Digital Techniques to aid Pipe Organ Voicing and Regulation
by Colin Pykett
Posted: 1 September 2007
Last modified: 23 December 2009
Copyright © C E Pykett
Abstract. Voicing and, particularly, regulating a pipe organ is fraught with difficulty. Not infrequently the instrument is unsatisfactory in terms of the way its registers blend with each other, its mixtures might scream in the treble or its mutations might be bass-heavy. Some recent organs are widely regarded as total failures because of such problems, which is scandalous in view of their cost.
This article suggests ways in which digital techniques might be used to guide the voicing and regulating processes so that the probability of getting it right first time with pipes is increased. It combines the use of modern digital music technology with the regulation procedure used by the well-known voicer Anton Gottfried, adopted by Ralph Downes for the organs for which he was the consultant in the mid-twentieth century. The most famous of these instruments was that at the Royal Festival Hall in London.
(click to access the desired section)
Voicing and regulating a pipe organ requires not only a high degree of skill but it is beset by many problems, some of which continue to defy solution by organ builders today. If this were not so, we would not find so many organs which are either disappointing tonally or complete failures. Some examples of the problems encountered are:
It verges on the scandalous that such problems continue to appear, given the enormous cost of new pipe organs or major rebuilds of them. For example, a new concert organ installed in England a few years ago for a seven-figure sum in pounds sterling is widely regarded as a failure by experts, and they have said so – in writing. Or the Choir organ in a cathedral in the south of England is so ineffective, even in the quire, that it might as well not be there at all. (It is of passing interest that the mechanical coupler action of this same instrument, by one of our leading builders, is virtually unusable because of the way it reacts to humidity and temperature – the electric coupling option has to be selected instead).
At the other end of the scale one only has to visit smaller churches at random to become amazed at the widespread follies and incompetence on the part of organ builders and their advisers, both in terms of the organs themselves and their placements. That litigation is not resorted to more frequently by customers can only be because they are inhibited by a fear of further expenditure, maybe the embarrassment of publicity, or most likely because they are usually rather gentle, stoical and mild-mannered individuals.
Yet putting the problems right can be next to impossible once an organ has been built because it will involve still more disbursement but, again, with no guarantee of success. To see why this is so, it will be useful to review first of all the processes used by a voicer when voicing and regulating a pipe.
Voicing an organ pipe involves the adjustment of two main parameters: the tone quality or timbre, and its loudness. The adjustment of loudness on its own is termed regulation. If these two factors are correct for every single pipe in an organ, then that organ will be successful. But what is “correct”? What are the practical boundaries and constraints within which the voicer has to work? Can he compensate for a non-optimum pipe scale for example? How were those scales chosen in the first instance? In other words, what guides the designer of an organ before a single pipe is made, let alone voiced? And what guides the voicer towards his goals of correct timbre and loudness when the pipes have been made? Does he have a sufficiently precise mental representation of the sounds he wants to conjure up? These are difficult questions, but we need to be able to answer them before arriving at solutions to the tonal problems which still beset organ building today.
A major difficulty is that virtually all the operations available to the voicer mutually interact. To illustrate this, let us consider what he does when presented with a newly made Principal (flue) pipe. Often this will not speak at all at first, or only with a windy and unsatisfactory tone. Therefore his first job is to “put it on speech” after first ensuring the wind pressure matches that chosen by the organ designer – the pipe must “stand on its wind” before it can be voiced. Then he might adjust the position of the languid and/or the lower lip so that the narrow sheet of air which it directs towards the upper lip meets the lip correctly, rather than dissipating too much of its energy outside or inside the pipe. Then, to get the tone right with the right proportions of harmonics, he will also increase the cut-up of the mouth (the height of the upper lip above the lower). Higher cut-ups result in fewer harmonics, and new pipes are usually delivered to the voicer with very low cut-ups for obvious reasons. Of course, if he cuts up too far he cannot go back again, and the pipe then has to be returned to the melting pot! But after a cut-up operation the languid might then need adjusting again because the upper lip has been raised. When the timbre is finally OK, the pipe might then be too loud, so the wind pressure at the languid will need to be reduced by constricting the foot hole, but this then makes the pipe sound flutier because the higher harmonics get weaker when the air stream gets slower. Unfortunately this would require the cut-up to be reduced to compensate, but we have just observed that this is obviously impossible. And so it goes on ...
As well as all this, a modern voicer will often “nick” the pipe by cutting small slots in the languid, among other things to modify the way the pipe comes onto speech – this affects the initiation transient of the pipe among other things, thus a nicked pipe will tend to suppress chiff for example.
All this has to be done for every pipe. A successful voicer needs an exceptionally precise and rigid aural memory, absolute rather than relative. In other words, he must have a detailed idea in advance of how each pipe is to sound, and he must not allow himself to be led astray from this mental sound image by how the pipe he has just voiced turned out. This is what is meant by the voicer’s “ear”, and few lay people have anything like this capability or can conceive what it means.
Because of all these problems, it is perhaps not really surprising that many organs just do not sound right. It is difficult enough to get it right first time, and often impossible to do much about it if it’s wrong. However, with enough skill and patience (and therefore enough money as well), dramatic improvements can sometimes be made to an organ just by re-regulating it, that is, by readjusting the powers of stops and pipes which are either too loud or too soft.
For understandable reasons, a voicer is unlikely to seek assistance in regulating an organ. This rare scenario would be akin to that of a surgeon who routinely casts around for help in performing operations. However some voicers in the recent past have been honest enough to admit that they faced a daunting task in certain cases, and Ralph Downes mentioned some of them, and the instruments concerned, in his book “Baroque Tricks” . He also described how he himself arrived at an understanding of how to regulate an organ. He then applied it to the organs for which he was the consultant, including his magnum opus at London’s Royal Festival Hall. To my mind the regulation technique he described was far more important in practice than the study of pipe scales with which he developed a near-obsession . Downes was taught how to regulate by the celebrated voicer Anton Gottfried, who applied his method to a large, new but apparently unsatisfactory Skinner organ in the USA at which Downes was presiding at the time. As Downes himself said, he learned truths so simple that he wondered why he had not thought of them before.
In the Gottfried/Downes method one starts with a “datum” stop (my term, not theirs), such as an 8 foot Diapason on the Great organ, and lays down “loudness bearings” across it (my term again), much as the bearings are first laid in frequency on a similar stop when setting a temperament. First, the C’s of each octave are adjusted in power for equality of impact and output (Downes’s phrase this time), then the remaining pipes within each octave are adjusted to suit. The same process is then applied to the 4 foot Diapason, firstly C on the 8 foot versus the corresponding C on the 4 foot, then all the other intervening pipes on the 4 foot. It is repeated for all the other stops making up the diapason chorus on the Great, and then the whole procedure repeated for the other departments. A Table of Balances will assist the process if drawn up in advance, containing as many examples as possible of which stops, and combinations of stops, should balance against which others (e.g. Great Diapason 8 versus Swell Geigens 8 and 4, etc). Examples of such tables are given by Downes in his book .
This technique encourages the realisation of a better-regulated organ which blends well within itself, and it is particularly well matched to those neo-Baroque organs which try to imitate the style of Arp Schnitger’s instruments or similar ones. In these organs nothing overwhelmed anything else, and the addition of each successive stop to a chorus resulted in about the same increase in subjective effect. The method would also have prevented, almost by definition, undesirable phenomena such as the screaming mixtures and bass-heavy mutations which arose frequently in the 19th and 20th centuries and which still appear today. However, I question whether it should be used to the exclusion of any other method. Applying it too rigidly to today’s clones of Gottfried Silbermann’s instruments, for example, could corrupt the feisty upwardly-voiced effect of his Principals, whose impact does increase as one ascends the keyboard. However an obvious modification of the technique should solve this problem – you simply set the upward-voicing (or any other desired prescription) on the “datum” stop, and then regulate the others from that.
If you prefer another method, then use it. However I know of no other which has been so minutely described and which (in skilled hands) can turn an acoustic monster into a musical instrument of beauty. Through his tenacity and persuasiveness, Downes demonstrated it several times on both large and small instruments which can still be heard today, and Gottfried used it routinely throughout his career. Even so, the central problem remains – who is going to pay for this amount of tonal finishing work to be done by someone of great skill on a new or unsatisfactory organ, particularly these days? What is preferable and needed is an inexpensive method of getting it right first time so that expensive remedial work is unnecessary. This is where digital techniques come in.
Imagine that it was not time consuming and expensive to run through the voicing and, particularly, the regulation of an entire instrument, such that you could do it several times in a day or two if necessary until you felt everything was right. Obviously this cannot be done with pipes, but it can with electronics. If you have a suitable digital test bed on which a simulation of the desired pipe organ can be set up, you could do exactly this, and in the building itself if desired (and usually it would be highly desirable). You could then proceed to voice and, particularly, regulate the target pipe organ in the same building using quantitative results and experiences gained from the simulation in the way I shall describe later. Additional confidence should result that you will get it right first time. The digital method could also be used to assist in re-regulating an unsatisfactory pipe organ with, hopefully, a reduction in the time, cost and risk involved.
But the digital test bed or organ which you use (call it what you will) cannot be any old commercial product – it has to have certain characteristics including those following:
My Prog Organ digital organ system satisfies all these criteria, and it is described in more detail on the Prog Organ pages of this website. It has been used to develop several simulations of the pipe organs described in other articles elsewhere on the site . A picture of the current experimental console appears below.
This "collapsible" console is currently used for playing the Prog Organ digital organ system. It can be readily dismantled and transported to other venues, and it can simulate any pipe organ within the constraints of the number of stop keys available. For the purposes of this article it matters little that only two manuals are used, because stops belonging to the other departments of a larger organ can be assigned at will to either of the keyboards. The stop names are not permanently engraved on the tabs; they are attached as required to wooden strips which hang on pegs above the stop keys. This enables any stop list to be simulated in a simple manner.
I shall now take each of the above points in turn by relating them to Prog Organ.
As a consequence of the simulations already developed , I have an extensive library of sound samples covering many stops and several different styles of organ building over the last three centuries or so. The most prized and important (at least to me) are stops in the styles of Arp Schnitger, Gottfried Silbermann, J F Wender and A Cavaillé-Coll. Other styles, probably of less interest in the present context, include those of Hope-Jones, Wurlitzer and Hill, Norman & Beard as well as simulations I have invented myself from scratch. Therefore if it was desired to build a pastiche pipe organ in any of these styles using the techniques described in this article, these sound samples could be used to develop the initial digital simulation.
Otherwise, an organ builder who is interested in using this procedure would need to build up a library of digital sound samples taken from his own existing piped instruments and others of interest. This can be a time consuming exercise, but one can take one’s time over it as the library only needs to be built once. A point of incidental interest to some organ builders might be that complete sound samples of good pipe organs can be copyrighted and sold under license for substantial sums to digital organ manufacturers or enthusiasts. This might cover or exceed the cost of building the library.
Assuming the necessary sound samples are already available in a computer library, it takes me typically a few hours to set up a first-cut simulation of a two manual organ on Prog Organ. My current system allows up to 14 speaking stops on each manual, 10 on the pedals and up to 8 couplers, tremulants or other non-speaking stops. Note that these are not limitations of the Prog Organ system as such, rather they are set merely by the prosaic limits which apply to the number of stop keys on the console I currently use to play the simulated organs.
If a larger pipe organ is under consideration, a two manual simulation could still be used to assist the realisation of a voicing and regulation strategy. One of the simulated departments would be kept the same (e.g. the Great organ), while changing the stops used by the other keyboard to represent successively the other departments. Keeping one department the same should enable the main “datum” stops to be available at all times for reference.
In Prog Organ some parameters can be changed very quickly – in a matter of seconds. These include the overall loudness of any stop, a most important parameter in the context of regulation, blend and balance. In addition it is possible to replace completely any stop with an alternative just as quickly, assuming the alternative already exists within the simulation. Thus, for example, one could have not just one Diapason but several, and select the one which sounds best in the building. Again by way of example, the variants could represent diapasons with different simulated pipe scale progressions (uniform scaling such as Töpfer and derivatives, non-uniform such as Dom Bédos, an organ builder’s own house scale, etc).
Other more detailed parameters, such as the loudness and tuning of individual notes within a stop, can be changed in a minute or so. A wide range of other parameters can also be changed just as quickly on a note-by-note basis.
All changes are reversible. If one should make a pig’s ear of the job while tweaking the simulation, which can happen surprisingly easily in view of the speed and ease with which changes can be made, it is possible to revert to any previous state within a few seconds (provided, of course, that that state had been saved in the computer!). Otherwise you would have to spend rather longer backtracking and re-setting your changes one by one. But no matter what you do, there are no unrecoverable dangers analogous to that of having cut up a pipe too far!
The loudspeakers in Prog Organ are separate from the console. Typically I use 6 independent sound channels, though more could be used. Again, this limitation reflects solely my current set-up. Each channel can either have its own amplifier and loudspeaker(s) or it can be combined with another. Thus, for example, it would be possible to use all 6 sound channels but only 4 amplifiers and loudspeakers, because sometimes it is convenient to cut down the number of loudspeakers for practical reasons in view of their size and weight.
Any stop can be sent through any channel, or it can be split between two channels. The latter enables the spatial variation of sound across a soundboard to be simulated (e.g. the C and C# sides). In this case each note (simulated pipe) of the stop can be split in different proportions across the compass. Being able to split the sound in this way is also useful when simulating the pronounced spatial variation across en chamade stops.
The loudspeakers of the simulation would typically be placed where the pipes of the target pipe organ will be located.
The current system is as easy to transport from one location to another as can be envisaged. The skeleton console of my current set-up is purely functional rather than an elegant piece of furniture, and it disassembles into flat pack form if need be in less than an hour. Thus the whole system – loudspeakers, bench, pedalboard, console, etc - will fit into a small van (smaller than a Ford Transit for example). I have transported it using a medium sized hatchback car towing a small trailer. A picture of the system being demonstrated in a church is shown below.
This shows Prog Organ in a church where it was being played by a gathering of organists and electronic organ builders. Probably the main point of interest for this article is that the entire system was transported to the venue in a medium sized car towing a small trailer, and it took about an hour to assemble it at the beginning of the day and a similar time to remove it at the end. The loudspeakers were separate from the console and were placed in various positions in the building during the course of the day. Nine different simulated organs with entirely different stop lists were demonstrated. Such a system seems well matched to the problem of optimising the voicing and regulation of a pipe organ, using the method suggested in this article.
The following procedure is proposed when using the digital test bed to provide data for voicing and regulating a pipe organ yet to be built, or for dealing with an unsatisfactory one:
A digital method to assist the voicing and, particularly, the regulation of pipe organs has been described, in conjunction with Ralph Downes’s regulation technique which he learnt from Anton Gottfried.
Although the method described here depends on quantitative SPL measurements, there would be little point in presenting lots of SPL tables and graphs because the effects of different organs in different rooms are so great that no purpose would be served thereby. It could also expose me to the charge that I believe that science, method and numbers automatically make for a better organ, whereas in fact I do not hold that view at all. Artistry, experience, craftsmanship and skill are what matter. However I cannot see why a skilled voicer should object to the concept of using today’s digital music technology to assist his work, any more than his cabinet-maker colleague who makes organ cases would eschew the use of today’s power tools. Another analogy is the use of digital photography by artists to capture a scene or a sitter, to facilitate their recall when completing the painting back in the studio.
I am attracted to the simplicity of Downes’s method of regulation, and I admire his tenacity in insisting on its use at a time (c. the 1950’s) when most British organ builders seemed to regard it as a mortal insult. He was unique among British organ advisers in being a professional musician of the highest calibre, having achieved this pre-eminent position through nothing beyond sheer talent rather than an accident of birth or influence. In this respect he was refreshingly unlike the independently wealthy and underemployed dilettantes (superannuated army officers, lawyers and - horror of horrors - real estate agents among them) who for some reason dominated organ consultancy half a century ago. I am only proposing to add to his method of regulation the use of today’s digital music techniques to enable it to be applied to pipe organs more expeditiously, hopefully at lower cost and with an enhanced prospect of a successful outcome.
An example of the value of the technique would be to demonstrate to an architect the folly of enclosing a pipe organ in an inadequate chamber. Putting the loudspeakers of the digital test bed in such a chamber would illustrate its undesirable effects on the sound of the instrument. On the other hand, if no alternative to the chamber existed, experiments with the test bed could provide useful voicing and regulation clues for when the time came to voice the actual pipes. In both cases the use of digital techniques would hopefully reduce the uncertainties involved in building a pipe organ, but at a relatively low cost compared to the total outlay.
Some organ builders’ clients might object to what they see as their pipe organ being designed using an electronic one. To counter this, it could be pointed out that the method described is but a logical extension of the accepted methods which have been used by acousticians and organ builders for decades using electronic test gear. For example, the pipe scales for the Tickell organ at St Barnabas, Dulwich were developed after the frequency response of the building had been measured electronically . A digital simulation of a proposed new pipe organ, or of an existing but unsatisfactory one, is but a sophisticated signal generator which generates signals much better tailored to the requirement than the white noise signals (or similar) which are usually employed by acousticians.
1. “Baroque Tricks”, Ralph Downes, Positif Press 1983. ISBN 0 906894 08 5. (If you want this book, I recommend you consider purchasing it through avenues other than the publishers themselves. Speaking as I have found, they are among the most unresponsive and awkward firms I have ever dealt with. Amazon or eBay will often be the better bet, cheaper of course, and you might well get the original hardback edition instead of the inferior paperback which is the only one now in print from the publisher).
2. Despite a phenomenal amount of research into pipe scales, Downes unconsciously succeeded in demonstrating in his book  that it mattered less than he thought. Apart from anything else, the scales used by Arp Schnitger on which he tried to model much pipework of his organ at the RFH “... exhibited such variety that ... I despaired of getting much further on any such basis ... “ (his words). And he either knew nothing about the scales used by Gottfried Silbermann or he ignored them, because not one example is presented. On the other hand, his method (actually Anton Gottfried’s) of regulation is of considerable importance.
4. “St Barnabas Church, Dulwich, the tonal concept”, W McVicker, The Organbuilder, vol 16, 1998. ISSN 0264-4746.