Tuning into the music capabilities of the BBC micro with Lizzy, Gary Herman and Mike Hilton.

BBC Model A and Model B micros are currently second only to the Commodore 64 in their music capabilities.

They are less immediately credible as instruments because there's no audio out — the sound output is via a small internal speaker capable only of the most rudimentary performance. This is no reflection on the capabilities of the sound chip itself which is potentially as pleasing to listen to and considerably more versatile than, say, a VL-tone. The sound source consists of three independent oscillators generating square waves. There's a noise generator operating as a separate channel and giving eight types of noise, and three oscillators and noise source can be mixed. ADSR and pitch can be independently shaped by a series of four envelope controls. There's a range of 5½ octaves. If you compare this type of equipment even with the most basic synth the most conspicuous absence is that of any kind of filter, but fairly complex sounds can be produced additionally by setting the oscillators to varying pitches and envelopes.

No LFO of course, but you might get a tremolo by tortuous use of the pitch envelope. At the time of writing software is limited (as at all times and everywhere in the field of computer music) but the wealth of BASIC routines for music production makes D.I.Y. programming practical and even inviting. Anyone handy with the soldering iron might consider invalidating their Beeb guarantee by wiring an audio out from the internal speaker connexions: we understand the same effect can be achieved by devising a programme to route the audio signal to one of the existing output ports.

THIS IS THE BBC NEWS...

It must be a good five years since 'The Computer Program' and BBC computers were just a gleam in Aunty Beeb's eye. In her search for suitable manufacturers she engaged in wild flings with ardent suiters Sinclair and Nubrain but finally settled down with outsiders Acorn. The expensive and overterm pregnancy that ensued seemed worth it when the impressive facilities of the new machines were announced in 1981. The advanced structured BASIC, well thought out graphic and sound capabilities and, for the period, a massive 16/32K or RAM, soon established the infant as the Apple of teacher's eye in schools throughout the country (a government subsidy to schools covering 50% of the cost price of computers manufactured in the UK didn't hurt either!).

Widespread school use boosted BBC appeal to home users and by the end of Christmas 1982 they were established as one of the leading home micros, saving Acorn's bacon in the process. In the meantime, thousands of clever games have been written and a vast amount of books and articles have appeared, all explaining some aspect of things BBC, all telling us how wonderful they are. We would be right in thinking that as far as the British micro world is concerned, the sun shines out of Beeb user Ports. So — where's all the software and hardware that makes good use of the music capabilities?

I know, I know. You can point at the Wimpsoft BBC Synthesizer and the Aarghprog BBC Composer but these things are thin on the ground and are inevitably written as if the programmer's experience in music stretches to second triangle in third form torture classes! Music applications for micros had been thought a minority interest and the current frenzy of digital jiggery pokery in music caught the micro world, pants at the ankles. Most computer music gear is specialised and expensive though MIDI points the way for marrying low cost micros with relatively low cost effects and synthesisers but we need the means to make the most of the versatile built-in sound facilities of sensibly equipped micros like the BBC — and at next to no cost.

Here at ES&CM we applaud the appearance of any software or hardware that promotes the use of these facilities. Hopefully there will soon be useful, easy to use professional examples; but really we want to encourage you to marry drum machine and BBC (or multiple BBCs) providing programmable rhythm tracks, or write comprehensive composing programs, or anything that aids in music making, because computer musicians themselves will be doing most of the experimenting!

To achieve worthwhile results (and when you do, we want to hear about it) you'll need a little help. ES&CM and many other magazines have reviewed and explained the BBC SOUND and ENVELOPE commands at length. We will continue to supply and explain programs and projects but many will be designed to help in the writing and construction of your own systems. As we're currently dissecting the BBCs we can suggest ways of improving your Beeb music potential immediately.

Potential Help

Two major headaches in music making with BBCs are —

1) Built-in speaker only.

2) No means of synchronising BBC to drum machine or tape, and vice versa, at little or no cost.

While 2) is common to most home micros, 1) is almost unique to the BBCs and this seems pretty silly considering the extensive control over sound provided. It appears even more ludicrous when you discover that, as some of you may already know, there is a small connector on the main PCB that outputs unamplified audio suitable for input to HiFi, amp or home recorder! As Acorn have been a mite slow in providing details we investigated ourselves and outline the poking around and soldering necessary in fig 1 but bear mind that hacking things about may affect your guarantee.

1. Remove the crosshead screws on the rear interface panel.

2. Remove the two front most screws on the underside interface panel.

3. Lift off the top case.

4. Remove the two nut/bolts on the extreme right and left on the keyboard.

5. Swivel the keyboard backwards, away from the main PCB.

6. Unscrew the nut holding the UHFTV connector to the rear of the case remove the tag washer and pull the connector clear trom the rear.

7. Locate and remove the four screws holding the main PCB to the lower case.

8. Ease the PCB forward to clear the connectors at the rear and carefully swivel it upwards and forwards to the limit of the cable length.

9. Referring to the photograph and inset diagram, locate connector PL6.

10. Strip three centimetres of insulator from a length of small gauge shielded audio cable (guitar type lead).

11. Unpick the tubular braid from the conductor. Strip five millimetres of insulator from the core and tin the protruding inner conductor and five millimetres of the braid.

12. Solder the inner conductor to the rear most solder filled socket of PL16 and the braid to the other.

13. Run the lead out through the 'analogue in' socket on the rear of the case or alternatively, push out the 'econet' socket cover and route it through that.

14. Re-assemble the computer following steps 1-8 in reverse.

The protruding lead can now be fitted with a five pin DIN plug or standard jack as required, to provide audio out to amp or hi-fi.

WARNING. It is easy to damage a computer and if you are unsure of yourself refer to someone experienced in electronic modifications.

This little modification alone dramatically increases the effectiveness of BBCs as musical instruments. Your lovingly enveloped sounds can now be EQ'd, flanged, chorused, etc., amplified and recorded. This new scope draws attention to the quality of BBC sound output so we should take a good butchers at its limitations.

Assuming that background noise levels on your particular micro are acceptably low, the first thing to consider is the fact that the square wave signal generated by a BBC oscillator is a crude waveform in musical terms. The harmonic structure of a square wave confines us to harsh, raspy and relatively thin sounds. The closest acoustic equivalent is the sound of a hard blown clarinet but, that isn't very close. Bearing in mind that envelope shaping merely gives control over the volume and pitch of a note, the only way to significantly alter the character of the raw square waves is by using a range of filters for each voice. As these aren't supplied, dodges can be used such as graunching the signal through guitar type effects pedals in different combinations and artfully twiddling amp and/or portastudio EQ. Unfortunately, this of course, affects all voices. Don't forget that a bit of controlled noise can do wonders for a lack lustre sound and forms an essential part of most percussion. The eight different BBC noises are particularly useful for creating acceptable shaker, tambourine and cymbal type sounds.

Now the sounds are clear, loud and interesting, what can we do with them? This brings in headache number 2. It could be argued that a sync to tape or drum machine is unnecessary and it's true that by accurately timing your programs you can bang out effective rhythm tracks with a number of options as to configuration. For example, two voices plus noise can make a simple drum pattern while the third provides bass; but this is very limiting and sync to drum machine is an attractive proposition. I should explain that synchronisation of one digital gubbins to another is achieved by connecting them together and causing the gubbins intended to control to emit a clock pulse proportional to its internal timing. The gubbins to be controlled abandons its own internal timing and uses the clock pulses received instead — Voila! zey ar seenchronized.

A gubbins can also record relevant portions of its clock pulses on audio tape, say, one track on a multitrack recorder, and take its cue on playback from the recording. So, it'll warble or bash along with the tape, at your command, without you having to record the output of said gubbins until final mixdown — sync to tape. Of course, if you combine sync to tape and sync to gubbins you can have a whole host of gubbinses warbling and bashing away to the loss of one track and the gain of clean signals on mixdown. BBC computers and drum machines are both digital gubbinses so...

(Click image for higher resolution version)

Computer Problems

The main problem here lies within the computers. They can operate using external timing but they can't cope with raw drum machine sync pulses and a little tricky programming and some simple hardware is required to do the job. Be assured that we will be providing the relevant information (hopefully for all popular micros) in a future project feature.

So now we have a vision of (pay attention you students) 20 Beebs all in on the one-drop with 60 oscillators roaring as one through a 10K PA! Those of you who have had intercourse with Econet may be unimpressed by such abilities but a project for last year's PCW show took many weeks of effort on the Part of Acorn experts to arrange for 2 BBCs to distribute programs over Econet to 9 others that played the 6 1/2 minute 3rd movement of the 3rd Brandenburg concerto using all 27 available oscillators as 9 violin parts, 9 viola parts and 9 cello parts. The main problem they had was in arranging synchronisation of notes as the Econet clock was not suitable for the task.

This was solved by having one of the 9 give a general RUN order to itself and the remaining 8 trusting to the individual accuracy of the nine internal clocks to keep time. As the piece repeated within 1 minute the 9 clocks were reset by a second RUN and so on, until the end of the piece. It worked — Just!

This article is intended to spark off ideas and action; so see what you can come up with. You're the ones with the special interest — the powers that haven't woken up yet!

CLEF COMPUTER MUSIC SYSTEM

Clef Electronics, who have produced kit electronic pianos, rhythm machines, and so forth for some years, have devised an instrument or system designed to do for the BBC what the Alpha Syntauri has done for the Apple.

Strictly, it does not develop the BBC's musical potential because the oscillators and sound circuitry used are provided in the CMS itself. What it does, instead, is employ the BBC as a control and storage device which works in conjunction with the CMS.

This is sensible because versatile as the BBC sound chip is, nobody would think of comparing it with a DX-9 or a Juno 106. On the other hand, its data processing capability creates a much wider potential than could be associated with a conventional dedicated poly. Before we go any further it's worth noting that the CMS should retail for around £425. It has a full size 5½ octave touch sensitive keyboard (if you want to know more about that check out a Clef electronic piano because it's the same one) and even if you had to buy the BBC specially this makes it very competitive.

Plus and Minus

The Clef works on the principle of additive rather than subtractive synthesis (as in the majority of synthesizers and particularly the cheaper ones) you start with a VCO which generates a simple waveform with complex harmonics. The harmonics are filtered out to modify the sound, and the end product is characteristic of the kind of VCO's and VCF's used. Additive synthesis, on the other hand, builds sounds by adding harmonics to one another.

On the Clef there are 32 oscillators available and each of these oscillators can be set to generate a different waveform. The amplitude (equivalent to power) of each individual waveform can be varied so once you've selected the harmonics you want you can 'mix' them at varying levels. As well as listening to the results anything up to the 15th harmonic will be displayed on the screen to give a visual indication of what's happening. You can go on adding harmonics up to the 25th.

Having constructed the sound in this way you can then use the 6-point (8-point?) envelope generator to shape it. The standard software, which is stored in EPROMS as part of the hardware circuitry, allows the envelope generator and the allocation of oscillators to be varied according to keyboard dynamics. This you could arrange for a note played harder to have faster attack on longer sustain, or you could programme extra harmonics to come in so that the sound changes according to how hard you hit the keys.

The current software will only allocate four oscillators per note but this is a limitation of the program and not the system: as with the AlphaSyntauri users will be encouraged to get stuck in and programme their own developments. With the four oscillators/note the Clef is 8-note polyphonic, 3 oscillators per note will give 12-note poly, 2 will give 16 and 1 will give 32. The envelope can be used on individual oscillators so that not only can you mix different oscillator sound but also different envelopes of oscillator sounds, allowing sound changes within the duration of a note. The oscillators can also be de-tuned relative to one another allowing a whole range of chorus type effects. Sounds can be stored and accessed simply by pressing numbers between one and 16 on the micro: more than 16 and you use tape dump.

The system has a noise generator so envelopes on unpitched sound can be mixed with the oscillator output (a surprising number of acoustically produced sounds have a 'noise' component and adding this in a synth can give vividness and authenticity). Each envelope can also be 'panned' left, right or centre for stereo imaging. There is also a real time sequencer accessed by a simple 'record' command. This will store dynamics information as well as pitch and duration, and the sequence can be overplayed as often as required — but only with the one sound.

So let's step back and look at the kind of instrument Clef are offering. Clearly it's more sophisticated than the Microsound 64 which is currently being offered to go with the Commodore 64. It's also more expensive, and has a full-size keyboard. What does it sound like? In theory at least owing to the high number of oscillators should provide us with a decent organ, synthesizer, string machine and piano sounds.

Cosmetic Changes

Clef CMS has been a long time in the making and we're still only talking about a 'pre-production' run — availability should be soon... but there's still the possibility of cosmetic and detail changes. More important, a product like this is never 'finished' because it's software based. No it can't allocate 32 oscillators per note, allow composition in step time or interface with MIDI gear — but all these are possible if someone writes the appropriate software.

Clef were faced with a choice of perfecting the ideal software package by summer of 1985, or bringing the product out now and throwing it open to input from creative computer musicians. The option they chose is calculated to get users more involved, and still gives an extremely versatile and obsolescence-proof musical instrument. The sounds are varied and complex, the dynamic sensitivity effective and the fact that you can store 16 sound programmes makes the Clef a very practical keyboard.

LOOKING FOR THE SOFT OPTION

Perhaps the most disheartening thing for any micro owner is the realisation that acquiring the beast is only the beginning. At the simplest level, it requires software (that is, programs) to demonstrate any of its much-vaunted marvels — and software isn't all that easy to come by. You have got to sit down and laboriously type in programs. Sometimes you have to write them yourself. Or, worse still, you have to search through rack upon rack of versions of Chuckie Egg and Hunchback in your local software shop before coming across the program that answers your legitimate needs.

Not, you understand, that I've got anything against games. On the contrary, I'm as partial to taking a potshot at any passing alien as the next person. However (and it's a big however), I don't want my computer to idle away its precious youth in electronic billiard-halls. Dammit, I want it to work for its money.

I muse quite a lot — you know, sit around and come up with ideas on how to save the world and what should be done for the general betterment of human life. Sometimes, I think to myself: 'If only I could get a computer to do such-and-such!' Since this occurred to me, I figure it must be a pretty obvious idea. So, pausing only to pack my trusty cheque book and my faith in humanity, I make my way to my neighbourhood program palace, there to unearth, all being well, the ready-written program that will make such-and-such a working reality. Needless to say, all is very rarely well.

So here I sit, waiting for some serious music software to come my way and hoping that all you software houses out there are still adventurous enough to realise that there are enough of us who want to use the marvellous facilities of the micro to their fullest extent to make it worthwhile to produce and sell the programs.

Music is, of course one of the chief areas where hardware technology leads software production by miles. Take, for example, the chip used to produce sound by the BBC micro. This is called the 76489 and it is produced by Texas Instruments. It's a tiny little thing — if you open up your BBC you may not even be able to find it lurking down by the loudspeaker — and, as befits its size, it's by no means the most remarkable sound generator chip on the market (that title, I'm sure, belongs to Mostek's 6581).

All the same, the 76489 is a minor miracle of technology: a 16-pin chip with three music channels and a noise channel featuring programmable frequency and programmable attenuation. What is most remarkable about this chip (and about the others featured in many home micros) is that it is entirely digitally controlled. To lay people that may not mean a lot, but to those of us involved with computer music it means just this: the grail of computer composition and performance is now within the reach of everybody who knows what a chord is. Say goodbye to the room-sized computers at RCA and Bell labs, laugh at the names of Milton Babbitt, John Cage and Karlheinz Stockhausen. With a home micro and a portastudio, you can be in the vanguard of modern music.

The BBC micro uses two BASIC instructions which make pretty much full use of the 76489's facilities: SOUND and ENVELOPE give more-or-less comprehensive control over queuing, synchronisation, frequency and attenuation. Major deficiencies are due largely to the chip itself — which provides only a pulse wave-form and includes no facilities for digital filtering or modulation. That said, BBC BASIC's sound commands are difficult to operate for the non-specialist computer user and so we look for some ready-written software to create a 'musician friendly' environment. We look, I'm sad to say, in vain.

Independent software houses seem to be more on the ball. Quicksilva and Bug-Byte, in particular, have both produced programs for the BBC which go some way to producing the sort of software a musician might want.

The two programs are diametrically opposed in their approaches. Bug-Byte's BBC Music Synthesizer is aimed squarely at composition, while Quicksilva's MuProc is performance oriented. MuProc attempts to simulate a synthesiser and mixing desk on your VDU screen and allows you to 'record' on one channel and effectively overdub on the three others. Inevitably, the facilities of a true synthesiser and mixing desk are not all available and, while the program allows fairly simple control over the BBC's sound output, I am reasonably convinced that MuProc's approach to the task of music production is, in principle, untenable. The principle is simply that no home micro can match even a cheap synthesizer or mixing desk for facilities. True, the program only costs around £8 (as against £30-£40 for a little Casiotone, say) but the musician is not likely to want to produce a whole work on just a micro.

It is more probable that the musician will want to use a micro to control existing musical devices or to use it as one among many sound sources or to use it as an aid to composition and arrangement. So far, software for control purposes (for example, using the MIDI interface) has been remarkably thin on the ground. Composition aids, however, do exist. Bug-Byte's BBC Music Synthesizer is one such.

What's for Dinner

This program is menu-driven (that is, it allows you to select options and move through a series of selections until you reach the final stage of sound production). The options include envelope shaping, writing and editing melody and noise effects channels, arrange phrases and saving pieces on tape or disc (which can also be done with MuProc). Regrettably, there are just too many restrictions on the choices (for example, all notes have the same duration) to make the program a serious contender for the musician's interest.

What is really required is a music composition language. A number of these have been developed for use on mainframe computers (you know, the big ones) but, as yet, there seems to be no software for the micro user. Ideally, a music composition language would allow you to specify all the significant features of a note and to change them at will. It would allow you to insert and delete notes and phrases, to arrange phrases, to change tempo and meter — in fact, to added benefit of letting you hear your work as you do it and letting you store all the information as data so it can be as easily recalled and manipulated as any computer data. In effect, a good music composition language should use all the facilities offered by digital sound production devices while making those devices completely invisible to the user.

From the performance or recording point-of-view, the musician is most likely to want a computer as one among many instruments. None of the software I have seen succeeds in enabling easy use of a micro's facilities in that area. Instead, programmers have been overwhelmed by versatility. They have tried to make the computers do everything (which results in a general level of mediocrity) whereas they should, for example, be trying to make the machine do one or two things very, very well.

All the same, brave efforts from Bug-Byte and Quicksilva. I must make one brief digression in order to point out that Bug-Byte also include a program called Auto-Composer in their package.