pixelatron

This one’s from 2001 or 2002. Factor 5 had just unleashed the GameCube’s astonishing-looking Rogue Leader, so it was the ideal time to talk to them about the making of their first Nintendo 64 Star Wars game, Rogue Squadron. This isn’t the edited feature – it’s the previously unpublished full original transcript with president Julian Eggebrecht.

Julian was always very kind to N64/NGC magazine with interviews. And he helped make Turrican too! What a dude.

How did you get started programming for Nintendo 64?
Julian Eggebrecht: We were working on PlayStation while the Shadows Of The Empire team were working on the N64. There was this whole mystery around the N64, that it was the wonderconsole and it could do everything. We were back in Germany at the time, and we had one discussion with the Shadows team where they filled us in on the machine. They were coming from the SGI world, approaching us from a different view than we would have – we were coming from SNES – and what they told us just didn’t make sense!

We began pushing Lucasarts the whole time that we wanted to do something on N64. We liked carts, for one thing. And we’d had a terrible time switching over to 3D, as did many people. And the thing about the PlayStation that was nasty early on was that you clearly saw Japanese developers got more information about the machine than the Western ones. That was frustrating for European developers like us.

So we pushed heavily for N64. We loved the first level of Shadows, and the next logical step – now we’d mastered 3D finally – was to extend that arcadey level and avoid the rest as were a bit scared of character-based stuff. So we thought let’s do a flight game. We were huge fans of the old vector Star Wars Atari game.

What else helped Rogue Squadron take shape?
In fact, Rogue Squadron came from an old idea to do a sequel to the old Lucasarts classic Rescue on Fractalus on the N64. The problem with fractal technology is you need a strong CPU. The PlayStation had a weak CPU, the N64 at least on paper had this 90Mhz MIPS CPU which seemed to be very powerful. In truth the whole unified memory artchitecture destroyed that, but we weren’t to know that! We liked the idea of doing a fractal-based game, and were working on landscape technology for a long time.

So the Shadows Of The Empire guys went off to work on Episode 1 Racer, and Lucasarts needed another team to do an Star Wars N64 game for Christmas – they said why don’t you turn your fractalus game into a Star Wars game.

The funny thing is, the original idea was to do a ‘best of’ the movies, a ‘greatest hits of IV, V and VI’ – basically exactly what Star Wars: Rogue Leader on Gamecube is. We pitched it to Lucasarts exactly like that at the time, but back then Lucasfilm licensing said no games should be set in the Star Wars movie timeline, you had to concentrate on backstories and sidestories. So our pitch got shot down. That turned out lucky, because we’d never have managed to do those scenes justice on the N64 – Gamecube turned out much better!

What were the early days of the project like?
We sat down to pick apart the Star Wars timeline, and find an area where we could use landscapes and craft. The initial idea to use the Rogue Squadron comic book timeline came from Jon Knoles, the project leader on Shadows Of The Empire. There were lots of original ideas we could use. Funnily enough, we chose to stay completely away from the movies at first – no Luke or Wedge, for example, But Lucasfilm said at one stage, why don’t you use Luke Skywalker? They’d changed their mind completely about using movie stuff. And it fit perfectly – before that we’d had to go to great lengths to write around the movies.

On the Lucasarts side we had the person who’d done most of the first level of Shadows Of The Empire, he helped our programmers. We focused on the landscape early on to get the game up and running quickly, while Mark Haigh-Hutchinson designed the X-Wing and our coders did the rest of the craft. A couple of other Lucasarts guys helped out too, helping get the Star Wars feel right.

How did you find coding on the N64?
Rogue had a sketchy history. It took us a long time to get to grips with the N64 – the version we showed at E3 one year into development was really terrible. The game really came together in the last four months – because we got access to microcode. We’d heard about this from the Shadows Of The Empire guys, who knew what it was (direct access to a coprocessor) but hadn’t touched it. So we knew it’d be perfect for making the landscapes much faster than using the CPU alone. We justified this to Nintendo with a wiritten presentation – saying we’d use microcode to do this and this and this. That convinced them – they gave us access to it, but that was a year in, when we’d already spent 12 months working on the game engine.

We were very ambitious with the sound – we had these grandiose plans for a complete sound system, which became MusyX, which took forever.

What was the first level to be complete?
We had the first Snowspeeder flying around landcape in Feb ’98 – it was getting a bit late. It was really ridiculous. We had no level editor til later, and the March E3 version barely made it. Between E3 and October the whole game together.

We did get good feedback from E3, though, because we showed the X-Wing – the controls were our main focus. Back at the time console games never had accessible space controls, you were controlling in the same way as PC games, which never worked. We needed a flight control that felt halfway like a flight control, but also felt like a car. The X-Wing controls took eight or nine months to work out. The player doesn’t realise what’s going on there – the controls help you quite a bit in terms of flying. It flies easily, as opposed to flight simulation.

What was the feedback like?
The feedback on the landscape was good. We just hadn’t built a game around it yet! That was our big scare – we wanted to do a mission-based flight game which is free-flight, unlike Star Fox. That sounds a nice concept – but to get a structure in there is very tough. All the credit goes to the level designers who pulled it together.

The radar cone was the big thing – you take it for granted that it leads you to the next crucial bit, but that wasn’t in there until July ’98. Until then, everyone was always getting lost. Even on Tatooine, people didn’t get what you had to do, and we were really panicking. We knew we needed a visual clue, but didn’t want pointers or arrows that cluttered up the screen, or take away control from the player. I had the idea for the radar cone after seeing watching a bit of Star Wars where they’re gathered around a table at a hologram display and you see an orange cake-like wedge approaching the Death Star. We decided to take that and use it as the guidance thing.

How was your working relationship with Lucasarts?
We’d been working with Lucasarts for so long, we always had producers, and one of their main jobs is to check in the back of his head whether our stuff will contradict or clash with the movies. But we’d done so much Star Wars stuff, we pretty much knew what we could and couldn’t do. We had to go back and forth to the Lucas ranch, checking everything, not only against what already exists but also what George was planning with the new movies.

Our main idea with Rogue Squadron visuals was to top Shadows Of The Empire bigtime. We knew we could do way more. We wanted to try out to get a PlayStation look on the N64 – we wanted subdued colours, get rid of muddy look. The Star Wars universe lent itself to some tricks on the N64 – everything is grey, so you could circumvert some texture problems by using grey textures, they can be larger than colour textures. We wanted to do the best-looking Star Wars game ever. Shadows Of The Empire was a step in the right direction – but we wanted to get as close to the movies as possible. In terms of visual design, we watched the movies to get stuff looking right

What were your other influences?
We were heavily influenced by some bits of Star Fox, especially the free-roaming bit with the mothership and all the ships swarming around. The cart helped to do free-roaming, as we could drag new bits of landscape off the cart quickly – Rogue Squadron wouldn’t have been possible on PlayStation.

Were there any sticky problems?
Scale is a big problem, we couldn’t keep landscapes the same scale as the movies. Gameplay-wise it’s a problem, as it’s like when you’re flying from point A to point B in a jetplane, it’s terribly boring. You have to muck around with scale. You start out with the best intentions. For example, if you’ve got the A-Wing to scale, it’s so tiny you’re laughing at it. The A-Wing in Rogue Squadron is completely out of scale. But even in the ILM mech paintings, they’re screwing around with scale.

The landscape texturing was a nightmare – now we can do multi-texturing on GameCube, but on N64 it was just one layer, so it was really hard to do convincing landscape textures, really few polygons. We had an R&D phases of three months on this, and there were ten different versions of Tatooine – sand is a nightmare, it’s so boring. The skies were raved about, too, but they were a nightmare. I think there were about ten different approaches to that, because the N64 was very problemantic about doing subtle gradients in sky. That took forever.

How did you achieve the impressive music and sound?
For us, the sound was just the logical technological progression from the Amiga days – we had a sample-based machine, just 4 channels, but we squeezed 20 voices onto the N64. Chris Huelsbeck had been around since the C64 days, doing little orchestral pieces on the Amiga, so when you tell him he had 16 voices to play around with, he’s pleased.

We surprised Nintendo with that – they were very sceptical and didn’t want to use streaming music like in Shadows Of The Empire. Nintendo was getting sceptical about their own sound hardware, we said trust us, and they were convinced.

Why no space battles?
In Rogue Squadron, we didn’t do space battles – everyone was sick of space battles, and there was the whole problem of a point of reference, as  it would add to design time of level. Nobody has yet done a landscape-based SW game, so we focused on that.

How come there was no multiplayer?
The lack of multiplayer was a time thing. We had plans to do multiplayer. First of all it was performance – on framerates below 20 and 15, especially in tiny windows in a split-screen, you can’t control it anymore.

There have been criticisms of the basic AI.
There was practcally was no AI in Rogue Squadron – we started out with high hopes, but it’s a huge hassle to write good AI, it bogged down the CPU. A good example would be how slow Perfect Dark became, largely due to AI. What we ended up with was two or three battles with TIE Interceptors who had specific AI, everything else was pre-scripted.

AI slows things down because ships have to avoid obstacles – every single craft has to calculate a ton of stuff and look ahead. So we ended up with one scene with two free-flight AIs that was barely okay, so we said let’s kick it out. Let’s do some nice pre-scripted stuffthat would draw in player much more.

How did you manage to keep the Naboo Starfighter cheat quiet so long?
None of the cheats were ever submitted to Lucas licensing – it’s so last-minute, they’re not advertised features. The 1968 Buick that you can fly with one cheat was the car belonging to Rudi Stember, our artistic director. We put it in every game now, it’s a cult thing.

The Naboo Starfighter – that was a nightmare to convince Lucasfilm licensing to put it in, they didn’t believe we could hide it properly. In the end, only two people, me and Lucasarts, knew the code. Even the programmers didn’t know. We decided to put it in after seeing it as the first picture of Episode I that Lucas ever released, in the summer of ’98. We were confident no-one would find it – there wasn’t a Gameshark out there at the time. It wasn’t in the PC version either, just to be safe. We did a patch later.

Why did the Stormtroopers explode?
Pure laziness, or more like a lack of time. We were worried that the little stick figures wouldn’t work out, but it turned out you were zooming so quickly over them, it didn’t matter. Lots of stuff came about by luck – the hi-res Expansion Pak option only came in last 6 weeks, we heard that Turok team was doing it, tried it, and it worked fine.

What would you change in retrospect?
The biggest weakness with Rogue Squadron was it had a terrrible terrible learning curve. Because it came together so last-minute, there was not time to get the curve right, there were huge spikes anywhere and everywhere. The medals went from there – originally they were to work out nicely, but they carried over that initial difficulty on some levels and made them ever harder. Take Raid On Sullust – even I cannot play that, I barely get through, let alone get a gold medal.

Did anything get left out?
In the very early design for Rogue Squadron, you could switch craft – that eventually made it into Naboo. The Chickenwalker code is a remnant of that, as you could originally switch to control of an AT-ST in-game.

The bonus levels came together at the last minute. The Death Star Trench Run was the worst – we really thought this would be our last chance to do a trench run level, so we went for it. But the engine couldn’t support it, we couldn’t make the trench long enough, lots of problem. Of course, in the end, we got to do it again in Rogue Leader anyway.

Screenshots from PC version