What can one say.
Bigus, I take my hat off to you a you called him on the farce quite correctly!
Let's see, what is best practice for performing an ETC responses?
How many sources are driven, and why do we limit them?
Why, because we intentionally limit the interference to a primary and its derivative virtual sources. If multiple sources are present we can certainly isolate the individual direct and indirect signals, but you are left with a practical mess where instead of 3X=6 and solving for X, we are stuck with 3X + 4Y + 2Z = 15 - Solve for X.
Is it difficult to construct a mess? Absolutely not! What is more difficult is to construct an experiment that more clearly illustrates the isolated cases in a clear, more easily understandable manner..
And where there is a source in the near field where the time differential between arriving signals is great relative to the distance from the source (such as 3 way speaker in the nearfield) they are treated as individual sources with regions of overlapping passbands that behave differently than a discreet source reproducing the same passband. Only when the distance from the source is great relative to the interdriver spacing (read asymptotic) do we treat them as a single source.
Any surprises here? Hell no. But to amirm who posits an experiment that violates this premise, he thinks this is instructive.
And what does he discover? Amazingly EXACTLY WHAT WE PREDICTED. That closely spaced sources result in significant polar lobing. And what , pray tell is polar lobing? Regions of spatial distribution where active noise cancellation occurs!!! Regions of full gain output interspersed with regions of NO effective output!
And how is this manifest? As you move about the space you move into and out of the regions of polar lobing. And this polar lobing is not constant. Meaning that as the frequency increases, the number of poles and nulls increases and the spacing between them decreases! This is the actual physical behavior that exists, whereas comb filtering is simply the picket fence pattern that appears on a frequency response but which does not actually exist - that indicates that polar lobing occurs.
And notice in the paper that the focus is on how "similar looking comb filters" can sound differently. Hmmm??? So apparently the focus ins on the FREQUENCY DOMAIN comb filtering and NOT on the spatial polar lobing which is the REAL PHYSICAL MANIFESTATION, whereas the 'comb filtering is simply a frequency response ARTIFACT.
Any wonder we start looking at a means to evaluate the actual signal offset in terms of TIME instead of frequency? Any surprises that the frequency domain simply indicates derivative symptoms but does not indicate causes?
But our learned friend inadvertently provides yet one more classic case of why the time domain is considered causal, while the frequency domain is simply derivative. And he still doesn't get it! After all, he still maintains that the frequency domain provides all the information necessary to describe what you hear - despite its failure to do so. Despite the ETC clearly showing two closely spaced sources reproducing the same passband that by definition result in spatial polar lobing and frequency response patterns called comb filtering.
So, if we know that polar lobing is in evidence, what happens as you move from spot to spot across the field where the spatially distributed lobing is in evidence?
We'll allow a few moments here for our esteemed friend to catch up.
You move from regions of full gain reproduction to regions where that frequency is GONE. Nada. A NULL! Can you hear this when the spatial distribution of the nulls is so great that you can physically move pone's head in and out of them? Ya think??????
Ironically this same concept is also what makes diffusion so wonderful!
But here we have a well mixed field of interference with a high number of virtual sources - the higher the better - such that the comb filtering that results is VERY narrow and closely spaced. It is less spatially distributed than our ears and the density of the lobes becomes such that you cannot detect it as a 'difference'. It become so dense that the effect is one of an average - a very smoothed average.
And then he shows us an example where the reflection occurs within a very small time differential - amazingly enough within the Haas intervals where the two sources are effectively fused by the ear-brain's inability to discern discreet signals as independent discreet events! Where we do NOT hear then as two separate events but instead we hear them as one smeared even, greater in gain, but suffering in 'articulation'!!!!
Are we surprised?
Not if you are familiar with the basic tenets of the Henry Precedence Effect and the Haas corollary, or even more fundamentally, SUPERPOSITION!!!!
Which means that out learned friend should be shocked!
And there's a term about which he should spend some time doing word searches!
Thus far, except for the use of MULTIPLE sources reproducing the same passband there are NO surprises and the effects were called without being provided ANY of the pertinent circumstantial data!
What is hilarious is that what the test REALLY proves is that multiple spaced sources reproducing the same passband is detrimental, as it results in polar lobing an comb filtering!DUH!
Let's see, we earlier posted this diagram explaining exactly such behavior - the exact same behavior that was accurately predicted.
Below you see the actual spatial polar lobing distribution based on the degree of separation relative to any given frequency with respect to its wavelength.
Above you see how this spatial polar lobing distribution is manifest in the frequency response as comb filtering.
And he thinks you need 2 mics to discern this??????? I don't think 2 brains would help in this case.
Folks, the problem is not how many mics that you employ, it is in not recognizing the fundamental issue at hand. And anyone who requires multiple mics to discern this is #^$%& clueless, as this should be just as readily apparent BEFORE measurements were taken as it was to me from the onset.
And this entire phenomena can be reproduced and illustrated MUCH more dramatically by simply stacking to 3" Auratone style speakers and applying identical tones to both as the speakers are slowly physically offset (the top one slowly slid back out of vertical alignment, as you listen to the polar lobing that results in the near field where it literally reproduces what seems like a civil defense or air raid warning generated by a rotating directional horn - which can be simultaneously displayed in terms of time based speaker offset and a concurrent frequency response with comb filtering that changes as the offset changes!
All in all, can we say acoustics 101 - what is SUPERPOSITION and why do we care?
Or, in the case of at least ONE - What IS acoustics? And why should they bother to learn it?
And in the meantime, return the certificate of participation for that extensive "several day" long class.
.Edited by dragonfyr - 7/5/12 at 6:23pm