Properties of speakers that creates a large and precise soundstage

[jim1274]

I have come to the same conclusion as you have. My room is acoustically treated and I have a short distance of 2 meters to my speakers in an equilateral triangle, this gives me a high ratio of direct sound and I very much prefer the sound inside that critical distance, which you often talk about. This makes significant variations from recording to recording when it comes to soundstage width and depth, it's simply more of a high-fidelity sound where I have a clearer view "into" the recordings instead of a "listening to music in my room" type of sound.

Currently my BMR “box” speakers are a little closer to an equilateral triangle from the listening position, 7’ apart and 8’ from my seat, Leaning forward to an exact equilateral position, i can get to an exact triangle. Well, I see exactly what you mean on the song that just came up in my demo/test playlist, Pink Floyd’s “Time”, noticeable in the early part of the chimes. The clarity improved. I’m going to need to rethink this go back to the drawing board….also same effect on Fleetwood Mac “Gold Dust Woman” that came up next. I switched back to the AMT dipoles and did the same forward and back listening position adjustment and the clarity did not change much if at all. Scratching my head on that. I have probably 400 hours in this soundstage evaluation “project” and just realized the work has just begun…

Some recordings have a very tight and narrow soundstage, while other recordings can have a wide sound with envelopment almost as if it was a multichannel recording. I like these large variations between different recordings which also often suit the structure of the songs well, some will simply work better with a wide soundstage while others will work better with a tighter and more intimate arrangement.

The difference in recordings is large for sure.

 

[jim1274]

I like reading about your, and other people's goals and approach. It's interesting to also see if there's any contrast to how I approach the issue too. I try a balancing act between "accuracy to the recording" and "achieving my preferred sonic presentation." To that end I generally don't like heavy toe in on most speakers, ones I've owned especially, because it tends to squeeze the sonic images and soundstage too tight. I like either pointed straight ahead, or at least some toe-out, until for instance central vocals sound to me more naturally sized and "relaxed." And toeing out also "relaxes" the soundstage (probably due to a little more sidewall reflections and slight off axis to tweeters?) and so I find the soundstage expands and feels less stuck in the speakers. I seek a sense of "instruments and voices occurring in real space" and a too tightly squeezed version of recordings for me works against this.

Good points/observations here. You have made a lot of comments over on the Omni thread on soundstage aspects, more related to my observations on size and envelopment characteristics. The Omnis and AMT dioples with far more reflective energy create a different “sound” than a forward firing speaker, and I have tried 2 other “box” ones besides the BMR. I think it’s a personal preference thing and also more recording dependent than previously thought.. I mounted a vigorous defense there on the merits of Omni dispersion, maybe too vigorous now with the addition of direct A/B comparisons and AMT dipole proving stiff completion.

If I had to pick between the Duevel Omnis, AMT dipoles, and the BMRs right now, it would not be easy. I still lean toward the dipole or Omni, the BMR (or any box tried) just lacking that characteristic of envelopement, immersion, spaciousness, 3-D, spaciousness (or whatever the right word is) that the multipolar dispersion designs bring.

On the other hand, I want a good level of accuracy, because I love the sound of different recordings (especially because I love a genre of music that specializes in really extreme swings of production styles during the 60's to 80s). So I don't want everything sounding the same, and want to maximize the difference between recordings. If a pair of drums were recorded very strangely in an old recording, off in some isolation booth, panned hard left in the back corner of the soundstage with a dollop of idiosyncratic reverb added, that's what I want to hear. So my ideal is to be taken to the event, like I'm "in" the particular recording, hearing real sounds, not just sort of seeing pictures of different recordings in a detached manner. (Which means getting a "live" timbral quality too).

Tonal accuracy is very important, but how does one know what “accurate” means in terms of the soundstage characteristics noted above? If you mean accuracy in terms of spacial localization precision, I will say that the Omnis may not match others. The AMT dioples do better than the omnis on this, as good as the box speakers in my early comparisons.

Fanatical attention to this particular goal has, I find, really paid off. I brought a recording over to a pal's place to listen to on his KEF LS50s which was really fun. It was an old late 60s/early 70's style all vocal harmonies record, very well recorded. It sounded really nice on his set up, very much like a "good recording." But played on mine it's just a different experience, the sense of the space behind the speakers melting away, no particular boundary to the recorded space, like the room just becomes the recording studio, with voices sounding so naturally "there"completely irrespective of the speakers, occupying that space.

Was that near field listening in both instances such that differences in your room acoustics played no part in your assesment? Just curious. The room acoustics have significant impact on the final sound for any speaker.

 

[tmuikku]

I have come to the same conclusion as you have. My room is acoustically treated and I have a short distance of 2 meters to my speakers in an equilateral triangle, this gives me a high ratio of direct sound and I very much prefer the sound inside that critical distance, which you often talk about. This makes significant variations from recording to recording when it comes to soundstage width and depth, it's simply more of a high-fidelity sound where I have a clearer view "into" the recordings instead of a "listening to music in my room" type of sound.

Some recordings have a very tight and narrow soundstage, while other recordings can have a wide sound with envelopment almost as if it was a multichannel recording. I like these large variations between different recordings which also often suit the structure of the songs well, some will simply work better with a wide soundstage while others will work better with a tighter and more intimate arrangement.

Cool cool

I must admit most of my listening is beyond, mainly due to living room arrangement and practical position of sofa being beyond the transition. Occasionally I get sit closer up at the transition, when there is dedicated time for listening. I don't know why I'm drawn to the sound with stream separation, or you, while some seem to be repelled by, perhaps previous exposure to sound in general which has stuck, after all there is learning process attached to sound perception which makes listening skill evolve through or lives. Or perhaps it's just how our brains work bit differently, could there be overwhelming alertness or something for some people? or perhaps some loudspeakers just aren't very nice without room blending it? hard to say, and perhaps even unnecessary, I think it is enough to be aware auditory system has it's part on perception, try to listen for it and then do what ever each feels best for each, take advantage to get better sound for you and also to understand others, they might listen other side than you!

 

[tmuikku]

On the other hand, I want a good level of accuracy, because I love the sound of different recordings (especially because I love a genre of music that specializes in really extreme swings of production styles during the 60's to 80s). So I don't want everything sounding the same, and want to maximize the difference between recordings. If a pair of drums were recorded very strangely in an old recording, off in some isolation booth, panned hard left in the back corner of the soundstage with a dollop of idiosyncratic reverb added, that's what I want to hear. So my ideal is to be taken to the event, like I'm "in" the particular recording, hearing real sounds, not just sort of seeing pictures of different recordings in a detached manner. (Which means getting a "live" timbral quality too).

It's good thing we can adjust the positioning isn't it, to our liking!

 

[tmuikku]

Accidentally hit reply in middle of my response..

So, are you saying clarity and “washed relaxed sound” are necessarily a trade-off? Before I swapped positions of the AMT and BMR from inside to outside, I moved my listening position up to an equilateral triangle position while BMR was still in the inner narrower separation position. I really did not notice much difference in instrument position on width, just more a reduction in depth, the speakers not disappearing as much as in a more farfield listening position. I also didn’t notice a distinct loss in clarity.

Hi,
yes, to the extent that is due to auditory system and stream separation. It is hard to communicate this stuff over without shared experience. I've used adjective hazy instead of washed previously, not sure where I picked the washed from. By washed I mean the clarity is not there, or could be improved.

The thing is, on my sofa it's quite nice sound, everything is nice and dandy, but it gets even better closer up. Like having a bit deteriorated eye sight and putting glasses on, you kind of see everything and understand everything without, but it just gets more detailed with glasses, focus. And this can be tested repeatedly by moving over the transition, toggle clarity on or off by leaning forward or backward, or taking a step in or out if you are standing. Everything kind of stays still within the image where they are left-right, but instead of kind of frontal flat kind of localization within the room the image perhaps expands or becomes more 3D kind of feel to it, this is also part of the haze or washyness I'm referring to.

The effect is very clear with mono noise for example, getting closer than the transition the speakers seem muted even if I look at them, all the sound is concentrated on the middle between, very sharp and clear phantom image of mono noise. Taking step further, the image expands and there is sound around the speakers as well in a way, everywhere in speaker direction, it's very distinct in a way, eerie feel almost to get closer when the focus happens. I can listen how the surroundings sound even, like whats the sound behind and so on.

If you experiment with this and find it audible and meaningful difference, then we could perhaps communicate more accurately about all the stuff. At least that's my hypothesis, this is about only common thing I'm aware of we have with our sound setups, we share very similar auditory system through common evolution, which gives some chance to understand each others perception. Well, a recording could be common as well The stream separation is either or, with very little sliding to it so it is very clearly defined in a way, so we should always be able to resort to it as reference point, it's not a sliding concept like image width or bass leanness or treble air which all vary with speaker and room and positioning and by the person who's telling the story.

 

[jim1274]

Hi,
yes, to the extent that is due to auditory system and stream separation. It is hard to communicate this stuff over without shared experience. I've used adjective hazy instead of washed previously, not sure where I picked the washed from. By washed I mean the clarity is not there, or could be improved.

Clarity, tonal accuracy, timbre and such are critical aspects of “good” as much (more?) than soundstage aspects of size, spatial precision, and “envelopment”. I did notice a lbetter clarity when moving closer to the BMR box speakers, but can’t quite figure out why that was not so with the AMT dipoles.

The thing is, on my sofa it's quite nice sound, everything is nice and dandy, but it gets even better closer up. Like having a bit deteriorated eye sight and putting glasses on, you kind of see everything and understand everything without, but it just gets more detailed with glasses, focus. And this can be tested repeatedly by moving over the transition, toggle clarity on or off by leaning forward or backward, or taking a step in or out if you are standing. Everything kind of stays still within the image where they are left-right, but instead of kind of frontal flat kind of localization within the room the image perhaps expands or becomes more 3D kind of feel to it, this is also part of the haze or washyness I'm referring to.

I experienced the same with the BMR box speakers. I don’t have much control of the 4’ minimum side wall distance spacing due to room furnishings, but easy enough to adjust distance to listening position from front wall. I did find that
moving the Omnis well into the room and closer to listening position made the rear wall reflections stronger and sooner arriving, making it sound like sound was coming from the rear of the room like my side (and rear to lesser extent) surround speakers on the seperate 7.4.1 system were playing themselves. Seemed too reverberant that way.

I’m not nearly done with speaker placement location variations testing of all 3 types of speakers
dispersion patterns, documenting the differences more systematically.

The effect is very clear with mono noise for example, getting closer than the transition the speakers seem muted even if I look at them, all the sound is concentrated on the middle between, very sharp and clear phantom image of mono noise. Taking step further, the image expands and there is sound around the speakers as well in a way, everywhere in speaker direction, it's very distinct in a way, eerie feel almost to get closer when the focus happens. I can listen how the surroundings sound even, like whats the sound behind and so on.

If you experiment with this and find it audible and meaningful difference, then we could perhaps communicate more accurately about all the stuff. At least that's my hypothesis, this is about only common thing I'm aware of we have with our sound setups, we share very similar auditory system through common evolution, which gives some chance to understand each others perception. Well, a recording could be common as well The stream separation is either or, with very little sliding to it so it is very clearly defined in a way, so we should always be able to resort to it as reference point, it's not a sliding concept like image width or bass leanness or treble air which all vary with speaker and room and positioning.

 

[tmuikku]

Before I moved to this phase of direct A/B comparison between the 3 types of speakers, mostly with the Omnis and then the dioples and box later, I did more extensive placement location testing. I adjusted distance from side and front wall and listening position closer or further until I found the right balance (reflective delay time?) for best sound. Is that what you are getting at, the point where the direct to reflected energy and time affords best soundstage and clarity, timber and such? I did find that finding the best positioning made a big difference.

Yes it is important to position things that sound the best, but that alone is subjective of course and I cannot relate to it remotely as we do not share the same space and time and auditory experience, hence trying to promote using the transition that happens with stream separation, switch in the auditory system, and not subjective preference. You could like either side, or both and thats cool, all we need is the transition as reference to be able to communicate over I could better relate what I hear with my system, to what you describe you hear with yours, if you described everything in relation to the transition. If you don't know about it then I think it's hard to know which side you are at. Description of speakers far away from walls and relatively dampened room makes me want to think you are having stream separation but not sure unless you notice it and communicate over.

Besides helping with communication between each other, the transition is important to each of us alone to be able to reason about the sound we perceive with our systems, which enables to tweak positioning to our liking more effectively if we know what we should be hearing, and what not, and then reason with logic how to optimize.

Is there a negative in having a wider and deeper soundstage while retaining instrument spacial localization precision? I mean the individual instruments spread further left to right and front to back as more centrally crowded together.

The thing is that if you have stream separation going, more accurate localization, the instruments are spread between L/R speakers like the mixing engineer wanted so varies by recording. If it feels too grouped in the center you have your speakers too close together, and could just spread the speakers out more to increase azimuth angle they are from median plane. Pick a recording which you think has extreme left and right panned sounds, now widen the base width between speakers as much as you dear, as much as you find feels natural to you.

Example, I feel equilateral listening triangle is too narrow with stream separation, and when speakers are roughly 40deg azimuth or something like that it feels natural to me, bit wider apart than equilateral listening triangle. To maintain stream separation, I need to maintain distance from ear to speaker, so basicaly what I've done is make roughly equilateral triangle for listening in sofa, and when I take a chair in front of the sofa to get over the transition for better listen the speakers are now naturally wider apart. I could draw an image later on if it's to help.

The point is, just do what ever feels right to you! The whole idea is to be able to root logic somewhere, and I find the transition very natural place to root my logic on as it doesn't depend on speakers or acoustics but my own perception which I always carry with me and can rely on, from which I can just span out and adjust the system to my liking as I understand what I hear, confusion is removed as I have my auditory system rooted. I could imagine having dipole and omni speakers at hand, comparing them and quite quickly learn how they compare for what I want to hear in my living room and be able to pick the suitable one. I'd know exactly what to listen for, and how they relate, both at the sofa and at the MLP in front. Actually, I would like to extend the MLP to sofa! that would be my goal. I would just adjust the setup and try to listen where the transition is, and figure out which one extends furthest.

 

[tmuikku]

Currently my BMR “box” speakers are a little closer to an equilateral triangle from the listening position, 7’ apart and 8’ from my seat, Leaning forward to an exact equilateral position, i can get to an exact triangle. Well, I see exactly what you mean on the song that just came up in my demo/test playlist, Pink Floyd’s “Time”, noticeable in the early part of the chimes. The clarity improved. I’m going to need to rethink this go back to the drawing board….also same effect on Fleetwood Mac “Gold Dust Woman” that came up next. I switched back to the AMT dipoles and did the same forward and back listening position adjustment and the clarity did not change much if at all. Scratching my head on that. I have probably 400 hours in this soundstage evaluation “project” and just realized the work has just begun…

Get them about 7' from ear to speaker, and speakers about 9' apart from each other, or something like that and try if it got too wide stereo image now Perhaps it would be better to keep speakers 7' apart from each other and get yourself closer so that you have roughly ~6' from ear to speaker. Try to listen when the image is natural to you, not too wide, not too bunched up. You don't have to adjust the speakers much if you just move your chair closer or further to adjust this, or listen standing up so moving back and forth is easy.

After experimenting you might find it out that it's the early reflections you want and that's absolutely fine. Point with all this is to help you to learn your own auditory system functioning, and know which state you want it to be to perceive the sound you want to hear with your speakers in your room. Basically, to improve listening skill, understand what your system actually sounds like.

 

[beefkabob]

In my untreated living room with the 708P several feet apart and a huge sub in the middle, with everything processed through Dirac, I could pick out every member of the band in front of myself. It felt live. In my cave, with or without Dirac, it never gets close. The acoustics aren't the same.

So, a lot of it is probably just a crap shoot.

 

[tmuikku]

Tonal accuracy is very important, but how does one know what “accurate” means in terms of the soundstage characteristics noted above? If you mean accuracy in terms of spacial localization precision, I will say that the Omnis may not match others. The AMT dioples do better than the omnis on this, as good as the box speakers in my early comparisons.

Relying on Griesinger papers, sound is more accurate with stream separation, and less accurate without, due to auditory system, your internal processor! This doesn't mean anything absolute, it just means that if you find what you perceive is not accurate, first make sure you are having stream separation going on by finding where the transition is and then moving closer to speakers than that. If it's still not accurate, you can logically reason it's the recording, or something in your setup as it is not your auditory system. No matter what the state of recording or playback system, you have it now more accurate than further away anyway! Point is, you now know it's not your auditory system that makes it less accurate, but something else, there is no confusion and logic can be applied.

Not knowing your auditory system affects, you might grave for more clarity for example, go and buy new speakers or amplifier in order to improve clarity only to notice it's still not there what you heard on the shop floor. You could have had stream separation happening in the hifi shop as it's different acoustic environment and likely optimized for good sound and mistakenly think it is the amplifier you are listening to, when in reality its your own auditory system giving more everything, paying attention! Then get the gear home, and have no stream separation and so on as you are not aware of any of this, might perceive no clarity and be disappointed to the purchase. While the problem could have been simply your auditory system not being able to provide clarity all along. Another, go buy cables, swap them in and go take a good listen and just to make sure unknowingly lean forward to hear better, and sure enough hear a better sound! and conclude, Yes, cables matter! Except they likely sound exactly the same as previous cable, you just made stream separation by leaning forward and perceive a change without ever knowing it.

Importance is being aware that auditory system matters and affects what you perceive regardless of gear and room, it's always with you. Learning to listen your own auditory system makes it possible to reduce confusion about what you hear, and instead enable use of logic.

Clarity, tonal accuracy, timbre and such are critical aspects of “good” as much (more?) than soundstage aspects of size, spatial precision, and “envelopment”. I did notice a lbetter clarity when moving closer to the BMR box speakers, but can’t quite figure out why that was not so with the AMT dipoles.

Yes, everything matters and at least some of it is not mutually exclusive, for example tonal accuracy could be off in absolute terms, but as long as both speakers match you could still have strong phantom image and make the "clarity" of your auditory system happen. Or, even if tonal accuracy was spot on, you might not have (max) clarity if your auditory system doesn't have stream separation going on.

If you find two speakers differ be aware that polar pattern and edge diffraction could make sound tonality vary a lot with listening axis, here is why you'd want good smooth polar response and optimize edge diffraction: sound would stay tonally correct no matter toe-in, and toe-in could then be utilized just to optimize spatial quality of (room) sound. If polar response is bad due to what ever reason, there is about one good listening axis where the toe-in must be locked, spatial quality cannot be adjusted otherwise you'd compromise the tonality.

Auditory system is separate adjustment layer basically, your speakers and room could be anything, ideal, yours, mine, and you could still change the sound by changing listening position as your auditory system locks in or not, stream separation or not. This is why I think auditory system as a subconscious thing in a way, it determines what I actually perceive with my conscious mind, it is like a filter layer between ears and mind. In general the speakers, room and positioning should be adjusted so that the filter (auditory system) gets into a state which provides sound you want to hear with your conscious mind, the one you logically think about the sound, the one you listen with. Auditory system is always there doing it's thing and you cannot swap it to another like speakers, or manipulate directly like EQ knob, but you could learn how it works and affects perception, and then indirectly exploit it to your advantage by adjusting positioning for example.

 

[tmuikku]

In my untreated living room with the 708P several feet apart and a huge sub in the middle, with everything processed through Dirac, I could pick out every member of the band in front of myself. It felt live. In my cave, with or without Dirac, it never gets close. The acoustics aren't the same.

So, a lot of it is probably just a crap shoot.

You could try shrink the listening triangle in the cave if you wish, to reduce room sound in relation to direct sound. It might be get impractically small though.

 

[sigbergaudio]

I just replaced the Saranna prototypes (cardioid down to ~200hz) with our SBS.1 speaker (traditional sealed) in the same room in the name of science (not really).

The sense of envelopment / immersion is similar but a bit less pronounced with the sealed speakers compared to the cardioid.

The cardioid speakers have notably better clarity and ..separation of instruments and whatever is going on. Easily noticed with complex / messy productions like for instance rock and hard rock. While still sounding good, the soundstage more easily bunches up into "wall of sound" with the traditional sealed speakers, while it stays separated with the cardioid speakers.

With "extreme" panning/phase effects like in the "Stole the night" track I referenced earlier, the directional cues are perceived significantly different. It's not perceived as panned as hard to the right, and it also doesn't shift position during the sound but rather stays more put. I can only assume the out of phase sound from the cardioid vents alters the effect. Hard to tell which one is more accurate (if there is such a thing), but they definitely sound different.

 

[tmuikku]

^Could you try if you find the cardioid speaker sound gets closer to traditional ones if you listen bit further away? I mean, your description sounds bit like the traditional speakers do not make stream separation happen, while the cardioids do. If so, they then have the transition at different distances in the room. Could be that you are on same side of transition with both, but the perception is still different. It would be very interesting information regardless.

 

[Brian like sushi]

I am a mixing engineer, and I am very curious about what creates the 3D sound or depth of field when listening to speakers. For example, when we are mixing or mastering, we use mid-side tools to make the sound more three-dimensional. However, many speakers, such as the well-known Yamaha NS10, have a very flat and forward sound. Of course, this is to help mixing engineers better hear the midrange or vocals. On the other hand, some speakers, like many series from Kef, have a very layered sound. Many people believe that speakers with a layered sound may not be suitable for mixing because it's difficult to accurately judge the volume of different instruments. I remain skeptical about this. So, I want to know what design principles create the 3D sound and depth of field in speakers? Is it DSP, amplifier, or speaker units? Do these speakers with a strong sense of soundstage have anything in common?
Happy dragon new year and Welcome to discuss together!

 

[Blockader]

I am a mixing engineer, and I am very curious about what creates the 3D sound or depth of field when listening to speakers. For example, when we are mixing or mastering, we use mid-side tools to make the sound more three-dimensional. However, many speakers, such as the well-known Yamaha NS10, have a very flat and forward sound. Of course, this is to help mixing engineers better hear the midrange or vocals. On the other hand, some speakers, like many series from Kef, have a very layered sound. Many people believe that speakers with a layered sound may not be suitable for mixing because it's difficult to accurately judge the volume of different instruments. I remain skeptical about this. So, I want to know what design principles create the 3D sound and depth of field in speakers? Is it DSP, amplifier, or speaker units? Do these speakers with a strong sense of soundstage have anything in common?
Happy dragon new year and Welcome to discuss together!

controlled directivitiy = layered & 3d soundstage.
good channel matching is also important which is something that most of the good DSP speakers and high-end analog studio speakers have.

Our brain processes location of sounds based on the SPL and phase difference between 2 ears. ILD and ITD: https://en.wikipedia.org/wiki/Sound_localization
If a speaker has good controlled directivity, it means the reflections bouncing off walls and reaching your ear from every direction will have a sound level similar to the direct sound of speakers. I am simplifying it a bit too much but if the reflections sound different than the direct sound of speakers, our brain can't figure out where the sounds are coming from.

As the difference between direct sound and reflections gets larger, the imaging quality of the speakers also diminishes.

 

[ppataki]

Based on my experience it comes down to the speaker being (or appearing to be) point source
The more point source you have the better holographic representation you will get
( I guess those Kefs that you mention were coax designs = more point source)

 

[Crosstalk]

controlled directivitiy = layered & 3d soundstage.

Do we have tests showing the controlled directive speakers were voted by the listeners as more 3D consistently? What speakers were used for testing this ?

 

[Blockader]

Do we have tests showing the controlled directive speakers were voted by the listeners as more 3D consistently? What speakers were used for testing this ?

This is a long topic to talk about, Floyd Toole's book goes into details of this and yes there are a lot of blind test results backing this up.

 

[Propheticus]

Perhaps this thread is of interest to you:

Properties of speakers that creates a large and precise soundstage

There are some similar threads about what imaging and soundstaging even is, and some threads about how to measure it (we're not great at that). In this thread I thought it would be interesting to have a look what properties is likely to enable a speaker to be good at this. My (!) definition of...

www.audiosciencereview.com

 

[tmuikku]

Yeah to me, in my place and speakers there is both 2D and 3D kind of presentation available, which changes with listening distance, and quite suddenly. This also means I can change at will, which means speakers need preserve what is on the recording well enough, not some particular thing like linear phase or point source, I don't have either.

I think 3D perception of stereo sound is due to auditory system mostly, when some threshold snaps in auditory system, because if I start moving myself slowly closer toward speakers all room early reflections as well as direct to reflected sound gradually changes with the movement, but at some particular distance (ear from speakers) there seems to be quite sudden change in perception where perception of frontal 2D kind of stereo sound changes into 3D space in a way, like stepping inside the sound. Also clarity happens, it literally feels like stepping inside the sound with one step. When inside, the phantom image doesn't localize into local room as much anymore, but is more the entity that is on the recording and also all around me rather than just on front.

While I do not know any details why it is actually so, like delays and levels of particular reflections and all that, everything that needs to happen happens since the perception shifts. My speakers have low edge diffraction, relatively high and flat DI and so on, I've tried to make them problem free, so in general I think speakers just need to be fine enough to preserve what is it on the recording. Then well matched between left and right, and then just get close enough to speakers so that local room sound gets out of the way enough in order to brain to lock in to the recorded sound and not to localize the sound into your speakers or the local room, into 2D plane at the portion of room where the speakers are.

Look for Auditory Proximity by David Griesinger for more, perceptual change is very much like he refers as Limit of Localization distance so I believe it's the same thing, property of auditory system and not particular property of speakers for example.