Heresy or....

[dbe]

Roman,

Glad to be of service some times.

I created heresy on the old rec.audio.tech board 15 years ago, when I "came out of the closet" and stated my philosophy that accuracy didn't matter, only the illusion of accuracy. Audio after all is only an illusion.

The flat frequency response target is a bit of a canard at times for several reasons. There is of course the recording to consider as you stated. But there are also inherent flaws in stereophony.

For example, for "purist" classical recordings with stereo mics, the diffuse sound field that occurred in the recording venue is then presented frontally over the speakers. This completely changes the tonal balance of the recorded diffuse field, adding far more 3 kHz emphasis due to the difference in the frontal HRTF, compared to the diffuse field HRTF.

There's also the power response hump from a 2 or 3 way's tweeter (centered at, you got it, 3 to 5 kHz) we discussed at length in the past couple months. Which then further accentuates the playback room's actual diffuse field response in the same range.

Then, there's the excess 7 khz ish energy that a frontal center phantom image is presented with by speakers, compared to real life. In real life, the frontal center source diffracts around the head differently than when its presented simultaneously from 30 degrees left and right from the speakers (ie with a more direct high frequency shot into the ears).

This is just scratching the surface. I believe strongly that the science of speaker design should spend 80% of its time determining what the right targets are, and not how to best realize some arbitrary flat graph.

Dave

The psychoacoustic phenomena that creates the illusion of stereophonic presentation is dependent upon our brain processing the information in a manner that is convincing enough to be credible. You have hit upon the many of the flaws seemingly inherent in many design executions. At the risk of sounding like a broken record I would like to add my pet peeve to the mix.... so I will.

Every time I see the frequency plot of a small 2-way that is flat to 20K I shudder at the thought of subjecting myself to listening to that speaker for more than 10 minutes. There has been a misconception that "flat on axis is right" in the DIY community for years. When The Albuquerque Speaker Society was in full swing (before Ken Martinez died, Sheldon Stokes moved away and Keith Kidder found bicycling) we built some of the speakers that were popular at that time. Every one of them exhibited the same characteristic - they were ALL bright on axis. This voicing is exacerbated by the obsession of even power response that is pervasive in the DIY community as a whole, which in itself is a whole other discussion (this type of power response exacerbates the 7K issue). Some of them were so bright as to be virtually un-listenable unless the listener was 20 to 30 degrees off axis. I realize that many designs are listened to in this fashion (pointing straight out into the room), but doing so will lose some of the micro-detail that creates the illusion of space and localization within the soundfield.

A by-product of having to listen to a loudspeaker @ 20-30 degrees off axis (pointed directly into the room, not at the listener) is the unintended consequence of HF side wall reflections being a major contributor to the diffuse field sound. An entire industry (absorption panels) has been built around this issue.

This brings me to my point. The 400,000 rule works. I have explained it on many occasions, so I will let Thorsten Loesch do it this time:


"Thus first be aware of the limitations of the speaker. Then consider the classic "rule of the 400000". This states that if the speaker is 3db down at a given lower frequency it should be 3db down at an HF point that will {should read: "when"} multiplied with the LF limit give 400000. So if your LF cutoff is 80Hz and we will equalise the system to be flat at 80Hz (or 3db down) then we should retain a flat response up to 5KHz and then roll off at an equal rate as the LF rolloff."


This works. It works every time (think Auratone 5C). If we want flat frequency response on axis out to 20K make sure the speaker is flat in-room to 20Hz. If not, then be prepared to adjust the frequency response at the listening position by listening off axis and live with the reality of not getting the best performance from the loudspeaker in the context of soundfield reproduction.

I have thought about the popularity of open baffles a great deal recently. I am of the opinion (obvious to most, I think) that the sound field radiation pattern of the speaker is the strongest selling point of the design. The frequency dependent off-axis cancellation eliminates much of the room splatter that screws mightily with spatialization from the speaker. This characteristic may, and I say 'may' because I have no data to back it up - just a hunch, ameliorate some of the lower mid phantom center issues that plague stereo soundfields.

All of this ties into what Roman is talking about in the context of design targets. Are we making proper initial choices in the design phase? Does the resulting speaker sound "right"? WTH is "right", anyway? How do we fix the problems? These are valid questions that fit somewhat loosely into the framework that the other, other Dave brought up at the head of the thread. As we attempt to create a more balanced sounding small speaker, non-linear distortion can become more of an issue due to the absence of the masking effects that wider bandwidth produces. Distortion is bad in any form and the loudspeaker is the biggest contributor in the system (other than the room - again, another discussion).

What a convolution building a good sounding system is, eh?

Enough of the SoC. I need to do some actual work.

Dave

[DDF]

Dave,
What you're posting about really intrigues me. I know you have done a lot of research in the field, and you're probably privy to a lot more of the information on human hearing than the rest of us are.

My wife (a physiology teacher) often says that speakers should be designed as a counterpoint to the way the human ear works. Problem is: while she can tell me precisely all the working components in the human ear and how they work, she has a hard time explaining how any of that translates to actual perception.

Our brain does a LOT of "filling in the blanks" when it comes to our senses. Take for example the "blind spot" we all possess in our field of vision in the space just between our eyes; somehow, our brain fills it in with SOMETHING. Likewise, our brain does a LOT of error correction with what we hear. I started noticing this back in college when people would make an XY-pair stereo recording of a show and it sounded awfully different than what I'd remembered.

Anyway, if you have any material on human hearing perception to share, I'd really enjoy reading it.

Paul,

Back in university, our team created a sophisticated computer model of the basilar membrane's mechanical response to external stimulae, taking the input at the oval window. A number of non-linearities had to be factored in such as hair stiffening adding damping on transients, adjacent coupling beteen hair cells etc etc. It was a blast, but application to audio was a bit limited. It did provide a better understanding, by way of illustration at the mechanical level, of masking, ear distortion generation, ear/brain integration time due to basilar membrane standing waves etc etc.

I guess my point is that to gain a real understanding of how we hear and listen, you have to go beyond this into how the brain processes the raw nerve firings. This is why I find audition and psycho-acoustics so facinating, useful, and why I'm so confused that it's largely ignored in the larger context of the hobby.

We're all like engineers. We just want to run out and solve a problem statement and build stuff. But just like in the real world at work, how many bother to take the time to really understand the questions they're trying to solve?

To be honest, for me this is really where most of my interest in audio now lands, because its the great uncharted territory. Crossover types, topologies, box loading etc, its all been heavily studied. What hasn't are:
a. audio performance targets that matter:
- measurements that matter;
- targets against those measurements that matter
- how we perceive sound
b. radiation patterns and room loading/interaction, and how they influence "a."

As topics come up that apply, I'm tickled pink to chime in on how human hearing works and how it can help make better speakers. In fact, I've been doing this for months but much of it goes uncommented on, I think there's just not much up-take for this sort of thing. I'll keep at it though, I just find the whole topic too cool to ignore.

Your example of "filling in" is a case in point. This filling in can lead to listening fatigue, but who measures that today, and how would you? Another small reason why you can't remove listening from the verification equation. But wouldn't it be cool to be able to measure this?

Dave

[DDF]

The psychoacoustic phenomena that creates the illusion of stereophonic presentation is dependent upon our brain processing the information in a manner that is convincing enough to be credible. You have hit upon the many of the flaws seemingly inherent in many design executions. At the risk of sounding like a broken record I would like to add my pet peeve to the mix.... so I will.

Every time I see the frequency plot of a small 2-way that is flat to 20K I shudder at the thought of subjecting myself to listening to that speaker for more than 10 minutes. There has been a misconception that "flat on axis is right" in the DIY community for years. When The Albuquerque Speaker Society was in full swing (before Ken Martinez died, Sheldon Stokes moved away and Keith Kidder found bicycling) we built some of the speakers that were popular at that time. Every one of them exhibited the same characteristic - they were ALL bright on axis. This voicing is exacerbated by the obsession of even power response that is pervasive in the DIY community as a whole, which in itself is a whole other discussion (this type of power response exacerbates the 7K issue). Some of them were so bright as to be virtually un-listenable unless the listener was 20 to 30 degrees off axis. I realize that many designs are listened to in this fashion (pointing straight out into the room), but doing so will lose some of the micro-detail that creates the illusion of space and localization within the soundfield.

A by-product of having to listen to a loudspeaker @ 20-30 degrees off axis (pointed directly into the room, not at the listener) is the unintended consequence of HF side wall reflections being a major contributor to the diffuse field sound. An entire industry (absorption panels) has been built around this issue.

This brings me to my point. The 400,000 rule works. I have explained it on many occasions, so I will let Thorsten Loesch do it this time:


"Thus first be aware of the limitations of the speaker. Then consider the classic "rule of the 400000". This states that if the speaker is 3db down at a given lower frequency it should be 3db down at an HF point that will {should read: "when"} multiplied with the LF limit give 400000. So if your LF cutoff is 80Hz and we will equalise the system to be flat at 80Hz (or 3db down) then we should retain a flat response up to 5KHz and then roll off at an equal rate as the LF rolloff."


This works. It works every time (think Auratone 5C). If we want flat frequency response on axis out to 20K make sure the speaker is flat in-room to 20Hz. If not, then be prepared to adjust the frequency response at the listening position by listening off axis and live with the reality of not getting the best performance from the loudspeaker in the context of soundfield reproduction.

I have thought about the popularity of open baffles a great deal recently. I am of the opinion (obvious to most, I think) that the sound field radiation pattern of the speaker is the strongest selling point of the design. The frequency dependent off-axis cancellation eliminates much of the room splatter that screws mightily with spatialization from the speaker. This characteristic may, and I say 'may' because I have no data to back it up - just a hunch, ameliorate some of the lower mid phantom center issues that plague stereo soundfields.

All of this ties into what Roman is talking about in the context of design targets. Are we making proper initial choices in the design phase? Does the resulting speaker sound "right"? WTH is "right", anyway? How do we fix the problems? These are valid questions that fit somewhat loosely into the framework that the other, other Dave brought up at the head of the thread. As we attempt to create a more balanced sounding small speaker, non-linear distortion can become more of an issue due to the absence of the masking effects that wider bandwidth produces. Distortion is bad in any form and the loudspeaker is the biggest contributor in the system (other than the room - again, another discussion).

What a convolution building a good sounding system is, eh?

Enough of the SoC. I need to do some actual work.

Dave

I recall J Gordon Holt years ago (30?) complaining about the exact same thing, that many flat speakers just sounded wrong. Bright, compressed and strident. (side bar: isn't that interesting how tonal balance leads to the perception of compression, completely unrelated mechanically; hmmmm...)

I just built a tiny sealed speaker with a Tang band 4" and a D26NC, for the TV. I gave it very flat mids, overdamped Q and extended high frequency because I wanted to accenuate voice intelligibility. Even with full BDC it sounds awful in a typical set-up for pure stereo listening, just unbalanced towards the upper mids: the mids are too flat and the Q too low for the high fs. Jammed into the corner on either side of my CRT, its a tiny bit upper bass heavy but the low Q/high fs works as does the tonal balance and intelligibility.

I've been thinking alot about radiation patterns lately myself. I surprised myself a few months ago. I started turning the TV speaker set backwards and turning on the sets internal speakers, so that a Wii controller doesn't get launched through the cone.

Intelligibility went way up compared to either the internal Tv speakers alone or the TB/Vifa set facing forward alone. All by adding less direct sound and more diffuse sound.

Its having me rethink bipoles....

I think you're dead on the money about the side wall and dipoles. have you ever had a long thin room and then placed the speakers along the long side? I lived with that set up a couple years and it was some of the best sound I ever had. The side walls were at least 15 to 20 ms away, as was my wall behind the listening spot. With a thick carpet, it was dreamy and not 'dead' by any stretch of the imagination.

I guess this topic's meandered a bit! Sorry for the thread jack Dave3!

Dave

[dlr]

I think you're dead on the money about the side wall and dipoles. have you ever had a long thin room and then placed the speakers along the long side? I lived with that set up a couple years and it was some of the best sound I ever had. The side walls were at least 15 to 20 ms away, as was my wall behind the listening spot. With a thick carpet, it was dreamy and not 'dead' by any stretch of the imagination.

This has been my preferred orientation since I first bought DQ-10s. Later with my Apogee ribbons, it seemed evem more important. I had no idea at the time (I knew nothing of speaker design) other than recognizing the benefit of delaying/removing side wall reflections. My only problem now is that my small house (maybe more appropriately, the wife) does not allow for such an extravagance.

dlr

[Paul Carmody]

To be honest, for me this is really where most of my interest in audio now lands, because its the great uncharted territory. Crossover types, topologies, box loading etc, its all been heavily studied. What hasn't are:
a. audio performance targets that matter:
- measurements that matter;
- targets against those measurements that matter
- how we perceive sound
b. radiation patterns and room loading/interaction, and how they influence "a."

Wow, my neck is sore from agreeing so much. (sorry, stole that quote from the Simpsons)

[biff]

This has been my preferred orientation since I first bought DQ-10s. Later with my Apogee ribbons, it seemed evem more important. I had no idea at the time (I knew nothing of speaker design) other than recognizing the benefit of delaying/removing side wall reflections. My only problem now is that my small house (maybe more appropriately, the wife) does not allow for such an extravagance.

dlr

I have occasionally been taken to task here for espousing that configuration , but i am a firm believer that the early reflections in the lateral plane have serious detrimental effects on the perceived sound. My room is approximately 30 feet wide with the speakers roughly on the third marks, and I had the luxury of playing around in the room prior to moving in and being "anchored" to a setup.



As you can see, the room has other problems, but the imaging is nice, with ductboard panels covered with fabric along the wall behind. But it exhibits no "hole in the middle" or lack of life. This arrangement also allows me to not use the center for anything other than movies with dialogue and yet the point of origin appears centered at the TV for appropriate material. Glad to know I'm not alone in the wide screen audible.

[dbe]

I have occasionally been taken to task here for espousing that configuration , but i am a firm believer that the early reflections in the lateral plane have serious detrimental effects on the perceived sound. My room is approximately 30 feet wide with the speakers roughly on the third marks, and I had the luxury of playing around in the room prior to moving in and being "anchored" to a setup.



As you can see, the room has other problems, but the imaging is nice, with ductboard panels covered with fabric along the wall behind. But it exhibits no "hole in the middle" or lack of life. This arrangement also allows me to not use the center for anything other than movies with dialogue and yet the point of origin appears centered at the TV for appropriate material. Glad to know I'm not alone in the wide screen audible.

Early sidewall reflections DESTROY soundstage. I can't imagine someone ripping you for talking about setting up a system with side walls waaaay the heck out there. Makes perfect sense to me.

You have a big room there Biff. The plants do help, don't they?

I really miss our home in the mountains of NM. Sitting on our deck during the summer evenings was glorious.... sigh..............

Dave

[MSaturn]

Man, that is a biiiig room. That's like my living room is, minus the nice view. I've found that ceiling reflections are very important, too - I get the best sound up there, where I'm far from the sidewalls and very far from the ceiling.

All of this seems to lend creedence to simply using speakers in free space - which is impractical at best ;p

[dbe]

I recall J Gordon Holt years ago (30?) complaining about the exact same thing, that many flat speakers just sounded wrong. Bright, compressed and strident. (side bar: isn't that interesting how tonal balance leads to the perception of compression, completely unrelated mechanically; hmmmm...)

I just built a tiny sealed speaker with a Tang band 4" and a D26NC, for the TV. I gave it very flat mids, overdamped Q and extended high frequency because I wanted to accenuate voice intelligibility. Even with full BDC it sounds awful in a typical set-up for pure stereo listening, just unbalanced towards the upper mids: the mids are too flat and the Q too low for the high fs. Jammed into the corner on either side of my CRT, its a tiny bit upper bass heavy but the low Q/high fs works as does the tonal balance and intelligibility.

I've been thinking alot about radiation patterns lately myself. I surprised myself a few months ago. I started turning the TV speaker set backwards and turning on the sets internal speakers, so that a Wii controller doesn't get launched through the cone.

Intelligibility went way up compared to either the internal Tv speakers alone or the TB/Vifa set facing forward alone. All by adding less direct sound and more diffuse sound.

Its having me rethink bipoles....

I think you're dead on the money about the side wall and dipoles. have you ever had a long thin room and then placed the speakers along the long side? I lived with that set up a couple years and it was some of the best sound I ever had. The side walls were at least 15 to 20 ms away, as was my wall behind the listening spot. With a thick carpet, it was dreamy and not 'dead' by any stretch of the imagination.

I guess this topic's meandered a bit! Sorry for the thread jack Dave3!

Dave

Bipoles!?! I must be seeing a Mirage



I just couldn't help myself.


I had to do it.

Dave

[romanbednarek]

Dave (DDF),

I understand what you mean about presenting posts on certain topics and getting limited feedback because this happened with one of the posts that I introduced a few months back. However, I think that it has to do with the diversity of the contributors on this forum because it consists of participants that span a wide range of experience levels. Even though some of the "newbies" might not respond to more advanced topics, I'm pretty certain that they can learn from these discussions because I recall being in the same situation when I started out and found that I was drawn to topics like these yet didn't have the confidence or experience to contribute even though I was learning. I suppose that the only harm that can result is when misinformation is assumed to be fact where only the experienced contributor might recognize the error yet a beginner might blindly agree (I also know this from experience as well by adopting theories that I later found to be false).

Anyway, my point is that even if the interaction may be limited, I'm pretty sure that there are many others who might not be posting but enjoy these discussions and can learn from them.

I think that your knowledge of psychoacoustics and how it relates to human physiology as well as how we actually perceive sound is very critical to answering some of the questions related to areas in audio which have yet to be proven including some that I have yet to understand. I've always been interested in learning what properties of a speaker design influence the soundstage presence and in general spatial presentation of the signal sent to the speakers. I understand that the room plays a huge role, but if we assume that the room parameters to be constant and the only variable is the speaker I still feel that there are variations inherent to the design of the speaker that vary from one design to another and influence the spatial presentation regardless of the room interactions to some degree. I'll add that I've had a few speaker designs that I did not like simply because they lacked depth and sounded "crowded" with music that had a lot of instruments and this resulted in a higher level of listening fatigue despite the use of low distortion drivers. On the other end of the spectrum was a pair of bookshelf speakers which used a polypropylene cone woofer that lacked an advanced low-distortion motor yet sounded extremely three-dimensional (almost too spacious) and smooth but lacked some detail probably due to an excessive amount of damping... yet this design resulted in minimal listening fatigue despite exhibiting higher levels of linear and non-linear distortion than the previously mentioned design.

I guess what I'm getting at is that there are certain characteristics that I value in a speaker design, like those mentioned in the previous paragraph, but don't quite understand exactly what "ingredients" are required to achieve those goals. I think that these are areas that span beyond the concepts that most DIY'ers are familiar with (like a focus on low distortion and balanced frequency response) yet are at more critical at defining the parameters required to achieve a great sounding speaker and not a speaker that looks great based on specifications.

I apologize for taking the thread further off topic but I'm intrigued by these concepts and I feel that while they defy convention, they are closer to defining what is truly important when designing a speaker to sound "right" based on the listener's preference or beyond the scope of what is commonly thought of as "how to design the best speaker".

[MSaturn]

I don't believe speakers - that is, the ultimate response of a speaker, including the room - should be trying to make up for errors in recording or presentation. IF we modify the signal at every point along the chain to try and get somewhere we're clearly failing to get, more modification is not the key.

There are many bad recordings out there, but most are music we -as a community - do not consider "good" or "listenable" music. Regardless, even on inferior records, there is information presented that should not be further modified, if only because it's already moving far away from what we first intended.

Most here probably enjoy jazz records, because they are so well-recorded. They seek a level of transparency and complete lack of processing that allows our systems to truly do what they were intended. I always demonstrate systems with classical, jazz, or well-recorded classic rock for a simple reason: these are the records with the dynamic range and intelligent processing that do not interfere with the stereo system. On very bad records - see any pop record, probably, released in the last ten years - there's probably scant little to be gained from a hifi system, because they're already compressed, EQ'ed, and otherwise ruined to the limit.

I still believe a speaker and its system should "Get in the way" as little as possible - that is not to say I'm a fan of crossoverless full-rangers, I simply mean limiting distortion, both linear and non-linear, is our most vital task.

[DDF]

Bipoles!?! I must be seeing a Mirage



I just couldn't help myself.


I had to do it.

Dave

Pfft! Ya, but mine will sound good (hopefully!) (No disrespect to Voecks!)

[DDF]

Early sidewall reflections DESTROY soundstage.

Early sidewall reflections can do one interesting but completely false thing: they really can give you big left of left and right of right. Neat if you're into that detail safari hunt sort of thing (I ain't).

[DDF]

I have occasionally been taken to task here for espousing that configuration

The imaging with a far-away sidewall set up is spectacular. With Q-sound I was getting unbelievably realistic images 25 feet left of left and right of right, which was exactly what was called for by the movie.

Biff, BTW, you have your priorities dead right: have to respect a man who's BBQ is bigger than his TV.

Dave

[biff]

Thanks Dave, I find a Tv to be useful with the OSD while configuring playback on the Oppo, and as a candleholder. hehehe

[DDF]

Dave (DDF),

I understand what you mean about presenting posts on certain topics and getting limited feedback .

Hi Roman,

Thanks for the kind words. FWIW, I agree with your position, its these finer details that can take the good to great.

Imaging is a very difficult one. There are several aspects I've been thinking about such as wondering if low order crossovers make the images seem bigger, due to the enhanced overlap. We localize height by the vertical HRTF and pinna cues. Do the different heights of the drivers smear the image height between them because we can uniquely localize both? However, vertical reflections are mainly interpreted as timbral, so maybe not?

Olive found that natural room reverb heightens the detection of reverb buried in the recording. That'd be an interesting area to explore. Much is known about the optimum amount of reverb for intelligibility, no one probably knows the optimum amount for ambience retrieval.

The standard "classical" view is that one requirement of good imaging is a clean impulse in the first 2 ms.

Spaciousness and image size are usually related to lower inter-aural cross correlation. I always wondered if this could be enhanced by lateral velocity sources.

J Gordon Holt wrote a good article (Space the final frontier) with an experienced hand's view of the requirements for good imaging. It might be on the Stereophile site still.

Dave

[Deward Hastings]

J Gordon Holt wrote a good article (Space the final frontier) with an experienced hand's view of the requirements for good imaging.

SL has been all over this in recent years also:

http://www.linkwitzlab.com/rooms.htm

http://www.linkwitzlab.com/stereo%20reproduction.htm

[romanbednarek]

Imaging is a very difficult one. There are several aspects I've been thinking about such as wondering if low order crossovers make the images seem bigger, due to the enhanced overlap. We localize height by the vertical HRTF and pinna cues. Do the different heights of the drivers smear the image height between them because we can uniquely localize both? However, vertical reflections are mainly interpreted as timbral, so maybe not?

Dave

I spent some time thinking about the imaging aspects of low order crossovers and generated some theories based upon my thoughts which seem to be related to your statement.

I was actually trying to figure out how critical perfect phase alignment was with even order crossover slopes and how it might relate to imaging. My idea was based on the idea of the phantom image between two stereo speakers when sent a mono signal. I wondered if the same phenomenon occurs between a tweeter and a woofer (for example) within the crossover region when those drivers are in phase from the listening position. This would imply that the image should be centered between the drivers if they are in phase within the crossover region but may shift to the woofer for frequencies below the crossover point and towards the tweeter for frequencies above the crossover point. This might infer that phase misalignment within the crossover region where both drivers are still audible might result in a blurred image due to the relative delays between the drivers (although the threshold of perception for both delay and relative SPL would come into play).

Also, I wondered if an odd order crossover slope, which has the drivers 90 degrees out of phase, might result in a slightly expanded or blurred image between the two drivers within the crossover region... once again using the example of a mono signal sent to a pair of stereo speakers but one has a phase shift of 90 degrees (not quite as bad as them being completely 180 degrees out of phase, but still exhibiting a delay relative to the other speaker).

It seems that this thread is full of tangents but includes topics that are fascinating to me and rarely discussed.

[DDF]

This would imply that the image should be centered between the drivers if they are in phase within the crossover region but may shift to the woofer for frequencies below the crossover point and towards the tweeter for frequencies above the crossover point.

It seems that this thread is full of tangents but includes topics that are fascinating to me and rarely discussed.

Let's keep tangenting until we come full circle.

This topic of woofer/tweeter overlap effects also facinates me.

We tend to hear height based on spectral notches, caused by the pinnae. A single tone played over a speaker in front of you will (unless swamped by the room) fluctuate in apparent height as the tone is swept.

From my analysis, the tone's position is most stable if the tweeter is located under the woofer.

Dave

[romanbednarek]

Let's keep tangenting until we come full circle.

This topic of woofer/tweeter overlap effects also facinates me.

We tend to hear height based on spectral notches, caused by the pinnae. A single tone played over a speaker in front of you will (unless swamped by the room) fluctuate in apparent height as the tone is swept.

From my analysis, the tone's position is most stable if the tweeter is located under the woofer.

Dave

Could it be true that this type of "tonal spatial expansion" which allocates different tones to different places within the soundstage based on driver location and crossover slope/frequency could result in the type of speaker that I prefer as it relates to a less "crowded" soundstage since different tones will be eminating from different points in space. I used to postulate that phase shift/group delay might have the same effect when comparing lower frequency tones to higher frequency tones and would also reflect the type of crossover employed but I suppose that our ears/brain might interpret those types of phase shift related delays differently due to the nature of the frequencies involved as it relates to relative wavelengths of the two tones. I suppose that since different types of instruments have different profiles of harmonic content beyond the fundamental tone (musical note) that any phase related delays or spatial distribution between the drivers might result in a different perception of each instrument within the soundstage (since the fundamental tone and the harmonics will be distributed differently depending on the frequency range that they fall in.... and I suppose that a flute would represent an acoustic instrument that is likely to have fewest harmonics).

I hope that we didn't scare everybody else away because it would be nice to get a few different points of view on these subjects.