A talk on Electronic Organs given by

Michael Anthony 7^th December 2002

• pipe sounds and harmonics
• the human ear and its response to sounds
• the development and application of electronic technologies to simulate pipe sounds
• don’t forget the caveats

The subject of electronic organs is a wide and, for some people, a very emotive area. For the purposes of this talk, I shall be concentrating on those instruments designed for use in leading church worship in all its varied forms, and for home practice, principally by those involved in that vocation.

This is in no way to dismiss the genre of instrument designed and equipped for the field of "entertainment", but these deserve a separate occasion, venue (and speaker) to do them full justice in their own right.

We shall explore the nature of pipe organ sound generation and composition, the workings of the human ear in relation to frequency response, and the development and application of the technologies used in these instruments with the appropriate consequences for music making.

The evening will be essentially a "top level", steadily progressive overview, with references given for more detailed personal reading of the in - depth use of these technologies.

Before we can start to examine the electronic organ and our expectations of it, we must understand the nature of the sound generated by a pipe organ so that we have a grasp of what we expect technology to deliver - i.e. what are the targets?

We can see that the wave has a repeating pattern in time, but how do we find out about its component parts? Fortunately, about 1800, long before we had the audio technology to access the wave, a French mathematician, Jean Fourier, produced a very elegant mathematical process to analyse any periodic waveform.

Harmonic analysis:

Applying his technique to our sample wave, we see that it can be reduced to a succession of very simple "sine" waves, each having a different frequency (i.e. the number of repetitive cycles per second - the more cycles the higher the pitch, the fewer cycles the lower the pitch).

We can now see that each flue pipe has up to the sixth harmonic present (and also higher ones at much smaller levels which may extend into or beyond the limits of our human hearing) (q.v.).

Considering the arithmetic, we see that for the note Middle C (256 c/s*) at 8’ pitch the harmonics are 2x - 512 c/s C, 3x - 768 c/s G, 4x - 1024 c/s C, 5x - 1280 c/s E, 6x - 1536 c/s G etc., i.e. a minimum equivalent keyboard compass of two and a half octaves of harmonic content above the fundamental.

Notes:

String stop pipes and reeds have an even higher harmonic series extending into the teens of harmonics, and beyond, with the consequent increase in equivalent keyboard compass e.g. the sixteenth harmonic of the Viol D’orchestre for middle C is 4096 c/s, which is also a C, four octaves above the fundamental (or even for the top manual note of fundamental 2048c/s up to 16,386 c/s).

So, a pipe organ having basic pitches at 32’, 16’, 8’, 4’ and 2’ with a five octave manual compass generates at least 5 + 4(harmonics) + 2 (32’ & 16’ pedals) + 1.5 (4’ & 2’ manuals, since some of the harmonics are already needed and counted for the string stop) i.e. in excess of 12 octaves of frequencies, with much of the characteristic timbres vested in the higher harmonics.

The actual frequency band is from 16c/s for a 32’ stop to beyond 16,000c/s for the string harmonics.

If our electronic instrument is to come close to emulating a pipe organ, then its technology must be capable of producing this quantity of individual notes for the mixing process.

A more detailed examination of flue pipe speech is available in an article on Colin Pykett’s website

Human hearing is a very personal and individual thing; no two people will perceive a particular timbre in an identical fashion. In a concert hall the acoustics will play a very important part in conditioning the sound wave as it leaves the instrument on its way to our ear. Ideally there should be an equivalent experience in all areas of the auditorium, but we know that this does not happen - for instance the bass line of an orchestra may sound tight and unified in one location and less well defined and more diffuse in another. The same will pertain for an organ. Where we sit, how many others are present and ultimately the precision of the performance will all affect what information arrives at our ears.

The ear is a wonderful creation and we are still learning more about its exact method of working (see article "The Power of Hearing" in "Physics World" ^(2a)). It can detect pressure signals from very loud and painful levels of sound to a whisper, with an pressure ratio of about 18 million to one.

We are all aware that our hearing falls off as we grow older, both in loudness sensitivity and also in the curtailment of the perceptible frequency range. At a recent touring lecture for schools promoted by the Institute of Physics an unscientific "straw poll " test of the audience’s hearing capability showed that the younger members present could hear up to about 14kHz whilst for the adults present the upper limit was only about 10kHz^(2b).

Do we know what our own personal hearing pattern is? This is what makes our sound perception unique, and, therefore, any subjective assessment of sound sources only personal and not definitive. This is particularly significant in relation to organ tone synthesis since, as we have discovered, a lot of crucial information is contained in high frequency harmonics for which we may well have impaired sensitivity, and thus incomplete information to form a reasoned judgement.

Furthermore, our wonderful ear plays its own tricks as we listen to sounds. If it feels that a sound is incorrectly balanced e.g. weak bass, it will manufacture its own amplified bass from the information it receives by creating a difference tone to boost the incoming weak bass. So again, what we perceive may not be an accurate representation of the original sound balance.

It is immediately obvious that there is a suppressed bass sensitivity and also a high frequency loss for each ear (some of which may be age related), but also a significant sensitivity imbalance between the ears themselves.

All of these factors could give this person a very difficult task of identifying some of the high frequency tonal nuances associated with tone synthesis.

The first requirement for waveform synthesis is that we generate at least the twelve notes of the chromatic scale at individually and mutually stable frequencies, according to the selected temperament (q.v.). Here we shall only consider electronic instruments having distinct pitches, not the continuously variable variety e.g. the Ondes Martenot.

There were many attempts at this with several early, unwieldy instruments being created, but the first compact instruments to be commercially available were the Hammond family. Who has played these war horses and not been initially frustrated by having to hold the switches for long periods as it winds itself up? But what is this "winding up"?

Let us recall our early school science lessons with a bar magnet and some iron filings. We place a piece of paper over the magnet, sprinkle on the filings, agitate the paper, and, Hey Presto!, a pattern of lines appears. These, we are told, are lines of force, and, the closer they are together, the stronger the magnetic field around the magnet. Furthermore, if we introduce a second magnet with an opposite pole (N - v - S), the lines become much more tightly packed between these poles, indicating a very strong magnetic field.

It was that great genius and the Father of modern electrical engineering, Michael Faraday who, in amongst all his ground breaking experiments in the early 19^th century, showed that if an iron bar with an insulated coil of wire around it is passed through such a strong magnetic field, then an electric current is produced in the coil. Faraday had discovered the phenomenon of Electro - Magnetic Induction. In 1821 he also used the reverse of this process by applying a current to the coil and causing the bar to rotate in the magnetic field - Faraday had produced the first electric motor.

This technique of induced current generation was refined over the years, and another engineer called Barlow showed that if a toothed wheel was rotated in a magnetic field and the induced current tapped off, then pulses of current could be produced, with the speed of rotation controlling the frequency of the pulses and the shape of the teeth on the wheel governing the shape of the pulse.

In 1933-4 Laurens Hammond designed the first of his organs using the rotating wheel system, and this appeared in public in 1939. It had 91 discs with different profiles of peaks and troughs which were used to generate 91 individual frequencies. A quick division shows us that this corresponds to seven and a half octaves of notes.

These notes were then combined internally in predetermined ratios to give the preset "stop keys" such as Tibia, Flute and Diapason.

Who could tell the difference between them all? With only seven and a half octaves of harmonics available for synthesis, these tones were always going to be deficient in harmonic development and missing some of the core individuality conveyed by higher harmonic content, and thus would all sound very similar. The real innovation of the Hammond system was the array of user adjustable drawbars, where the player could select and mix personalised harmonic composite tones. The drawbar feature is still used on their sophisticated instruments of the 21^st century.

These instruments use an analogue process of Additive Synthesis.

These instruments have become icons and many are still around today in a very playable condition - critics of electrical obsolescence please note! However, for the reasons given (and a few others which will emerge later) they fell short of the target of true pipe emulation.

Following the Second World War there was a great surge and improvement in audio technology. In particular, better microphones became available with a wider frequency response, thus allowing better wave "pictures" to be taken and hence giving a more refined knowledge of harmonic content. Also, soon after the War the semiconductor was born, but it was many more years before this invention would revolutionise the world of waveform synthesis.

Early semiconductor devices were based on treated versions of a material called Germanium, but this was not sufficiently thermally stable to give reliable frequency generation, and it was not until the appearance of Silicon based devices that designers could achieve the inherent stability necessary for musical purposes.

Why was the advent of the semiconductor so special, other than for its stability?

Unlike valves (those glowing glass bottles so familiar in our old radios) these devices were physically very small, and operated on very low voltages (up to 10 - 20 volts only, compared with 200V or so for valves). This meant that there was a great reduction in the weight and bulk of electronic hardware and a substantial increase in safety (who remembers accidentally touching a valve top cap connector and having a glowing nose!),and also that new signal processing capabilities existed.

Because of the small scale of these transistors it was now possible to generate a whole new range of waveforms with sharp profiles, e.g. a square wave and a sawtooth (ramp) waveform (sketches).

Why are these so important? A square wave contains a full range of harmonics in precise proportions and a sawtooth has a pattern of harmonics closely resembling that found in reed pipes.

Thus, by using special circuitry to filter out the unwanted harmonics or to emphasise selected ones, a closer representation of the pipe compositions could be achieved.

This is "Subtractive Synthesis", and still an analogue process.

It was claimed by many that the individual timbres were good (acceptable, passable?) depending on your personal polarisation, with the reeds being especially improved over earlier systems. However, the ensemble of the various choruses always seemed to fall flat and was not a successful sum of the (reasonable) component stops. Why?

So far we have considered only our "simple" waveform, an instantly switched on and switched off sine wave, and applied it to a single note (pipe). Clearly, this does not represent the true behaviour of a pipe nor the practical minimum of music making.

A pipe is a mechanical system with airborne inertia along its length. From the moment we press a key and admit the air, there is a finite time before the steady note is established - there is a period of instability known as the starting transient; also when we release the key there is a finite time as the sound decays, albeit not as long as the starting time. So these are two aspects of pipe speech which our electronic synthesis to date has not yet addressed.

Again, for more information, see the referenced website article on flue pipe speech.

Furthermore, if we play a four - part hymn tune on a single stop there will be four such waves created, and if we play on, say, an 8’+ 4’+ 2’ chorus there will be twelve, and if we add a pedal line at 16’ then we will have thirteen waves in the chamber, each causing some measure of interference with the others, principally as a subtle variation of the steady state volume of the rank. This interference is unique to those stops (i.e. pipes) in that specific soundboard configuration and in that specific location. It is a unique ambience and ensemble effect.

An electronically synthesised organ using analogue techniques can only hint at producing this effect.

So, here are just three areas which our simple wave model to date has not considered. It is possible to add circuitry to each stop to disturb its switch-on and release characteristics, but the ensemble is much more elusive, and this explains why many analogue instruments produced unrewarding tutti effects. An early attempt which should be mentioned is the Leslie tremulant. This consisted of a fan rotating over the front of the loudspeaker cone to interrupt the steady flow of sound energy into the auditorium, but with limited success.

We have one of the largest analogue instruments ever built at the Royal Concert Hall in Nottingham. It was provided, amid great controversy, by Copeman Hart, and remains a "piéce de résistance" of its genre, certainly Ernest Hart’s largest analogue instrument.

Electronic Synthesis using Digital techniques.

Whilst all this development of analogue instruments was evolving, the Allen Organ Company had been experimenting with a new form of tone synthesis - digital, and were so convinced of their processes and the results that, in 1971, they patented their techniques, thus preventing anyone else from replicating them for almost 20 years; hence the surge of digital sound sampled instruments appearing in the early 1990s once the patent had expired

However, a group of researchers at Bradford University devised the basis for a completely different and alternative method of digital tone synthesis which came to the commercial market in the early 1980s, frequently bearing their name.

But what exactly is digital synthesis? As has already been hinted, there are two distinct systems in use, one based on "sound sampling" and the other on the "Bradford System" and its evolving derivatives.

Let us return to our now not - so - simple waveform with its starting transient, release pattern and the superimposed ensemble/ambience effect on its steady state amplitude. Let us also recall during our childhood days when we drew on sheets of paper clowns juggling or horses galloping, each image marginally different from its neighbours, and then flipped them past our eyes. Our clown magically tossed his apples (or whatever) into the air and the horses legs moved. Each picture was a snapshot in time which became part of a whole moving image. The more pictures we drew in each repetitious cycle, the smoother was the apparent movement.

Well, digital sampling of pipe waveforms is exactly the same idea. Instead of leaving the microphone permanently powered to record a continuous (analogue) waveform, we now "fire" it (pulse it on and off) many times a second to record on appropriate instrumentation/computer the unique progression of time/amplitude co-ordinates of our wave (sketch). The commonly used sampling rate is just over 44,000 times per second. In order to capture the true tone of a pipe across its compass many sampling points will be taken, thus capturing the variations in timbre in different pitch bands. For example, with middle C at 256 c/s this would give in excess of 170 sampled co-ordinates per cycle of the wave.

This is the starting reference point for all digital synthesis - a set of computerised codes for each stop, thus enabling the manufacturer to build a reference library of tones. The difference between the two types of commercial organ stems from how each manufacturer treats this coded information. Digitally encoded signals have the great advantage that, unlike their analogue counterparts, they are not corrupted during processing, so integrity of tone information is preserved.

"Sound sampled" organs are offered by such companies as Allen and Viscount.

Here they take their code library and implant it into memory chips so that when a stop is drawn and notes played the appropriate codes are released, converted back to an analogue signal, amplified by conventional audio amplifiers and played through the loudspeaker. As samples take up a lot of memory space only comparatively short time samples are stored, so that when a long note is held, the steady state tone is produced via a synchronised repeating loop of this sample until the note is released. The number of samples and the length of the loop have a direct effect on the final tone quality of the sound and the accuracy of reproduction, with, for example, a short loop and many repetitions leading to poorer quality than a more detailed sample code offering a longer loop time. All harmonic content details are inextricably contained within this encoded information.

"Real Time" (Bradford System) organs are offered by such companies as Copeman Hart, Eminent and Norwich. (Apologies to all other companies not named in either category!)

They use a completely different philosophy for their synthesis. In the simplest terms, they consist of a series of digital number generators which produce individual codes for each harmonic of each pipe tone for each note. When a stop is drawn, the generators spring into life and create the codes appropriate to that stop, and when a note is played, these harmonics are mixed in the predetermined proportions as derived from the original pipe samples for those notes at their respective pitches. This is a return to the additive synthesis process we first encountered with the early Hammonds. This process allows suppliers to come along with their computer and change these harmonic codes and their proportions to vary the starting transient, steady state and release characteristics.

Again, these mixed digital signals are then converted back to analogue for audio amplification etc.

Further reading sources are given which explain these processes in more detail for those who wish a deeper insight⁽³⁾.

There are also other factors which affect the final sound.

The electronic architecture employed greatly affects the quality of the sound. The codes are made up of "bits" of data. For example, 20-bit processing allows for over one million variations of a single encoded string of data (2²⁰). The higher the "bit" value, the more information that can be encoded, but with the consequent increased requirement for processing hardware.

Once the codes are converted back to analogue signals they become subject to the many similar and various problems we encounter at home with our own audio systems. The amplifiers must be of adequate power output to operate in the selected auditorium below the point where overload distortion becomes an issue, and the type of loudspeakers used must be able to handle the power delivered over the full frequency spectrum required with minimum (zero) distortion. Further, it is desirable to split the multiple signals for each division of the organ into several audio channels to retain clarity of the sound and reduce other forms of distortion which can occur if all signals are passed through one amplifier/speaker. The placing of the loudspeakers in the auditorium is also of prime importance in achieving a spatial separation of sound channels.

Remember, in order to communicate effectively in its environment, the audio systems of the organ must be capable of moving air, and plenty of it!

In the last few years the sampled sound instruments have taken another step forward and incorporated further voicing options into the consoles so that these are now available to the player at any time, without the need for "the little man and his magic box" to pay a visit. This means that it is now possible to independently access the three principal parts of the waveform code, namely the starting transient, steady state and release along with independent rank tuning and ensemble effects, so long the preserve of the "real time" instruments only.

Indeed, there is now becoming so very little to choose between the realism of the sound technologies of either camp that if one is called upon to select a particular instrument for home, church or public auditorium then the inherent quality of the sound synthesis almost goes without saying; the choice may well come down to differences in the range of console management features and other facets of the total package.

Electronic instruments can also offer the option of selecting various Temperament settings. There is a good mathematical article (accessible via this Society’s website) which spells out the arithmetic of the various systems. Note that this article uses concert "A 440" pitch reference, with the consequent values being long decimal numbers.

Also, look before you listen! If you see few speakers or small speakers then be aware that for all the advanced technology within the organ, the delivered sound could be very compressed and not a realisation of the full system potential.

This is like buying a widescreen television and viewing through a pinhole in a piece of card.

I could take you to such an installation very near here. Listening to it one could easily make a sweeping statement and say that all electronic organs are rubbish, even though this case represents a "bad hair day" for one supplier. Further persuasion of the customer should have led to a slightly reduced (cheaper) specification instrument and a better external sound system within the budgetary considerations.

There is a rule of thumb which indicates that approximately one third the cost of the organ should be spent on the external sound system. So, for example, if the selected organ cost £12,000 then a further £3000 - 4000 should be spent on the sound system.

However, if the budget maximum is £12,000 then about £2,500 - 3,000 of this should be given to the sound system, leaving up to £9,500 to spend on the instrument itself. This option would clearly have implications for a revised choice of instrument.

Recently, some members of this Society visited Peter Collins’ workshop for a discussion and demonstration of a hybrid instrument for a church in Trono, Sweden, in collaboration with Allen. This approach has been used in America for many years, but its use in this country is still regarded with much suspicion and, indeed, vilification from some sectors of our fraternity. Whilst it was stressed that this was a one - off venture for Peter Collins, future collaborations were not ruled out. The day proved to be very informative for those with open minds and ears.

Remember that tonight’s talk has been about the science and technology of the genre, and only a limited area of that, it was never the intention to open a debate on the issue of the appropriateness of choosing a pipe based instrument versus an electronic, although a few thoughts on the subject may be of interest for those of you who visited this church ten years ago. You will notice a significant difference in the appearance of our Sanctuary.

When this church was faced with the scenario of an ailing pipe instrument in 1995, many factors of the Church’s worship life had to be considered. These included the reducing size of the traditional but versatile Choir and the consequent need for a reduced level of accompaniment, their remote location viz - a - viz the organist given the physical layout of the Sanctuary, the changing pattern of service music and the changing nature of service attendances i.e. different large and small congregations which required a completely different approach from the organist on the same limited instrument.

In that complex situation, at that stage of the Church’s worship life moving forwards, the choice of an electronic instrument enabled all of the criteria to be addressed (successfully it seems some six years on!). The removal of the still functioning or reasonably repairable parts of the pipe instrument to various good homes also opened up considerably more flexible space in the Sanctuary itself and an adjoining vestry.

There is no simple blanket answer to that ultimate choice. I believe that each situation must be carefully and prayerfully examined on its own merits and an appropriate decision taken for that fellowship at that time and again, most importantly, moving forwards.

A further article accessible via this Society’s website also gives an in - depth discussion of the physics of the pipe organ.

Having arrived at our technological marvel, we shall now explore what we can do with our instrument, other than simply play it.

There is one word, or acronym, which dominates the world of electronic music, and that is MIDI - Musical Instrument Digital Interface.

A useful introduction to the MIDI system and its applications can be found in the referenced publication⁽⁵⁾ .

MIDI is simply a system of command codes devised and agreed by the principal manufacturers in the popular music business in 1981-2, which allows several instruments to be connected to each other and, with one Master keyboard, all the other Slave keyboards can be controlled in terms of voice selection, timbre, attack volume etc. This allows many sounds to be layered, which was particularly important in the earlier days of electronic keyboards and synthesisers, when most keyboards were only monophonic i.e. only supported a single part.

There is the option to connect our organ to a home computer and, via different software packages (e.g. Sibelius or Soundforge), to compose our own music and amend it on the computer, or to play our repertoire into the software and reproduce our efforts on the screen, thereby highlighting any performance shortcomings or deviations from the printed score. Perhaps this is a possible subject for expansion at another meeting, again at a different venue with a different speaker (selective glances!!)

I would also refer you to another excellent article on Colin Pykett’s website^(1b) where he explains the MIDI system with particular reference to pipe organs, and discusses its use for controlling some pipe organs with detached consoles, but also points out the potential problems of applying a simple operating system to a complex instrument, sometimes playing complex music.

As we are all aware, the introduction of new music and songs into our Worship has brought with it many alternative instruments, some offering live performances and others recorded reproduction.

We shall consider a few of the recorded systems available.

It is also worth saying at this point that there are several recorded hymnal libraries available which are straight organ recordings onto CD, some being aligned with specific hymn books, whilst others simply cover the most widely used tunes.

There are other libraries recorded in MIDI format which require a MIDI player. These have a built - in store of synthesised instrumental sounds which are then selected and recorded as MIDI reference codes onto floppy disc. On replay, the codes activate the appropriate voices in the instrumental library and a fully fashioned song accompaniment is produced. These songs are self-contained within the MIDI player which usually has its own audio system, but could also be connected to a church’s own sound system for better leading of the congregation - Sample.

A different approach is possible with the disc recorder whereby an organist can record his instrument onto floppy disc, including all registration changes, so that accompaniments are available using a locally recognised sound source played in a familiar way, without the organist being present at the actual service e.g. holiday - Sample, with audience participation!

There are, of course, some provisos for using this method, mainly upon the preacher to have the hymns available, and not to change his/her mind about the number of verses on the day, since the organist will have recorded the hymns with registrations appropriate to the mood of each verse. Imagine singing a triumphant final verse to pianissimo shimmering strings!

Yes, service music does require co-operation all round for all of the time.

As we have found tonight, singing to a "black box" is an acquired art which is not as flexible as some manufacturers and vendors would have us believe. However, some churches have no other alternative, so please do not let us dismiss applications of technology as unwanted intrusions, since they could be the musical lifeline for some fellowships.

As already hinted, the instrument here (Viscount Domus 1332) provides good, versatile worship support, and so I shall now play some excerpts of pieces which illustrate the more extreme options of registration which you would most probably not find on a short session of familiarisation. I have been playing it for almost six years, and I’m still learning some of its hidden ways.

(Selected) Variations on "My young life hath an ending" Sweelinck,

Noel No. 10 in G Daquin

A Christmas surprise, as you’ve never heard it before!!

Tuba Tune Porter - Brown (Dive for cover!)

I trust that this evening’s talk has proved informative in some respect, and that you can each go from here having felt an insight into the world of music technology and the many possibilities for its application in your own sphere of playing.

Grateful thanks are given to all original authors for use of their information.

1 a) and 1 b) Colin Pykett's two articles on pipe speech and MIDI
2 b) and 2 b)  Physics World, May 2002 articles
3 Organists’ Review, August 1999. p275 "Electronic Technology" from Colin Peacock of Wyvern.
4 a) and 4 b) Articles referenced via this website Alternate temperament / The physisc of the organ (see Links page) 
5  "What is MIDI?". Basic Theory and Concepts by Helen Casabona and David Frederick published by Alfred     Publishing Co., Inc. ISBN #0-88284-364-8

Colin Pykett also wrote two further articles on Electronic Organs for Organists' Review, published in August 1998 and August 1999, the latter issue also being that referenced in 3 above.

To return to Original Articles close this window.