Michael McNeil has designed, constructed, voiced, and researched pipe organs since 1973. Stimulating work as a research engineer in magnetic recording paid the bills. He is working on his Opus 5, which explores how an understanding of the human sensitivity to the changes in sound can be used to increase emotional impact. Opus 5 includes double expression, a controllable wind dynamic, chorus phase shifting, and meantone. Stay tuned.
Editor’s note: Part 1 of this article appeared in the March 2025 issue, pages 12–16. The Diapason offers here a feature at our digital edition—two sound clips and two videos. Any subscriber can access this by logging into our website (thediapason.com), click on Magazine, then this issue, View Digital Edition, scroll to this page, and click on each soundclip or video in the text.
Prologue
In Part 1 we explored Hartmut Ising’s equation and the voicing factors that affect timbre. We saw how this equation could be used to build a voicing model and how that model was validated in the real world of voicing. In Part 2 we apply that model to the voicing of a Fourniture and show how it relates to a principal chorus. We will see what the model can teach us about the sound of a principal chorus and how it redefines the sound of Gottfried Silbermann.
—M. McNeil
Voicing the new Fourniture
Figure 12 shows the bench and tools used for voicing the Fourniture. There are probably as many methods of voicing as there are voicers. The process used to implement the new model is illustrated in Figures 13.1 to 13.9. Voicers separate this process into “regulation” and “voicing,” where regulation deals with adjustments to power from toes and flueways. Hartmut Ising shows us that regulation is an integral part of the voicing process because toes and flueways interact with cutups to affect timbre.
Figure 13.1: We first fit the tuning slide, mark the expected final length from the top of the languid, and then cut the pipe a few millimeters longer than this mark. If the pipe is left too long we will be voicing at a lower pitch, and if we raise the cutup until the pipe sounds right, the cutup will be too high and the timbre too dull when the pipe is cut down to the tuned length (mild profanity erupts when I have to lower cutups).16
Figure 13.2: We next open the toe just short of the modeled value with a reamer (the tool with the red handle in Figure 12), deburr and chamfer the opening, and then make careful final adjustments to the diameter with a tapered gauge (tenths of a millimeter are important).
Figures 13.3 and 13.4: If you have an unvoiced pipe, you may need to find a way to open the mouth for raising the cutup. There are many ways to do this; I opted for a very small 0.042-inch drill bit on a Dremel tool (about 1 millimeter and roughly half the modeled cutup of the smallest mixture pipe). I scribe the cutup from the modeled value on the upper lip. I then drill and make cuts a bit below this value because it is always easier to cut metal away than to add it back. A magnifying glass on a headband, seen in Figure 12, is useful for voicing mixture pipes; tenths of a millimeter in cutup are critical here, especially on the smallest pipes.
Figure 13.5: Trim the cutup and sides of the mouth and deburr the upper lip (a few swipes with a small lip raiser works well).
Figure 13.6: Adjust the flueway depth to the modeled value with a lip raiser and check it by inserting a tapered brass or steel wedge through the toe. Keeping the flueway depths to a precise value was a new departure for me, but it turns out to be critical if you want to control the interactions between flueways and cutups. The tapered brass rod used in Figure 13.6 has three scribed lines at 0.4, 0.6, and 0.8 millimeters of depth. In Figure 13.6 the flueway depth is between the second and third scribed lines at 0.7 millimeter (the third scribed line is hidden below the lower lip).
Figures 13.7 and 13.8: Adjust the upper lip so that it is aligned with the lower lip or slightly protruding. Check the speech. If the pipe is slow to speak or easily overblows, check the height of the languid. A correctly adjusted languid will be close to the position where you can just see the lower edge as seen in Figure 13.8 or slightly lower. Using a steel rod through the toe or the top of the pipe, tap the languid up or down at each side by very small amounts and check the speech. Finer adjustments can be made by adjusting the upper lip in or out. These adjustments align the vortex issuing from the languid to allow the upper lip to intersect it, producing the pulsations of pressure that drive the resonator to sound. Dropping the languid moves the vortex into the pipe, and raising the languid moves the vortex out of the pipe.
When everything is correctly adjusted, there will be a position of the languid and upper lip that makes the speech fast, slow to overblow on higher pressure and clear in timbre. There were instances in which no positions of the languid or upper lip would prevent overblowing on slightly higher pressure, and in all cases these were due to my errors in the cutups or flueway depths. Remember that I first scribed the cutup from the bottom edge of the uncut upper lip. If the pipemaker allows the upper lip to overhang and descend below the lower lip, the scribed cutup will be too low!17 This happened on several occasions, and the pipes would easily overblow. Remeasuring the cutup height from the lower lip and raising the cutup to the modeled value solved the problem.
In the very rare cases where the cutup did not solve the problem, I found that the flueways were too shallow. “Eyeballing” the flueway depths may save you time, but I learned that this does not always work well. Making the flueway deeper would normally make the pipe more powerful and more ready to overblow on fully open toes, and this is Ising’s assumption. But the areas of the toes and flueways in the lower-pitched pipes of this mixture are nearly equal (see the row “Ratio, A toe/A flue” in Figure 7 of Part 1). When you make a flueway deeper with more restricted toes, the pressure in the foot vents and the pipe becomes more stable and less likely to overblow. Restricted toe diameters and deep flueways are common characteristics of classical French voicing, and this is why I used the name Fourniture on this new mixture. Germanic voicing, with its much more open toes and more restricted flueways confirms our intuition, and the pipe will indeed be more ready to overblow as we deepen the flueway.
Figure 13.9: We are nearly done. Cut the pipe down to the tuned pitch and then cut off about 1 or 2 millimeters more to allow some tuning margin for the slide, burnish down the burrs at the top of the pipe on a mandrel, and adjust the tension of the tuning slide. The voicing worksheet allows me to voice all three of the pipes speaking together on a single note as a group, and we can now tune all three pipes on the chest. Figure 14 shows some of the voiced and tuned Fourniture pipes on the windchest.
Some reflections on voicing the Fourniture
Ising shows us why we encounter problems when we use recipes of cutups that are ratios of mouth widths. Cutup recipes force us to adjust the toes and flueways to produce the desired timbre, and this affects power balances and speech. Ising teaches us to “regulate” toes and flueways to get the power and the speech we want, and then raise cutups to “voice” the timbre we want.
As voicers know, there are many combinations of toe diameters, flueway depths, and cutups that yield similar timbres and power, but the speech onset is different. This is why I used Silbermann’s generous flueway depths and more restrained French toe constants as a starting point—I wanted Silbermann’s blend and gently articulate fast speech. Figure 15 shows this voicing on the lower pitch of a façade pipe with Andreas Silbermann’s bolder and warmer nicking. Nicking quickly disappears above 1′ in this chorus.
I want to stress that my use of the term blend is subjective. I hear it as the seamless fusion of the sound of two or more pipes, and it appears to be associated with the deeper flueways of Romantic voicing and a uniform promptness of speech in the chorus. The depth of the flueway has a strong effect on the form of the vortex. When toe and flueway areas are approximately equal, deepening the flueway will yield more flow and more energy in the vortex. But the pressure and velocity of the vortex will be lower, resulting in a stronger fundamental with a lower Ising value and a warmer timbre.
The sound of a principal chorus
Figure 16 shows the Ising timbre values of the Fourniture and its principal chorus. Although my scales and power balances are different, the timbres are similar to those of Silbermann’s principal chorus shown in the dashed blue line. The basic trend is a bass and tenor with Ising’s fastest-speech timbre of 2 and a steep reduction in timbre brightness above 2′ pitch.
The bottom of Figure 17 shows the spectra of the 16′-8′-8′-4′ foundations in a D-major chord spanning tenor D to middle D, and in the image at top we see the addition of the Fourniture. It is a splendid sound in meantone, and D major features one of its eight pure major thirds.
The sound of D. A. Flentrop
I applied the voicing model to the principal chorus of the 1965 Flentrop Hoofdwerk at Saint Mark’s Cathedral in Seattle, Washington, to understand the instrumental brightness of its sound.18 The pressure of the Flentrop chorus is 80 millimeters and close to the model assumption of 85 millimeters. All of the pipes measured in Figure 18 sit on the windchest with no pressure losses from offset tubing, and I raised the pressure compensation in the model at 4′ pitch from 88% to 100%. The Ising timbre values in Figure 18 show that the instrumental quality of Flentrop’s sound is concentrated below 2′ pitch and above 1⁄2′ pitch. Flentrop’s foundations in Figure 18 are much smoother in timbre between 2′ and 1⁄2′. The higher pitches of the 8′ Octaaf have a decidedly vocale quality at the console, and the low 1.47 Ising value at 1⁄2′ pitch confirms what we hear.19 The smoothness in the 2′ to 1⁄2′ range is the key to the absence of overbearing harshness in Flentrop’s exciting instrumental sound.
The sound of vocale voicing
The vocale sound is typically characterized by a smoother timbre from higher cutups. Contemporary vocale voicing appears to show a preference for controlling power at the flueways, but we find deep flueways in the ancient vocale prototypes in the village church of Krewerd (circa 1531) and the Reformed Church of Oosthuizen (1521), both in the Netherlands.
John Brombaugh provided me with his notes on the pipework at Krewerd, taken in June of 1971 when the organ was dismantled for restoration. He noted that “The windways [of the Gedackt] are quite wide in general, and the sound of the pipes ‘pops’ on in a most special way. The trebles of the principals are extremely high cutup and have wide windways. Many of these pipes do not overblow in the usual way and develop the very intense vocale sound. Sometimes with extreme pressure these pipes will overblow.”20 The only flueway he measured, the tenor B pipe of the 2 2⁄3′ Quint (a 2⁄3′ pipe), was noted with an exclamation point—it was 2 millimeters deep, three times deeper than Silbermann’s very generous flueways at this pitch, and off the top of the graph at any pitch shown in Figure 3 of Part 1.
Photos of the organ at Oosthuizen, posted briefly online at the time of its most recent restoration, were taken from a vantage point at the top of the case, looking straight down into the flueways of the façade pipes in each flat, all of which were very deep. Scaling and voicing data of the façade pipes is shown in the table below.21 The wind pressure was measured by Flentrop at a robust 87 millimeters. The pipe lengths suggest a pitch at just over 1 HT from A=440, and the Normal Scale cutups (NSMH) in the table are adjusted by +1HT.
Pitch Dia. MW Flue Toe Tc MH NSMH
8′ F 84.0 70.0 1.40 10.7 1.28 30.0 +6 HT
4′ 67.0 58.0 1.20 10.3 1.45 19.5 +3 HT
4′ f# 55.8 50.0 1.20 10.8 — 15.4 +3 HT
2′ 42.0 36.3 1.20 10.0 2.15 13.0 +5 HT
2′ f#′ 39.5 30.0 1.00 10.4 2.83 10.4 +6 HT
1′ 27.0 22.5 — — — 6.5 +1 HT
1′ a′′ 19.0 16.5 1.00 6.0 1.70 6.0 +9 HT
[Flentrop data (1966), Fisk data (1974?), extrapolated values for calculating timbres]
These deep flueways were a significant factor in E. Power Biggs’s praise of the “brave tones” heard in <Soundclip 1>.22 John Brombaugh remarked that the sound of the Oosthuizen organ on this recording strongly motivated his interest in organbuilding. We can use the data in the table to derive Ising timbre values for the façade pipes (taking care to adjust the pitch frequencies by +1HT).
The Ising timbre values in Figure 19 will drop with the application of nicking, and the façade pipes of the Oosthuizen organ are nicked. Charles Fisk noted 11 nicks in the tenor C pipe, 7 nicks at tenor F-sharp, 6 nicks at middle c′, 5 “big” nicks at f-sharp′, and 3 nicks at a′′ (with the notation “yes” for nicks in the other pitches). Nicks help the vortex to form more quickly at the languid edge, giving it more distance to expand toward the upper lip with an effectively higher cutup. Ising timbre values diminish as nicking becomes bolder, and the timbre values in Figure 19 might drop by perhaps as much as 0.2. Even with this drop, the brightness in the timbre of its “brave tones” is clearly evident.
Note that the brightness tracks the toe constants. The flueways of the Oosthuizen pipes are deeper than Silbermann’s, and they are clearly not used to control power or timbre. The mid-range brightness increases with lower Normal Scale cutups and more open toes. The sound of this organ has been characterized as “intensely vocale,” and the intensity of the sound is a brightness that far exceeds contemporary vocale voicing. This raises an interesting question: Why is this sound characterized as vocale? The answer and the appeal of this sound to Brombaugh and Fisk lies in its very deep flueways. In 1975, about a year after taking this data, Charles Fisk wrote passionately about the musicality of deep flueways.23
How do we characterize the sound of Gottfried Silbermann? The voicing model shows that the deep flueways, restricted toes, high cutups, and lower Ising timbres of Gottfried Silbermann’s powerful voicing have strong affinities to the ancient vocale style. This differentiates Silbermann from the sound of Arp Schnitger and D. A. Flentrop.
Slow and fast speech, or slow and fast voicing?
Slower and faster speech is not to be confused with the common voicer’s term of “slower and faster voicing,” which is an unstable inclination to overblow to the octave, not from the cutup or pressure, but from the languid and upper lip positions. “Faster” voicing with lower languids will cause a pipe to overblow “faster” on slightly higher pressure. To illustrate the difference, Gottfried Silbermann used very high languid positions with stable and slower voicing, while his deep flueways and high cutups produced faster speech.
Flueway depths, toe diameters, cutups, languid heights, and upper lip positions all interact to change the power, promptness, timbre, and stability of the speech. Knowing which of them to adjust and to what degree is not always obvious. The Ising equation helps us understand what needs to change.
The flow of wind
There is a school of thought that believes that closing the flueway with open toes will achieve any desired level of power and timbre, and while this is largely true, it has limits. A constricted flueway will slow the formation of the vortex that issues from the edge of the languid. As a case in point, we hear slower speech in neo-Baroque voicing when it employs extremely closed flueways.
Romantic voicing is characterized by warmth, excellent blend, and very deep flueways, but it constricts the toe, and this, too, has limits. Romantic speech is fast until the area of the toe becomes significantly smaller than the area of the flueway it feeds, i.e., when the toe flows much less wind than the flueway. As a case in point, we hear very slow speech in theatre organ flue pipes voiced on exceptionally high pressures with extremely closed toes.
These examples show why speech will not “pop” on with a restricted flow of wind in either the toe or the flueway, and this implies upper limits on the pressures that will permit toes and flueways to be opened enough to produce fast speech.24 Gottfried Silbermann’s manual pressures ranged from 55 to 97 millimeters, varying by the size of the room and the absorption of reverberation.25 Wind pressures will steeply rise, of course, in the acoustical black holes of outdoor pavilions and large rooms with no reverberation.
The vortex and the resonator
Ising tells us that speech becomes progressively slower as timbres ascend above 2 with more brightness. This is the result of cutups that are too low, and it has nothing to do with resistance to the flow of wind. Ising’s slower speech is caused by a mismatch between the impedance of the vortex rising from the flueway and the impedance of the pipe resonator that the vortex drives to sustain sound. Building on Ising’s work, John Coltman used the electrical concept of impedance in 1968 to describe how organ pipes produce sound. Coltman’s math is difficult, and I made a short video using a mechanical analogy to show how the impedances of the vortex and resonator interact <Impedance video>. Cutups are the key to matching the impedances for faster speech. For those brave souls who want to dig deeper, Coltman’s original work can be accessed.26
Neo-Baroque voicing occasionally produces a slow, gulping speech caused by both sources of slower speech: high and bright Ising values on the verge of overblowing from a very low cutup and high resistance to the flow of wind in a severely closed flueway. Figure 20 illustrates a pipe on the left with fast speech and a pipe on the right with slower, gulping speech. The speech of both pipes can be heard in the <Fast and slow speech video>. The first example in the video is the pipe on the left in Figure 20, and the last example has the slower, gulping speech of the pipe on the right. Each example has three repetitions, and the last repetition is held to show the harmonics on the spectrum analyzer. If you closely watch the spectrum analyzer on the pipe with slower speech, you will notice that the fundamental harmonic rises more slowly than the higher harmonics. This is a graphical demonstration of gulping speech and the difficulty of a resonator trying to respond to a vortex that has been cut too low. Compare these sounds with the cutups, flueways, and toes that produced them in Figure 20. These pipes are what I had on hand and do not sound the same pitch, but they are similar in scale and demonstrate differences in speech quite well.
Promptness of speech in a chorus
D. A. Flentrop appears to have found the minimum flueway depths that would preserve fast speech with a good blend, and his voicers clearly understood the Ising interactions between toes and cutups.27 Andreas and Gottfried Silbermann opted for much deeper flueways than Flentrop, and their speech is fast as well.
Slower speech can be very musical, but the promptness of the speech, whether slower or fast, must be uniform for a crisp chorus attack. My analysis of the organ at Saint Maximin in France showed that the Isnards used lower toe constants and shallower flueways than the Silbermanns, and they went to great pains to equalize the toe and flueway areas of each pipe in the Grand Orgue foundations in the 4′ to 1′ range. The relaxed and elegant speech of this voicing is the result of a mild resistance to the flow of wind in the toes and flueways <Soundclip 2>.28 In the row labeled “Ratio, A toe/A flue” in Figure 6 of Part 1, we see that Gottfried Silbermann also used virtually equal toe and flueway areas in the 4′ to 1′ range, but those areas are larger with less resistance to the flow of wind, and Silbermann’s speech is faster.
§
Hartmut Ising has given us a powerful tool to understand some of the complex interactions we encounter during voicing, drawing back the veil on much of its mystery. Readers are welcome to email the author for copies of the voicing model and/or John Brombaugh’s elegant pipe length model.29
Notes and references
All images are found in the collection of the author. The numbering of the notes and figures is continued from Part 1.
—It is strongly recommended to use Sony MDR 7506 headphones for the soundclips. Earbuds will not generate bass sound.
16. Charles Ruggles, who lives near me in Colorado, worked for John Brombaugh in Ohio. He related that he became adept at cutting off the resonator of a pipe, trimming it, and resoldering it with a lower cutup at a time when John was experimenting with higher cutups, testing its limits. Brombaugh considered higher cutups an essential feature of the vocale sound.
17. This can happen when the upper lip is not burnished in enough and does not align with the lower lip (it overhangs the lower lip). If the toe is slightly tilted when it is soldered to the resonator, the upper lip can descend below the lower lip by just a fraction of a millimeter. Although rare, this is more likely to happen on pipes with very wide mouths. This is of no consequence on larger pipes, but on small mixture pipes in the 1⁄4′ to 1⁄8′ octave, just a few tenths of a millimeter can be enough to make the pipe easier to overblow when the cutup is measured from the bottom of the upper lip. Blowing by mouth lets you test for this instability by varying pressure and the attack.
18. Michael McNeil, “The Sound of D. A. Flentrop,” The Diapason, volume 115, number 9, whole number 1378 (September 2024), pages 14–20.
19. Ibid.
20. John Brombaugh, unpublished typescript, 1971. The term “wide” here unequivocally means flueway depth. Brombaugh and I discussed this, and I pointed out that the flueway’s “width” is the same as the width of the mouth. The term “depth” would more logically refer to the degree to which the flueway is opened, and John agreed.
21. The voicing data on the Oosthuizen façade pipes by Charles Fisk (1974?) and the scaling and wind pressure data by “JAS” (Johannes Steketee) of Flentrop Orgelbouw (1966) were kindly provided to the author courtesy of Blair Batty just in time for the publication of this article. Mr. Batty was a pipemaker for Charles Fisk at the time the voicing data was collected by Fisk.
22. <Soundclip 1> [00:26] of the Oosthuizen organ, Psalm 106 by Maistre Pierre (John Calvin), was played by Johannes Steketee of Flentrop Orgelbouw in the circa 1969 recording by E. Power Biggs, The Organ in Sight and Sound, Columbia, KS 7263.
23. Charles B. Fisk, “Pipe Flueways,” Music: The AGO and RCCCO Magazine, December 1975, page 45. A reprint of the article can be found on the website of C. B. Fisk, Inc., www.cbfisk.com/writing/pipe-flueways/.
24. Michael McNeil, The Sound of Pipe Organs (CC&A, Mead, Colorado, 2012). See pages 114–116 for the math describing the combined effect of resistance to flow from the toe and the flueway on the promptness of pipe speech.
25. Frank-Harald Greß, Die Orgeln Gottfried Silbermanns (Sandstein Verlag, Dresden, 2007). Greß has carefully researched original wind pressures and noted later changes.
26. John W. Coltman, “The Sounding Mechanism of the Flute and Organ Pipe,” The Journal of the Acoustical Society of America, volume 44, number 4, 1967, pages 983–992. Coltman, a concert flutist and an engineer at Westinghouse Research Labs in Pittsburgh, Pennsylvania, makes many references in this paper to Ising’s earlier work. One will find Coltman’s model for Ising’s fastest-speech timbre of 2, where the impedances are opposite in sign and matched in magnitude, and Ising’s overblowing timbre of 3. Coltman addresses only the fundamental harmonic, while Ising’s equation addresses timbre and is useful for voicers.
27. McNeil, “The Sound of D. A. Flentrop.”
28. McNeil, The Sound of Pipe Organs, pages 165–166. <Soundclip 2> [00:32], Isnard organ, Saint Maximin, Louis Marchand, Tibi Omnes Angeli, Bernard Couderier, BNL A, SCAM/BNL 1995.
29. Email: mcneilmichael83@gmail.com. Permission to provide the Brombaugh pipe length model was kindly granted.
