Last updated: 27 November 2015. Click About This Website for update list.
This website does not invade your privacy by inserting cookies into your computer (more >>)
For over sixteen years the most stable and extensive resource on the Internet for pipe and electronic organs
The hub of this site is the Complete Articles page which gives you instant access to many detailed articles dealing with numerous technical aspects of both pipe and electronic organs. Use the Google search box below to quickly identify areas of interest. While browsing, why not also listen to over 4 Ĺ hours of music played on the three manual organ below and the Prog Organ virtual pipe organ here?
SEARCH THIS SITE
!! NEW ARTICLE !! - Some novel observations on organ pipe sounds and their frequency spectra
This article shows that the four classes of organ pipe tone - flutes, diapasons (or principals), strings and reeds - have frequency spectra characterised by significant amplitude scatter which often obscures their systematic structure. It is therefore remarkable that our ears have little difficulty in deciding that the sounds from a given rank of pipes belong to the same stop on an organ. However groupings of harmonics have been identified in the spectra analysed over some forty years of research, and each can be associated with a linear trendline fitted to the harmonics of its group. Using additive synthesis it has been found that the subjective tone quality of a pipe is little changed when it is reconstructed from its trendline values rather than from its actual harmonic amplitudes. In any case, such differences as do exist between the actual sound of a pipe and its reconstruction are often eclipsed by those occurring naturally between adjacent pipes owing to the gross variations in their spectra. In most cases only two trendlines are required in each spectrum.
This finding might be significant in aural perception because it is a means of reducing the confusing welter of data in real spectra to a much simpler generic form which nonetheless retains the essentials of the sound of an organ pipe which satisfies the ear. Because the ear and brain process both amplitude and frequency logarithmically, the trendlines themselves on a similar log-log spectrum plot take the simplest possible form, namely straight lines rather than curves. This being so, each trendline is defined using only two parameters - its slope and its intercept on the amplitude axis. Thus a spectrum with an arbitrary number of harmonics can be reduced to just four numbers which nevertheless seem to encode its essential subjective features. This represents a considerable dimensionality reduction in a spectrum with many harmonics which has implications for acoustic pattern recognition, whether implemented by brains or machines.
It is well-established that brains extract linear features when processing visual sensory inputs, and it is therefore possible that similar neural mechanisms might operate when processing the auditory sensory data arriving from the inner ear. An amplitude-frequency spectrum is merely a 'picture' which must be projected within the auditory cortex in some way, and it is conceivable that linear features within it - such as the trendlines discussed here - are extracted as part of the acoustic pattern recognition processes in the brain. There would seem to be implications for the mechanisms of musical perception in these suggestions.
The picture above is of a test rig used for experiments on pipe organ valves, such as those described in the articles entitled Calculating Pallet Size, Touch Relief in Mechanical Actions and Response Speed of Electric Actions. These can also be accessed from the Complete Articles page where summaries are also available.
This electronic organ is a dual purpose instrument containing both "straight" and "theatre" voices, designed and made by the author. It is tuned to the author's Dorset Temperament with the addition of some impure octaves as described in Keyboard Temperaments with Impure Octaves. A full specification is available for download here (PDF file, 117 kB).
The things they say:
These recordings span some years and they were made in various rooms and auditoria. The older tracks were made using analogue equipment and some were recorded acoustically using microphones, hence the occasional noises due to piston thuds and page turns, etc. Other tracks were captured electrically. All are of real players performing in real time - no synthetic MIDI 'performances' here. I have not got round yet to normalising the volume settings of all the tracks so they are compatible with each other, therefore you might wish to adjust the volume between tracks depending on which ones you select. Do not be alarmed if some tracks appear to start with an excessive noise level - this simply means they were recorded at a higher level than others. Just turn the volume down to suit. In any case, it is a wise precaution to always begin playing each track at a low level to protect your audio equipment and your ears from unexpectedly high signal levels when the music begins. Although the instrument has 13 ranks of theatre organ voices in addition to its 'straight' sounds (see specification), copyright considerations preclude the inclusion of much theatre-style music here. Playing time 1 hour 35 mins approx.
Website copyright © C E Pykett 1999-2015. See About this Website.