Well, we do suffer from time misalignment during live concerts. I was a few times at the first rows in extreme left and the entire sound was practically unlistenable as the orchestra never sound synchronized.  If was bad but it would be even more horrible if it was recording.  In the recordings it would be different, depends of the recording techniques ere used, amount of tracks/microphones, their positioning and the way everything is mixed. Sure, in case of two microphones properly positioned (and I personally do not know what would constitute properly positioned as I have zero experience in it) the arrival would be more or less equalized and it is what “better” sound engineers do…..

Still, the biggest question that I see in all of it is: why time discrepancies in “live” sound are more forgiving then the time discrepancies during sound reproduction phase. I can go into discussion about monaural filtering, distance localization, Doppler effects, judging distances with reflective context and about  the effect of overriding of one sources by other sources… however I think that that the answer might be simpler and form a slightly different dimension.

Live Sound and Reproduced Sound are NOT the same entities.  Live Sound (as a sequence of pressure waves), no mater how “wrong” it is, is Sound that out hearing is accustomed to deal with and out brain knows how to interpret or even to filter-out the problems with Live Sound’s. The Reproduced Sound is a not Sound itself, not the primary Reality, but an interpretation of Live Sound via language of mathematical algorithms (transverse waves). Human hearing is absolutely disabled to operate in the environment of transverse waves and therefore we need mechanisms of conversion from longitude waves to transverse and then back to longitude. In the language of longitude waves distance is a Reality but in the language of transverse waves distance does not exists and described ONLY at a calculateable equation. However, any equation is just approximation to reality, successful ONLY in context of a given hierarchical coordinate system. So, only by the means of conversion of Live Sound into Sound Reproduction the already are “loosing” timing as a humanly immediately referenced ingredient. It so, then brain has less capacity to comprehend time discrepancies in Sound Reproduction then it does in Live Sound.  Add to it the harmonic distortions of Sound Reproduction that screw up distance localization and along with time delays misrepresent everything and it would be understandable that the time alignment in Live and Reproduced Sound are very different.

The best evidence that time alignment in Live and Reproduced Sound are very different is the fact that misalignment in Live Sound affects volume very minor. However, in Reproduced Sound any misalignment is not only highly affects volume but also it highly affects ability for listeners to DISCRIMINATE VOLUMES. I do not even mention imaging that got absolutely destroyed by time misalignment….

So, to conclude what I proposed I would like to point out again that in Live Sound we do not deal with alignment but rather with “management of events arrivals”. In Reproduced Sound we deal with equating of mathematical derivatives of multiple functions. What is difference between “live event” and reproduction of “derivatives of functions”? The difference is that the first is random but the second is hieratical. A human awareness is alien to hierarchal perception and therefore any, even minute, discrepancies in hierarchal description of Realty our mind recognizes as a “huge error”,  the errors that our awareness begin to interpret, compare, analyze, correct… in other words …waste our “CPU time” while we should be wholly dedicated to music listening.

op.9 wrote:

One of the dark arts of playing in (and conducting) an orchestra  is the question of judging when to play. This is a huge subject. For example, often the backs of sections are playing slightly behind anyway - If you play exactly on the speed of light signal from the conductor when you are at the back - you will most likely be early. Its very tricky. I prefer being at the front! Percussionist have it hardest - timing a 'ding' exactly where on the beat you want it (i.e. on or just after) when positioned right at the back is a miracle. Only the very best musicians get it right. Conductors who try to solve this with clever leads or flicks of the wrist tend to cause more problems than they solve. When you listen to a really great ensemble its wonderful how all those musicians find a way to really listen and to use the acoustic and delays in the hall to create the perfect sound for the listener or microphone. (Horn players playing a low note have to play a good quarter of a second or more in front to compensate for the length of their instrument and also because their sound is best reflected off the back wall of the hall) Interestingly, in the film music studios most instruments have their own microphone and you simply play with the click track. Comb filtering heaven! Then they do what they like with it in the control room.

However, and it is a big however!

op.9 wrote:

Yes, I agree they are different issues. However, quite a big however too... It all depends on where we believe 'reproduction' starts. We are very rarely simply reproducing the feed form one single microphone. Choices are made very early on in the recording chain that might as well be called 'performance' choices. Choice and positioning of players and mics, mixing, delay of spot mics, multi-band compression and multiple waves plugins - and that strange buisness of 'mastering' too. As I see it our 'hi-fi' reproduction is just the tail-end of those choices. Lets not kid ourselves that we are reproducing something absolute. Its all pretty chaotic by the time it get to us!

…I would differentiate, in fact would recognize as completely irrelevant, the arriving discrepancies from instilments to microphone and the arriving discrepancies form loudspeakers to listener.

Let make a mental experiment: We have a single mono microphone, 6 feet form it we have a double-bass player and in another 10 feet from the double-bass we have a flute player. We record them: let call it “Result A”. Then we put them at the same distance, record them and call it “Result B”. Then us to play  the “Result B” on a playback where (let simplify everything) a woofer is 6 feet form a listening position and the tweeter is 16 feet – Let call it “Result C”. The point that I would like to propose is that “Result A” and the “Result C” would be very different. BTW, I did not try it and it is purely my speculations…

In reproduced sound any fundamentals will have harmonics and overtones across entire frequency range. Therefore some of particles of a same note will reproduce by two very different drivers and reach a listener at different points in time due to the different axial positions of the drivers. Our hearing is not accustomed to us as it does not exist in nature. In live sound it should not take peace as ach instrument fundamentally and harmonically sit at its own place and we use it...

Rgs, the Cat

be wrote:

The paradox of the orchestra sound would be resolved if we asume that the audience (in the good seats) mostly hear the reflected sound, primarily from the wall behind the stage and that the players at the front follow the visual lead of the conductor and the players behind, play accordingly to the audible lead of the front players, the sound would then be time aligned, when reflected from the back. This would obviously only function if the orchesetra is not to widely spaced and the the listener is not to much to the side. That it is indeed the reflected sound and the reveberation that is dominating seems to be correct since it is imposible to pinpoint the instruments at least from a seat that is not to close to the stage.

 The reverberation radius usually is very small, I believe in Carnegie Hall what it is 3.2M reverberation radios is a distance at witch the ratio between direct sound is reflected sound is 50%:50%) Also, the live performances usually take place in the lager rooms where RT-60 is very long. (RT-60 is a test that indicted how long time a given frequency decays in a give space for 60dB).

So, what I would be very interested to learn is how RT-60 affects the timing discrepancies in playback stage. Let pretend that we have for instance an arrival miss-alignment between drivers measured in half-wavelength. We recognize some subjective auditable negative consequences of that miss-alignment listening it in a room that has RT-60 = 0.6 second at 60Hz (it means that 60Hz reach a point of minus 60dB at 0.6 second -  a typical number for a good size audio room). So, would our “subjective auditable negative consequences” be at identical amplitude if we listen the same playback at the same distance but in the room that would have RT-60 = 1.3 seconds at 60Hz? (A larger room)

In a larger room that reverberation radius is smaller, phases get randominized faster and therefore I would presume that our negative reaction to time discrepancies in a large room would be way less annoying to us…