The first computer I got was a Sinclair ZX81, a friend had gotten the loan of one during the summer and we spent the entire school holiday typing in games from magazines learning how to program BASIC. Sadly the ZX81 didn’t have a sound chip… So when the ZX Spectrum came out a year or so later, I got one… The Spectrum had a single channel beeper (you could play one note at a time). I remember typing up basic programs to play back simple melodies. Some of the games like Manic Miner and Zombie Zombie managed to simulate 2 channel music by playing very short beeps close together. That was all done in Machine code though, there was no way for the end user to manage that without writing the specific code to do it.
The BBC Model B and the Commodore 64 came out just after that, and I had borrowed a BBC over the summer holidays, and had spent time playing with a music program for it. (It was called The Music System). The BBC had the capabilities to play 3 notes at once, but the sounds were square waveforms (beeps), or white noise. I remember spending several days in a row inputting Bach’s Toccata in D Minor into it. The Commodore 64 had a huge edge over the other computers though, while it only had 3 channels, it had the capabillities to play several different types of waveforms.. While there were a few commercial music packages, people like Rob Hubbard (famous C64 composer) had taken the time to write machine code music drivers, this gave them amazing control over the soundchip, allowing them to do fantastic effects like arpeggios and changing waveforms rapidly so you could add a little white noise sound to a bass sound for example. Which could give the listener the effect of having a hihat play (and thus, make it sound much richer than just 3 channels). Once I heard Rob Hubbards early demo’s I was amazed.. and I was determined to do get involved in doing audio on computers.
You have to remember at the time, this was groundbreaking stuff, previously in order to create music of any form, you had to have the ability to play instruments well enough to record them being played. Somewhere to record them (studio), and usually other band members.. The concept of one person on their own creating a full melodic piece themself at home was truly cutting edge and a very exciting prospect.
Anyway, that machine code driver became absolutely key to developing great audio on the C64. As I couldn’t program, I was left using a commercial package called Electrosound, which was very basic. I started creating lots of demos, mostly cover tunes of existing pieces of music.. At the time people would meet weekly at computer clubs, and there was a pre-cursor to todays internet at the time called compunet, where people could share demo’s that they had made over the phone lines.. (think like an old BBS system - except before those).
A friend bet me 5 pounds that I couldnt write my own piece of music, which got me started composing, After that, I started sending all my demos to game companies and pestering them all to use my music.. In time I had a programmer friend program my own music driver for me which really helped ;)
A couple of years later the Commodore Amiga came out, which had 4 channel sample playback, which was great, you could actually have the sound of a real sampled guitar in there. Or if you wanted you could use very small (Chip) single waveform samples.
The Atari ST along with the Amstrad, and Spectrum +3, maybe the Sega Master System, all used similar chips which were very similar sounding to the BBC, so they didnt really break any new ground audio wise. By the time those came out in the late 80’s early 90’s I was working for Imagitec and we had developed our own music drivers across all formats, and they shared a unified data structure.. So I could write the music easily on the Amiga in a tracker program, then simply type the note data into code on the ST, create instruments for it using the driver adding effects like arpeggios and the like, and then be able to hand that data off to a programmer who could implement it on several other systems without much in the way of tweaking.
To some extent we were able to use the same data for Gameboy, NES, SNES, Saturn,Genesis , and even Turbo Graphics. They all just needed some massaging to the instruments to get them to sound good on each platform. Whenever the sound chip offered superior audio features than the amiga, I would try and take advantage of those, like withTop Gear, I added the echo effect, and extra channels to the melodies.
On the PC, the Adlib, and Roland MT32 cards appeared, and the adlib was very similar to the sega genesis.. The MT32 was just amazing.. at last, you could create real sounding music, it didnt sound like traditional video game music did as it was literally its own synthesizer.
I dont remember much about it.. other than it had 8 channels and that was amazing at the time… later on we discovered the SNES ICE development kit had a sound card that was removeable, it was an ISA card, so we were able to take it out of the development kit and put it in a regular PC, we had some software I had designed, MEDIT that we used to compose and we were able to use the same software to develope music for several music platforms at once from the same data.. MT32 Adlib, Snes, etc.. I dont think we ever got it working for the genesis though. When I first started working on Top Gear, I spent the first day or so seeing what the music driver could do. Most of the manuals were in Kanji, so we had no clue what any of it meant, but the data structure was fairly obvious.. After some trial and error I discovered the echo effect the SNES had, which was mind blowing at the time.. it was a real time audio DSP effect.. No other system on the market had anything like that.. but it came at a price.. it used memory for every millisecond of echo you used.. This is why the top gear music had an echo of exactly one 16th note, it was all we could afford memory wise :)
So I imagine you were one of the first professionals to work with this chipset?
I think previously I had done American Gladiators and wheel of fortune, although I cant remember which came first.. I know I was sent to Seattle to convert sound samples to the SNES at Nintendo HQ for a week. I had a massive library of amiga sound samples, and nobody had the development kit to convert them to the SNES in the UK.
-Barry Leitch (http://www.barryleitch.com)
Aquiris Game Studio is a Brazilian independent studio, mainly focused on multi-platform games for midcore audience. We publish our own games and also codevelop some of our projects with a publishing partner (i.e. Cartoon Network).