt the three-legged RC droid, since it's missing the rivets around the manipulator arms that were a feature of the three-legger by this point in the filming. Instead one of two fasteners are just visible above the right-hand (from R2’s point of view) compartment - a hallmark of one of the two-legged Kenny droids.

One of the things that George was upset about was R2-D2 wasn’t working well mechanically on the set. The movement was pretty short because it kept breaking down. The way it was shot mechanically, it couldn’t move that quickly, so you might get just one rotation of the head from right to fit. But just by using a cheap analog device in film printing, we would reverse the print [of his head moving] back and forth a couple of times ... so it would allow us to swivel the head of R2-D2. It’s not human movement, so you could skip-frame it too, so you could speed up the pivoting [of the head]. When we added the voices later, it would cover over the jump cuts.

– Richard Chew, film editor, 2007

Conjecture: It seems this shot was filmed with a two-legged R2 leaning back to resemble a three-legger. It was resting on a wheeled stand or something similar. The head could be manually made to turn, and the droid could then roll away. The jump cut edit is due to some alignment issues between the head turn and the roll. Or else some time passed, and excess footage was edited out.

The body was likely one of the Kenny droids, as suggested by the two rivets. But the dome was the Identities dome used for a number of pickup shots throughout the film, as suggested by certain details on the prop.

Arguing in the desert

In the Tatooine desert, the two droids bicker on the dunes. In this shot R2’s head clearly be seen spinning back and forth in response to C-3PO’s remonstrations. Photos exist showing how part of this scene was filmed. A two-leg R2 was set up with Kenny Baker leaning back inside. His feet rested on an apple box, and the droid was tilted back to resemble the three-legger’s pose. The photo below with producer Gary Kurtz has a number of giveaways that a two-legger was in use – there's no middle foot, the outer feet have the two-leg rollerskates on them, and the side of the body has the two-leg motion restrictors. Baker simply turned the head from inside.

However, other reaction shots were filmed using the RC R2. These were pickup shots filmed later in California's Death Valley (see the next chapter), in order to extend the sequence slightly and to give more of a sense of a conversation between the robots.

I classify these pickups as puppetry, because someone was crouched behind the RC droid, turning the head by hand. A scan of the original 1977 release print shows six frames where a pair of very pink human fingers are visible on the side of the dome! Oops.

Dude! Fingers in frame!

I find it pretty amusing that every single time the film has been shown in history the literal fingerprints of one of the crew have been right up there on the screen for all to see! They're just essentially unnoticeable during normal viewings, because they flash by in a split second, and because the fingers are pretty close in colour to Death Valley’s Mesquite Flat sand dunes.

If you want to see these digits for yourself, pause the movie at C-3PO’s line, “What mission? What are you talking about?" and then hover over the pause button for the R2 head turn that immediately follows.

A moment later another reaction shot shows a couple of frames of something dark blue at the left edge of the dome; possibly a coat sleeve. The sleeve is gone in the Special Edition and later, but the fingers are still there. Even the streaming version on DisneyPlus has the fingers.

This coat sleeve is now missing from the revised digital releases of the film, but was present in the original film prints, as seen here.

Rolling through the arroyo

In the scene where our plucky mechanical buddy nervously trundles through a rocky canyon at dusk, shortly before being zapped by the Jawas, he stops and looks around. This brief moment was filmed as a pickup in California’s Death Valley after principal photography in England and Tunisia was complete. And even though it’s the three-legged droid that was used, you can see head motion at one point.

But look closely and you can also see a pair of dark shadows moving behind his middle leg. Conjecture: one of the young children employed to play Jawas during the Death Valley pickups crouched behind the stationary RC R2, and turned the dome from behind. The human legs seem extremely slender and tiny, suggesting a child and not an adult little person. Note that the droid is never seen rolling during this shot.

Incidentally it almost certainly wasn’t Kenny Baker back there. His work on the film was long complete by that point, and it wouldn't have made any sense to fly him over to the US for a couple seconds of film footage. Also, this was the RC R2, which was packed with electronics and mechanisms, and not an empty droid that he could have operated from inside.

Done with string

One shot - aboard the Death Star, when R2 is plugged into a computer console and our heroes are about to be crushed by the compactor - shows another approach. If you look closely there's a thin piece of fishing line or piano wire on each side of the dome, and that was probably used to turn the head.

Shuffling without Kenny: droid puppetry

A lot of behind the scenes photos exist showing technicians lying ignominiously on the floor or ground, judiciously shoving R2’s feet along – a very simple form of puppeteering, really. There were numerous instances when either Baker was unavailable, or the droid was required to do something that couldn’t easily be done with Baker inside, and this was an expedient solution.

The first shot, of course, is from Empire. But I include it here because it's kind of funny. The glamour of filmmaking...

Note the second photo above - there are large strips of gaffer tape holding the back panel on! Sometimes you see these strips on the bottom edge of the dome as well. This is probably why one of the domes has a large rectangular bare patch of metal that should be blue - the gaffer tape must have ripped it off. I believe that it's John Stears in the red sweater in the last shot of the three above, and he can also be seen pushing R2's feet along in the cantina sequence. He may have been the effects supervisor for the film, but he clearly wasn't above getting down and dirty with the rest of the techs when necessary!

That said, sometimes the motion wasn't all that convincing. The hive of scum and villainy scene features an R2-D2 that's sort of wobbling about in an odd fashion, almost like he's vibrating. The droid on an unevenly-balanced washing machine spin cycle.

This view shows the prosaic truth of the matter. John Stears was lying on the ground, jiggling the fibreglass-body droid with one hand. Note Gary Kurtz in the background, taking a photo.

This is a screengrab from footage broadcast on RTS (Radio Télévision Suisse, the French-language Swiss network, though it was then known as TSR: Télévision suisse romande) in late 1977. Gary Kurtz fields most of the questions via an awkward translator, and Carrie Fisher looks like she's fighting to stay awake. The show contains a bunch of Peter Shillingford's 16mm Tunisian footage that I've not seen elsewhere, including this view.

Finally, the infamous pink shorts photo - beloved across the Internet for showing late boom operator Ken Nightingall (1928-2020) recording on-set audio in Tunisia clad in nothing but gloves, boots, and tight pink shorts - shows how R2 got off the landspeeder. 

In the film R2 magically sort of wobbles his way off the back of the speeder, seemingly unassisted. But once again, the reality was banal. Somebody was kneeling down behind the vehicle, out of frame. And this person simply lowered R2 off the vehicle when required. Kenny Baker wasn't inside the droid at the time. In fact, you can see him relaxing in a folding chair in the upper right corner of the picture.

Non-Kenny two-leggers

Two of the R2 units marked on the Stears diagram were lightweight shells. It's not known for certain if they were metal or fibreglass, but it's presumed that since the Stears diagram was for the metalworkers at Peteric they would have been metal as well. These droids were not intended to be used by Kenny Baker, and couldn't roll around, but were general-purpose effects shells.

One of the shells was marked as being for the Sandcrawler, and the other for the Landspeeder. In the film, R2 is seen being sucked up into the Sandcrawler via a large tube. Since antigravity technology in 1976 wasn't very advanced, and vacuum suction would have been really expensive for something that big, they actually used some simple mechanical tricks to achieve this effect.

In his autobiography Anthony Daniels wrote, I watched as they set up the shot where the Jawas would suck Artoo up into the sandcrawler. I learned that they would actually drop the unit down and then reverse the film.” A fascinating clip in the Disney Gallery: The Mandalorian (the making-of Mandalorian TV series) shows this sequence being set up. You see the droid clearly swinging upwards on wires so thin that they're effectively invisible. Conjecture: it appears that the robot was gradually lowered down on the wires, and then the film reversed and sped up by skip-framing. If you look carefully there's a pair of tubes in the background emitting smoke or steam, and the cloud appears to go rapidly the other way when the robot rises up.

From original documentary footage filmed by Peter Shillingford, recently re-digitized. You never knew R2 could levitate!

The Landspeeder scenes, where R2 is sort of lying on the back of the vehicle, used the lightweight shell tied down with thin wires over the shoulders, as can be seen in the photo below. Note how the middle vents of the R2 are dented inwards; presumably damage from a fall.

The lovely photograph below, which was taken upon completion of the cantina arrival scene (note how the shadows are longer than they are in the film, suggesting the picture was taken later in the afternoon), shows the same lightweight shell R2 that was strapped to the Landspeeder. You can see that it lacks the leg-concealing hoses and rollerskates of the Kenny droid, along with a lot of random details. And the dented-in vents are apparent.

I think the famous photo of Alec Guinness being presented with a birthday cake was shot shortly thereafter on the same day. Incidentally, where did this cake come from? Did they ship it in from England? Was it made locally in Tunisia?

The funny thing is I think this is the same droid shell that was used in the toppling scene, once R2 gets zapped by the Jawas. Here's a shot from that scene, brightened considerably so details are more obvious. The dome looks different from the one above (note the holoprojector isn't in the same exact location), but the body has the same inwards-dented middle vents, the slightly darker covers over the battery box holes, and a small hole where the restraining bolt goes. There are also missing octagon port details. In fact, there shouldn't be any restraining bolt marks here at all, since R2 has yet to be equipped with one. Continuity!

Colour-altered screenshot from the film. The dome for this droid had bolts, rather than simple gravity-reliant pins, so that the head wouldn't fall off during the scene.

Another thing of note is that the shells did not appear to have either Kenny rollerskates or wheels. Instead, the soles of the shell units feet were given textured tracks, suggesting that R2 rolled around on sort of magic tread mechanisms. These are visible in the Jawa photo below.

The Jawas are seen carrying the stunned R2 towards the Sandcrawler, and the lightweight shell with treads is clearly being used. A heavy droid would not have been much fun for the little people and children who played Jawas to carry! However, this shot was not used in the film for technical reasons (though two shots taken shortly thereafter were), and a substitute pickup shot, filmed in Death Valley with a matte painting Sandcrawler, was used instead for the initial approach. For details on this, see chapter III.

Incidentally the picture below shows the effects model used for the brief scene where Luke, Ben, R2 and 3PO enter Mos Eisley. At least in the original release of the film - the shot was replaced with a CGI version in the Special Edition. The model contains plastic action figure toys dressed up as Luke and Ben (Ben is a remade Six Million Dollar Man – they had the technology), and carefully crafted droids.

You can see R2’s magic foot treads in this shot, though they have two halves in the model, where they didn't on the full-size props. Note also the pencil scribe lines weren’t finished on the left side, presumably since the model was meant to be filmed from the right and back only. They even added some pencilled-in circuitry details on the very bottom of the body for fun – ones similar, though not identical, to the ones built into the full-sized Identities droid prop. Amusingly the feet are on backwards.

Finally, one of these shells may have been repainted green for a brief shot in the Massassi Temple, where a copilot droid is being loaded into an X-wing fighter. This droid seems to have a third foot, suggesting it was the one built for the Landspeeder scene. (that way there’d be a middle foot visible rather than an unlikely looking hole) Also the rear power coupling thing has a strangely-angled tube.

Photos in the Japanese Star Wars Chronicles book show the battle-damaged R2 following the attack on the Death Star. This seems to show that the beaten-up and paint-spattered droid used in this sequence was the same one used in the landspeeder scene above. You can see the characteristic dented-in front vents, and the retracted third leg is visible.

How did R2 negotiate stairs?

This is kind of the Dalek problem any rolling robot is going to be stymied by a simple staircase. In R2’s case, he was supposed to have remarkable two-legged capabilities. So, in the scene when our heroes enter Docking Bay 94, R2 is seen bravely clambering his way down a flight of steps.

This behind the scenes clip shows the awkward machinations involved. I really don’t see how Kenny Baker in the R2 suit could possibly have balanced on the stairs like this unassisted.

Conjecture: there must have been a pair of thin wires under each shoulder, holding the weight of the droid, and allowing the empty shell to be operated like a marionette. If you look carefully at the finished footage in the movie, R2 seems to lurch and swing about in an odd kind of way, suggesting that he was suspended. The camera then pans away to the right. Interestingly, a Kenny R2 with leg hoses was used, not the lightweight Landspeeder R2, say.

A second two-legged R2 is then shown trotting into the bay afterwards, and moves up into the frame in another Texas switch. This is clearly Kenny Baker inside a second R2 shell. The funny thing is that he’s actually just walking along normally in this shot, with his legs protruding from below the barrel, as shown below.

You can actually make out the walking motion in the final film, since the dome bobs up and down quite a bit. I suspect the scenes of R2 climbing into the escape pod, and our heroes entering the cantina, were done the same way, since R2’s shoulders can be seen swinging far more dramatically than at other moments in the film, and normally the leg motion restrictors prevent that.

There’s a similar staircase moment in ROTJ when C-3PO and R2 enter Jabba's audience chamber, except one R2 is shown rolling to the top of the stairs, and a second one glides into frame at the bottom.

Lighting

R2 had what appears to be really simple lighting in the dome, but it was actually implemented using very fiddly mechanical means in some cases. The lights had to be turned on manually as they were on their own separate circuits, and the animation mechanisms were very noisy for performer Kenny Baker. They also had to be turned off when not filming, to preserve battery charge.

This diagram from 1977 shows the internal lights. "QI" stands for “quartz iodine”, ie: a quartz halogen tungsten light bulb. Nowadays, of course, you’d use a bright LED or two, but back then the energy-guzzling heat-spewing halogen lamp was the best technology for the job!

The main dome lights were in two basic forms, now known as processor state indicators (PSIs) and logic lights.

The process/processor state indicators (PSI)

“PSI” is the name commonly given by fans to the two recessed round lights on the front and rear of R2’s dome. I've no idea where the “process” name came from, since the lights don't seem to show any processes per se – arguably the logic lights (later in this section) do that. But that's what the droid building community tends to call them, though they’re simply “pulsating lights” in the original droid plans.

The appearance of the lights varies from droid to droid in the case of ANH, and from movie to movie. Sometimes they simply glow, sometimes they pulse (ie: change brightness, dimming and brightening), and sometimes they flash alternating colours. Generally the front PSI shows red and blue light. The rear PSI typically shows yellow and green.

For more in-depth information, have a look at the PSI section in the R2-D2 Detail Catalogue.

A rear PSI from a 1976 droid. Note how beat-up the holoprojector to its right is!

The logic displays

The “logic displays” or “logic lights,” as they're referred to by fans, are the sets of metal-framed light dots which shimmer away on R2's dome. They're supposedly meant to show that his electronic brain is busily processing data, but of course they're simply little animated lights designed to look cool. They’re called “flashing optics panels” on the 1976 blueprints, and were basically like the cheesy fibre-optic spray lights sold by US maker Fantasia in the 1960s and 1970s. R2’s lights nevertheless looked convincingly high tech and computery for the time, though.

John Stears wrote in 1978 that “in order to give Artoo-Detoo a visible change in mood, I used fibre optics.” This is a bit strange – the fibre optic lights really only flicker and shimmer semi-randomly – so I don't know if the concept of robot moods was ever seriously considered. Today, of course, you could set up a bunch of smart LEDs and control them remotely to display different colours and speeds, but that was simply a dream when Star Wars was filmed!

The irregularly-arranged fibres of an ANH dome.

There's another plate visible on the front, though it seems to be held in place by three standoffs, or bolts with spacers. It doesn't look like a round disc, so how the mechanism worked isn't clear.

In terms of how it was meant to function - there is test footage shot in England prior to location photography, showing an early black-painted R2 dome that's probably the original lamp reflector used as R2's head prototype. This has some test fibre optic lighting in it. Funnily enough, the lights twinkle and colour-shift far more effectively on this early test than they ever do in the actual film. Another interesting point is the condition of the dome. It’s really dented and beaten-up, and two marks indicate where a handle was likely removed. This suggests it may be the actual first dome found by Roger Christian which served as a basis for all subsequent R2 domes! I’m not sure about the meaning of the markings, as it obviously wasn’t a 20 watt light bulb in there. Presumably it didn’t refer to power ratings at all, since I think it was a 1000 watt bulb.

An early 1976 lighting test of the fibre-optic array for the rear “logic lights”.

Once the ends of the fibres were glued together into a block they were cut flat, sanded, and polished to form a smooth surface. Some logic lights have a slightly dished interior, however, owing to the resin used to make the fibre optic bundles not hardening flat, thanks to surface tension.

The shot below shows a technician working on an Episode I R2 head. Although it was a prequel film, the dome appears to have been one built for the Original Trilogy at some point. The bundle of logic indicator fibres is visible on the left side. The light would have shone into the assembled cable end.

For more details on the logic lights and how they were done for each film, please have a look at the logic light section of the R2 Detail Catalogue.

Holoprojectors

The R2 domes were all equipped with three “holoprojectors” for delivering flickering, floating 3D recordings, such as the desperate plea from Princess Leia. These were actually passenger cabin reading lamps, salvaged from old Vickers Viscount airplanes by set decorator Roger Christian. The Viscount was a 1950s-era British turboprop that would have been not uncommon in aviation wreckers’ yards in the mid 1970s. These lamp parts are one of a couple of documented “found” greebly types on the entire droid – the other main one being the shoulder buttons.

A Vickers Viscount passenger cabin, showing the overhead reading lamps in situ. A plane from the days when the windows were huge and curtained, and the overhead luggage bins didn't have doors.

Incidentally the grasps or “nozzles” were, in fact, moulded from Bakelite (an early hard plastic, invented in 1907 by Leo Baekeland) and not stamped from metal like the rest of the lamp. This is because the original aircraft lights contained tungsten light bulbs, which obviously got very hot. Bakelite was therefore used to protect the user from fingerburn when adjusting the light.

A pair of original Vickers Viscount passenger reading lamps, courtesy Robert Jackson of OpenR2. It seems there were two iterations of the lamp used on the planes. One version had no flanges, no circular row of tiny holes, and appears to have often been painted green. The other had flanges, a ring of very small ventilation holes, and was painted gloss grey. It's the latter kind, shown here, which matches the appearance of the holoprojectors seen on-screen. Incidentally parts of these lamps were also used, upside-down, on the Millennium Falcon and X-wing fighter dashboards.

The plastic nozzles appear to have been painted an industrial gloss grey at the factory, to approximate the colour of the metal components, but were sometimes painted black on the ANH R2 props. Modern hobbyist-built R2 units often have all-metal holoprojector nozzles, partly because I don't think a lot of people know that the original parts were Bakelite, and partly because a machined aluminium nozzle does look pretty cool.

The frontmost knurled “nozzle on the dome contained a bright light for the hologram sequence in Luke's garage. There doesn't appear to have been a built-in fixture for this lamp - it just seems a bright light was temporarily stuck inside the dome for the duration of the scene. The lamp also had its normal lens replaced with some sort of diffusing material, like paper or a translucent plastic sheet, presumably so that it would photograph better.

The holoprojectors were, on the ANH droids, a source of considerable continuity errors. The ring colour (seen painted black below) varied from dome to dome, scene to scene. All the ESB droids had three identical gloss grey holos. For more details, check out the section on holoprojector colours.

Dyeing the blues

The R2 units used in ANH have a very specific look to them. The white is just plain old paint, but the blue is something of an elusive and mysterious colour. From straight on it looks blue, but from certain angles it appears sort of purple, and at off-axis angles it even looks nearly black.

The finish is also pretty fragile. The droids in ANH have blue panels that are scratched to bare metal in many places, especially around panel edges.

The exact colour process used to make the ANH R2s blue was a mystery for some time, but in the past few years there's been a growing consensus in the R2 building community that it was probably a dye and not a paint. The theory is that layout or marking blue – Dykem or an equivalent UK brand – was used. Astromech member darth_myeek may have been the first to propose the idea. Chris Reiff on the RPF documented his research into layout blue as well.

Layout blue is a special dye, also known as layout fluid, used for scribing metal objects prior to cutting or welding. Technically you apply the blue liquid to a metal surface, where it forms a thin film, and then scratch the line you want to cut into it. As such it's not really intended to be a permanent colour, and indeed it fades to a kind of purple or pale blue when exposed to UV: it's a “fugitive colour. This means it would have required continual touchups during filming. It's also intended to be easily removed, unlike paint that's supposed to stay put.

The three-legged R2 had transparent blue dye over aluminium panels.

Here's a photo of one of the original R2 builders at work. Notice how R2's head has the various cutout positions marked in blue dye prior to cutting. That's layout blue.

The original droids: handcrafted by English engineering professionals in ties, NHS glasses, and woollen jumpers. This man is Jack Swinbank, the metalworking expert at Peteric who was responsible for the bulk of the sheet metal work done on the original ANH droids.

This transparent blue dye makes a certain sense for a prop. When applied to an aluminium surface it yields a lovely intense blue, since the underlying reflective metal can be seen. There can be a real depth to the panels, since the colour varies from light to dark, depending on the angle of light. Colour longevity in 1976 wasn't really a concern, since a film shoot lasts for weeks, not years. They did have to touch up and recoat the blue as the shoot went on, however, so sequences filmed later on do tend to feature darker blue panels, reflecting the build-up of dye.

Unfortunately for most hobbyist R2 builders who aren't making a movie, layout blue is kind of a lousy choice for droid panels. First, the colour isn't lightfast and can turn pale blue or purple, even with UV-blocking clear coats. Second, it scratches off easily. And third, it only really works if all your blue components are made from aluminium. If you have any resin castings or 3D prints or whatever, you can't really use Dykem dye raw, since the parts won’t have that subtle reflectance of metal. Instead you have to paint the part silver first, and that always looks a bit different from a metal part, dyed.

I haven't seen any firm photographic evidence that layout blue was used as the final blue dye on the English R2s, but here are some interesting photos on the topic. First, this appears to be Leslie Dilley, art director for ANH, working on a fibreglass R5. It looks like he might be touching up part of an ankle with engineering blue or something similar. (note how the Series III Land Rover in the background has yellow headlamps. Tunisia was a former French colony, and carried on the French tradition of mandating selective yellow headlamps for many years.)

They appear to be using these bottles. Anyone recognize the products or brands? Presumably they were commonplace in British engineering shops in the 1970s.

Next, we see ILM model maker Steve Gawley at work on the miniature R2-D2 that sat inside the model landspeeder that was used for some effects shots. Note how he’s handpainting the blue details on the legs of his tiny replica, and he’s doing so using Dykem-brand blue. (there’s a spray bottle there, and he's obviously sprayed the dye into a can bottom or paint lid)

Finally, here’s a shot of ILM's Grant McCune (note the early McQuarrie-designed Star Wars logo on his apron!) touching up the blue on the RC R2 before a pickup shot. Note the wooden cage behind him - that's the box that the crew built to truck the R2 props out to Death Valley.

The Dykem is turning purple in the drips on his little container or bottle cap. Now, the latter two photos were taken in California and not England, and after principal photography was complete, but they're still interesting.

Radio-control (RC) technology

The radio-controlled droid's head was a mass of wires and electronics. Back in 1976 remote control systems were fairly limited, and nothing like the digital devices available today.

There was a small nascent market for RC cars and planes, and that basic technology was used to command the RC R2. It was thus possible to have a remote control box equipped with switches and levers. Each control would send simple analogue information (this was a pre digital age) on a specific frequency range determined by a radio crystal. The frequency used would thus dictate an independent channel of information. Servomotors were also used to physically flip switches in order to control other functions.

Photo by Chris Casady.

A set of matching crystals and receivers were built into the R2 body, and small servomotors were hooked up to those receivers. Servos are motors that don't normally go round and round, but which turn in response to commands from the remote unit. On a model plane, for example, a servo might adjust the position of an aileron. Similar servos were used to control R2's motors, and the steering mechanism.

This photo of effects supervisor John Stears struggling with R2's leg deployment shows he was using a Ripmax Futaba (Futaba was the Japanese manufacturer; Ripmax was the UK reseller) M series 6-channel 27 MHz radio transmitter designed for model aircraft.

This was advertised as a “digital proportional unit, though it wasn’t a modern digital system with error correction and compression. It was effectively an analogue system with pulsed AM (not FM) signals. An M6 transmitter cost over £100 back then; equivalent to about £700/$900 today.

Note the blue ribbon on the antenna. That was a pennant used to colour-code channels between devices, so operators could remember what device controlled what. Blue suggests he was using a 27.255 MHz crystal, or channel 6.

Other photos and film footage of Stears operating RC droids, including the Treadwell, always show a Futaba Ripmax M-series control box. Behind the scenes footage suggests the left lever was used to control the leg descent mechanism, and the right lever was used for movement and steering.

Note for completists: if you want to buy a Futaba Ripmax M6 yourself, note that the ones used to make Star Wars were mid-period units. They had black trainer buttons (back upper right side; clearly visible below) versus the silver trainer buttons used in earlier models. And they had blue volt meters, versus the green/red meters used in later models.

John Stears operating the RC R2 in Tunisia. Closeup shot from documentary footage by Peter Shillingford, showing an M6 with a blue battery meter.

Sadly the radio control technology of the day was not eminently reliable, as the filmmakers discovered to their frustration. The modern era of digital error detection and correction technology, that has made RC tech considerably more dependable, was many years to come.

Stears did himself no favours by constantly blaming the remote control problems on interference from taxi radios - even in the middle of the Tunisian desert! (an anecdote related by a number of people who worked on the production) Though to be fair, when they were filming in the UK the 27 MHz range was extremely susceptible to interference from then-illegal CB radios imported from the US, which is why later UK model aircraft standardized on 35 MHz. Stears also blamed "radio stations" and the desert ground conditions, mica in the sand or whatever, though whether all that has any actual basis in engineering fact or was just him bullshitting, I don't know!

Personally I think a more likely theory was the fact that the RC R2 was constructed from solid aluminium. Even if the servo receiver antenna was removed and taped to the outside of the body, as we can see in some behind the scenes footage, the whole thing still functioned as a Faraday cage that blocked radio transmission. The fibreglass droids built for Empire no doubt had better reception for this reason.

When the RC R2 was shipped to the US for pickup shoots the UK-licensed equipment was removed and replaced with US frequency gear. Equivalent Futaba stuff was used - so an FP-T6FN transmitter and an FP-R6F receiver were installed by William Shourt of ILM.

The voice.

An utterly essential, and indeed transformative, element of R2-D2's design was the voice. Without his plaintive beeps and cries, R2 simply wouldn't have come across as effectively as a truly empathy-inducing character.

INT. REBEL SPACEFIGHTER - MAIN PASSAGEWAY

The little dwarf robot makes a series of electronic sounds that only another robot could understand.

For his remarkable work, Star Wars' sound designer Ben Burtt deserves full credit. From inquisitive beeps, to lonesome whistles, to derisive farts, to pathetic whimpers, R2's voice is unmistakable.

R2-D2 was probably the most difficult voice to work on, because it was the most abstract. Here we had supposedly a machine that was going to talk, that was going to act, that was going to draw on our emotions.

– Ben Burtt

Ben Burtt, at some point in the mid to late 1970s and probably working on Empire, surrounded by open reel tape machines and wielding a parabolic microphone. A lot of the field audio for Star Wars and Empire was recorded with portable Nagra open reel machines, but Burtt seems to have a Sony Densuke-type/TC-D5 cassette recorder slung over his shoulder.

Burtt, like nearly everyone on the production team for the first film, was a young man at the time; just out of film school at USC. In fact, he was recommended to producer Gary Kurtz by his professors at the school.

He worked with Lucas to define the sound of the film, focusing on real-world sources and collecting noises even before filming began. At the time – the early 1970s – science fiction films tended to go for highly synthetic and electronic sounds to convey the other-worldliness of the images on screen. Burtt and Lucas rejected this approach, and instead worked with original field recordings of real vehicles, animals, and machines.

Blaster fire, for example, was created by tapping a radio tower's guy wire to get that characteristic "pew pew" sound. The memorable sound of the light sabre was created by mixing various electric humming noises together (an unshielded mike cable next to a TV; an idling film projector), then synchronizing the sounds to the on-screen blade motion by moving a microphone mounted on a stick. These real organic sounds lent a subliminal verisimilitude to the film. In fact, this emphasis on creating a realistic soundscape was one of the many ways in which Star Wars broke new ground in filmmaking.

Ben Burtt, at work in his editing suite. The Darth Vader inflatable toy, and the Joe Johnston sketches of Boba Fett on the wall, suggest that the photo was likely taken between 1978 and 1979 during pre-production of the Empire Strikes Back. It seems this was probably the basement of George Lucas' house at 52 Park Way in San Anselmo, prior to the move to 321 San Anselmo Avenue.

With R2, however, some electronic aspect was required, and so the robot's voice wasn't created from found sound. Instead, Burtt used an analogue synthesizer - an ARP 2600 - and mixed and spliced the bleeps together with his own voice, making baby-like sounds to convey R2's personality.

The voice of R2-D2 was the hardest sound effect I ever had to make, because there was really no precedent for it. Most robots in movies spoke English, or they spoke some language that you understood. They had a moving mouth, so you had some idea (when they were speaking). The script of Star Wars just said that R2 made electronic sounds, and didn't really specify exactly when R2 would talk, or the quality of the noise. It took me six months to figure that out. And I was really nervous, because R2 had to talk and act with Alec Guinness, of all things. Yet you couldn't understand a word that R2 would say, but you had to understand the meaning behind it.

It ended up being a problem solved by combining a keyboard - an ARP synthesizer, which I had at the time, which made electronic sounds, and my voice, making funny little R2-like noises. And performing them together, and spending endless hours editing and clipping little snippets out of that sound, stringing them together to give them sense of meaning.

– Ben Burtt, 2008

The ARP 2600 was an interesting choice of synthesizers. Released in the early 1970s by American manufacturer ARP Instruments, it was fairly complicated to master and a bit unstable in use. It had three VCOs (voltage controlled oscillators) and was monophonic - it could only play one note at a time. But it also had a lot of knobs and sliders that would allow you to alter pitch and tone to simulate inflection in real time.

Critically, it also had an “envelope follower that you could plug a microphone into. This innovative feature would allow you to make sounds into the mike, and have the synthesizer respond in time. In other words it would take incoming amplitude from the microphone and convert it to a varying voltage which could be used to control the behaviour of the various sound-generating parts of the synthesizer.

A lot of people have attempted to make similar "robot" beeps over the years, even using the same equipment. Most sound nothing at all like R2. It really was Burtt's ingenious use of his own voice, and his audio artistry, that made R2 into an actual and believable character.

The first effect that Ben Burtt laid in was the exclamation of Artoo when he got hit by that Jawa shot and fell over, and it just broke me up. The film really came to life.

- Richard Chew, film editor

Not to be forgotten are the motor whirs that R2 and C-3PO made when they moved. This must have been tedious to edit into the soundtrack - 3PO's many head and arm motions having to be accounted for - but were essential to the effect. After all, the R2 prop was just a metal can that made buzzing noises for the dome lights to work. C-3PO's mostly fibreglass costume would make clattering noises as Anthony Daniels walked. These on-set sounds had to be eliminated and synchronized sound effects laid in over top.

R2-D2’s motors covers every single move it does. They got buried most of the time, but when they do surface it helps keep a consistent texture that tells you that it really is a robot.

– Ben Burtt in Film Sound Today

Star Wars losing to Annie Hall for 1978's Best Picture Oscar was probably a mistake, but Ben Burtt's Special Achievement Award for Best Sound Editing definitely wasn't.

We thought it would never work. We thought that audiences wouldn't understand R2 at all, and be lost. But boy, were we proven wrong! Now it just seems part of the culture, so I'm thankful for that.

– Ben Burtt, 2008

Artoo's reply is a rather rude sound. He turns and trudges off in the direction of the towering mesas.


On to part III: After Elstree

Copyright

This text was written entirely by and for 3Dsf.info. Feel free to make copies for your own use, but I ask that you not repost it for download elsewhere. The reason is I'm updating these pages all the time for accuracy and development purposes. So the most up to date page should always be available at 3Dsf.info!

The majority of the photos are copyright their respective owners. They're either the property of Lucasfilm or of various people who have posted images online; specific provenance is mostly not known. They're reproduced here for the purpose of criticism and research.

Contact

If you have any corrections or comments, feel free to drop a line:

contact@3dsf.info