The Objectives of a Loudspeaker in a Small Room

[graaf]

The standard approach, which is 99.9% of all speaker application, is to throw omni directional speakers in a room and turn them on.

ehem, are You joking?
perhaps I should clarify what I meant when writing "omni directional speakers"
I was thinking of the likes of MBL Radiahlstrahler
and this is by no means a "standard approach"
as an element of some "standard approach" we can surely consider "speakers with largely uncontrolled directivity" as opposed to e.g. Summa
but this is something completely different

Even if this were so, what difference does that make if what I wrote is correct?

Or perhaps incorrect? How can You be so sure?
Because I don’t know

(1)They enjoy the euphonic distortion
(2)They think, incorrectly, they are getting all the musical information contained in the recording
(3)They don't know any better

I’m not talking about "the general public" or "Bose enthusiasts" and what They like
I’m talking about most discerning audiophiles and about what they judge as accurate and realistic that is about what was written thousand times of space-recreation abilities of MBL Radiahlstrahler and the likes

It was not clear and simple before he and Geddes did their research. I don't think for one minute that Toole believes loud direct early reflections are desirable

Then tell us please what are the results of Toole’s research?
Have You read the "controversial" Toole’s paper or an abstract of it?
I don’t know anything about Toole’s beliefs but the scientific hypotheses he makes public seem to be contradictory with what You believe he believes in.

BTW I find it symptomatic that Dr Geddes has started to talk in this thread about "very early reflections" instead of earlier just "early reflections" and that You have started to talk in this thread about "loud direct early reflections" instead of earlier just "early reflections"

Can it be that "early reflections" as such are not bad at all? That the only bad "early reflections" are "loud very early reflections"?
Can it be that very rarely we have those bad "early reflections" in our living rooms?

All we know now is that "Toole has concluded that normal reflections in a typical small living room seem not to interfere with perception of the recorded space"
This is not direct quote form Toole and it is not very precise but this is precisely all we've got for now.

He prefers the masking behaviour of flat lateral power response because he gets lots of ambience, he likes this but he gets it at the cost of lost and distorted musical information

How do You know that he "prefers masking" and accepts that musical information is "lost and distorted"?
I think it is very unfair to say that of Mr Moulton
He seems to be very keen on sound perfection and on accuracy
Check this: http://www.kiqproductions.com/index.html
"Speak Accurately About What You Hear"

He says a LARGE room. We are talking about small rooms: Specifically, the objectives of operating speakers in small rooms.

I know exactly what we are talking about
But You seem to have overlooked that this "large room" in Moulton interview is just an example and that he says that "it should get better in small rooms"

Moulton then generalizes to a small room

not just "generalizes" - he says: "I've done enough of this now and designed enough rooms to say that yes, it does work."

His room treatments are not much different from Geddes's prescriptions, nor from the characteristics of Tool's own home living room.

I agree
But do You really mean a typical home living room which Dr Geddes criticizes so much as "badly designed"?
Do You mean that there is from audio reproduction point of view negligible difference between home living room and Dr Geddes' "well designed room"?

(Speculation: Moulton's "wide dispersion speakers" probably have really interesting narrowing directional and power distribution characteristics at higher frequencies.)

no need for speculations - You can just visit B&O and listen to critically acclaimed Beolab 5 loudspeakers which were co-designed by Moulton and quite critically acclaimed

There are two inter-related topics here: one is objectives of using speakers in a small room. The other is characteristics of a small room as used for listening to recorded music.

I agree
abstract of Toole’s paper seem to suggest that characteristics of a typical small listening room i.e. a typical living room is just fine

You should find this article by Floyd Toole interesting:
...
particularly pp22 and following.

thanks for the link
I know this article
It seems to predate his paper on "The Acoustics and Psychoacoustics of Loudspeakers in Small Rooms" and not to take into account the results of this research

I wonder are we going to ever know what is in this apparently "secret" paper.
"Secret" because I’m sure that many of diyaudio users have access to AES papers, Dr Geddes included, and I can’t understand why they choose not to refer to this particular paper

 

[fcserei]

I think You know the cause of my dissatisfiaction

anyway
Do You think that it is really necessary to "create the early reflection and reverb tail pattern" when we have natural or artificial "reverb tail pattern of a large room" already in the recording of "big music" and we have also "early reflections" already in our listening room?
Perhaps all we need is to remove the room's own reverb (i.e. everything after 50 ms) as Mr Moulton suggests?

What do You think?

Yes, it is absolute necessary. What is recorded is just a small fraction of all the ambient info you perceive in reality. It is more like window dressing on the record.
If they would really record the correct power ratio of direct and reflected sound, every record would sound like amateur home videos, where one mic on the camera makes every indoor sound like it was recorded in a bathroom.
I have lots of ambience channels and it is pretty interesting when you play a record of a single instruments ( just because it is easier to demonstrate with that) and you can "swap" perceived venues with changing the ambience signals and not changing the main speaker signals at all. And YES, CHANGING THE AMBIENCE (especially early reflection) SIGNAL SIGNIFICANTLY INCREASE OR DECREASE INTELLIGIBILITY, or even change perceived tempo. Playing back Klemperer in longer reverb halls will make his "slow" performances to be the right tempo, just weightier. Also most of the perceived bass punch and dynamics is coming from the ambience.

Removing the room's own sound completely is not necessary, because it can be effectively masked as long as the room you want to put the performance in is bigger than the listening room, which is about 99.9% true.

2 speakers will never do that, let it be the directionality of a pretzel.
The only achievable goal for 2 speakers to reproduce realistically in a small room is music, which would sound good originally in a room with similar acoustics. Even in that case wider dispersion, lots of diffusion and reflection, and sitting in the far field helps.

 

[gedlee]

I may have a completely different question to Mr Geddes.
I understand pretty much completely your reasoning behind Summa, and it indeed makes a pretty compelling case.
What I haven't seen so far discussed (or mentioned in Summa's papers) is how is box coloration addressed. I have to say here that I have never heard Suma, but I have heard reasonably large number of so called high-end box solutions, many of them universally accepted as well engineered (ATC, B&W, Avalon, Thiel, Vandersteen, Duntech/Dunlavy to name just a few). They all now sound colored to me after hearing boxless speakers (Linkwitz Orion, Gradient etc.). This coloration is combination of box resonance and reflexed sound emanating through the woofer membrane. It is very characteristic, and once you hear it you always hear it (yes, it is a curse).
We all know about waveguide and its magic in Summa, but is there any special trick to adress box colorations as well ?
Because as it stands now, I am inclined to view boxless speakers as preferred in-room solution.
At least until I perhaps get a chance to hear Summas in Melbourne (unlikely as it seems).

Bratislav

I am not completely convinced that boxes are as evil as you claim. I have seen and done scientific studies of them that showed little to nothing of note.

But at an rate, the Summa boxes are unique in their construction. They are composite with fiberglass skins - made like a sailboat. They are extremely rigid and have very low well damped vibrations. In the early models there was a double walled constrained layer damping system used, but in the ESP models that we will sell we have deleted this last feature as we did not find box re-radiation to be significantly affected by its removal.

As to sound coming back through the cone, we do fill the boxes with sound absorbing material. I have not seen any definitive studies that show that sound re-radiation through the cone is significant - and I have personally done tests on this very topic - and found nothing.

In conclusion, even though I have taken measures to combat these "effects", I have not been able to confirm what you claim to hear and I have looked. I suspect that there is something else that you like about OB systems and from this experience you conclude that Closed boxes have flaws. I would not doubt for a minute that OB and closed boxes sound different. I would claim that there is no evidence to say that OB are more accurate, and I would claim that because of the edge diffractions, the likelihood is that the OB is, in fact, less accurate than a closed box. SL made some claims several years ago that "dipoles" gave a better bass response "because they excited fewer modes", but this has been pretty widely discredited.

 

[gedlee]

BTW I find it symptomatic that Dr Geddes has started to talk in this thread about "very early reflections" instead of earlier just "early reflections" and that You have started to talk in this thread about "loud direct early reflections" instead of earlier just "early reflections"

Can it be that "early reflections" as such are not bad at all? That the only bad "early reflections" are "loud very early reflections"?
Can it be that very rarely we have those bad "early reflections" in our living rooms?

But do You really mean a typical home living room which Dr Geddes criticizes so much as "badly designed"?

Do You mean that there is from audio reproduction point of view negligible difference between home living room and Dr Geddes' "well designed room"?

no need for speculations - You can just visit B&O and listen to critically acclaimed Beolab 5 loudspeakers which were co-designed by Moulton and quite critically acclaimed

The level of a reflection does have an effect, sure, but "no reflections" is best. I will state it again - very early reflections are never beneficial to sound imaging and coloration.

I refer to living rooms not as being "badly designed", but simply "not designed" for sound reproduction. Usually doing nothing positive for the sound quality is a negative thing.

I have heard the Beolab 5 loudspeakers - I have not changed any of my opinions as a result.

 

[jlo]

Not sure about measurements, but this is a fairly credible source ...

thanks, I knew this one but it gives no facts and no values

 

[graaf]

Playing back Klemperer in longer reverb halls will make his "slow" performances to be the right tempo, just weightier. Also most of the perceived bass punch and dynamics is coming from the ambience.

Removing the room's own sound completely is not necessary, because it can be effectively masked as long as the room you want to put the performance in is bigger than the listening room, which is about 99.9% true.

thank You for really very interesting comments

especially because that remark that "removing the room's own sound completely is not necessary, because it can be effectively masked" is very much or perhaps even exactly what Mr Moulton says

still I am not totally convinced that the result of such a reverb processing are completely psychoacoustically consistent.
I would expect that incoherent data received by the brain (eyes can see the walls and ears can't hear them) has to result in a kind of unnatural quality of perceived sound.
Such was the conclusion of research done by Boothroyd-Stuart (Meridian) some time ago.

besides with such a reverb processing we find ourselves completely on the subjectivist side - i.e. it is the personal preference that matters - as we have practically no recordings made with something like this in mind

but isn't "high dfidelity" about accuracy and not subjective preferences?
on the other hand Linkwitz - designer quite fanatical about accuracy - has a surround system for "Ambience recovery from a 2-channel CD" at home because "doing so increases the apparent acoustic space around the listener and significantly enhances the realism of reproduction and its enjoyment"

well, so in the end it is "realism vs accuracy"

 

[graaf]

or should I rather give an
or at least

anyway
still I believe that there can only be realistic AND accurate
and that nothing but accuracy of the equipment + recording made with realism in mind can result in realism in sound reproduction

artificial reverb processing looks a bit like prosthesis

 

[chrisb]

I am not completely convinced that boxes are as evil as you claim. I have seen and done scientific studies of them that showed little to nothing of note.

But at an rate, the Summa boxes are unique in their construction. They are composite with fiberglass skins - made like a sailboat. They are extremely rigid and have very low well damped vibrations. In the early models there was a double walled constrained layer damping system used, but in the ESP models that we will sell we have deleted this last feature as we did not find box re-radiation to be significantly affected by its removal.

The timing of this thread with the removal of information from your former website is unfortunate, as it probably provided much technical data regarding the Summa products in particular, and your philosophy in general. Can we look forward to updated website on the new venture, or in the interim do you have non-proprietary information from your own research available that can fill in some of the blanks for the uninitiated?

As to sound coming back through the cone, we do fill the boxes with sound absorbing material. I have not seen any definitive studies that show that sound re-radiation through the cone is significant - and I have personally done tests on this very topic - and found nothing.

In the case of the very light materials used in construction of cones for some full-range drivers, is it unreasonable to postulate that some low level intermodulation might occur- particularly in the case of shallow, very lightly damped chambers as one might find in a back loaded horn?

In conclusion, even though I have taken measures to combat these "effects", I have not been able to confirm what you claim to hear and I have looked. I suspect that there is something else that you like about OB systems and from this experience you conclude that Closed boxes have flaws. I would not doubt for a minute that OB and closed boxes sound different. I would claim that there is no evidence to say that OB are more accurate, and I would claim that because of the edge diffractions, the likelihood is that the OB is, in fact, less accurate than a closed box. SL made some claims several years ago that "dipoles" gave a better bass response "because they excited fewer modes", but this has been pretty widely discredited.

Not all closed box systems are well enough designed (or constructed by "cloners") to avoid flaws of their own.


p.s. note that the sarcastic tag-line below certainly doesn't apply to most of the posters in this thread, but I don't think it's possible to selectively remove it for a particular audience.

 

[Pano]

There is some very interesting info on speakers in small rooms here: http://www.gradient.fi/design/stereo/index.html

The "Bypass Test" on page 3 is very surprising! Can flat FR speakers in an anechoic chamber be near perfect? There are some obvious faults in the test, but they are discussed a bit on the site and also in the paper below. Maybe we could discuss them here.

I'd like to hear Doc Geddes views on the bypass test.

Also some views and info on the subject in this PDF by Robert Green. The Gradient Bypass test is sited here. Not all of the document is about our subject, but there is some good stuff starting at page 9.

Thanks to jlo for pointing these out.

 

[Bratislav]

I am not completely convinced that boxes are as evil as you claim. I have seen and done scientific studies of them that showed little to nothing of note.

But at an rate, the Summa boxes are unique in their construction. They are composite with fiberglass skins - made like a sailboat. They are extremely rigid and have very low well damped vibrations. In the early models there was a double walled constrained layer damping system used, but in the ESP models that we will sell we have deleted this last feature as we did not find box re-radiation to be significantly affected by its removal.

As to sound coming back through the cone, we do fill the boxes with sound absorbing material. I have not seen any definitive studies that show that sound re-radiation through the cone is significant - and I have personally done tests on this very topic - and found nothing.

In conclusion, even though I have taken measures to combat these "effects", I have not been able to confirm what you claim to hear and I have looked. I suspect that there is something else that you like about OB systems and from this experience you conclude that Closed boxes have flaws. I would not doubt for a minute that OB and closed boxes sound different. I would claim that there is no evidence to say that OB are more accurate, and I would claim that because of the edge diffractions, the likelihood is that the OB is, in fact, less accurate than a closed box. SL made some claims several years ago that "dipoles" gave a better bass response "because they excited fewer modes", but this has been pretty widely discredited.

I don't have means to test this, so in order not to be completely unscientific (which bothers me a lot!) I'll propose a mind experiment.
I am no Feynman by any strech of imagination, but please bear with me and feel free to poke any holes in my theory.

Let's assume we have a perfect anechoic room next to our house. In fact, we share a wall with it. Let's cut the hole in this wall so one Summa fits in there perfectly. Only front of the box is in our living room, the rest sits inside of the perfect dead room next door (let's just forget for the moment that most anechoic rooms have walls meters thick).
Now for blasphemy - we jigsaw the back of Summa and fit exactly the same high efficiency 15" driver there. We keep the front disconnected and then blast the back driver with 100+ dBs sweep.
The "front" lobe of this driver is perfectly absorbed by our ideal anechoic room. The only sound it will make will be inside the box.
Now, intuition is terribly unreliable and unscientific, but what it tells me in this case is that sound we are going to hear in our living room (though the original Summa bass driver) is far from insignificant. In fact, at certain frequencies it may even approach double digit percent mark. Why ? Because absorbing material isn't terribly effective at all frequencies, especially bass (as everyone with dividing wall problem will atest), and speaker cone isn't much of a barrier for sound (otherwise we'd all build houses with impregnated paper walls )

Any obvous holes here ?

 

[soongsc]

...
In the case of the very light materials used in construction of cones for some full-range drivers, is it unreasonable to postulate that some low level intermodulation might occur- particularly in the case of shallow, very lightly damped chambers as one might find in a back loaded horn?

...

Light cones with little absorbing capability will suffer from enclosure resonances and back waves, even the cavity resonances in the driver itself becomes very critical.

 

[fcserei]

or should I rather give an
or at least

anyway
still I believe that there can only be realistic AND accurate
and that nothing but accuracy of the equipment + recording made with realism in mind can result in realism in sound reproduction

artificial reverb processing looks a bit like prosthesis

There are experimental mch recording methods for realistic and accurate reproduction. I heard one of the most promising, and have to tell, it is a completely different realm. Sadly, untill they will be widely available, we have to have the prothesis.
But they are very good prothesises.
When Gran Teatro La Fenice in Venice burned down in 1996, its sonic behaviour was at least partially saved, because several acoustical measurements had been performed short time before, employing the binaural impulse response technique. These measurement later used to compare the acoustics of the rebuilt opera house to the original. These kind of measurements are the base of ambience systhesis too.

 

[fcserei]

still I am not totally convinced that the result of such a reverb processing are completely psychoacoustically consistent.
I would expect that incoherent data received by the brain (eyes can see the walls and ears can't hear them) has to result in a kind of unnatural quality of perceived sound.
Such was the conclusion of research done by Boothroyd-Stuart (Meridian) some time ago.

Also inconsistent because it is missing the beer in my hand and the brunette on my lap if I play R&R...

But let's say you have a nicely decorated HT room. At low lighting it's ambience ( the colors, lighting, seating, forward orientation etc.) is not that far from a concert venue with 2000 seats. Playing big room in a small one is very much possible and very convincing.

but isn't "high dfidelity" about accuracy and not subjective preferences?
on the other hand Linkwitz - designer quite fanatical about accuracy - has a surround system for "Ambience recovery from a 2-channel CD" at home because "doing so increases the apparent acoustic space around the listener and significantly enhances the realism of reproduction and its enjoyment"

There is NOTHING accurate about putting 2 speakers in an arbitrary position in a small room with a completely different acoustic properties than the recording space.

 

[terry j]

Any obvious holes here ?

In the wall???

 

[Bratislav]

In the wall???

Thinking about this experiment again, it may be better to have two Summas (with no backs) joined. Thus soundwave from the test driver travels through two enclosures, mimicking the attenuation of the reflected sound more closely.
But I don't think it would change things drastically. We'd still hear this sound emanating from the driver in our living room at rather significant SPLs.

QED ?

 

[mikey_audiogeek]

Let's assume we have a perfect anechoic room next to our house.

And into this room we put a cat...

 

[gedlee]

All

I have been very busy here in BKK (Bangkok, Thailand) and I have not had a chance to answer all of the (mostly good)questions. Here are some brief answers to many of them. (By necessity I have to be brief in my responses here and this is the problem with these kinds of forums since often a short answer does not always tell the whole story. I try and reference where someone can go to get a better answer, but almost never will I actually go and do any research - hence, I have not looked up Dr. Tooles papers.)

Regarding titles, I prefer Not to use them, but people should be aware that using a title for one person but not another is impolite. In an achedemic environment it is improper to ever call a Ph.D. "Mr." or "Mrs." A little consideration for etiquet goes a long way.

Among the really well trained and experienced acousticians there is almost no disagreement on the effect of very early reflections (VER). Until our recent research people had not really considered diffraction, but in light of our recent paper it simply cannot be ignored. In most, if not all, tests on the audibilty of VER and diffraction the sound playback level was held constant and was usually on the low side (70s dB SPL). Unfortunately, in light of our recent data, these results must now be questioned. This is because we found a very strong effect of audibity with playback level. Thus one can only talk about prior results IF the playback levels are held at the same level. This severely limits the applicability of this prior data. Please keep this very important new insight in mind when discussing these previous test results.

Regarding the test done in an anechoic chamber, I don't doubt these results, for a single low playback level. But I would claim that they would fail at some higher dB SLP level when the diffraction became unmasked or seriuos nonlinearities became evident.

As to the thought experiment on sound transmission through the cone, there can be little doubt that such a thing does occur, I myself have had these same thoughts. But in a real world test where there is a sound field masking these effects, one cannot say that these effects would be audible. That is the issue. Just as one cannot deny that cables have an electrical effect, one that is probably measureable, there has never been a scientific study that has shown that these effects are audible on real life source material. One can chase these small, intuitively obviuos effects until they are exausted (both physically and financially) but I have to question the logic of doing that if in the end they do not represent a significantly audible phenomina.

Note the word "significant" in the above claim, because it is paramount to understanding my point. One must rank the "audible" effects and deal with them in priority order if a reasonable design is the goal. Because, I have seen, all too often, designs that minimized one "pet pieve" effect of the designer while missing the "big picture" in the design goals. In my experince axial response falls into this category, because, while it is important, it is only a small part of the overall design problem. Just getting the axial response right is not going to yield an adequite playback in a real room (not an anechoic chamber). Far more important is the control of the directivity for smooth and constant coverage, even if the axial response has to suffer, and low diffraction, which hardly shows up on an axial frequency response plot at all.

I have also found that nonlinear distortion - in the drivers - is not a problem that needs to be fixed. All too often designers conclude that what they don't like in a design is the "distortion" without really sorting out that this is due to a nonlinearity in the driver itself. In every test that we have done to determine the subjective effect of driver nonlinearity we have shown that it is not an effect that listeners can detect. It is measureable, and logically it should be reduced, but doing so does not appear to resuolt in better subjective perception.

I should note that there will always be designs for which nonlinear distortion and/or cone re-radiation are significant. I am not speaking to how to do designs wrong, I am only saying that in my designs these things are not significant issues. Trying to push a 1" tweeter to 110 dB SPL IS going to cause a nonlinear problem (or burn out the tweeter), but the solution is so obvious - use a compression driver, you will never detect the nonlinear distortion in one of these in any application in a small room. But then of course if you use this driver on a poor horn then there will be diffraction issues, etc. to worry about. It's all about putting the whole design into proper perspective in terms of performance expectations and price.

The Summa data was removed from my website because I no longer sell Summas. The technology was purchased by Ai, my company here in Thailand. Ai will soon have a website and all of the old Summa information, as well as a whole lot more newer data, will be posted there.

 

[Bratislav]

As to the thought experiment on sound transmission through the cone, there can be little doubt that such a thing does occur, I myself have had these same thoughts. But in a real world test where there is a sound field masking these effects, one cannot say that these effects would be audible. That is the issue. Just as one cannot deny that cables have an electrical effect, one that is probably measureable, there has never been a scientific study that has shown that these effects are audible on real life source material. One can chase these small, intuitively obviuos effects until they are exausted (both physically and financially) but I have to question the logic of doing that if in the end they do not represent a significantly audible phenomina.

Note the word "significant" in the above claim, because it is paramount to understanding my point. One must rank the "audible" effects and deal with them in priority order if a reasonable design is the goal. Because, I have seen, all too often, designs that minimized one "pet pieve" effect of the designer while missing the "big picture" in the design goals. In my experince axial response falls into this category, because, while it is important, it is only a small part of the overall design problem. Just getting the axial response right is not going to yield an adequite playback in a real room (not an anechoic chamber). Far more important is the control of the directivity for smooth and constant coverage, even if the axial response has to suffer, and low diffraction, which hardly shows up on an axial frequency response plot at all.

Earl,

thank you very much for taking time to answer these seemingly trivial questions. Point is, these aspects of speaker design are not 'out there' in the web/library space and it is very hard to get any reliable data. I also couldn't conceive a measuring jig to DIY measure these effects.

Now, it is completely possible that my preference of 'dipole bass' has nothing to do with reradiation of sound through the cone and box colorations. I'm trying hard not to fall for 'novelty'; but, I have also to say that I've been fond of dipole/boxless bass from early days of panel speakers (Apogees, Maggies and in particular some really large electrostatics). Now we're talking completely different principle of operation here, but modern dynamic equivalents (SL Orions etc.) have brought the same "rightness" to my ears. I'm biased in a way that I mostly use classical music as a reference (nowadays I visit live classical concert of some sort at least twice a month), so things like 'slam' or 'punch' (normally associated with horns and high efficiency speakers) don't really register much on my scale of importance.

It is shame that I can't find Summa nearby to have a listen. Maybe now, being made closer to Australia they will start to appear in larger numbers here, who knows.

best,
Bratislav

 

[gedlee]

Over the years I have found bass to be very hard to get right. It seemed like sometimes it worked and sometimes it didn't. The room was always a BIG factor, in fact, probably the biggest factor. But over the last couple of years I have found that the multiple bass sources approach virtually always does a decent job, but the room still ends up being a really big factor. Even with multiple subs, getting the bass right is a real task. It must be audible, without being boomy - even the slightest boominess (accentuation of a single frequency) is highly objectional to me. Some people take boominess to be good bass - they don't seem to get the fact that they are only hearing one note all the time.

I have found that attempts to extend the bass downward are always the most difficult since the modes get more widely spaced and they tend to not be very well damped. In the systems that mentioned, they virtually never go all that low in frequency, so they tended to avoid the real room problems due to the simple fact that there just wasn't any response down there to object to. Dipoles are simply NOT going to go very low in response without a monopole to take over. And once you have a monopole you have all the problems that you mentioned.

 

[graaf]

There is NOTHING accurate about putting 2 speakers in an arbitrary position in a small room with a completely different acoustic properties than the recording space.

well, but that's exactly the accuracy of the control room!
in which the recording is being made - monitored, mixed and so on to finally become something we get and that is all that we have

isn't the accuracy of our HiFi system about giving us exactly what the producer intended?
this is the accuracy as a goal of our HiFi system - neutral transparency - just like in Dr Geddes' "aquarium" metaphor

questions of accuracy of the recording itself and of realism are separate from questions of accuracy of our HiFi system

accurate HiFi system with inaccurate recording will bring inaccurate sound
and with unrealistic recording it will bring unrealistic sound

"garbage in - garbage out"

and signal processing - like deriving surround reverb channels from two channel CD - is a prosthesis to unrealistic sound of unrealistic recording

sometimes it can improve subjective realism of perceived sound perhaps but NEVER accuracy
it just cannot be done!
there can be no more accuracy in a HiFi sound - i.e. high fidelity to the physical reality in front of the microphone - then it is in the recording

what is lost cannot be regained with signal processing

so there is rahter NOTHING INACCURATE about putting 2 speakers in a small room with a completely different acoustic properties than the recording space
because this is how our music is made in control rooms

minus "arbitrary" of course but there is nothing arbitrary in positioning of the stereo loudspeakers - read Blumlein's stereo patent, everything in it has theoretical grounds