For a number of years sampling technology has been viewed as the best available technology and sound in the digital organ world; indeed, to the average church goer, more or less indistinguishable from a good pipe organ. Put very simply, with this technology a number of notes from a pipe organ are ‘sampled’ (i.e. recorded) and then digitised and stored on the onboard organ computer for playback from its memory when requested.
Of course, as you would expect, with sampled sound there is great variation in the quality, with much being determined by the original quality and length of the recording and the amount of computer processing subsequently involved. In essence, the longer the sample the better, with as little computer processing in use as possible. Makin and Johannus have very much led the way in developments in this area and customers have watched how year on year the sound quality is enhanced.
With both a pipe organ and sampled organs there are three definite phases to the sound: the start of the note, the holding of the note, and the release of the note. In each phase there are definite sound characteristics which play a part in providing a realistic sound, and indeed this is a very complex subject where many nuances can be taken into account, such as fast repetition of notes where wind is already available in the pipe. Makin and Johannus have very much dealt with this particular aspect of sound generation and realism in recent years with many technological advances
However, perhaps the most important aspect to date is that a ‘loop’ of the sample is required for when a note is held. Sample loops are very varied with cheap and cheerful organs only having a sample loop of a second or so, which as you can imagine become very wearing on the ear. For sometime now with the Monarch technology used in Westmorland Custom organs our sample loops have been a minimum of five and in some cases ten seconds. Such samples provide incredible realism and thankfully, since computer memory is now much cheaper, are now within the financial reach of our customers.
Makin organs do not share samples between different stops, indeed for our mixtures we have separate samples for each rank! In the pipe organ world, unless it was an extension organ, you would not expect the Swell Open Diapason to use the same pipes as the Great No 2 Diapason. Therefore, if it is not usually the case in the pipe world, why should it be done in the digital organ world? The answer, of course, is that this will save the manufacturer some money; hence this practice may be found at the cheaper end of the market. But it is the customer that loses out since two different ranks of similar pipes, such as the Diapasons, should not be voiced alike!
One other dubious practice sometimes employed by certain organ manufacturers is to use computer algorithm to convert the sample of one stop to another, for example a Dulciana into a Gemshorn. This can be used to fill in the gaps where a company doesn’t have good pipe samples of a particular stop, rather than taking more samples of the right stop which is expensive to do. From the customer’s point of view this is very much a false economy, since it is so difficult to achieve a convincing sound this way, and a well trained ear will easily be able to spot this.
One subject that is not mentioned by most manufacturers is how many samples are actually used for each individual stop. The reality is that, in most cases, for a 61 note stop such as an Open Diapason, there are only one or two individual notes sampled per octave; so the 61 notes of a rank are probably made up of only 10 or so actual samples, with the other notes being generated by computer algorithm. In the past this was seen as an adequate solution to providing a good sampled sound. However, with the advent of very cheap computer memory, this is perhaps one cost-cutting short cut that is no longer needed; and indeed its removal would dramatically enhance the realism of the overall sound.