Pipes versus digital

[Southern Comfort]

the non-directional quality of pipes

Well, that will depend on whether they are mounted in a proper case or not. If they are, the sound will certainly be directed and focused in one principal direction.

I still maintain that the essential difference between pipes and loudspeakers is that with pipes a very large surface area of pipework is vibrating and sending sound waves through the air in various directions. With speakers, a much smaller area of loudspeaker cone is sending waves in whatever direction the cone is pointing in. But the important point is the comparative area of vibration, not the direction.

[manniemain]

the non-directional quality of pipes

Well, that will depend on whether they are mounted in a proper case or not. If they are, the sound will certainly be directed and focused in one principal direction.

That's certainly true because with a Theatre Organ where the pipework is all enclosed and sometimes ducted too (yuk!) there's still a strong difference between pipes and digi.

[mcb]

the essential difference between pipes and loudspeakers is that with pipes a very large surface area of pipework is vibrating and sending sound waves through the air in various directions. With speakers, a much smaller area of loudspeaker cone is sending waves in whatever direction the cone is pointing in. But the important point is the comparative area of vibration, not the direction.

What follows from this, SC? In the end the sound waves have to make their way into to the outer ear, i.e. into a very confined space indeed. If the listener walks around it's probably easy to hear the spatial separation of pipes relative to loudspeakers (though our choir enclosure at floor level has a C-side and a C#-side just like ranks of pipes). But if the listener is stationary should we expect there to be an audible difference to do with the spatial origin of the sound (assuming a sufficient distance from the source)?

[Southern Comfort]

At last someone has spotted that my initials are the same as those of the Latin title of the Liturgy Constitution! Well done, mcb.

I think the answer is that the characteristics of sound waves emanating from ranks of pipes are in fact very different from those emanating from banks of loudspeakers (thought I had said this earlier in this thread), so the fact that all sound waves eventually end up in the ear of the listener doesn't really enter into the debate. The point is the nature of the sound that is coming over. A good digital can reproduce many of the characteristics of pipes, but, the sound-producing agent being different, the sound itself can never be identical. This can easily be demonstrated with the right kind of scientific equipment.

[mcb]

I think the answer is that the characteristics of sound waves emanating from ranks of pipes are in fact very different from those emanating from banks of loudspeakers (thought I had said this earlier in this thread), so the fact that all sound waves eventually end up in the ear of the listener doesn't really enter into the debate. The point is the nature of the sound that is coming over. A good digital can reproduce many of the characteristics of pipes, but, the sound-producing agent being different, the sound itself can never be identical. This can easily be demonstrated with the right kind of scientific equipment.

I don't think this is true, at least not true within the bounds of human hearing. You can reproduce a sound digitally so that it sounds identical to the original - any irreducible differences (given top quality recording and playback devices) will stem from quantisation, that is resolving a continuous sound pressure wave into a stream of numbers. The human ear can't hear above 20 kHz or so (a great deal lower by the time you're my age!) so any sampling rate above 40 kHz will provide all the frequency information the ear can make use of. If the signal is sampled at 16 bits (CD quality) that gives each sample 2 to the power 16 (= 65536) different possible degrees of loudness, and the human ear can't distinguish that finely.

Then there's the quality of the hardware doing the digital-to-analogue conversion. But that comes down to how expensive your kit is, and with the best ones the output will, once again, be indistinguishable (within the bounds of human perception) from the 'real' source; at least, once you've added the resonance effects of the building acoustic.

So what does that leave, beyond hi-fi buffs' pieties?

[Southern Comfort]

I think I'm talking about the actual shape of the wave-forms, not their pitches or amplitudes. It is interesting that the most sensitive recording engineers have now returned from digital to analogue recording because their ears tell them that the sound just isn't the same - they use descriptions like "sterile".

I suspect we're going to have to agree to differ on this one, mcb

[mcb]

Hmm... One man's sensitive recording engineer is another man's pious hi-fi buff, I suspect. I'd be very surprised if you were right about a general return to analogue recording.

I think I'm talking about the actual shape of the wave-forms, not their pitches or amplitudes.

You mean timbre? In the end it comes down to the same thing - wave shape is determined by the frequency, amplitude and phase of the component sine waves, and there ain't nothing else.

But this is definitely a subject where loyalties and convictions run deeper than religion! You're right about agreeing to differ.

M.

[contrabordun]

In the end it comes down to the same thing - wave shape is determined by the frequency, amplitude and phase of the component sine waves, and there ain't nothing else.

This is true of course at the ear - the problem has always been creating an array of speakers that produce in their cones, frequencies, amplitudes and phases that, once the sound waves have reflected all around the building, generate the frequencies, amplitudes and phases at each listener's ear, that the listener would expect to hear from pipes.

It's relatively easy to produce an 'exact' copy of a pipe organ at the loudspeaker, because as mcb says, all you need to do is to reproduce the F,A & P experienced by the original sampling tool, but as the sound spreads out into the building, the reflections and so on become a more significant part of the listener's experience. (This is why most digital organs sound best through headphones - certainly mine does!). The key to creating something 'realistic' is having lots of channels of amplification, and speakers so that you get lots of independent sources of sound.

Now of course, pipe organs are themselves notorious for problems of balance within and between divisions, and also for sounding different in different parts of the building, and the larger the organ and building, the worse this gets, so it's not all one way traffic.

[Southern Comfort]

I think Contrabordun has put it in a nutshell. What arrives at the listener's ear can never be identical when the natures of the sound sources are so different. Thank you, cb !

At the risk of not yet agreeing to differ with mcb, I am also convinced that, while the digital folks have got the initial transients of organ pipes down to a fine art (although artifical chiff is added across a range of pitches, rather than each pipe having its own characteristic chiff - this is a problem when you sample only one note in each octave and then use the same sample for the rest of the octave; only a very few manufacturers claim to sample every blessed note of every rank) it is quite obvious that they have not yet really tackled the final transients of notes. The decay of notes on almost all digital organs is so artificial that it screams "I am not real" at you. Add artificial reverb and you exacerbate the problem - natural reverb in the room can ameliorate it to a certain extent, but not completely.

[mcb]

What arrives at the listener's ear can never be identical when the natures of the sound sources are so different.

It's the "can never" that's out of place. There's nothing in the science, so to speak, that makes it impossible. You may be right or wrong as to how close the best of the digital crop comes to achieving it at present.

[contrabordun]

It's the "can never" that's out of place. There's nothing in the science, so to speak, that makes it impossible.

I agree with this - a hypothetical digital organ that had an independently channelled amp and speaker for each pipe, spatially arranged around a large case, would come very close to the pipe sound. (While being approximately as expensive as an actual pipe organ of course).

So if you think of a continuum of quality with the above at one end and today's state of the art a the other, you just have to decide how much you want to spend...

There's an interesting technical discussion here: http://www.pykett.org.uk/voicing_electronic_organs.htm

[manniemain]

If (and I do mean "if") the pipe organ has a certain something that the commercially produced digi has not, is the pipe organ "mark up" a good use for the finite resources of the church? In other words do we really need it to be this good? (This is not a rhetorical question. - I really can't make up my mind and I do like the pipe organ I play on a Sunday!)

[Benevenio]

This BBC news item is an interesting discussion of digitising the sounds old instruments - in this case harpsichord - and gives some reasons as to why you might want to go digital. It also makes it clear that the one and the other should not be seen as interchageable but each has merits that ought not be dismissed out of hand.

Not being an organist, this discussion is quite fascinating. The only parallel for me, a clarinet player, would be to be asked to play a digital clarinet and have the sound produced through a sound box. I suspect that I should come down on the side of the analogue instrument every time, playing the digital only for amusement. Perhaps the fact that the digital instrument is known as a 'wind contoller' might have something to do with my reluctance too…

[Southern Comfort]

There's an interesting technical discussion here: http://www.pykett.org.uk/voicing_electronic_organs.htm

Yes, very interesting indeed.

Reading it, it becomes even clearer that, say, 9 notes played simultaneously on 21 stops through 3 loudspeakers cannot reproduce the same effect as 189 pipes. Individual notes on one or two stops - perhaps yes. Massive chords on full organ - certainly not. Not without 189 channels and 189 speakers.

The author fails to explain why the Allen technology produces sounds which are so much blander and more artificial than the Bradford Computing system, when at least in theory (according to him) it should be better.

[mcb]

it becomes even clearer that, say, 9 notes played simultaneously on 21 stops through 3 loudspeakers cannot reproduce the same effect as 189 pipes.

OK, but it would be misleading to conclude from that line of argument that a good digital organ is somehow inferior for liturgical use. Picking over the fine detail of pipes vs digital is fine for conservatoires and concert halls, but the unequivocal message for churches ought to be that the sound of a good digital organ is wholly worthy of the liturgy. There's a risk that being over-precious about the merits of the pipe organ will make someone who's never heard one think that a good digital organ doesn't sound (for most people, in most contexts, for most purposes) like a really good organ, that happens to cost a tenth of the price of a real one and never goes out of tune.

In the end it's a moral argument, not an aesthetic one. I'll stick my neck out and say I can't think of any circumstances in which a church should buy a new pipe organ rather than a digital one.

M.