This is a very interesting question and one that I have been working on ever since Jon Dahlquist helped me set up my DQ-10s in 1974.
He showed me that proper speaker placement, and room interaction can produce the psycho-acoustic phenomena of "soundstage" and "images" at specific spots within that stage.
Ever since then, I have pursued the "holy grail" of gettting a more realistic image and soundstage. Over time I did find that the room interaction part was actually a detraction, but I will cover that, read on.
Much of what Frank says is true, and stereo will never give you "perfect" realism, but I have acheived an incredible degree and find it the best I have ever heard.
So what produces the best image and how can it be done?
First you need to understand that "NO" speaker/component system will image well in every room. There are no speakers that will "turn off" the room interferances that destroy the images that can be produced.
Acheiving a soundstage and imaging is easy. All you need do is place your speakers away from room boundaries, and sit equidistant between them in a triangle that would be considered "nearfield.
All that is needed is that the "signal" you are receiving is absolutley "EQUAL" from both speakers and both speakers are equal in amplitude.
That is it. You will now hear things that come from the left, on the left, and things that come from the right, on the right and things that come from the center, exactly in the center.
But, it can be better.
What makes and places an image?
The image is placed in the soundstage in to ways.
1) is simply by dividing the signal between the speakers in such a way as to "pull" the perception to the left or right to the degree nessessary to acheive placement
2) is by perceiving specific phase information to acheive placement goals
Better imaging is also the product of reproducing the subtle matrix, of delicate ambience from an original live recording.
Many things can compete with and totally overpower this information and without it, your imaging will "NOT" be anywhere as good as it could be.
During the live recording process, the microphone will "pick up" not only the sound of the performer, but also the sound of the perfromer, in realtion to the venue or performance space.
That is, the sound that "actually comes" from the performers original signal interacting with the room and space they are in. Needless to say, this sonic information is "delicate" and subtle. It is the perfromers reflected sound energy from the rooms boundaries, reacting on the mike. This sonic is subtle and dleicate, yet is the essense, of what you are trying to produce, if you want the maximum in imaging.
Only the most highly resloving systems and least interactive rooms can reproduce it.
To achieve the highest level of imaging, you need to consider the following:
1) Sources, and equipment that offer the highest degree of resolution and detail
2) Speakers that are "limited in dispersion" so that they provide maximum signal clarity with minimum room interaction
3) Reduction and elimination of "all" mid and high frequency reflections and dispersion.
What speakers and speaker design will give you the best (meaning most accurate) imaging?
To answer this, first you have to look at what might "damage" the subtle matrix I mentioned earlier.
Now please realize I am not "slamming" some brand or type of speaker, and also realize that there are many "very pleasing" sonic presentations that will give the impression of presense, air, fullness, and other things, that are not representitive of what the original sounded like, but never the less are beautiful sounding and prefered by some.
So that said, on the speaker side, any sonic information put out by the speaker that does not arrive at the listener "directly" from the speaker "will" have the potential if not dealt with, to damage the signal.
This includes dipolar information (above certain frequencies) which reflects off any room surface.
Any information reflected off the speaker baffle is damaging.
It includes any "large" speaker that does not operate as a point source in relation to the listener in the upper frequency ranges.
Also any speakers that don't operate with minimum phase will not be able to accuratley re-create the "ambience matrix" well since they will smear it.
What room sonics will detroy imaging?
In the mids and highs, "ALL OF THEM"!!!!
Tips for imaging:
1) Set up speakers away from all room boundaries. When possible make the largest triangle that can be made with the "rule of thirds".
That is divide both the length and width of the room by three and place the speakers in a triangle within the center third of each dimension
2) Sit as nearfield as possible
3) Treat your room for maximum reduction of "all" reflected sound
4) Assemble gear that has the resolving power to recreate the ambient sonic matrix
5) If room treatment is not possible, try to block/absorb baffle reflection/refraction and driver dispersion at the speaker, by the "hooding" your speaker tweak.
And a few "after thoughts".
A large room will generally "image better than a small room, when sitting nearfield, but never the less, will still be much better when treated.
These comments are based on a 2 channel stereo system, and not multichannel music.
My comments are not to promote or critique a specific brand, design, or line of speakers, treatments, or components, but more to relay my experiences in acheiving the best images with the tools and environment we have.
It should also be known that there are other ways to create, sonic "mirage" images which may even be preferable to some.
And finally the question that generally makes imaging a difficult little guy to pin down.
Why try at all, since the recording engineer makes the images for us?
And this is a good question.
I might add that even the perceptions at a "live" perfromance will vary. That is, if you are front/center at an "unamplified jazz quartet", you will hear more specific imaging than the person(s) a few rows back.
When this is recorded with a multimiked set up, and reproduced, as I suggest, the imaging you will get is the front/center image, not the rows farther back.
In fact, in my current set up, I can not only "hear" the rooms of each live performance, but in studio recordings, I can hear the damped booths and even the fake ambience sometimes thrown in by the engineer.
I like it.
