The idea is sound enough, but does the world need another Commodore-based sampling system? After looking at the £300 Microvox, we conclude that it does.

While dedicated samplers of the keyboard and rack-mounting varieties seem to be multiplying faster than you can say 'n-n-nineteen to the dozen', computer-based samplers remain few and far between.

That's not too surprising. Many computer-owning musicians are more interested in using their machine for sequencing, while the data-intensive nature of sampling doesn't sit too well with the limited memory of today's eight-bit home computers.

None of this has daunted British software house Supersoft, who recently brought out the Microvox Digital Sound Editor for the ever-popular Commodore 64. It's an impressive system, more versatile than Microsound's sampling package for the same micro (reviewed in E&MM June '85), and it represents an affordable introduction to sampling for those who can't contemplate buying the likes of an Akai S612 or a Mirage. It's playable via MIDI, so you can use it from a MIDI keyboard, sequencer or drum machine. But Supersoft have also considered non-MIDI musicians (known as 'aliens' in the trade); samples may be played over a three- to four-octave range from the 64's QWERTY keyboard, and this option has been made compatible with Commodore's overlay music keyboard, too.

But before you get too excited, I'll point out that unlike the above-mentioned dedicated instruments, the Microvox doesn't allow polyphonic playback of samples. What it does allow you to do is store as many as 16 samples in memory, and assign individual samples to any notes over the maximum four-octave playback range. Thus the Microvox can be played multitimbrally from a MIDI keyboard or the 64's QWERTY job — enabling you to create all sorts of multisamples, or double up percussive sounds with a MIDI drum machine. As well as being able to assign a sample to a note, you can also define the actual playback pitch of the sample — obviously useful for percussion sounds.

You can also select MIDI Mode 4 on the sampler. This allocates each of the 16 samples to one of the 16 MIDI channels (sample A to channel 1, and so on), all triggerable selectively — useful for sequencing.

It's worth bearing in mind that the Microvox's necessarily limited sample memory (just under 40K) won't magically expand as the number of your samples increases. As with any sampler (to one degree or another), you have to sacrifice sample duration and/or bandwidth in order to fit an increasing number of samples in. But then, if it's megabytes you're after, you have to spend megabucks, too.

Microvox is available in two versions: 1.2 and Pro, on disk and cartridge respectively. The former includes a 2000-note, single-track sequencer and a digital delay; the Pro version removes the sequencer but adds slightly longer sample times, more special effects, individual fine-tuning and volume setting of samples, response to velocity-sensitivity over MIDI (affecting sample volume and timbre independently or together) and auto-looping of samples. The Pro version is actually a £70 'upgrade' to 1.2, so you actually end up with both versions.

All the facilities in the world are no good if the sample quality is low. Fortunately, this isn't the case with the Microvox — though if you know your spec sheets, a maximum sample rate of 42kHz (giving 20kHz bandwidth) and companding that effectively gives 10-bit resolution should lead you to expect something reasonable.

There are eight preset sample rates in all, giving bandwidths ranging from the above-mentioned 20kHz to a mere 1kHz. Sample duration varies from just under a second with maximum bandwidth, to just under 20 seconds for the minimum (always assuming you're using the total sample memory). Sounds appear quite acceptable at less than the maximum bandwidth, while speech is acceptable at 5kHz, giving you something like 3.5 seconds of sampling time.

A low-pass filter is fitted to help avoid aliasing, and this is automatically adjusted when you select a different sampling rate. Supersoft have thoughtfully included user-selectable offsets to the filter cutoff frequency at both the sampling and playback stages, giving you a valuable extra degree of control over the sample, including the ability to brighten or darken the timbre of a sound.

The 1.2 program disk also contains a few sound samples of its own, including some weighty percussion voices that range from the orthodox (bass drum, snare, claps) to the unusual (beer bottle, metal box and tupperware), together with such delights as 'popping note', 'aaagh note' and 'chopped guitar', which offer an effective — and amusing — introduction to the sampler's capabilities. There's also a rhythm sequence which shows off the percussion sounds to good advantage. Supersoft could usefully follow the likes of Ensoniq, Sequential and Akai in offering a library of different sample disks — a move they're considering as I write this.

"Background - The data-intensive nature of sampling doesn't sit too well with the limited memory of today's eight-bit home computers — but this hasn't daunted Supersoft."

The sampling unit itself looks like something from the wireless age — full marks for aesthetics, chaps. In addition to the essential audio in and out, it houses four rotary controls governing input gain, repeat amount (for the DDL), mix amount (governing the output mix of incoming signal and Microvox-generated sample) and output volume. Single MIDI in and Out sockets are to be found on the back of the unit; the In allows you to play the Microvox's samples from a MIDI keyboard, sequencer or drum machine, but the Out doesn't do anything as yet — though the manual says enigmatically 'future software packages will make use of the MIDI Out socket'. We can but wait.

On power-up you're presented with a menu offering nine pages (eight on the Pro version). Each page lists its own options, and you step through and alter these using the 64's Function keys ora joystick.

The Sampling page displays threshold and peak level meters, together with the eight sample bandwidths and their associated durations. Threshold level is user-selectable from the software, while the strength of the incoming signal is adjusted from the sampling unit. The sample — or work — area is defined as whatever sample memory isn't occupied by the other 15 samples.

Once you've selected your sample rate/bandwidth and filter offset (if required), and decided whether or not to use companding (it's advisable), you're ready to sample. The screen blanks while you're sampling (and, incidentally, whenever you're playing back samples). For the more technically-minded, this is necessary if the 64's video circuitry isn't to steal cycles from the main processor and compromise the sampling process.

Once you've taken a sample, you can select a Listen option which allows you to hear your sample played back at the recorded pitch. This provides virtually instant — and invaluable — feedback. Almost as quickly, you can go to the MIDI or QWERTY playback pages, whence your sample can be played back over its full range. If you've previously selected (on the Keyboard Assigner page) which samples are to be triggered by which pitches, you can hear your sample in conjunction with any other currently assigned sounds.

The Microvox can do much more than just play back samples, however. Included is a Waveform Editing page which displays the contents of the current work area, and it's here that you can do all sorts of weird and wonderful things to your samples.

There are two types of display available: Dynamic and Segment. A segment is a block of 128 samples, and while the Segment display allows you to step through (surprise, surprise) individual segments, the Dynamic version 'summarises' each segment and displays the whole sample at once, allowing you to take in the overall envelope of the sample.

This is about as fancy as the Editor gets. Operations you can perform on the sample are limited to manipulating the sample data as it stands, rather than extrapolating from that data (à la Fast Fourier Transforms) for the purpose of manipulating individual harmonics (à la the Fairlight). In a sense, the Microvox options are closer to traditional tape manipulation techniques.

Scanning through the sample is achieved by moving a cross-hair cursor across the display. You can set colour-coded Start, Loop and End flags (clearly visible, these) that define which portion of the sample is to be played and, if you wish, looped. The Pro version's auto-looping facility searches out an ideal loop point for you. It's not a quick process, and can take several goes, but you can at least sit back and watch the loop and end flags move through the waveform display in their quest for the perfect loop.

The effectiveness of this depends on the sample itself (not all samples lend themselves to glitch-free loops), but the manual honestly states that not every step has been taken in software to ensure the perfect loop. That said, I was able to achieve some pretty reasonable loops, and this option is certainly an improvement on the trial-by-error tactics required by the 1.2 version.

"Sampling - You can store as many as 16 samples in memory, and assign individual samples to any notes over the maximum four-octave playback range."

As the portion of waveform between the start and end flags is equivalent to sample A, you can chop a single sound up into several different samples — useful in some cases. And by moving the start and end flags around, you can reverse, reposition, invert, reflect, erase, fade in or fade out the whole or any part of your sample. As with the sampling page, you can immediately listen to the result of your endeavours at recorded pitch, or go to the MIDI or QWERTY playback pages for the full range.

The Voice Handler page is where you get to assign your samples to any of voices A-P (a voice here meaning a sample). In order to avoid overwriting a sample, you have to transfer it to one of voices B-P, and conversely, to work on a sample in the Waveform Editor page you have to transfer it to voice A. The Voice Handler page also allows you to listen to individual samples, delete and rename samples, save to disk individual samples or all the samples currently in memory, and — particularly useful, this one — merge together any number of samples. Merging is a clever way of layering sampled sounds, but more than that, it has the advantage of freeing memory for further samples. If you decide you've made a dreadful mistake, you can always use the Invert facility of the Waveform Editor to remove an offending sample.

Loading of samples, songs and patterns is accomplished on the Disk Handler page. On entering this page, your current disk is automatically catalogued on screen. Samples and patterns can be loaded individually or as a group, depending on how they were saved; if they were saved individually, they can be loaded into any position. The complete sample memory takes just over 90 seconds to transfer between disk and computer — just about enough time to make the proverbial cup of tea. Samples already in memory aren't overwritten by this. Instead, the Microvox informs you if there's not enough room left to house an entire new file.

While Microvox samples can obviously be played from a MIDI recorder, the 1.2 version's internal sequencer is a must for anyone not possessing such a machine. Yet it could also form a valuable adjunct to a MIDI sequencer or drum machine, and Supersoft have included internal, external and MIDI sync options, the latter allowing the internal sequencer to be slaved to a MIDI device. There's just one problem with MIDI syncing: if you send a MIDI Stop code from a sequencer or drum machine while a sample is playing back, the sample continues to play for its full duration. This isn't going to bother anybody using cymbal samples, which sound better if left to decay naturally, but it'll doubtless bug users playing back samples of longer duration, which carry on regardless of the rest of the MIDI machinery.

The Internal option doesn't, at present, drive anything but the 1.2's own sequencer, while for External syncing, the manual makes vague reference to the use of a sync box and the 64's user port, and refers to an Appendix which didn't exist in my copy. Supersoft have included a dock-dividing facility for syncing to the standard clock rates, so non-MIDI syncing is a serious proposition.

The internal sequencer adopts a pattern/song approach familiar to anyone who's ever used a programmable drum machine. Recording can be in real or step time, with real-time recording being from either a MIDI instrument or the 64's QWERTY keys. In real time, the screen blanks out and you get a metronome click from the 64's SID chip; the pattern loops during recording until you stop it.

You can input up to 24 patterns, and then chain these together into a single song with a maximum 170 links. Each link can be given a repeat setting of up to 255 repeats(!), and can be transposed up or down two octaves in semitone steps — handy for adding variety to existing patterns. Each pattern can be given an eight-character name, which is helpful, too. Tempo setting is global.

Patterns are confined to a 4/4 time signature, which is in turn divided into 96 steps (corresponding to MIDI's 24 clocks per crotchet). These are represented as a grid which can be scrolled across the screen when recording or editing in step-time, with the selected step being highlighted. For each step you can enter sample (A-P), pitch, whether the sample is to be reversed or not, and whether the step is a note on or a note off. The latter facility allows you to cut samples down to the duration you want, and is a very useful inclusion. Individual patterns or complete songs can be saved to disk, and editing facilities allow you to insert and delete steps in both patterns and songs.

If there's one area where the generally user-friendly software falls down, it's in numbering the 96 steps; it gets to be a bit of a headache when you're working in step time, and trying to relate the numbers to more familiar and meaningful notation. I'd like to see Supersoft dividing up those 96 steps graphically in a way that makes more sense to musicians, as opposed to adding machines.

"Sequencing - The 1.2 version's internal recorder is a must for anyone not possessing such a device — yet could also form an adjunct to an external sequencer or drum machine."

One word of warning: before using the sequencer, you need to ensure that all your sounds are edited first. Why? Because the sequencer and the waveform editor use the same area of the 64's memory, and thus can't be used side by side.

The Special Effects page on version 1.2 consists of a digital delay. Duration and delay time depend on the amount of free memory and the selected bandwidth, while repeat amount is controlled from the sampler unit. The Pro version adds a harmoniser(+/- an octave in semitone steps) and chorusing, phasing and flanging effects; the latter three are achieved with appropriate settings of modulation waveform, speed and depth parameters and repeat rate, and some example settings are provided in the manual. This page is intended to act on any incoming audio signal, not on the samples themselves, and there's no doubt it's useful if you don't own equivalent effects units.

All in all, the Microvox is a fine budget-level entry into the world of sampling for a broad range of people. It's best suited to recording and sequencing applications, with its pattern-based sequencer making it especially useful as a sampling drum box; those wanting a performance-oriented sampling instrument will probably still need to look upmarket.

The Microvox is capable not just of recording high-quality samples, but of serving as a flexible musical tool with the help of multi-sampling, sound editing and sequencing facilities, and a healthy degree of MIDI control. It's an extremely attractive package, and, rare among home computer-based systems, it features user-friendly software design and an accessible manual to complete the picture.