Hi Tom, I agree that we don't want the speakers doing things which draw attention to themselves as the sound source, and ime such things are often (if not always) easier to pick out in mono than in stereo. But beyond that I'm not totally clear about what you mean.
What are these "differences between the left and right ear" that you refer to? Is this in the first-arrival sound, or in the reflections (in which case room acoustics comes into play), or both?
Can you provide an example (general or specific) of a speaker which "has complex features that produce differences between the right and left ear that allow your brain to easily localize the depth "as the source""?
And, can you provide an example of the sort of desirable "simple radiation pattern" which enables the speakers to "not even be noticeable in the stereo image"?
Thanks!
Hi Duke
This issue opens a can of worms and there is more to it than i alluded to.
A friend and former co-worker Doug Jones was involved with research in the 80's that revealed more about how we hear and localize sounds.
The result of that work was a set of recordings which capture / mimic our localization. To be clear this Head Related Transform Function and our ears unique Pinna responses are a key part of localization, it is more than just loudness and precedence in time.
Our outer ears shape produce a frequency response which changes with direction and elevation AND there are two notches which move in frequency with those changes in source direction.
When Doug and friends place tiny microphones deep inside people's ears they saw this comb filtering and plotted it. We do not hear that "combing" as a flaws as it appears to measurement minded, we associate this stuff with source direction.
Here are some recordings Doug Made in the day which exploit this additional aural cue.
https://www.audiocheck.net/audiotests_ledr.php
Unlike an amplitude panned stereo image, these signals can extend well outside the normal speaker coverage as they trick your brain into associating that sound with a location in space. The better your speakers are at preserving stereo information, the more real this sounds.
Funny, now days you even hear TV commercials which exploit this effect, sounds that "sound like" they are floating above or outside your speakers.
While popular lore says we can't hear very fine combing (combing is what you get when you add a delayed signal to an un- delayed signal), THAT kind of combing can "mimic" or interfere with the location combing your ears need to identify locations in space and can create the kind of new radiation that makes it easy to hear or localize the physical depth (with eyes closed).
A few years ago Doug and I taught a class at Avixa / Infocomm, the challenge I was given was make two pairs of speakers that looked and sounded "the same" (even used the same drivers) but radiated and sounded very differently when in stereo. In the class demo, one pair made an ambiguous but appealing "wall of sound" out of Diana Krall and with the other, she was right there in front of you like you could walk up to her.
One pair had as many things as i could think of to make combing in this time / frequency range the other had as little as possible with the same drivers.
There are too many things to list here that can cause these image disrupters but the idea is that until the expanding wave is "large enough" to be self supporting (like how a baffle or horn can confine the angle and project it at more or less the same angle per Keele's pattern loss thumb rule), until that point, it can re-radiate at any point there is an abrupt change in the rate of expansion.
In a typical small speaker this would often be the transition around the tweeter, the edges, the near by drivers and often non-pistonic behavior where some parts are moving one way while others are going the other way etc.. In many multi-way speakers, you have to be far away enough to not localize separate sources but often enough the separate sources make the depth easier to hear as the source .
The up side is that if you want to hear speakers that do a better than average job disappearing into the stereo image, you can do this yourself.
Obtain a pair of these drivers (I know these are good drivers and work for this);
https://faitalpro.com/en/products/LF_Loudspeakers/product_details/index.php?id=401000100
Then cut out a mounting hole in the center of a baffle (use 1/2 inch plywood) at least 2 feet square.
Cover the front side of the baffle with 1/2 inch foam padding (or actual acoustic foam) stuck on with super77 or other casual adhesive.
Then mount the driver from the front side so that it sits in a slight recess of compressed foam around it.
This unusually good 3 inch full range driver is small enough to radiate like a simple piston from a single location in time and space all the way up to several Khz, above that it is a mixed mode (traveling wave something like a soft dome).
The driver mounted this way and eq'd flat up high is more free than most of this "spatial identity" as I think of it and can produce a VERY compelling stereo image without standing out in it....up to the SPL runs out of gas or LF extension. Up close these are very good and at times one is hardly aware of loudspeakers.
You can enclose the back and calculate an "alignment" but if you like them and want to add a subwoofer / bass speaker, use a sealed box for these. Also, there is a 3 element RLC network that takes care of a rising hf although i don't see that in this newer computer. If i run across it I will post the values
At work, these same radiation related aspects of radiation on SQ apply even more so, I believe and I think some of our large systems show it is mostly these radiation / interference issues and not as much driver limitations that limits the effective range and fidelity of large arrays.
Best
Tom