Is that a stereo system vs a mono system?
Hi David, all
I was not there but believe it was 1 channel of each that the customer was comparing for an installation he was considering.
The issue is when you have two identical signals that are NOT arriving at an essentially identical time (compared to the wavelength), they are adding together constructively and destructively and produce comb filtering. This was the issue discussed here;
http://forums.prosoundweb.com/index.php/topic,155341.60.htmlWith loudspeakers it is a bit more complicated as they are located and radiate in 3 dimensions and so the comb filtering isn’t in 1 dimension like with an electrical signal but is a 3 dimensional phenomena that changes with position, such a radiation pattern is an interference pattern, it is combing in 3d.
If one were to take a high resolution measurement of array of two or more sources like two monitors or an array, even made of absolutely 100% perfect drivers, you would see the pattern of combing & lobes and nulls moves in the measurement as your measurement position and /or their spacing changes.
The marketing argument is you can’t hear this IF the pattern is complex enough (with a lot of separate sources but time distortion is usually increased as well) but that 3d combing or interference pattern is what you hear if the wind blows across a large system or you play pink noise, music or sine waves and move around such as in the video. Or, if done with one source, not present at all as discussed here;
http://forums.prosoundweb.com/index.php/topic,155289.0.htmlThe other way to look at it is Time. It is more difficult to discuss but is MUCH more related to what happens to music or voice intelligibility, much of which has transient information (where signals start and stop suddenly). Here an easy to imagine test signal might be a single pop or tick from an LP or record which is a very short naked transient.
If fed to all the speakers in an array, what arrives at your ears first is from the closest speaker, then the next closest and so on until the radiation from the last speaker arrives.
Here the original short impulse is transmogrified and stretched out in time because of the different path lengths from each source to your ears and instead of being a fraction of a millisecond long, what may arrive at your ears lasts for several or more milliseconds in time instead of a very short event. This is Time distortion and directly tied to clarity / voice intelligibility.
Added to the mix is unless floating off the ground or the ground covered with people, each source also has a nearly equal loudness floor bounce which is delayed by the differences in the length of each path length as well.
The marketing argument why “this can’t happen” is that each element only has say a 5 degree radiation angle but like other things, there is an element of fiction here which is revealed when you examine the vertical beam width the more reputable mfr’s give.
Alternately, one can use Don Keele’s pattern loss thumb rule to figure out approximately where that pattern expands / looses pattern control. That pattern loss Frequency (using inches) is approximately 1X10^6 / (height X angle) and each octave below that the pattern expands by a factor of 2. For example, with a 12 inch tall source and 5 degree nominal angle, the pattern control is lost below about 16.6Khz, at 8.33K it’s 10 degrees, 4.16K its 20 degrees and so on.
Being next to other horns also does not mean they alter the pattern of each other, they continue to radiate as independent sources to produce that combing / interference pattern as Huygens principal calls for (and heard in that video). The other argument is that with DSP one can align the arrivals to be in synchrony at one spot and this is possible to do BUT only for one spot as everywhere else has different path lengths and the same kind of time distortion.
In the really simple case of just two sources like the floor monitor question, if they are essentially identical and each preserves time well enough (and not distort the time component individually) then the two should produce a center or phantom image (the voice sounds like it’s between the two) when playing the same signal like some stereo speakers do.
A last comment is about video or other recording processes and hearing. Your hearing process takes two separate inputs from your ears and in the process of making one “image” seeks information and in the process rejects noise without your knowledge.
For that reason, it is usually easier to hear coloration and other loudspeaker issues with a measurement mic and headphones (which bypasses much of that automatic process) .
Find a speaker that still sounds ok after just 2 or 3 generations in a generation loss recording and you have a speaker reasonably faithful to the audio signal.
Loudspeakers are by far the weakest link in the chain.
Best
Tom Danley
