I’ve been quite fortunate with computers – During my early childhood I owned both the 48k ZX Spectrum, and Commodore C64, before finally delving in to the foray of the Commodore Amiga. My first distinctive taste for music finally surfaced when I watched a set of 10 demo disks that came along with the Amiga A500 – contained upon these 880kb behemoths were various demo’s demonstrating the various capabilities of the Amiga A500, and it was this that started my fascination with early computer music composition using software often referred to as “trackers”. The use of music trackers soon developed in to a heaving community of Amiga modders (composers who use 4-tracks of 8-bit audio in tracker software to make music).
When you compare music composition in today’s gaming world it’s rather like comparing Joseph Stalin to Richard Simmons, the similarities are far and wide – Today we enjoy the use of surround-sound capabilities, so-called ‘high definition’ audio, and digital audio encoding, however during the early years things were exceedingly different. This article will draw on my own personal experiences with each of these computer systems, their sound capabilities, and assorted interviews using information gathered from many different sources.
The ZX Spectrum
The ZX was a quirky little machine released in 1982 by Sir Clive Sinclair. It was often praised in the UK for being the first personal home computer. Technically it only had 48k of memory, and a 3.5 Mhz processor (today most computers use dual or quad 2.4 Ghz). The machine had severe limitations, but nurtured programmers who took pride in abusing the hardware to squeeze as much functionality out of it as possible. Whilst the 48k machine didn’t really have the audio power to do anything other than “beep”, it did encourage people to write music for many games (called “beeper music”). It might not be the most impressive of sounds, but for a young kid like me it was an amazing accompaniment to the various games I had lying around at the time.
2 years later a new version of the ZX Spectrum was released with an increased ram capacity of 128k, and included a new sound chip, the memorably named “AY-3-8912″ chip. In fact, there were 3 such chips, but for the sake of avoiding utter confusion the rest of this article I will refer to these chips as “AY”. The 3-channel audio chip wasn’t restricted to the ZX Spectrum range, but can also found within the MSX, Atari ST, and Amstrad computers, allowing very basic sound generation. I say basic, because the waveforms were only capable of generating square waves with a frequency range of 4096 different pitches, and the 3 envelopes for volume shaping were quite rudimentary. However with 3 audio channels and a noise generator (typically for drums) it really opened up a wealth of options for the budding games musician. It is interesting to note too that Yamaha went on to produce a new chip under licence (the YM2149) which was adapted from the AY chips – although this is nothing new, subsequent models went on to be used within the DX/TX synthesizer range which dominated the music industry for over 10 years.
So we now know that the AY chips within the Spectrum could produce 3 channels of square waves at one time, which gave us the ability to make simple tracks, but what if we wanted other waveforms, or the ability to use filters?
The C64
In Popular Culture
Of those that have been listening to the radio recently, you may have recently heard of a band by the name of Crystal Castles.
Although the name takes after the Atari computer game of the same name, the duo used both the SID (C64) and AY/ZM chip (Atari) as a sound generator on a number of popular tracks (Crimewave, Courtship Dating), and is a pleasant surprise that I hope will surface with more bands in the near future – especially as an accompaniment to indie based genres. Their sound is a typical waveform modulated via a squarewave LFO through the wavetable to generate a “telephone” sound used predominantly in the early 90’s Amiga demo scene. Among the chip bleeps you’ll also find vocal sounds written within Yamaha’s Vocaloid (see track Untrust Us).
Whilst the UK was enjoying the apparent boom of the ZX Spectrum market, the C64 went in to production 2 months later and dominated the video gaming industry for over a decade, selling over 30 million units from 1982 – 1994. What made these machines so desirable was cutting edge use of graphics and audio hardware in order to crack the gaming market. This was achieved by separating the processes in to different hardware components – CPU (Central Processing Unit), GPU (Graphical Processing Unit) and SID (Sound Interface Device).
Whilst it’s possible to talk about the subtle hardware differences between the Spectrum and C64, this article is only interested in the SID, and this is where the C64 really shines. The SID chip was designed by engineer Robert Yannes who later went on to co-create Ensoniq, a digital synthesizer company. The chip was originally designed within 5 months due to apparent time constraints (sadly this limitation created a problem with noise-to-signal ratio on the earlier 6581 chips), but the chip went on to fuel the C64 demo scene for many years. Today, companies such as Elektron still release synthesizers and drum machines based on the sound, and ongoing projects like the midiboxSID, or Hard SID use the chips 2nd hand within their own hardware interfaces.
Yannes was inspired as an electronic music hobbyist to devise a computer chip that would achieve all the typical elements of an analogue synthesizer where other computer chips of the day had failed. Again, like the ZX Spectrum, the chip could play three channels at one time, but as with analogue synthesizers you could intermix each of the channels independently or in conjunction of one another to create far more complex sounds. Furthermore, each of the 3 oscillators had a choice of 4 waveforms (triangle, sawtooth, variable pulse, and noise), 3 amplitude modulators, 3 envelope generators, sync capability, ring modulation and even a 12dB filter.
Whilst both the ZX and C64 could provide the user with synthesis generation, neither could play back samples appropriately. Both the ZX and C64 had to rely on the hacking of their noise generators to push sampled audio through, which was achieved by modulating the volume of the generator so fast it would give the ability to play back 4-bit sampled sound. Gritty to say the least!
The Amiga
8-bit Sampling Tip
Interested in retro sampling? A useful tip is to have a look on the internet for “MOD” tracks. MOD’s are essentially Amiga music files where the instrument waveforms are stored within the file and transferred in to memory when loaded in to tracker software such as OctaMED.
Using PC programs such as Awave Studio, it’s quite easy to extract gritty 8-bit beats, leads, and pads in less than 5 minutes. MOD’s were around at a time where sampling was mainstream, and whilst not every sample will be useful, it does give you the opportunity to add an extra layer to existing tracks and will certainly give your track something ‘different’.
The Amiga was first set upon the market in 1985, and aimed to spread itself across the entire spectrum from business applications to video gaming and multimedia workstation. It was regarded by many to be an inexpensive solution to heavier duty machines such as the Apple Mac. In short, the machine sported a custom chipset with advanced graphics and sound capabilities, and a pre-emptive multitasking operating system (Amiga Workbench) long before Microsoft’s Windows 3.0. The Amiga A500 & A4000 computers solidified my interest in both sampling and composition.
The sound chip (nicknamed Paula) took a different approach to the whole business of generating sound. Rather than providing us with typical wavetables (squarewave, sawtooth etc.), it used 4-channels of 8-bit PCM (Pulse Code Modulation) which enabled us to define 4 waveforms for our composition use.
As the Amiga Hardware Reference Manual states, to achieve the production of sound on the Amiga, a waveform has to be represented as a finite string of numbers. This transformation is made by dividing the time axis of the graph of a single waveform into equal segments. Each of the resulting points is called a sample. These samples are stored in memory, and you can play them back at a frequency that you determine. The computer feeds the samples to a digital-to-analogue converter (DAC), which changes them into an analogue voltage waveform.
Confused? Don’t worry – what this essentially meant was that Commodore Amiga gave worldly consumers a taste of 8-bit sampling, where you could choose your own sound definitions by simply recording a sound and setting its loop point/method. When you consider that the EMU SP-1200 cost in the region of £2200 at the time (£1000 today), it does make you realise the level of creativity at your fingertips, especially when you coupled sound recording in to the Amiga with Trackers such as OctaMED. The only downside really was the quality of the DAC’s, and lack of gritty filters such as the SSM filter chips.
By 1989, C64 users felt compelled to remain with the Commodore scene as the C64 dominance waned, and this was soon made apparent by the thriving community of music makers who studied the on-board chip to create fully sampled tunes within software trackers. Games Companies quickly took advantage of this and composers such as Tim Follin, Allister Brimble, David Whittaker, Martin Walker et al. strived to dominate the games market with an extra layer of musical creativity. C64 users weren’t left out, and many sampled wavesforms from the C64 wavetable to continue the foray of chiptunes, making the Amiga one of the most solid communities around until commodore’s demise in 1994.
The PC
Creative & Ensoniq
Remember Robert Yannes? His company Ensoniq, went on to produce many sound cards for the PC market during the late 1980′s, and was eventually bought by Creative Technologies.
Creative in turn had recently bought E-mu (makers of the Emulator series samplers) which ended up with a curious mix of divisions which inherited each others projects.
So Amiga fell in to disrepute, and consumers were left to decide which direction to go in. The PC was at this time coming of age with interchangeable hardware, and the release of Windows 95. Creative Technology looked in to cashing in on the growth of the IT industry, and although they had created sound cards for IBM since 1989 (incidentally with the YM3812 chip) it wasn’t until 1992 when Creative started the 16-bit line of sound cards that we really started to see a big difference in games composition.
The Sound Blaster Pro 16 was essentially two things. It supported 16-bit sound which meant it could play back CD quality audio, and also integrated native support of FM synthesis through Yamaha’s OPL-3 sound chip. Composers could either opt to record their tracks on to CD and playback the CD-Audio recording (NIN, for instance composed audio tracks for Quake), or use the inbuilt FM synthesis via midi programming (this can be seen in many early PC games such as Doom).
Over the years, Creative made many adaptations to their Soundblaster range, introducing extra features to enhance games such as EAX, where games programmers could enhance the atmosphere within a game by using the cards on-board processing power to control aspects such as reverb. In 2001 Creative finally broke in to the 5.1 market, giving software companies the opportunity to add full surround sound capabilities for games.
Of the Future
Today, many composers will render down all tracks to compressed formats. As software abilities grow in power, developers will often opt to programme their own functions to deal with surround and reverb effects. Music however, is still unchanged, and one larger question remains. If we’ve had 16-bit audio since 1992, why (as consumers or music lovers) are we still stuck at 16-bits for music? From 1982 to 1992 computer companies were obsessed with getting as much as they could from technology; however this process of increasing audio quality has ceased for over a decade now. Audio companies such as RME, M-Audio and Digidesign do have products that support 24-bit playback, but for the most part consumer level music is still being rendered at 16-bit, 44100Hz. Furthermore, compression formats such as m4a, mp3, and ogg are taking a step backwards in quality.
It could be argued that hard disk space is a key problem, however terabyte hardrives are now readily available (more than enough to cater for 24-bit 192,000 Hz audio). The film industry should be praised for their push towards clarity in order to enhance end-user experience with DVD-Audio sound quality at 24-bit 96 kHz and support for 5.1 sound. Whilst the gaming industry is starting to take steps on this front, the same could not be said for the music industry, where record labels see no interest in the DVD-Audio format, and instead opt for on-line store compression formats such as m4a, and mp3. Portable devices such as Apple’s iPod could quite easily support 24-bit audio playback, unfortunately they opt to convert audio down to 16-bit’s.
So the question remains, why are music consumers still listening in 16-bit stereo 16 years on?
