©Lin Yangchen
Home-made manometer for measuring wind pressure, using standard aquarium rubber tubing of internal diameter 4.1 mm and a ruler graduated in both millimetres and inches.
Click to sections:
Toe hole
Scaling
Cut-up
Flue gap
Nicking
Slots
Pipe quirks
or scroll down.
Toe hole
The toe hole diameters of a number of pipes were measured using a vernier caliper, the mean of two orthogonal readings being taken from each pipe. Toe holes mainly affect loudness by regulating the volume of air entering the pipe per unit time. Larger toe holes also breed more harmonics as the faster air is compressed into a narrower jet striking the upper lip (see Colin Pykett comm.). The three-dimensional shape of the hole also matters (Reiner Janke comm.). But the toe hole is not the only determinant of the rate of air flow at the languid; the dimensions of the foot are a factor, and the windchest itself may be delivering different flow rates to different holes depending on its construction.
The correlation between pitch and toe hole diameter is fairly universal despite the multitude of pipe nationalities, makers and varieties (denoted by different symbols in the scatter plot). A notable deviation is the metal stop of unknown provenance in the Flute 4′ (left), with its rather small toe holes displaced downwards from the main regression line. Curiously, the highest pipes of the 4′ have the smallest toe holes even though they are an octave lower and much fatter in scale than the Zimbeln at the top of the 2′. The overall range of toe hole sizes in the 2′ is smaller than that in the 4′: there is a limit to how small toe holes can get.
Another outlier is the disproportionately large toe hole of the experimental bamboo pipe made by Jerry Ng. The pipe isn’t greatly different in timbre than the pipe next to it, at least up to about 20 kHz (see acoustical analysis). It could be that it has a double-L-shaped duct between the toe and the flue that needs extra horsepower to negotiate. Another reason could be its considerably thinner flue which compensates for the size of the toe hole and vice versa. Or perhaps high pipes all sound the same anyway because no one can hear their harmonics above 20 kHz.
Scaling
Pipe scales in the Flute 4′ compared with Töpfer's Normalmensur (dashed line), which is given by the calibration point of diameter 155.5 mm at 8′ C and the equation dn = d1/2^((n − 1)/(h − 1)) where d1 is the diameter of the starting pipe, dn is the diameter of the nth pipe and h is a constant halving number which in this case is 17.
The metal pipes of unknown provenance are somewhat thinner than Töpfer’s principal scale, although a visiting organist regarded them as having an English diapason or German romantic tone somewhere between a flute and a principal.
The Klais pipes, stamped on the pipes as principals, are a little fatter than the Töpfer standard. The slight shortening of the Klais pipes by Navaratnam to move them up a semitone had little if any effect on scale, as shown above in comparison with four unaltered pipes (plotted at their intended pitches) from the same original set. When rolling the pipes, Klais appears to have stepped up the mandrel diameter every other pipe.
The Sauer Cornett is fatter still, as expected for this species of stop, but as a solo mixture V including a third-sounding rank it would have produced quite a different resultant timbre than that of the individual pipes.
The scaling of the Navaratnam wood pipes is not directly comparable with the others since the pipes are stopped. Nevertheless they are rather narrow flutes which are a better match for the higher pipes in the stop.
The diameters of the metal pipes of unknown provenance halve at the 17th to 18th pipe as determined by reading off the raw data. The halving number of the Klais pipes was determined by a polynomial regression to be close to 17. That of the Navaratnam pipes was determined from a linear regression to be between 22 and 23.
But you know what? Scaling need not be taken too seriously. According to Colin Pykett, Arp Schnitger’s scaling was quite haphazard. During his time, things weren’t so scientific. But he voiced his pipes into a beautiful sound anyway.
Measurement tools: Avinet precision spring scale made in USA, Stanley PowerLock, Bausch & Lomb 10× Hastings triplet loupe, vernier caliper. Micrometer screw gauge shown later. Raw data are available from the Auferstehungsorgel pipe database.
Cut-up
The lower the cut-up, the higher the mode frequencies of the mouth tone generated by the air jet on the upper lip. This influences the character of the attack transient, which tends to have amplitude peaks close to the mouth tone frequencies (Angster et al. 2004). See Steenbrugge (2012) for further discussion of the acoustical mechanics of cut-ups.
Mouth height is quite well-behaved across the Flute 4′, decreasing smoothly up the scale (left). The Navaratnam wood pipes have high mouths for their effective inner diameter (right) partly because they are stopped.
Flue gap
Measuring flue gap width using a micrometer screw gauge and strips of paper inserted into the flue. Paper is easy to find and work with and won't damage the flue; steel feeler gauges can't be bent and are too big for small pipes. When adding a paper strip takes the fit from too loose to too tight, half the thickness of that strip is added to the measurement. Thinner paper will give more accurate results. The anvil and spindle should grip the paper surface firmly without compressing it; the fingers can feel a gentle but definite onset of resistance if the thimble is rotated slowly and steadily.
Correlation between pitch and flue gap width. Steenbrugge (2012) discusses how flue gap width influences fluid dynamics.
Nicking
Nicking of the languid, and sometimes labium, dampens the attack transients, reduces noise from air flow around the mouth and prevents random mode switching (Angster et al. 2017, Colin Pykett comm.).
Number of languid nicks per millimeter on the middle pipes of the Flute 4′ showing two approaches to nicking: one as a controlled pipemaking parameter that progresses in a predictable series through the rank (metal pipes of unknown provenance), the other as a final voicing adjustment to compensate for random variations in timbre (Klais). The nicks on the metal pipes of unknown provenance are also considerably deeper and wider than those on the Klais pipes, supporting the hypothesis that the former was a design feature and the latter was a fine adjustment.
Languid nicking on the (left–right) pipe of unknown provenance G5, Klais G#6, Walker A#6 and Bevington B6. The Bevington pipe is an example of the extremely heavy nicking popular in the first part of the 20th century (Hendrickson 1976; also see earlier scatter plot). Even the pipes an octave higher on the Piccolo 2′ (not shown) are less densely nicked. The nicks were almost certainly all made at the same time, judging from their even spacing and similar shape. The small variations in nick shape and orientation appear to be from differences in pressure and direction of application of the same tool. It is unclear what tonal effect the heavy nicking has had on this particular pipe, which also has an unusually wide flue gap (also see earlier section). The higher wind pressures, desire for duller accompanimental timbres and prevalence of dead acoustics at the time of its manufacture are possible factors (Hendrickson 1976). Yet the attack transient doesn’t sound any smoother than those of the other pipes, and the stationary state is actually brighter and louder instead of the other way round as would be expected. The Bevington pipe has a similar scale and cut-up to the rest, but there could be other differences not revealed by the data that were compensated for by the nicking.
The Klais E6 of the Flute 4′. Nicks can have highly irregular spacing as the voicer adds more nicks between existing ones, sometimes with different-shaped tools making nicks of different shapes and sizes on the same pipe. Here each nick appears to have been made simultaneously on the languid and labium with either a blade-shaped implement or a triangular one.
Slots
Slots enhance the higher harmonics (see acoustical analysis). They were used extensively by Aristide Cavaillé-Coll in his principals, which have a “hardness” between the fluty principals of Schnitger and the stringy principals of Silbermann (Colin Pykett comm.).
Slots that don’t go all the way to the top leave a section of complete circumference between the top of the slot and the top of the pipe. The higher this so-called ring section, the more horn-like the tone (Reiner Janke comm.).
The slot widths of the metal pipes of unknown provenance in the Flute 4′ have a variable but roughly proportional relationship with pipe scale (left). The ring height, however, decreases more slowly than the speaking length (right).
Pipe quirks
Walls of different thicknesses in the bottom E of the Flute 4′. The thinnest panel is slightly tapered. The front of the pipe is on the left. There is no detrimental effect on sound.
The bottom B-flat is made of pine with a maple back panel.
Bullet hole in the side of the foot of pipe 20, with no audible ill-effects.
References
I am grateful to Jerry Ng, Alphonsus Chern, Chris Bragg and James Atherton for insights that helped improve the article.
Angster, J., Rucz, P. & Miklós, A. 2017. Acoustics of organ pipes and future trends in the research. Acoustics Today 13(1):10–18.
Angster, J., Wik, T., Taesch, C., Sakamoto, Y. & Miklós, A. 2004. The influence of pipe scaling parameters on the sound of flue organ pipes. The Journal of the Acoustical Society of America 116:2513.
Hendrickson, C. 1976. Nicking. The American Organist.
Steenbrugge, D. 2012. Flue organ pipe operating regimes and voicing practices. Proceedings of the Acoustics 2012 Nantes Conference pp. 2789–2794.
Toe hole
Scaling
Cut-up
Flue gap
Nicking
Slots
Pipe quirks
or scroll down.
The toe hole diameters of a number of pipes were measured using a vernier caliper, the mean of two orthogonal readings being taken from each pipe. Toe holes mainly affect loudness by regulating the volume of air entering the pipe per unit time. Larger toe holes also breed more harmonics as the faster air is compressed into a narrower jet striking the upper lip (see Colin Pykett comm.). The three-dimensional shape of the hole also matters (Reiner Janke comm.). But the toe hole is not the only determinant of the rate of air flow at the languid; the dimensions of the foot are a factor, and the windchest itself may be delivering different flow rates to different holes depending on its construction.
The correlation between pitch and toe hole diameter is fairly universal despite the multitude of pipe nationalities, makers and varieties (denoted by different symbols in the scatter plot). A notable deviation is the metal stop of unknown provenance in the Flute 4′ (left), with its rather small toe holes displaced downwards from the main regression line. Curiously, the highest pipes of the 4′ have the smallest toe holes even though they are an octave lower and much fatter in scale than the Zimbeln at the top of the 2′. The overall range of toe hole sizes in the 2′ is smaller than that in the 4′: there is a limit to how small toe holes can get.
Another outlier is the disproportionately large toe hole of the experimental bamboo pipe made by Jerry Ng. The pipe isn’t greatly different in timbre than the pipe next to it, at least up to about 20 kHz (see acoustical analysis). It could be that it has a double-L-shaped duct between the toe and the flue that needs extra horsepower to negotiate. Another reason could be its considerably thinner flue which compensates for the size of the toe hole and vice versa. Or perhaps high pipes all sound the same anyway because no one can hear their harmonics above 20 kHz.
The metal pipes of unknown provenance are somewhat thinner than Töpfer’s principal scale, although a visiting organist regarded them as having an English diapason or German romantic tone somewhere between a flute and a principal.
The Klais pipes, stamped on the pipes as principals, are a little fatter than the Töpfer standard. The slight shortening of the Klais pipes by Navaratnam to move them up a semitone had little if any effect on scale, as shown above in comparison with four unaltered pipes (plotted at their intended pitches) from the same original set. When rolling the pipes, Klais appears to have stepped up the mandrel diameter every other pipe.
The Sauer Cornett is fatter still, as expected for this species of stop, but as a solo mixture V including a third-sounding rank it would have produced quite a different resultant timbre than that of the individual pipes.
The scaling of the Navaratnam wood pipes is not directly comparable with the others since the pipes are stopped. Nevertheless they are rather narrow flutes which are a better match for the higher pipes in the stop.
The diameters of the metal pipes of unknown provenance halve at the 17th to 18th pipe as determined by reading off the raw data. The halving number of the Klais pipes was determined by a polynomial regression to be close to 17. That of the Navaratnam pipes was determined from a linear regression to be between 22 and 23.
But you know what? Scaling need not be taken too seriously. According to Colin Pykett, Arp Schnitger’s scaling was quite haphazard. During his time, things weren’t so scientific. But he voiced his pipes into a beautiful sound anyway.
The lower the cut-up, the higher the mode frequencies of the mouth tone generated by the air jet on the upper lip. This influences the character of the attack transient, which tends to have amplitude peaks close to the mouth tone frequencies (Angster et al. 2004). See Steenbrugge (2012) for further discussion of the acoustical mechanics of cut-ups.
Nicking of the languid, and sometimes labium, dampens the attack transients, reduces noise from air flow around the mouth and prevents random mode switching (Angster et al. 2017, Colin Pykett comm.).
Slots enhance the higher harmonics (see acoustical analysis). They were used extensively by Aristide Cavaillé-Coll in his principals, which have a “hardness” between the fluty principals of Schnitger and the stringy principals of Silbermann (Colin Pykett comm.).
Slots that don’t go all the way to the top leave a section of complete circumference between the top of the slot and the top of the pipe. The higher this so-called ring section, the more horn-like the tone (Reiner Janke comm.).
The bottom B-flat is made of pine with a maple back panel.
Bullet hole in the side of the foot of pipe 20, with no audible ill-effects.
I am grateful to Jerry Ng, Alphonsus Chern, Chris Bragg and James Atherton for insights that helped improve the article.
Angster, J., Rucz, P. & Miklós, A. 2017. Acoustics of organ pipes and future trends in the research. Acoustics Today 13(1):10–18.
Angster, J., Wik, T., Taesch, C., Sakamoto, Y. & Miklós, A. 2004. The influence of pipe scaling parameters on the sound of flue organ pipes. The Journal of the Acoustical Society of America 116:2513.
Hendrickson, C. 1976. Nicking. The American Organist.
Steenbrugge, D. 2012. Flue organ pipe operating regimes and voicing practices. Proceedings of the Acoustics 2012 Nantes Conference pp. 2789–2794.