I don't know if soundstage is a distortion, but I have noticed that usually when a soundstage gets wider, it also gets shallower or more forward.
68 responses Add your response
Nope. Soundstage is nothing more than the height, width and depth of the music sound field emanating from the speakers. To get the maximum soundstage is determined by a well executed amplifier design as well as a damn good crossover in the speakers. I don't buy your point that analog results in a better soundtage than digital. That would only be the case if your using a lousy, poorly designed CD player. If all your separates are top notch designs and synchronize perfectly together, then the soundstage should not be an issue.
Soundstage is the result of two things- good high frequency response and good phase reproduction (low phase shift). If both are correctly reproduced side information that appears to be beyond the spread of the speakers is normal.
The Soundstage concept is why stereo was created. The original principles were created by Alan Blumlien decades ago.
Roger Water's Amused to Death has extra phase encoding in it that takes advantage of this fact.
Note that when you listen to your system 'out of phaase' that while the sound may extend beyond the sides of your speakers that it also looses center image specificity. This is not a 'distortion' as such coming from your system or the recording. Some folks refer to the sound as 'phasy' and in the old days some audio stores (when you had them) intentionally set up some speakers 'out of phase' knowing that it creates the impression of big soundstage which could impress a beginner who would buy, take the speakers home and set them up 'in phase' and then wonder what happened to the big sound he heard at the store.
When your system is 'in phase' you will maximize its ability to produce a holographic soundstage, only limited by its design and set-up.
Unless your recording has some out of phase sounds in it, and many do (often intentionally put there I think for its effect - you get good specificity and the out of phase information makes it appear huge), you should have no sounds appearing outside the space between your speakers (on the plane of the speakers).
FWIW, multichannel sound effects -artificial though they might be, can be had on the cheap by putting 2 small speakers on the rear side walls of your listening room powered by a seperate amp and attenuator out of phase with the main system. I believe they called the a Hafler system. Not all that good really, but a very interesting thing which can be addictive if you are not really all that fussy. The whole room seems more energizied and the sound stage collapses when you turn off the small rear speakers. There have been many, and some very expensive, audio components developed expanding on that simple Hafler thingie.
Hope that helps you understand this subject a bit more. BTW,
I don't think my componenets are "poorly designed", as I use a rather nice Sim Audio CD player, and have an LP12/Ekos/Arkiv B turntable set up. I don't know whether the point of stereo is to spread the sound outside of between the speakers. I thought it was to create a sound stage as wide as the distance between the speakers, which of course was nothing when dealing with a single speaker in mono reproduction.
None the less, rarely do I get much of anything beyond the speakers unless the speakers are pretty much pointed straight ahead, but then I loose a solid center image.
Notably, the widest soundstage I can get at home sometimes seems to be through a pair of AKG 701's but I know some view their portrayal of soundstage as exeagerated.
Soundstage is determined by recording technique. For purist recordings the microphone setup is the determinate. For studio recordings soundstage info is created via electronic processors.
On the reproduction side it is not clear to me that audiophiles are interested in an accurate soundstage reproduction. Wider or deeper seems to be desired whether or not it is warranted. Dipoles, wide dispersion or rear firing tweeters and omnis all overlay additional soundstage info to what is already in the recording.
If your system can provide a reasonable facsimile of the real soundstage info present in a purist recording, it means your system a lot of other things really well.
Grab any Q sound encoded source material and you'll Very Clearly see (hear?) that soundstage is not only not a "distortion", but its also not purely a function of your system. If your program material is recorded in a manner that allows dramatic staging, a good system will pass that through in the listening.
BTW. Q sound isn't necessarily great in any other way, but the stage is dramatic. IIRC, Madonna's Immaculate Collection (Hits) is a Q sound disc
My London Decca cartridge on OTOH does staging I would describe as 'other worldly' in comparison to any other cartridge or digital I've ever heard. I used to think it was artificial but I can't see a cartridge having the ability to rearrange the sound stage. The only conclusion I can come to is that analog does in fact provide a much more defined and accurate as engineered sound stage. This has been my overall experience.
I agree with Csontos - analog recording recreates soundstaging quite a bit better than digital recording does. This is one of the main reasons why so many professional musicians still prefer it, and wish that it was still done. As much as digital recording has improved, it is still in the end a mix of way too many different mikes, and the end result, even though it can be altered and controlled and edited so much more easily, is not as natural sounding.
For me 'better' is simply more defined. The Decca cartridges have the ability to produce very sharply defined point source imaging. Should this be considered artificial because it does not follow general design convention? 30 years ago I bought an Audio Quest cartridge for $75.00 that had no mags or coils but only a tiny circuit board to which the cantilever was attached. This thing was rejected by the audio community as an inferior design. Funny thing is it performed similar to the Decca and one of the best I've ever heard. Ime, cartridges in general produce a relatively vague sound stage. The very reason why their are so many different camps trying to capture a semblance of reality in their systems. Before digital hit the scene, there was not the sharp division between the tube guys and the rest. Interesting.
"On the reproduction side it is not clear to me that audiophiles are interested in an accurate soundstage reproduction. Wider or deeper seems to be desired whether or not it is warranted. Dipoles, wide dispersion or rear firing tweeters and omnis all overlay additional soundstage info to what is already in the recording."
I never thought of it quite that way, but it's true - I absolutely prefer a wide soundstage as I find it easier to track individual players and singers that way than with a more condensed image. So in this instance I have to concede preference trumps original artistic intent.
I suspect the reason that the Decca does so well in Csontos' system has something to do with the effective arm mass, the speed stability of the table, and the short cantilever used on the Decca.
If the 'table has a slight (inaudible) speed variation, the soundstage will be less distinct as the arm oscillates over the cartridge cantilever due to changing skating forces. A shorter or stiffer cantilever will be alleviate this.
Newbee, IMO/IME tubes and analog have the most detail, hence the best soundstage definition.
Zavato, there are actually 2 "soundstages" involved, 1) the soundstage of the recording and 2) the other one that is independently projected by your system. If the system's soundstage is less than optimal, then the presentation of the recording soundstage is in some way limited or curtailed accordingly. Typically and, in particular, it's curtailed in terms of things like its overall size shape and dimensionality, depth-vs-width ratio, its ability to disappear from the speakers, forward envelopement, coherence, air, distances between performers and so on. All these things can be impacted by speaker positioning, xover quality, the relative quality of component power supplies, amp/preamp design, wiring and much more. Ideally, the point of a hifi is to "allow" (and I'm stressing that concept as much as possible) the soundstage that the recording engineers labored to create to come through to the listening room (via your system) as unaltered or as unimpeded as is practicable. In other words, hopefully and as much as possible, the recorded soundstage should 'fit cleanly within' the system's soundstage without the system acting to impose its own soundstage limitations onto that of the recording. For most of us, all that can be easier said than done, at least to start with...and I haven't even yet mentioned the complication presented by the acoustics of the room itself. I can tell you from direct experience though, that once all that's done, the recorded soundstage is very typically much larger than the zone between the speakers. In my setup there is very often as much information from about 4-5 feet outside edges of the speaker cabinets as there is between them. This is not "added" by sound engineers. Many folks do not realize how much their own system may be actively convoluting, or distorting, the soundtage of their recordings and, as a result, may from time to time infrequently experience 'random' or 'unexpected' musical sounds from outside the edges of their speakers, but without it appearing to be commonplace or otherwise natural sounding and believably a part of the presentation. The typical conclusion is that some recordings are in some way 'manipulated' to achieve this effect, but it plainly isn't so. Nor is it necessary. Good systems can allow either the entire, or the majority of, the recording's soundstage to come through well enough for everything to be properly coherent. I have a roughly $5k CD-only rig, but have coupled it to almost $6k of power conditioning. I'll be the first to tell you that's just a wee bit unorthodox, but in my case it has worked wonders without drawback, but on that, of course, YMMV. My components, while not state-of-art highend, are, like yours, not "poorly designed" either. But, the kind of conditioning I found not only kills electrical noise (the real enemy), but also makes bad power supplies sound acceptable, and acceptable ones sound good, good ones sound excellent and excellent ones sound off the hook. In the end, soundstage may be, at least somewhat of a degree of distortion - one that may be necessary to complete the audio illusion (at the very least, no 2 systems have the same inherent soundstage fingerprint), but, to that degree, I welcome it. But, I'm saying really the biggest challenge is undertaking the pursuit of not 'cramping its style' as much as possible. Regards. John
In my system, with later generation pseudo-omni OHM Walsh speakers running, digital/analog source does not matter for soundstage. Its purely a function of the recording.
Most decent to good or better quality recordings have wall to wall soundstage, often many feet out beyond speakers.
The sound emitted from the speakers fills the area immediately in front of and well behind the speakers. Actual location of speakers cannot be determined with eyes closed. Its more like the performers are on stage in front of you. Mono recordings even take on a presentation much like if the actual performers were sitting front and square in front of you but with a naturalness to the acoustics that makes it sound more like a live performance than a recording.
The OHMs are exceptional in this area, when set up well. WIth right amp, dynamics are top notch as well to boot, no fuzziness or other effects often associated with large soundstage due to phase artifacts occur.
Other good omnidirectional speakers I have heard, like mbl, also tend to do exceptionally well in the soundstage area, however the presentation is so different from what might be used to otherwise with more conventional dynamic speaker designs that some will take to it and never look back while others may never take the plunge.
I'm not sure you can call the Decca a cantilevered cartridge. They do make short cantilevered conventionally designed cartridges but I was referring to the original Decca Blue, Red, or Gold type. In the original versions, a line contact stylus is mounted in the end of a verticle post which is quite long and then bent to horizontally engage the coils. Said post is then damped/secured by a 'rope' which is tied around it near the stylus and attached under tension at the rear. There is no suspension. I think it's the positive downward impact of the stylus and subsequent resonance of the rope that yields this result.
We already have hybrid components but referring to analog/digital, I wonder what cartridge manufacturers would have come up with by now had digital remained a phenomenon or an anomaly.
Ralph, you made a very good point about tone-arm oscillation. All my records have a pencil line on the label of each side which serves as an alignment tool to precisely center the record before play. I've been doing this since my first system. The salesman who sold it to me threw in a Sure test record with a pencil line drawn out perpendicular to the spindle. It took a little while but it did dawn on me. Otherwise, only half the record is being accurately tracked. I know you were referring to something else but I think this is much more relevant.
Hum... I agree that a well-executed amplifier and crossover is beneficial to the sound-stage, but I believe those benefits are more apparent in the actual staging of the instruments and players within the sound-stage.
The room itself, and the way in which the speakers are placed has more to do with the overall dimensions of that sound-stage. Speaker placement is the key to achieving a wide sound-stage. Then comes room treatments. I have heard some mid-fi gear that produces exceptionally wide and deep sound-stages, but at the same time poor definition within that sound-stage; and it is that definition ('realism") that we all strive for... Which entails much more than just dimensions...
Engineers are very well aware of what can be expected at the outcome. Hence their title. They are compensating for every limitation they are aware of in the play back system. Their goal is the same as ours, ie: those who care about accuracy. They do after all listen to playback before production. So the onus is on the gear to replicate what was intended, is it not? Variations in the performance of the gear is what this pursuit is all about. If this is not the case, then what is the purpose of the sound engineer?
Hi guys - my post was specifically about the recording side, not the playback side. To answer the last post of Csontos - most recording engineers of today do not at all have the same priorities as those of the past. With digital recording, they can make anything sound pretty much however they want to, and they do. There is absolutely no attempt anymore to be faithful to the actual sound, in the actual space it is being performed in. I am speaking here mainly of orchestral recording, done in a hall or large church, though of course in a recording studio they have even more control over the sound.
The other part of it is the miking itself, before it is mixed and edited. Back in the old days, with the great analog recordings, the recording companies would hang just a couple of mikes either out in the hall, or high above the orchestra, ala Mercury. This way they captured the sound of the space as the audience heard it, and were as faithful as possible to it. Conductors were also much more heavily involved in the process back then - the vast majority of them aren't really anymore, as far as how the recording itself goes (they are of course very concerned with the musical side of it). Now, with digital recording, you see at least one mike in every section of the orchestra, as close as they can get it to the musicians, and then they mix it and tweak it however the hell they want later. There are sometimes mikes out in the hall too, but very rarely, and even when there are, they aren't used much in the mix. Almost never does the final result sound like it actually did in the hall, in fact most of them don't even seem to care about recreating the sound of the actual space whatsoever. In other words, they simply don't care about soundstaging. This is a major reason why so many recordings and orchestras sound more and more the same in this day and age - not only are players all over the world sounding more and more the same, and regional differences are being lost, the recordings themselves contribute to this sameness by the way they are miked and edited and mixed.
Another dirty little secret is that to this day, so-called "sound engineers" are by and large not systematically trained, and learn on the job, almost always from someone else who had to do the same, if they have any help at all. There are very few books on the subject, and most recording engineers I have encountered in my career have never read any of them anyway. It is really quite shocking how little most of them actually know about what they are doing. Unfortunately, the fact that it is so easy for pretty much anyone to make and edit a recording now only exacerbates the situation. There are still some good ones out there, but in general, it is really true that most of them really don't know what they are doing.
Audiophiles love to talk about having equipment that is faithful to the recording, and as an audiophile myself I understand where they are coming from; however, as a professional musician, I know that almost all recording done nowadays is NOT faithful to what we actually sound like, particularly in the soundstaging aspects. They are just creating a mix that sounds good to them, and the miking is done so that they can manipulate what happens as much as possible. And again, this is even more true in the recording studios - just ask any orchestral musician that works for the film industry in Hollywood if the movie soundtrack accurately reflects what they actually sound like. They will probably not be able to help laughing in your face. Some audiophiles don't like to hear this kind of talk, but I'm sorry - it's true.
"This is a major reason why so many recordings and orchestras sound more and more the same in this day and age - not only are players all over the world sounding more and more the same, and regional differences are being lost, the recordings themselves contribute to this sameness by the way they are miked and edited and mixed. "
Hadn't thought about that but rings true.
Does it matter?
I tend to think whenever everything starts to sound the same it does, but most people probably would not care.
The flip side is there are more recordings out there than ever each day, so the variety is still going up even if many newer ones tend to have more similarities due to production than in the past.
Also I suppose the fact that much litening to music occurs on portable devices using earphones has a major effect on the way recordings as a whole are produced. Gotta be a very small % of listeners out there who care about soundstage and have the tools to actually get it.
The good news is that most all recordings these days can have a big soundstage when played if desired, even if in many or most cases it is mainly a result of the production as opposed to anything resembling an original live performance.
Its a lifelike illusion of the players performing in YOUR room which is different than where recorded regardless. That's all one can expect and mostly what matters to me.
Having said that, I am a big fan of recordings that DO attempt to reproduce original live performance soundstage, Like old MErcury recordings from teh late50's and early 60's, and some current niche labels like Mapleshade and Dorian.
Commenting on Learsfool's most recent post -- multi-miking classical recordings is not necessitated by the use of digital recording devices. There are no technical reasons why a 2 or 3 track, a la 1950s Decca/Mercury/RCA, digital recording cannot be made. The reasons for miking each orchestra section and/or soloist are artistic, practical and ultimately economic. Multi-miked, multi-channel digital recordings give the audio engineer/conductor/record producer/record label far greater control and make it easier to sculpt the final sound of a recording than purist recording techniques can. Effectively, classical recordings have adopted the pop music paradigm of record production which doesn't place an emphasis on "realism". It's not about the capture of an event, instead it's more about the creation of one.
I have one old Mercury Perfect Presence lp where the liner notes and graphics provide exact detail on performer's locations on stage during recording, the goal for the recording being to reproduce that, and for the listener to have a clear reference for how well the mission was accomplished. Very cool! That makes that particular lp very special and useful as a reference recording for soundstage and imaging.
Im not sure whether were not talking about it because it is obvious or because theres some being obtuse afoot, but maybe worth retreating to first principles for a moment to try and find a common vocabulary about what soundstage is an how it happens.
As I understand it, you have to start with basic psycho acoustics. Youve got two ears, one on each side of your head, at about a heads breadth apart. This gives us stereoscopic hearing which is basically the ability to triangulate distance and direction of the source of sounds based on the micro-second time differentiations between when the sound waves hit one ear and then the other. Two receptor points (ears) permits triangulation of the third (origin). Simple.
In its purest and simplest form, stereo recording does exactly the same thing. Two mics, set up about the same distance apart as your average pair of ears, records two channels of what in theory would be exactly the same two channels worth of acoustic queuing youd hear with your ears were they in the same spot as the microphones. Thus, these two channels worth of information should be able to recreate the same ability to triangulate source and distance of sounds. Soundstage, at its purest and most abstract.
From there, this pure abstraction pretty much goes to hell. For a lot of reasons. Even assuming a perfect time and phase correct recording, the mission-critical micro-second queuing differences between the two channels are pretty small. Theres a whole lot of ways they can get messed up, just from reading them off of whatever medium is at hand, processing them, amplifying them, subjecting them to the whims of multiple power sources and AC lines, shooting them through all manner of wires, and ultimately off to a pair of speakers, which have to turn these electrical signals back into a physical process by vibrating just so in order to excite sounds waves. Good luck with that. But from there, things really go wonky, because these same sound waves emanating from whatever vibrating bit, or typically several distinct bits, which may or may not really get along, is/are getting its/their shake on then are loose in the room and have to find their way to your ears. They bounce off of stuff, stuff can get in the way, the room can resonate at weird frequencies, they can bounce into each other and either cancel each other out or get excited in strange and inappropriate ways, in short, they can get into all manner of trouble. Only then, if everything goes as planned and they get to your ears still carrying the same micro-second encoding as first recorded (again, assuming it was recorded in the first place) then you get to hear this intended soundstage. Its a wonder it ever happens at all.
All of this, in turn, assumes a relatively simply miced, straightforward stereo recording of, well call it, unamplified source noise. That, more often than not, is not a correct assumption. Multi-channel recordings, including separately processed, altered or even wholly-created electronic channels have no original soundstage to preserve. None. Rather, they present the building blocks for the sound engineer to create whatever soundstage they see fit. Maybe hes looking at a multi-track recording of a symphony and wants to try to recreate a sense of the original soundstage through the mix. But, even assuming thats the goal, its still using the captured channels to create an approximation, and then down-mixing into the two channels of a stereo recording. Fundamentally, what we might describe as the soundstage is frequently an arbitrary fabrication born out of the decisions of the recording engineer. (Which then have to make the odyssey back from the recorded medium, through the chain, to your ears. Again, good luck.)
Anyway, you get the idea. Soundstage is, in some regards, stupefyingly mysterious. But in others, its really pretty straightforward. There are all manner of things that can effect, mess with, or otherwise screw it up. But they are, by and large, knowable things. As for one medium of recording being inherently better than another at preserving/conveying/reproducing/etc this information in theory, that makes no sense to me. In practice, who the hell knows? In any event, that's my theory and I'm sticking to it (unless and until I change my mind).
Who the hell cares? If there is a sound stage, why do I care if it's faithful to the original? I'm far more occupied with replicating the final version 'accurately'. That's what I care about. Because along with that goes the entire sphere of the performance issues of my gear. My goodness! Who of you have systems so perfect so as to have the need to focus on outside parameters? We only have control over decisions we make about our gear.
I have this related question and hope some knowledgeable members, especially Atmasphere, who do his own recordings, can share their views!
It is relatively easy to understand why we can create soundstage width in our stereo system, as there are 2 channels having slightly different information.
Then, how do we get soundstage height, since there is no top and bottom channels? Is it an artifact of speaker characteristics and in-room placement? Or is there some hidden information in the recording that can create this effect?
In many of the well set-up systems, I could hear clear, and relatively consistence (across the different systems), soundstage height! Yes, a higher ceiling and taller speakers seem to help in this area. So, it is hard for me to believe this is just an artifact!
Appreciate for all your comments!
It seems you've answered your own question. Room geometry and speaker placement are clearly responsible for the effects we hear. A sphere is the ideal cabinet so any diffraction caused by a 'box' is an artifact. If you fastened your speakers to a wall as if it were the floor, what do you think you'd hear?
Thanks for the comment!
So, you are saying that soundstage height is not imbedded in the original recording, but only an artifact of room geometry and speaker placement?
The interesting thing is that, if these are indeed artifacts, they are pretty consistent across different systems. For example, say in a recording of a singer singing and in the same time playing a guitar, in all the well set-up systems I have heard, I could hear the guitar at a position lower than the singers mouth / chest, just like in real life! The same can be experience in live recordings like Belafonte at the Carnegie Hall, in which I could clearly hear some of the audience chanting in the higher balcony position!
This effect is also quite consistent across different type of speakers, i.e. conventional speakers with tweeter on top and bass at bottom, panel and ribbon speakers with their units physically extended from top to bottom, electrostatic speakers with a relatively full range diaphragm etc.!
You're kidding yourself if you think you're hearing frequencies coming from areas or the plane they're not being produced except for diffraction and room reflections which are by no means precisely directed. The best you can hope for is a coincidence. Drivers are typically placed on the plane where their specific frequency band coincides with roughly where you would expect to hear those specific frequencies for a natural production of sound. Sound engineers may be manipulators but not magicians (in the real sense anyway). Speaker design itself I think plays a much bigger role in the height of the sound stage than does the recording. I think that's true for the other components too. Does it not follow that there should be a relatively consistent outcome?
Hi Onhwy61 - your comments on my post are absolutely correct. I would add to them, however, that I once asked a good engineer why he doesn't go ahead and do a more "purist," as you call it, style of recording, with just a couple of mikes out in the hall, and his response was that although he completely agreed with me on every point as to why that would indeed be preferable, he said he would certainly be fired if he did so. It comes down to the sort of thinking that if we have newer technology and newer capabilities, that this must be better, and you better use it. Now I am no Luddite, but neither do I believe that new technology is always better. The digital recording technology is so much cheaper and so much easier to use and manipulate the sound you are recording - that's why it has stuck, despite the fact that the vast majority of listeners who actually bother to make a direct comparison (and granted, this is quite a bother nowadays) prefer the analog.
I had an interesting experience in the 80's.
I had a private one on one demonstration from Dave Wilson where he ran through several of his recordings, on vinyl, through the very first Wilson Watts ( no Puppies at that time) but they had some 2pi panels extending the baffle.
Sitting only 3 feet away from each speaker forming a triangle, in a nearfield listening position I could hear the microphone patterns as clear as a bell. The soundstage was massive - deep wide and high, behind the speakers.
The rest of the system was his own gear - Goldmund Studio, Goldmund arm & cartridge, Rowland Coherence pre, Rowland Model 7 power amps all MIT wired.
In my view microphones can capture soundstage, they capture the resonances and artefacts of the recording environment if placed correctly.
To me the soundstage reproduction is very dependent on the quality and placement of microphones. Listening nearfield helps to eliminate ones own room from imposing itself on the reproduction.
Regarding the question about soundstage height, while I can't say this with 100% certainty, experiences I've had listening to a test record containing half-octave warble tones have led me to believe that our hearing mechanisms perceive certain parts of the treble spectrum, especially in the 7 to 10 kHz area iirc, as originating from a point in space that is considerably higher than the actual source.
If so, musical notes having significant spectral content in that region would tend to be "pulled" higher, along a vertical axis, than notes that don't. That may explain, for example, why a female singer's voice may be perceived as emanating from a point in space that is above the level of the guitar she may be playing.
I haven't done enough experimentation to totally rule out the possibility that ceiling reflections at those particular frequencies were the cause of what I perceived with the warble tones, but I'm doubtful that the leaf/quasi-ribbon tweeters in the speakers I was using at the time had sufficient vertical dispersion to cause ceiling reflections to be responsible.
So in the absence of specific evidence, I would not make the assumption that our hearing mechanisms perceive all frequencies as emanating from the height of their actual source.
The question about soundstage height is a good one.
I am pretty sure this is more attributable to the playback system and room acoustics than how the source is recorded in that I do not think two channel stereo format is suitable for capturing height dimension as opposed to width and depth.
I think that relative vertical location of drivers in the speaker system is a factor. Tweeters tend to be mounted higher and high frequencies are more directional so those frequencies might tend to be perceived as coming from higher location than others, for example. OF course, room acoustics and distance from listening position to speakers would be a factor as well.
In general I am a fan of speaker designs that tend to have drivers closely spaced, that emulate a point source as much as possible especially in smaller or most typical listening rooms. Larger rooms where listening occurs from more of a distance will be more forgiving.
I suspect that the ambient information contained in the recording, you know, the reverberant decay, echo, etc. is three dimensional; thus, on well recorded material the size of the venue is identifiable on a reasonably good home system as Carnegie Hall or Boston Symphony Hall, for example. Thus as one's system evolves, one should observe a better reproduction if the three dimensional ambient information, and a more accurate representation of the recording venue. One should be able, with some persistence, to get the perceived height of the soundstage to be the actual height of the room where the recording was made. Think of the soundstage as an expanding sphere. I can certainly understand if you've never gotten soundstage height you might be a little mystified.
FWIW, I like Almarg's observations regarding ceiling (room) reflections of the very high frequencies as being a possible issue contributing to the illusion of height. I would add though that I think the volume (SPL's) might be as great a contribution as well.
At least this was my experience with QUAD 63'S a number of years ago, not so much today with my present stuff. I think this is because of the vertical dispersion differences between the speakers. (My present tweets are Dynaudio Esotar's).
Thanks, Newbee. To clarify, though, what I was saying is that my experiences listening to the warble tones on the test record I referred to led me to believe that it is the nature of our hearing mechanisms themselves, completely independent of reflections from the ceiling (or anything else), that can cause notes having differing frequency content and harmonic structures to be perceived as originating from different heights.
My reference to ceiling reflections was made to indicate that the listening experiments I had done had not been extensive enough to positively RULE OUT the possibility that ceiling reflections were responsible for what I perceived. But for several reasons I was highly doubtful that reflections were the cause, or at least the main cause.
Just now I have repeated the experiment, using my present Daedalus Ulysses speakers, which as can be seen here have a driver layout that is vertically symmetrical, with the tweeters in the middle. In addition to performing the experiment at my normal listening distance of about 11 feet, I also performed it listening very near-field, from about 2 feet directly in front of the tweeters.
The results confirmed my earlier belief. I perceived each half-octave warble above about 5 kHz as originating, to varying degrees, from heights that were WAY above the tweeters, even when those tweeters were just 2 feet away and exactly at ear height. Below about 5 kHz that effect did not occur.