Wolfgang Palm and Wolfgang Duren discuss the development of their PPG digital music synthesisers

Herr Palm: I'm really both a technician and musician. First my interest grew through music. I didn't learn to play at school, (except in a band like most pupils), but my father was a pianist. From '75 I studied physics and electronics subjects and also had a group and played organ. I remember hearing Keith Emerson on an ELP record use portamento and glide effects ('Hoedown'). I wanted to copy these sounds so I made a simple oscillator with some portamento — it was a frequency to voltage converter and then voltage to frequency, controlled from the electronic organ. By the end of my studies I had built my first modular synth (like the big Moog system). The Minimoog was successful in Germany at the time.

I then started work on developing synthesisers. First a very small monophonic, not unlike the Minimoog, but much cheaper than it cost to buy in my country. These early instruments did not use 1V per octave control for the oscillators, but were linear (under each there was a preset for tuning individual notes in a resistor chain!).

Soon after we started selling some of our synths I met Chris Franke of Tangerine Dream — it was funny because he had bought an old Moog system that he wanted to modify and was advised by Moog in Germany to go to me. We did provide a lot of design projects for them.

Early instruments

The 1020 Synthesiser was our first use of digital circuitry in an analogue instrument, for generating the waveforms. All our systems initially were of course monophonic. In 1978 we started developing polyphonic systems — actually, the first instrument was duophonic with some special 'hold' functions, but it was digital using a lot of CMOS ICs. Next I produced a 4-voice keyboard that only had one trigger output so you could play four simultaneous notes as chords. The circuitry employed keyboard scanning with 'next available' oscillator allocation as well as split-point setting. It was also one fixed connection to each oscillator (from our modular system) rather than the multi-options we now use on the PPG Wave 2.2.

Before the polyphonic microprocessor controlled instruments, there was another instrument — a micro-based sequencer. Klaus Schulze has a lot of our modular sequencers we've designed. The micro we started with was the 6800 because it was quite cheap at the time. It was quite a problem for me to find enough hours to grasp machine code, BASIC and other micro software plus hardware techniques, and like many others, the physics I learnt at school didn't cover this. We considered the sequencer to be very flexible, with its small keyboard that could play in realtime, add loops, run backwards, do auto transpositions, chain sequences etc.

Herr Duren: Many people have tried to build synth related instruments like sequencers without having thought about it before. Consequently, what the engineer may accept is often quite unsuitable for the musician.

Herr Palm: We'd probably sold about 30 or 40 synthesisers by 1979, and 20 or so sequencers, all mainly in Germany with a few in France and Holland. It had become my sole occupation and I just made enough to keep going.

Then some people approached me with a view to making a big firm out of PPG! They tried to get dealers in Germany — they actually got 5 but then it all broke down and I think they went bankrupt. One of the dealer's salesmen was Wolfgang Duren and so that's how we met. Our company had been established in May 78 as PPG (Palm Production Germany). Meanwhile the dealers owed us a lot of money which they couldn't pay and this did cause a setback financially for us.

Computer instruments

We also made a programmable duophonic synthesiser before the sequencer — it was a crazy machine with over 200 CMOS ICs. We did regard it as a flop, but it was probably the first machine to store programs (holding up to 50). Our first sequencer was boldly called the 350 Computer Sequencer — we think we were one of the first to put the word 'Computer' into a musical product title!

It was after the 1979 Frankfurt Fair that we decided to have a go at the polyphonic computer instrument. A lot of attention was being given to Oberheim's multivoice machines. My idea was to make the same functions and sound generation in a digital way. We experimented with various digital filters and so on, and eventually came up with a system using 'wavetables' — 64 waveshapes in each table so that you can move through the waveshapes by using envelope control, LFO and other CV output devices.

There was a problem at first — the prototype machine was virtually all digital — only VCAs after the oscillators but with no filters, because the filtering was done by changing the waveshapes. The sound is different of course — quite sharp and hard, and not as 'fat' as the Moog.

Herr Duren: We did like the sound, so did T.D. and certain musicians. Others said "do a Polymoog"! The Prophet also made it difficult for us to compete in the early days.

Yet we created a lot of interest as the PPG Wave instruments used an analogue interface for their digital system. In the first place we thought we would replace the filter by means of the wavetables, but a filter does put its own non-static character on an instrument's sound. The 64 complex waveforms making up a table are factory-generated and fixed with individual amplitude and shape.

Herr Palm: Then if you call up a wavetable, the computer calculates the waveshapes between two fixed waves. Of course, you can then vary the overall amplitude of the waveshapes selected. Only in the development system can the waveshapes be set up.

Herr Palm: We showed the Wave Computer 360 at Frankfurt 1980 with some slight changes to further production models — some sliders, no special LCD display with 7 segmented LEDs instead, a rather complicated tuning of voices that was slightly awkward to use plus a VDU display.

Herr Duren: Although there was no sequencer in the instrument, there was a facility that enabled clocking or stepping of the oscillators and you could detune each oscillator to make a sequence.

Herr Palm: The response was not so good because of this rather hard, metallic quality of the instrument.

Herr Duren: But always there were just a very few musicians who would use the new sounds, then after hearing their records, people started to take an interest in the PPG machines.

Herr Palm: About 40 of these instruments were sold and at the next fair we brought out the updated version (the Wave 2) with filters and a new display. It still took some time for the strong interest to be created — most people preferred to be able to synthesise traditional instruments rather than use completely new ideas.

And now we are producing the Wave 2.2. Our business is still very small with only 7 people working on the production as we use local companies to produce major parts for us.

Herr Duren: We feel that the Wave 2.2 is now the result of collaboration with many musicians — the position of controls, how it responds to continuous stage use — all those sort of things. For example, we've added a second wheel controller following people's suggestions.

Herr Palm: One of the big difficulties in producing software for the Wave is that there are many possibilities for wrong moves when positioning the cursor in the large display. To actually debug all these and other things so that the system will not lock up or crash has taken a long time. The development for the Wave 2 and 2.2 was done with a 6809 micro, which is quite powerful in combination with the Wave system. As we said before, we did our own work in Germany more or less alongside systems under development in other countries— mainly USA, and our Wave 2 came out before we even saw the Fairlight.

Herr Duren: One of the first instruments we heard about was the Synclavier.

This situation about analogue/digital choices is that if a musician wants to work on a computer system then he or she moves the cursor on the screen to various points of entry. With our system you have the analogue control panel, moving knobs to synthesise as most musicians prefer.

Herr Palm: There was another design that was not generally known about — it was a large system using 2 microprocessors (called 340/380 system). Thomas Dolby has one and it was a development that allowed sampling of sounds, making original wavetables. It had additive synthesis and a powerful sequencer which we call an 'event generator' because you can define all the time points exactly, as well as having different sound setting for each tone — just after that we heard about the Fairlight. That's really why we didn't make too much of this system.

Wave 2.2

The Wave 2 has since been very successful and is now replaced by the Wave 2.2 and this makes use of special software with the Waveterm VDU display. Most of the software preparation I do myself. I find that Wolfgang helps me greatly with establishing the direction in which we should be going.

Herr Duren: Looking back, the important point for us was our determination to sell our instruments outside Germany once we'd brought out the Wave 2. At that time I was on tour with T.D. and met with Desert Distributors who agreed to start promoting PPG in the UK. We also made contacts in France. One problem with these kind of instruments is that they do require more than just a 'sales' approach because of their complexity. So now we have PPG sold in the UK by Syco Systems. We've now produced a dozen or so of the PPG Waveterm which is an expansion unit for the Wave 2.2. It is a 19" rack system with a video monitor having graphic facilities, 8" floppy disc memory storage and an audio signal processor for recording and treatment of natural and other acoustic sounds.

We've put keys directly under the monitor screen for typing in multi-commands (indicated above keys on the VDU). New sounds can be created by individually programming the amplitude of up to 32 harmonic overtones and screen 'pages' makes specific tasks easy to set up. For example, computing waveforms graphically by additive synthesis (not just with sine waves but also with more complex 'pre-formed' waves); creating wavetables from up to 64 partial waveshapes; inputting transient sounds and then defining start/finish points of the sample to make new sounds — a segment of the waveform can then be inserted into a wavetable; in addition, the Event Generator allows complex sequences to be programmed, edited and recorded.

Herr Duren: It is also worth noting that sound samples, like the digital sounds, can be treated using the analogue controls. We've had to redesign the oscillator hardware in the Wave 2.2 to allow this to happen. Normally, the oscillators run in small loops for the wavetables making single periods of the sound. If you move the wavetable the loop changes to another position of the table, but if you playback sampled sounds, it goes through the whole memory in one sweep and therefore requires this new hardware.

Herr Palm: The total sample time at 16K bandwidth is about 1 second in the normal version. But any sample rate can be set to add more time at reduced bandwidth. You could record 5 seconds of speech (but not music) satisfactorily. Virtually anything can be sampled and instruments like the flute can be sampled well; grand piano and strings too; even a funky phrase on the bass! The big advantage of the sampling is not just to copy traditional instruments but that you can 'zoom' on to a part of the sound and then use it. First, you define a loop which is maybe 2 or 4 periods of the sound, then you can analyse this and get 32 overtones. The quality of the analysis means you get virtually the same sound back — of course it's only a static period created without dynamics.

So you could put a guitar attack with a piano decay and soon. By defining loops you can stretch a sound's time parameters. The Wave 2.2 controls can then alter sounds in realtime, while the screen commands may take a second or so to execute. Also specific parts of the keyboard can use different waveshapes. For example, the sound of a choir needs a different waveshape on each key otherwise you get high 'chipmunks' and low 'growls'!

Future developments

Unlike the Fairlight, which has eight voices with different sounds, our system has a larger memory with all oscillators able to select any position of this memory at any time. In other words, up to eight notes anywhere on the keyboard can be set to reproduce the correct sample for that note (or group of notes if a limited number of samples are analysed in the waveterm).

It's now 3 or 4 years since we started the polyphonic instruments. I'm now working on a lot more software for the Waveterm and the Event Generator. We're also thinking of doing an extra keyboard with normal acoustic piano keys that are both pressure and velocity sensitive (at the moment the present keyboard is just velocity) with wave control from the keyboard as well as other parameters.

All the new products we'll be bringing out will have a communication bus, a connector for an 8-bit parallel bus, so that a musician can add to his basic system step by step. At the Frankfurt '83 Fair we'll have a 19" rack containing a Winchester Drive where you can sample sounds directly on the disc (like the Synclavier) — even 15 minute sounds! Another part of the Waveterm will be a mixing console where you have 16 sliders, knobs and push buttons for CV in/out controllable through the Waveterm software. So you could have a function like the Linn Drum where you program the rhythms by pushing the knobs or you can do a computer mix etc. Our aim is to produce a range of 'instruments' that must be for all musicians — musicians working in groups can use the Wave 2.2 and maybe the piano style keyboard; people who compose at home could do programming with the Waveterm without the processor keyboard. Each musician can make his own package to suit his needs, and the Winchester disc (6 Mbyte) will give immediate access to 100 sampled sounds, opening up a new sound spectrum for the musician.

We acknowledge with thanks the co-operation of Syco Systems Ltd., (Contact Details) in the preparation of this article.