What's Left In Speaker Design To Reduce Distortion/Increase Detail Retrieval?

[goat76]

It seems generally accepted that a pair of microphones can suffice to capture a stereo recording intended for reproduction by a pair of loudspeakers
(or perhaps more credibly, with suitable provisions and processing, stereo headphones).
That sound pressure captured at 2 points in space should be acceptable for recreating an entire sound space
or the sound image from musicians in that space seems absurd to contemplate, but is generally accepted.
Mathematically, one would need to record sound pressure over the entire surface of the space to reproduce it.

Two capturing points are all that is needed for a distance to occur between them on the horizontal plane, and the distance to the sound sources will capture the distance that we hear as depth.

Those two distances will be enough to capture both width and depth and will together capture the three-dimensional space of the recorded venue. If we try to capture the space with even more capturing points, that's when the perceived distance information gets more diffused, and the reason why multi-mono recordings are losing that natural width and depth where we can otherwise pinpoint different sound sources in an ensemble to their natural place in the three-dimensional space.

The downside of a recording made with only two capturing points is unfortunately obvious, different types of instruments will not be fully captured at the same distance, and it’s pretty clear that microphones don't work like our sense of hearing.

 

[FrankW]

You mistook someone for making naive subjectivist arguments

I currently listen ..good "quality drivers"...gob-smacking sense of clarity and detail in to recordings.
Then I go over to my Pal's place and listen to a pair of big ol' Estelon speakers,..retail for something like $65K... they just seemed to obviously dig out more sonic information in the recordings..on my system would be well placed in spatial terms, and I can hear if the drums were placed in a reverb. But the Estelon speakers just seem to effortlessly carve out precisely where the drums are in the soundstage and the precise acoustics or added reverb around the drums...and exactly where that reverb "ends" is more vivid and obvious. Basically there is this constant sense of more sonic information, presenting more precision about what is in the recording.
Which had me wondering what accounted for these differences. Better drivers? The more heroic efforts that went in to removing the influence of the Estelon cabinets? The whole design?

No mistake. 100% dismissed all science to account for differences.
1) You.
2) Zero controls for biases, like price, materials, cabinet's, etc, etc, etc.
3) Zero cognizance of volume/spl affecting everything, including "detail retrieval" and other audiophile contrivances
4) Zero cognizance of vastly different rooms/subsequent modal exactment, possible +/-20db variation
The latter 2 having zero to do with the speakers per se.
The Klippel paper I linked (unlike all others) made clear controlled listening was mandatory for any "investigation" into speculations.

 

[FrankW]

True, but the underlying blind preference for a speaker with a gently sloping downward response, in room, still appears to hold for the vast majority of listeners

The confusion in this is measured in room you get a downward sloping curve of something that is flat on axis in an anechoic chamber. The downward slope is a measurement artifact. So in this sense downward sloping room curve and flat anechoic are the same thing.

Incorrect. Your confusion is clarified by the quote above I'm responding to (and quoted in my response). There is no underlying blind preference for a speaker with a gently sloping downward response, in room
The blind preference is for anechoic smooth. A possible result in room, could be a smoothly downward slope. Its not a guaranteed correlation. Thus it cannot be used as a preference, because it isn't. Easily demonstrated by non smooth anechoic speakers having similarly smoothly downward slope...but not preferred.
The audio forum belief that smooth downward in room as a "target", is nonsense.

 

[FrankW]

My "subjectivism" is different. I only want to know what I need, that's why I take myself as the subject to be investigated.

Ok Matt

 

[Galliardist]

Incorrect. Your confusion is clarified by the quote above I'm responding to (and quoted in my response). There is no underlying blind preference for a speaker with a gently sloping downward response, in room
The blind preference is for anechoic smooth. A possible result in room, could be a smoothly downward slope. Its not a guaranteed correlation. Thus it cannot be used as a preference, because it isn't. Easily demonstrated by non smooth anechoic speakers having similarly smoothly downward slope...but not preferred.
The audio forum belief that smooth downward in room as a "target", is nonsense.

What I find interesting here is that when I posted what you are now telling me in another thread here, it was also called nonsense - hence what I said this time..,

 

[FrankW]

What I find interesting here is that when I posted what you are now telling me in another thread here, it was also called nonsense - hence what I said this time..,

Can you link for me?

 

[Galliardist]

Can you link for me?

To be honest I’d prefer not to, since that thread was awkward to say the least, and I caused some bother with a different comment there which I don’t want to stir up again.

 

[fineMen]

One can measure distortion of all types “at the ears”. Currently we don’t have the technology to eliminate all types, so it is a trade off in terms of what matters more, but it is measurable and the causes (if not their solutions) are well understood.

To me, the OP was very clear about the terms he was looking to address. I thought I was as well.

The rest of this seems to be about the issue of the circle of confusion, which I agree we are still in (though less and less all the time). Until that circle is closed, I personally am not worrying about the other side of the recording. The recording is the recording.

Ja, lots of confusion. I leave it at that. Only an explanation regarding the idea of a communication (sic!). A recording is a piece of art. It is an abstraction of the real thing. Like a pencil sketch isn't the real thing but allows the informed viewer to understand. We are so far used to it now, but there was a time when cave paintings were the next hot thing ;-)

A recording communicates, it doesn't "have it all". Some aspects are subdued, others are emphasized etc pp ad nauseam.
That's what I wanted to remind you and others of.

To leave it: get yourselves real speakers, namely 3- or better 4-way designs. KEF's R3s are fine, especially when supplemented with subs. And then after just trust the excellence, because I told you so ... stop thinking ...

 

[IPunchCholla]

Eh? So the musicians do not exist? I think your intended meaning got lost somewhere.

When I listen to a recording of a Shostakovich symphony, I expect it to sound like I am seated in the audience, but without the sneezing, munching and farting. With rock music it’s a bit different as many instruments are recorded via the feed rather than via a mic.

Of course they exist. But even when I was in a speed metal band in the 90s, we were using a 4 track to record and weren’t playing together. The song was simulated. There was no single performance being recorded. We even put together parts out of multiple takes. Now, a lot of music isn’t even recorded, but generated.

When I am talking about the vast majority of music, I am thinking in terms of all of the music listened to. What percentage of music do you think classical makes up? Even rock is a small fraction. Apple Music has/had a feature that let you listen to the 25 favorite songs being listened to around the world. Almost all of it was rap, the last I checked. None of it was recorded as a single take of a performance where the listener was expected to feel there was just a flimsy curtain between them and the performance. In fact, I would say except for vocals, much of it wasn’t recorded at all.

The second part of it is that it isn’t being mixed to sound like a recording of a single event in time and place. The locations of the instruments aren’t being put in a precise place. Instead panning and reverb effects are being used as creative tools. Drums sometimes have different toms, cymbals, snares sounding like they are in places that make no sense in terms of where the other parts are And sometimes each sounds like it is in a different space, reverb-wise. I’ve even heard things like the attack on a guitar being spaced differently than the fundamental. DAWs are now as much a creative tool as folks playing instruments.

So, no, for most music I don’t think there is meant to be a sense for the audience of “being there”. The only ”there” is you the listener right now listening to the song.

 

[IPunchCholla]

Ja, lots of confusion. I leave it at that. Only an explanation regarding the idea of a communication (sic!). A recording is a piece of art. It is an abstraction of the real thing. Like a pencil sketch isn't the real thing but allows the informed viewer to understand. We are so far used to it now, but there was a time when cave paintings were the next hot thing ;-)

View attachment 288172

A recording communicates, it doesn't "have it all". Some aspects are subdued, others are emphasized etc pp ad nauseam.
That's what I wanted to remind you and others of.

To leave it: get yourselves real speakers, namely 3- or better 4-way designs. KEF's R3s are fine, especially when supplemented with subs. And then after just trust the excellence, because I told you so ... stop thinking ...

I would never claim to not be confused. I don’t think I am saying anything different than you, honestly. Nor am I going to blow $2k on speakers that are less performative set of speakers than my current setup, particularly if the difference isn’t audible to this particular meat sack anyway.

 

[FrankW]

The only ”there” is you the listener right now listening to the song....

...in different rooms, sitting in different chair, staring at different speakers interacting, at random volume, retrieving details.

 

[Pudik]

I'm sure there's room beyond speakers like the Genelec 8351/8361 and the speakers you heard in terms of sinad. However in term of real world improvements, ie those that we can hear, I'm more inclined to say it's in the quality of recordings and multiple speakers for exact placement of instruments instead of a stereo image by 2, than speakers with a higher sinad.

The reason I'm saying this is that I heard the 8351 and left with the feeling not much could be improved upon but for the varying recording quality. I thought the speakers were absolutely amazing btw.

An EXCELLENT comment. How true. IMHO, stereo on two channels, even w a sub is becoming obsolete, especially if one listens to orchstral/big band music, cuz no Stereo system meant for the average guy (such as I) could ever give justice to these genres. But, of course, no matter how many channels, the recording must be good. Garbage in, garbage out. Just my two cents.

 

[Pudik]

I think there are areas where improvement can be made to existing designs, but what changes will consumers tolerate? For example , Genelec’s “Minimum Difraction Enclosure”. Many here just plain don’t like the look, even though it offers measurable advantages over more conventional looking enclosures. Or read through the 8381A thread and note how many don’t like and aren’t interested in what could be the highest performing speaker on the planet simply beacuse they find it ugly.
Aesthetics trumps performance for most I guess.
The same for fully active designs, a lot of people (the majority?) aren’t interested because of fears over longevity, obsolescence or simply because they love to swap components. What is the ratio of passive to active speaker sales do you suppose? (Small lifestyle products excluded)

As far as i m concerned, a speaker could look like a ramshackle egg carton if it sounds good to my ears. Never cared abt aesthetics. Accuracy and transparency, though. Surely, this is all subjective, but what art isn't. I never ignore science, on the contrary, since i have engineering background. However, if i ask someone about 'La dolce vita', which i loved at the time, that someone may well answer 'i hated it!! Yeah, my comment is ignorable, since it may seem to be outside of the subject, but it also seems to me that this discussion malformed into a polemic. Would you guys, please, calm down and pursue a constructive exchange of ideas? Thx, P.

 

[kemmler3D]

So I'm coming to this thread very late, so I'll do that lame thing where I answer OP's question without reading the large amount of preceding discussion.

I think we all agree voice coil-based speakers are pretty good today, maybe close to the best they will ever get in terms of audible improvements. Purifi is putting out drivers with -60dB distortion across wide frequency ranges which for most content means no audible distortion at all.

However, I think we all agree that speakers using today's tech can't consistently produce a truly convincing illusion of reality. Leaving aside the question of whether recordings can or should do that, let's take that as the standard for maximal sound quality.

So what's needed to reliably fool the ear? The reproduced sound must:

-Come from the right direction
-Have access to the full dynamic range (up to 120dB SPL or something)
-Have access to the full frequency range
-Have no/low audible distortion - phase, harmonic, or otherwise

I think that covers everything. On that list, today all of those things have been achieved, or at least the tech exists if you can afford it. If you sit in the middle of a sphere of Genelecs playing a 240-channel mix of a 240 channel recording of an orchestra... or something - you're probably going to hear a very convincing simulation of reality. So I would say it's possible but not practical for most people.

"Come from the right direction" is the least practical requirement. Even if you have a perfect object-based Atmos recording, having more than 6 or 7 speakers in a room is a big stretch for most people. Most people aren't willing to run cables all over the place, fiddle with receivers, and so on.

So IMO In order to have an effectively ideal speaker system in a room, you're going to need drivers to be smaller, more powerful, and much more power efficient than they are today, maybe even wirelessly powered, and you're going to need lots of them, maybe 20+. Even if we relax and say subs can remain as they are, you're still going to need tiny units capable of 200-20khz @ 100+ SPL all over the place.

I don't think this is possible with traditional voice coils. Maybe some advanced form of DML could do it.

If we limit the discussion to 2ch then I would say the tech is basically mature, but fundamentally limited in how realistic it can be.

 

[FrankW]

However, I think we all agree that speakers using today's tech can't consistently produce a truly convincing illusion of reality. Leaving aside the question of whether recordings can or should do that, let's take that as the standard for maximal sound quality.

No, at least one who doesn't agree.
20+ years ago.
https://web.archive.org/web/20180521025158/http://www.onhifi.com/features/20010615.htm
https://www.stereophile.com/content/wheres-real-magazine-we-see-it-february-2001
Using rather pedestrian drivers (Vifa?) https://www.stereophile.com/floorloudspeakers/1194snell/index.html
The audiophile eyes saw plastic cones and plain vanilla wood boxes (at a lot less than $65k !). Yet the "detail retrieval" etc, etc. didn't seem lacking. At all.
Go figure.

 

[MattHooper]

The Klippel paper I linked (unlike all others) made clear controlled listening was mandatory for any "investigation" into speculations.

Well before you posted the Klippel paper, when you were incorrectly presuming I was putting too much confidence in my sighted listening impressions, I replied
to you:

"If we want to be more sure of any such conclusions, the ideal way to evaluate would be blind listening, for instance similar to the type cited by Floyd Toole."

And yet...you keep imagining that I need to be told this stuff?

I am so cognizant of the liabilities of sighted/uncontrolled listening impressions that I put in my tag line, attached to my every post a disclaimer should anyone, like you, feel they have to continually point out the problem of sighted listening:

"Of course...it could always be my imagination." (Do you see that at the bottom of all my posts? Wonder why it's there?) I have said over and over that my sighted anecdote in the OP is too unreliable to be put out as some fact, which is why I've said it can be ignored for the questions I've raised.

You keep brow-beating others for being un-scientific and yet cling tenaciously to your original errors interpreting my post, cherry-picking quotes and ignoring anything I've written that acts as counter evidence to your rash suppositions about my knowledge and arguments. The very definition of biased thinking.

BTW, about this, which you wrote:

"including "detail retrieval" and other audiophile contrivances"

I've explained what is meant by "detail retrieval" in the sense important to my question, which you have of course ignored, so here it is again:

There is sonic information encoded in recordings. This includes everything from the most conspicuous sonic characteristics of melody/song structure, voices, instruments, musical performance, down to the timbre of the chosen instruments, even the difference in timbre between various drum cymbals, down even to the subtlest bits of reverb or acoustics added or captured in recordings. The question then is are the best speakers capable at this point of reproducing all the sonic information in any recording, accurately, or is there still any ways to go? And IF there is still enough happening in the best speakers to distort the recorded information, or there may be yet more detail left unresolved, where would speakers need to improve. It's a totally open question on all counts.

So, FrankW, if you are going to wave away this talk of detail retrieval as some mere "audiophile contrivance," you ought to be able to answer this question:

What IS it that is on the recorded signal we are trying to reproduce? What kind of information is on a recording that we are trying to transpose back in to sound waves via loudspeakers?

Just like your previous resistance to the premise of this thread, if you ever bother to answer, you'll no doubt end up validating the assumptions I've made in making this thread.

 

[MattHooper]

An EXCELLENT comment. How true. IMHO, stereo on two channels, even w a sub is becoming obsolete, especially if one listens to orchstral/big band music, cuz no Stereo system meant for the average guy (such as I) could ever give justice to these genres. But, of course, no matter how many channels, the recording must be good. Garbage in, garbage out. Just my two cents.

I disagree.

With some hopefully obvious caveats: It seems dubious that even the most extravagant current two channel systems could reproduce a symphony indistinguishable from the real thing. Much less loudspeakers "you and I" could afford.

But that having been said, I'd disagree that one can't find affordable stereo speakers that do justice to orchestral music. That's partially because "doing justice" to any music is a subjective call. What may be good enough for me may not be for you.

But since I'm a soundtrack nut (and I enjoy classical music) orchestral music is among my favorite to listen to. I was listening last night and just blown away the the timbral richness I was hearing, the sense of scale, of sonic "distance" to the various orchestral sections. With my eyes closed it really wasn't that hard to imagine I was listening through a hall to a real orchestra. That is definitely my mind, my "will to believe" making up the difference there between this home presentation and the real thing, but those mere two channel speakers were giving my imagination a hell of a lot to play with!

 

[kemmler3D]

No, at least one who doesn't agree.
20+ years ago.
https://web.archive.org/web/20180521025158/http://www.onhifi.com/features/20010615.htm
https://www.stereophile.com/content/wheres-real-magazine-we-see-it-february-2001
Using rather pedestrian drivers (Vifa?) https://www.stereophile.com/floorloudspeakers/1194snell/index.html
The audiophile eyes saw plastic cones and plain vanilla wood boxes (at a lot less than $65k !). Yet the "detail retrieval" etc, etc. didn't seem lacking. At all.
Go figure.

I guess I should revise what I said there.


I mean a system that reliably, consistently, and somewhat independently of room reflections or any such considerations, (assuming a perfect recording with as many channels and 3D information as you might want) produces a truly illusory experience.

What I'm talking about: You are blindfolded... a jazz band plays a song in the room. They take the jazz band out of the room and play a recording of the same jazz band. You literally can't tell the difference. For the system to be considered ideal, IMO 90% or more of listeners need to fail an ABX between the recording and the real thing. This has never happened or even been attempted as far as I know, except maybe in very specific lab conditions.


AFAIK there is no such system today mostly because stereo recordings can't really do that, except in rare circumstances where the room happens to spatialize the recording in just the right way.

A truly ideal listening setup could do that every time, reliably, as long as the recording was of sufficient quality.

I think today drivers that are capable of making up such a system exist, but systems like that have been built rarely / never, nor have recordings like that been made very often, if ever.

I guess what I'm saying is I see plenty of room for audio tech to improve, but IMO voice coils are too big and power-hungry to build a consumer-friendly system that could do it.

 

[fineMen]

So what's needed to reliably fool the ear? The reproduced sound must:

-...
-Have access to the full dynamic range (up to 120dB SPL or something)
-...
-...

My point of view is opposite. The ear cannot be tricked. At least not mine. It has too many millions of years of experience up its sleeve for that.

The recording abstracts the event just to make it recognizable to the listener at home. An active process, starting with miking and ending with listening, not a virtual reality for ultrasonic bats (or cats, as someone wrote above). Stereo is not for the thoughtless. It doesn't fool the ear, it serves it.

Well, and then the dynamics - really 120dB? Background noise at home, the health of the ears, neighbors and other villagers?

Well then, stop with the late-romantic glorification of the not so young cultural technique of audio recording (as opposed to hearing aid).

This seems a bit like the miracle of cave painting. We have 4-D films today. But the abstraction from an image of reality to a drawing is still considered art. And some people don't even understand the principle.

Translated with www.DeepL.com/Translator (free version)

 

[kemmler3D]

The ear cannot be tricked.

My assertion is that it can, in theory... but the recording techniques and equipment needed to do it are rare or nonexistent today.

There's nothing in the laws of physics that prevents us from reproducing a complete 3D sound field with 100% accuracy. So the ear CAN be tricked, I am just not sure if it will take 10, 25, or 1000 years for us to reach that point.

Perhaps both recording and transduction will be done with lasers, somehow.