In Electronics & Music Maker magazine, march 1983, this interview with Chris Franke appeared. Unfortunately we don’t know the name of the author. Please note that all (c) are with the magazine.

Tangerine Dream last visited UK in November ’82 as one of their regular yearly appearances, receiving a wider audience appeal especially for younger electro-music fans. This last concert was reviewed in our January ’83 issue and used new compositions that were improvisations on themes, some for possible new albums. Much more keyboard work is involved, still with importance to sequencer lines, and significant equipment changes include the digital recording of natural sounds via a sample-to-disc system claimed to exceed the quality of the Fairlight’s 8-bit system. Real drum events are sampled, as well as original ‘voice’ effects, which are triggered by sequencers.

Chris Franke has been doing his own solo work and hopes to release an album by the summer. Tangerine Dream have produced several film scores, including three American movies (Soldier, Wavelengths and Deathbite) and a German TV crime series (the theme music reached the Top 10 of the German charts).

‘White Eagle’ and ‘Logos’ were the last Tangerine Dream albums to be released (both in ’82) and the band are enjoying their success and their ability to produce albums and scores very quickly from their own high quality studios.

Chris Franke is one of the world’s most experienced users of the sequencer and in this article reveals some of his knowledge of this important musical instrument.

There are two ways I can approach using sequencers — first, the one I have in the studio which is an electronic composition machine. Its operation is similar to systems like Synclavier II’s SCRIPT, the Fairlight Compositional Program, or the Roland Microcomposer machines. With this system I compose straight into the computer instead of writing down the music. I can be quite subtle and put in a large number of parameters and more detail, with immediate feedback — that’s very important for a composer. I’m not interested to say that this is ‘electronic music’ — this is just a composition played by electronic instruments. If I wish to play it with violins or brass that’s also fine — it’s just a compositional machine.

Early days

In 1973 I was using one of Moog’s sequencer modules which were the first available for stage use. I liked its access to triggering notes that let me ‘improvise’ with it — it wasn’t just a little pre-recorded ‘cell’ of music. There were also sub-programs so that rhythms could be varied. I used this for two years although its main difficulty was the tuning, done by rotary knobs for every note in the sequence. This made pitch changes impractical for real-time performance because it was stepless.

At the end of 1974 my first idea was to make this quantized — both for the pitch and the timing, so we invented a system with little switches, one switch for 1-12 semitones and another for 1-8 octave selection. There were also two more rows of knobs (1-12) for fractional selection of control voltages to use with VCF and VCA modules.

Then we did the ‘Programmer’ because we were not satisfied with sequencers that lasted just 8 or 16 events. Because we had a lot of the Moog sequencers by this time in a rack, we built the programmer to control these. Different sequences could be played in a selected order, let’s say, 4 times the first sequence, 8 times the second and soon. The programmer used electronic switches and various dividing circuits for allowing a number of events (or pulses) to be counted before triggering a new sequence of events (i.e. by switch to another control voltage channel).

The Moog system, as I’ve said, was nice because it had three CV channels so that you could control the cut-off frequency of a filter or the VCA volume in live performance.

During ’74 and ’75 before going to America, we tried to improve the stability of the oscillators by making our first digital oscillator, which was not voltage-controlled but used a switch matrix that couldn’t drift out of tune. Since the switches were also digital this made the pitch very stable.

We moved on to thinking about compositional machines and I bought the MC-8 microcomposer from Roland when it first came out. I tried different products from companies that were putting out so-called ‘digital sequencers’. One from EMS, for example, had a lot of notes, and there was also the ARP ‘graphic’ style sequencer.

A little later there were some German digital sequencers that recorded very well in real-time. I like to change my sequencers, you know, and like to modify them, but my main aim was always to look for a ‘live’ sequencer.

The Present

Now we are building two very large machines, one is a composition machine which can play 32 voices for polyphonic music. It has three different screens for reading separate pages of the music score and you can see the information for 4 different voices at once out of the 32.

The TV screen is maybe a transition to a new kind of interface — I don’t like them too much because it looks very ‘office-type’ for a musician and is not very practical. Other kinds of display will come very soon.

We have succeeded to build my new sequencer without a screen for the stage, because we have little ‘graphs’ where we can display words and information. The displays actually consist of very little squares with LEDs underneath. You can show words and numbers that move across for you to read an instruction sentence. The LEDs are the smallest square type you get in the most modern equipment now — it looks very good! They’re nicer than the PPG liquid crystal display in the Wave synthesisers because they’re illuminated.

So to play sequences on our next machine will be a big panel where you don’t use any rotary switches any more, but for every note you have a whole row of switches for semitone and octave pitch adjustment plus three types of trigger selection. Even though it’s a big panel, it’s easy to enter notes.

I much prefer mechanical switches to touch sensors by the way — I also hate touch keyboards — I think they’re too abstract. Even if switches beep or have an LED (like the Moog Source), I need the physical feedback of the switch as I play. Many people think it’s futuristic to make switches smaller and less mechanical, but that’s wrong. Next year I think people will be wearing old-style watches because they’re much more fun than digits (but still with the electronics inside).

Our sequencer is designed so that you don’t enter any rhythmic value any more — I call it a ‘trigger selection’ sequencer because if you want a full note you just use up a whole page of 16 rows of buttons. That means every page uses rows of knobs to represent one bar of 16 x 16th notes, although you can change page lengths and double it to have 32 rows of 32nd notes. There are all kinds of copy functions, including transpose. A quarter note (crotchet) uses up 4 rows, a sixteenth note (semiquaver) uses only 1 row. And so you don’t enter a value but just change part of the machine. Rests automatically take place if you don’t enter a trigger and note lengths and times can be also set. Sequences can be then recalled in any order and there can never be a rhythmical error, because the bar lengths are preset. Anytime signatures can be used from 3/8 to 15/16 — 7/4 is my favourite at the moment. I don’t have a very jazzy feeling but still use odd time bars a lot — it’s not just that I’ve played drums, but also piano, violin and trumpet for a long time.

When I tried out strange rhythmical bars they didn’t swing enough for me, but now I’ve found a lot of consolation by using 2 different pulse-structured bars playing against each other. I’m influenced too by Indian and African rhythmical events because they are using not only one measure at the same time. Somehow they make a polyphonic pattern structure. So now I’m experimenting a lot with accents.

So the basic difference between composing and live performance sequencers is that for the composing, you are going to the machine for an idea in your mind and you just programme it. And then, when you’ve almost worked it out, you do one or two changes to continue your composing.

The ‘live’ created sequence is different — you go to the machine and have nothing in mind maybe and just start to try one out — even though it is a very well known trial and error system, it can be a most unique thing because you will occasionally invent melodies and rhythms which you hadn’t thought of before. Normally, I play sounds on the piano or sing them and I get sequences so strange and interesting coming to me in a random way, like a spiritual thing, through an aerial on the top of my head! (laughs) — I get a gift, it’s very strange.

I work for a long time in this way — people do it in science with analogue computers, and there’s still good results from this fiddling with knobs in a creative way.

When composing with sequencers, it is very easy to construct a piece that may become very boring because it hardly ever changes. There are pieces we play where nothing has happened for a while and we still like it, but when the other parts with the sequence, the mood, the structure or just the feel is not right, it’s like a ‘cracked record’! But we now have enough sequencers to allow the music to be changing all the time.

Changing the Sequence

There comes a point where you have to decide what things to change as the sequence plays. Oscillator pitch is the obvious thing to consider, and next might be the filter cut-off to vary the tone. Various kinds of trigger — on, off (for rests), legato, or accent (by varying the VCA control voltage). We use 8 accent levels — we thought about 64, but 8 is just fine.

The latest fantastic parameters which nobody else has done yet is the ‘sound colour’. I can change the whole program for every note.

It’s very complicated to make it work musically because it’s not easy to find sound colours that fit together. If you take a trumpet followed by strings and brass sound it will sound like rubbish, so you must find sounds that fit together.

I am trying to have every note with exactly the programmed sound colours I require — something like using a program preset ‘advance’ button on, say, the Chroma or Memorymoog. This is becoming much easier to do since IC components are now cheaper and Winchester hard disks can be used for memory storage which are very reliable, ‘roadproof’, temperature stable — and small as well. The disks can store an amazing amount of information.

Another decision to make is how many notes to play together, whether from 8 to 32 polyphonic events. Many of the parts are used for drums with often just 4 lines for melodies — never using 32 at once (unless you play a Wagner piece!). By 32, I mean 32 voices with six different parameters. There are 16 trigger lines to the drum machine to play 16 different percussion sounds.

Sampling Sounds

The drum machine we first used to play sampled sounds had an 8-bit memory length of just 400 milliseconds, which was basically okay. This was the one we used for the ‘Choronzon’ piece and so on. Now we’ve moved on to a 12-bit digital memory, which is much better for cymbals and snare drums — it’s much cleaner. Cymbal crashes and other longer length sounds are no problem since we do them with a repeating loop around a short sample. This new unit can store 12 different events, each lasting up to 2 seconds.

Our biggest project yet, which should be completed by 1984, will have 16-bit resolution sampling, with 8 second maximum sample time for every note and every key will be able to have a different programmed sound. I think that sampling a different sound on every note of the keyboard will be a very important development, because you can use it as a drum machine, or play natural or synthesised sounds. We can modify the sound within with various filter and loop processing.

I do a lot of the sample sounds myself and tried at the beginning to use only real drums, but really my intention was not to have these — it was just an exercise, and I use it as a reference. If the real drum is as good as I know it them the system’s fine. But now I’m starting to use very crazy sounds in it — the ‘fun’ part is just beginning! It is possible to get a very good snare drum sound now, although if you’ve only got 8-bits, you should use something else — it’s much more original anyhow and I don’t care for real drums at the moment.

One of my most interesting ‘new’ drum sounds I’ve used is almost indescribable — it’s a mixture of voice, xylophone and other instruments for the one event. I don’t use cymbals at all — I never have done, because cymbal sounds are too traditional for me. When I started to use drums with our music I took out all cymbal-type sounds. I did have one real cymbal sample made and I recorded it for 8 seconds, with a repeat loop in the middle of the sound plus a VCA envelope for decaying the sound — then you have a perfect cymbal (but it’s too real for me).


The structure of a whole piece must be arranged first, otherwise you can get trapped easily as you hear things over and over until you lose the sense of how long a sequence lasts. Normally, we don’t play the same thing more than 50 times. This is where the ‘sound colour’ type of sequencer capability can be fantastic. If you don’t put the same sound colour on the same pitched note, then the same melody becomes quite different each time you hear it.

It’s interesting that people hear more than one melody, even though the pitch is the same — there’s a sound colour melody and a pitch melody.

It is very easy to keep repeating the same sequence for longer than necessary and we had this problem in the early days as the whole group had to decide together how long to play a sequence when we added other things to create a development. We did this by listening and watching on stage during the performance, but now we have more fixed and disciplined structuring that’s prepared beforehand in our rehearsals. I’m almost counting bars again!

One problem that still exists for us is how often to use the sequencer and we have to make sure there’s not too much repetition in a concert lasting some 3 hours.

The ‘master clock’ for running the sequencer comes from a real-time unit I have — it’s basically a fast running oscillator that’s programmable, because we may perform a concert that has individual pieces with at least 9 different tempos. My engineer designed the clock with an output socket and number selection as on a metronome, so 120 represents 120 beats to the minute and so on. All my drum machines, modulation effects and sequencers can be controlled from this to synchronise everything together.

Program Storage

We do not have a Winchester disk system for storage yet, so we use CMOS RAM memory with battery back-up. If I have a power failure or even a transient that causes loss of memory, I will reload the information via a tape interface, but that hardly ever happens. In the last 3 years, the Prophet has lost its programs only 3 times even though we’ve travelled extensively to many countries.

With the sequencers we are now using the Jupiter 8 and the Wave 2 (we also have the Wave 2.2, the Memorymoog and the Emulator in the studio). I am not particularly interested in instruments with built-in sequencers, although I think they are a very good idea for many people. Some of these instruments will tidy up your sequence recording and pull ragged notes onto the beats. I can also do this with a single sequence. I can enter by numbers all the rhythmical events, all the chords and things; and then I can add a lead melody for one voice from the keyboard. Then the computer can shift notes which are a bit off the beat onto the beat. Normally, I don’t use this facility because I like these little shifts that happen by accident.

I always need to ‘feel’ the music line — it can be loose, it can be late or early. If I’m composing a melody line by numbers I would shift the note to get that feel and give them slight delay or anticipation of the beat — that’s the freedom music needs.

We are now finding that our pieces are getting both longer and shorter! We do pieces that last 2 minutes and 50 minutes — that’s surprising really. You might think that when we use sequencers, the music will last longer, but it depends on the type of piece. If it’s rock or disco style, then it’s short. Others will be more meditative or minimal music with a longer development.

I don’t find it that hard to relate my developments to classical music — I’ve done quite a lot of ‘classical’ music sequencing in the past, but instead of using rallentando, crescendo and rubato, it’s more a matter of ‘dynamics’ which I do at the mixing desk. This gives a real fine touch of ‘classically-adjusted’ music of the 16th Century for choir and orchestra.

I also find myself using more and more foot pedals — the GDS which I play has a programmable foot pedal. This means it stores not only fixed parameters but also real-time curves like a sequencer. This is something that a sequencer maybe needs in the future, with the pedal acting like a fader on a computer mixer. As I play the GDS sequencer, I can then programme the volume at the same time (or other things like the filter cut-off). This would be interesting because it would enable much more control of a sequence as it plays, because your hand would still be free for the keyboard and so on.

Looking ahead

I have new units on the racks behind me containing a polyphonic sequencer that has real-time analogue access even though it’s digital, and in ’83, we hope to use the new system we’ve been developing over the last three years. The present sequencers that we have on stage are prototypes for this. With more complex music being performed and the need for more programmable sound colours, the demands on the computer require real-time setting within one 16th note of all the parameters of the sound colour program in use. So the sequencer has to know in advance when to change sound colours — that means it’s not really realtime.

The whole concert could now be put on computer and just played back, but that’s not ever Tangerine Dream’s intention. We aim to have everything eventually with realtime control as well as the possibility of editing at any stage many different parameters. So far we’ve used only one voice in this way and look forward to the new developments which should add another dimension to our playing with sequencers”.