3

Recently, I read about some major musical concert held somewhere in the USA to a record-large audience.

Since there were so many people, not all could see or even hear the music from the main loudspeakers.

Therefore, they put additional loudspeakers far away from the scene, connected with cables from the main equipment at the stage (presumably).

Since sound travels slower in the air than electrical signals inside cables, the music coming out from the far-away loudspeakers would have been out of sync with the actual sound waves, unless they did this "clever trick".

The clever trick apparently consisted of them spending a huge amount of money ("the cost of a new car per loudspeaker") on then-cutting edge electronics which delayed the signal exactly so that it would match the sound waves, and thus not cause disharmony.

While this made me think of how clever that sounds, I don't understand why exactly this would be so costly, even if it was in the early 1970s. In fact, I don't understand why it had to be so technical at all. It seems like this could've been accomplished in some analogue manner, very cheaply. The cost of a new car, many times over? Really?

What made it cost so much money? And had nobody ever held a big enough concert before this relatively late date, which would warrant a similar solution?

Did they have concerts for many years which sounded bad for the people far away, because the real audio waves were mixed (unsynced) with the local loudspeakers?

And if they barely heard the music so far anyway, did it really matter at all? Or was this more of a way to sell tickets by claiming a "perfect hi-fi-quality experience for every single participant"?

(Sorry I can't remember the name of the concert; I thought I had it bookmarked.)

Silven B.
  • 31
  • 1
  • 1
    Engineers capable of designing that work for free? Or do you think they should be paid? – Solar Mike Nov 03 '20 at 06:42
  • 2
    State of the art means they had to create it, build it, test it, ship it, set it up, operate it and make a profit. All made from a bunch of discrete transistors. A computer was an air conditioned room. And a new car was a couple of thousand dollars. So it makes perfect sense. – StainlessSteelRat Nov 03 '20 at 11:20
  • The Wikipedia article on delay lines lists a bunch of patents on analogue delay technologies, but all of those patents would have expired by 1970, so there is a puzzle as to why cheap generic copies weren't becoming available. On the other hand, the problem might be mutually-conflicting specifications for getting just the right delay time, impedance-matching with the speaker, safety of the audience (trip hazards due to long cables, possible presence of mercury, ...), and quantity of materials used? – Daniel Hatton Nov 03 '20 at 12:26
  • On the other, other hand, there were geostationary telecommunications satellites in the 1970s, so it might have been cheaper to work out the round trip time for a signal sent back and forth to such a satellite, position the speakers to make sure the required delay would be an integer number of times that round trip time, then send the signal back and forth the required number of times. – Daniel Hatton Nov 03 '20 at 12:38
  • @DanielHatton maybe nobody wanted to buy them so they weren't being made? – user253751 Nov 03 '20 at 18:09
  • An analog delay line that doesn't cause distortion and is milliseconds long is actually quite difficult. – Drew Nov 03 '20 at 23:39
  • "*It seems like this could've been accomplished in some analogue manner, very cheaply.*" How hard did you think about these words before you said them? Why does it seem this way? Because it sounds pretty tricky to me. If I asked you to slow down the wave in a pool, or a sound moving through the air, how would you, yourself, go about it? I'll even be generous and let you use an analogue method. In fact, I'll let you use the most expensive analog method you can think of. I'm guessing you have no idea how you would do it. – DKNguyen Nov 04 '20 at 00:25
  • @DKNguyen: If one built a device similar to a tape delay, but using a magnetic drum spinning at 1800RPM using a synchronous motor, every ~11 degrees of displacement would be about 1ms. My guess is that what would be expensive would not have been any single delay, but rather a means of producing *frequency-copensated* delay that was phase-accurate. – supercat May 28 '22 at 17:19

1 Answers1

4

The band was the Grateful Dead and the invention necessary was an analog delay line that was inserted between the main audio mix signal line and the power amps that drove the speaker arrays (which were co-located with the speakers themselves). Delays on order of a millisecond were needed and the amount of delay had to be adjustable to within a tenth of a millisecond and locked to that setting with less than one percent of drift on timescales of a quarter of a second.

In the early 70's this would require all-custom design as no off-the-shelf components were available, and digital means of processing audio signals did not exist. I do not know what the delay systems cost.

niels nielsen
  • 13,033
  • 1
  • 11
  • 30
  • The problem this was solving must have been something other than matching the delay in direct arrival of sound from the stage, though, right? In a large stadium, the latter would have been of the order of hundreds of milliseconds. – Daniel Hatton Nov 03 '20 at 20:29
  • 1
    It's odd that they didn't use a tape delay. \$ delay = \frac {distance\ between\ head}{tape\ speed} \$. At 15"/s (studio speed) = 375 mm/s you get 3 ms/mm of gap between the record head and the replay head. – Transistor Nov 03 '20 at 22:07
  • @DanielHatton, it was to exactly match the stage sound with the sound out in the crowd. They were playing for crowds so big that the remote speakers were hundreds of feet away from the stage speakers. – niels nielsen Nov 03 '20 at 22:18
  • 2
    That was my point. 'Hundreds of feet away from the stage', the sound coming directly from the stage would be delayed by something of the order of $100\,\mathsf{ms}$, so an electronic delay line that only produced a delay 'on order of a tenth to a half a millisecond' wouldn't be partcularly helpful in matching it. – Daniel Hatton Nov 03 '20 at 22:25
  • [Eventide DDL 1745 Digital Delay](https://www.eventideaudio.com/products/legacy/ddl-1745-digital-delay) – StainlessSteelRat Nov 04 '20 at 19:16
  • @nielsnielsen Did these attempt to compensate for dispersion as well as average time lag? the speed of sound can vary by about 20% over the acoustic frequency range on a humid evening if I was reading the (very small, weirdly formatted) graph correctly. If not, you would be cancelling some frequencies and reinforcing others. That seems to gybe with holding sub-microsecond accuracies (for phase matching) over quarter-second timeframes. – Phil Sweet Nov 04 '20 at 23:10
  • @PhilSweet, I have no idea. – niels nielsen Nov 05 '20 at 03:44
  • As Daniel Hatton said, the numbers don't make sense. What is the source for "Delays on order of a tenth to a half a millisecond were needed"? Do you mean 0.1 to 0.5 *second*? That would be 34-170 meters, which does make sense for a very large audience. Likewise, did the delay have to be adjustable to within a hundredth of a *second* (not millisecond)? It seems pointless to care about a hundredth of a millisecond, during which sound propagates only 3 mm. The answer should clarify the actual amount of delay and also whether the high cost was because of it being "too large" or "too small". – nanoman Jun 04 '21 at 21:39
  • @nanoman, "adjustable to within a hundredth of a millisecond" makes absolute sense. Consider the fact that the period of 15 kHz is about 0.07 ms!!! If you would really like to be able to synchronise sound up this frequency you should be able to delay in fractions (quite some of them actually) of this duration! Please keep in mind that this is rather impractical even in today's sound system design and optimisation but from a theoretical/ideal point-of-view makes perfect sense. As you point out in this duration sound propagates about 3mm which is about 45 degrees phaser shift @ 15 kHz!!! – ZaellixA Nov 20 '22 at 17:10
  • @ZaellixA I'm not convinced. Any attempt to synchronize the exact *phase* of sound waves from two sources (primary and secondary speakers) would be effective only for a listener in a specific, very small region. This is 3D (not 1D where everything is collinear). Once you consider a slightly different listener location where the relative distance to the two sources changes by a significant fraction of a wavelength (say ~10 cm), the phase coherence is lost. ... – nanoman Nov 22 '22 at 05:05
  • @ZaellixA ... The synchronization that matters to the listener is at a longer time scale, that of the "envelope" of the musical tones, where time shifts of less than a few hundredths of a second are no problem (this is why an orchestra can be spread over a ~10 m region without muddying the sound, and why recorded singing is sometimes superimposed with an echo ~0.01 s later for a rich ["double tracking"](https://en.wikipedia.org/wiki/Automatic_double_tracking) effect). ... – nanoman Nov 22 '22 at 05:06
  • @ZaellixA ... Surely you agree that niels nielsen's reference to "delays on order of a tenth to a half a millisecond" is low by 1000x (it would make sense only for speakers separated by a mere ~10 cm), as noted in Daniel Hatton's upvoted comment. Likewise, I have laid out why "adjustable to within a hundredth of a millisecond" is also low by 1000x for the precision of synchronization that matters for music. The obvious explanation is that niels nielsen simply mistyped millisecond for second in both cases, but unfortunately has not engaged to clarify in these 2 years! – nanoman Nov 22 '22 at 05:08
  • @ nanoman, my bad, you are right, will edit. – niels nielsen Nov 22 '22 at 06:48
  • @nanoman, thanks for the comment. I strongly believe that you are right in the fact that having a delay that has accuracy of hundredth of a ms is an "overkill" for speakers separated by hundreds of feet (or tens of metres). What I suggest is that if you were to align correctly to a very broad band you would definitely require it. If you miss the delay by some 0.5 ms then you will very well align frequencies up to about 1 kHz but from this point onward (frequency-wise) you'll create comb-filtering. Whether this is of practical use or not is another story though. Please correct me if I am wrong. – ZaellixA Nov 22 '22 at 14:09
  • @nielsnielsen Thanks for the edit, but you increased the timescales by only ~1 order of magnitude, not 3, so they still seem way too low. "Delays on order of a millisecond" would compensate for a distance of only ~30 cm (1 foot). I thought you should say "Delays on order of a tenth to a half a *second*...adjustable to within a hundredth of a *second*". You yourself commented that "the remote speakers were *hundreds of feet* away from the stage speakers". – nanoman Nov 22 '22 at 17:01
  • @ZaellixA Such alignment is hopeless anyway in 3D space since there's not a single listening point. Music perception isn't that sensitive, which is why we enjoy it just fine when muddled at the level of several meters or ~0.01 s between sources (including echoes). Music isn't a few pure sine waves where we'd notice those interference/filtering effects. Each musical tone is spread over frequencies due to vibrato and harmonics. Also the sound amplitude from two different sources is unlikely to be equal at the listener, so even if a frequency interferes destructively it isn't canceled completely. – nanoman Nov 22 '22 at 17:23
  • @nanoman, I completely agree with you and this is why I clearly stated that having extremely fine control of delay time would make perfect sense from a **theoretical** point of view. As for the amplitude of the two sources, this is a rather nice addition to the topic and the reason why most delay stacks/towers used nowadays are most often comprised of either less or smaller loudspeakers. Just to "conclude", we are definitely having the same opinion, it's just that I was referring to the conceptual theory behind the topic. Thanks for the constructive conversation, it's really appreciated :). – ZaellixA Nov 23 '22 at 12:33