Watchout, I'm gettin' wordy..
Tonedeaf makes some very good points. I especially like the 'just another tool in the kit' mentality to using the Stereo Zoom. He also makes good points about the difficulty in wrapping your head around how it works, the ways in which a sound source is localized, and also in raising the question 'what is accurate'?
I'd like to emphasize that all of this is all somewhat fuzzy, none of it is hard-fast, especially the accuracy part. Do we mean accurate sounding as in 'you are there'?, accurate in that you can point your finger in the direction of a sound and you would be pointing the same way you would have at the event?, accurate in the placement of the reverberant information on the recording, accurate in the feeling of space?, accurate timbre?, accurate dynamics?, accurate recording of the banal bar chatter?...
Recording and playback is always an illusion. We simply cannot get anywhere close to accurately reproducing the actual acoustic event with our simple equipment. I find it down right incredible that our brains are so easily fooled and that we can so easily suspend disbelief at will. Video and film is even more of a suspension of disbelief with rapid, flickering still images on a small rectangular screen. When you think about that, it becomes hard to believe it works at all, much less that it is able to suck us in so completely.
Some illusions are more convincing than others and as Datbrad mentions some illusions are even better than reality, which is especially rewarding for the recording maker. That, to me, is what elevates recording to art, beyond mere mastery of the technical craft. We can play around with trading one illusory part for another until we are best satisfied with the best resulting illusion, then make a personal mental list of what we consider most important for achiving it, like Richard has.
To clarify what I meant about accuracy, while the pure coincident XY is the most clinically accurate in terms of soundstage replication, it is not the most natural compared to near coincident (DIN, ORTF) where the time delay between the ears is also replicated.
I don't mean to badger the point, but I’ll touch on this again because it falls into both the
what is accuracy, really? category and the no hard-fast rules, but fuzzy world.
Excluding the possibility of ranking 'naturalness of sound' above directional pin-point reproduction in achieving the 'most accurate' recording in that personal artistic sense, certain coincident,
amplitude-only X/Y techniques can often be best at the ‘point-your-finger-at-the-sound-source’ aspect of reproduction accuracy which is the aspect which I think Datbrad is stressing. The Blumlein technique (90 degree crossed figure 8’s) is especially accurate in that sense for the front 90 degree quadrant directly in front of the mics. It is also one of the special cases where the angle formed between the two mics really is just about the same as the Stereo Recording Angle. Yet Blumlein certainly is not accurate in the same way for the other 270 degrees around the mics, even though it can create a particularly satisfying illusion. Angular distortion is covered in section 4.1 of the SZ.
In the example you give:
Recording with a pair of XY cards at 90 degrees setup around 8' from a group of musicians standing on a 20' line across the front of the stage, (piano on the right, drum kit on the left, horn player in the middle) would replicate exactly the same soundstage heard live from the same position.
If my trigonometry isn’t too rusty, the musicians in the example above form an angle of just over 100 degrees as seen from the microphone position. To “replicate exactly the same soundstage heard live from the same position” would require a playback image with the piano 50+ degrees over to the right and the farthest part of the drum kit 50+ degrees left. That clearly isn’t possible using two stereo speakers arranged a total of 60 degrees apart as seen from the listening position. If we forget
absolute angular accuracy and allow for an even angular compression of the recorded image to fit the whole 100 degrees into the available 60, then the relative positions of the piano, horn and kit will be pretty accurate in relation to each other. Yet there will still be a significant amount of angular distortion for any source that is located halfway right between the horn and drums or halfway left between the horn and piano. If you believe the Stereo Zoom charts, that angular distortion will push those mid-way sounds 6 degrees closer to the speakers than they should be. Would other configurations be more angularly accurate, yes.. and others such as spaced omnis would be worse. See page 14 for that exact example (X/Y cards at +/- 45 degrees and their 6 degrees of angular distortion).
Is that a deal breaker? Probably not. To quote the Stereo Zoom directly:
"It should be said however, that it is not obligatory
for the SRA to be equal to the angle occupied by
the sound source. Most sound recording engineers
in fact prefer the SRA to be slightly larger
than the sound source sector. This is equivalent
to leaving a little headroom in a picture or more
correctly in this case "sideroom" in the sound
image. The amount of sideroom is obviously a
matter of individual judgement, but is rarely
more than about 10° for a small group of musicians.
On the other hand, in the case of a much
larger orchestra it is quite often necessary to do
the opposite and place the limits of the SRA
within the orchestra (negative sideroom), the left
limit being within the first violins, the right limit
within the double basses. This allows more
space for the individual instruments (flute, clarinet,
oboe, etc…) in the middle of the orchestra.
But this is a question of individual preference,
and there are as many different choices as
there are sound recording engineers!"In my own personal list of trade-offs, the sense of depth and space is usually more important
for me than pin-point imaging. That’s one reason I now almost exclusively record in 4 channel surround for my own listening. I love hearing the room around me and the crowd behind. It just makes the illusion much more real for me. Yet, like I mentioned above, nothing is hard-fast. I still want to hear angular source information too and I’ve learned that I can still get very good spacial imaging with spaced techniques if I do it right. I’ve made spaced omni, stage lip recordings where tom rolls across the drum kit move convincingly across the playback stage and you can point out the high-hat on the right of the kit and the crash cymbal on the left.
I think where things get tough for some tapers is when we want the mics to do something beyond simply replicating the image at the placement location of the mics. This is where the art of recording with mics comes in. Using unconventional placement, patterns, etc. you can make a recording that sounds better than the soundstage did live at the same location in the room, and you really have to develop the ability to "see" what the mics will "hear" and know how to manipulate them to give you the results you are seeking. Experience is really the only teacher for this skill, though.....
Truth. Thanks for the great discussion.