Audio Science: Does it explain everything about how something sounds?

[microstrip]

(...)

I am interesting in learning more about taking my own measurements and how to interpret the results. I would like to get the left and right channels to more closely match each other and to get the response to be smoother without resorting to digital room correction. It should result in better sound, but I will have to listen to know if it sounds more natural or real to me.

(...)

Peter,

We would all love to know how to interpret the results of measurements, making the bridge to the subjective. However all that WBF readers have been given are a primer on how to connect a microphone to the computer , ultra simplistic recipes such as "if the measured response at seating position has less than X dB variation" you should be a happy man or "my response is flatter than yours, I do not care about your poetry".

I often use REW as diagnostic tool - looking for peaks at the natural first resonances of the room, at the nuls caused by interference with the front and back wall. But every time the best listening setups were not the best looking graphs.

BTW, I respect and trust in those people who, due to large experience, scientific and empirical knowledge, really know how to use measurements. We had examples of their successful work in this forum.

 

[Groucho]

A good question to ask is: as a designer, could I make two pieces of equipment that measured identically but sounded different? If the answer is yes, then 'measurements' are clearly not sufficient to describe how something will sound. Of course the argument is dependent on the measurements being incomplete, which they always are.

 

[jkeny]

They can measure everything in terms of accuracy and signal transfer beyond levels of accuracy the ear is able to do. I would say they are able to measure everything needed to specify the perfect hearing aid. They don't know the processing needed by the ear however. So they lack parameters to fully specify the perfect hearing aid. But they do pretty good, and are getting better all the time. The advancements aren't in measuring basic things.

OK, so you think they have done all the measurements needed to specify a perfect reproduction system? (makes me wonder why they aren't being used by audio designers?)
You say that they don't know the processing needed by the ear? I was under the impression that the functioning of the ear is pretty well known at this stage so I'm confused by your claim.

Why I'm using hearing aid manufacturers is because they have the financial incentives to do the measurements & the resources to do them. The excuse used for audio equipment designers not doing these measurements is that there really isn't enough incentive - they get by with the usual specs, etc. & a good story

I'm trying to tease out what these measurements are that fully characterise what's needed & you say hearing aid manufacturers have these measurements but yet aren't able to produce a perfect reproduction system (miniaturised or not)

 

[microstrip]

A good question to ask is: as a designer, could I make two pieces of equipment that measured identically but sounded different? If the answer is yes, then 'measurements' are clearly not sufficient to describe how something will sound. Of course the argument is dependent on the measurements being incomplete, which they always are.

Fortunately Dieter Burmester answered your good question.

SOUND AND MUSICALITY

As former professional musician Dieter Burmester is still driven by his passion for music which he lets run free in his home studio as often as possible. He is convinced that musicality cannot be described adequately with the means of measuring data.

Measurements are able to make a statement about the technical and mechanic quality of the piece of equipment. They cannot, however, predict the actual sound. Proof is given by the fact that it is possible to build two devices, which have exactly the same technical data but a completely different sound.

 

[Ron Resnick]

I am not sure that hearing aid design is an apposite analogy. The hearing aid discussion makes me think of my tertiary hobby, amateur radio, where we talk a lot about audio, but strictly in the clinical context of communications quality and intelligibility. I assume that hearing aid design also is focused on voice intelligibility.

It is not obvious to me that our sonic goals in high-end audio, such as dynamics and soundstaging, have much to do with hearing aid design and measurements.

 

[treitz3]

A good question to ask is: as a designer, could I make two pieces of equipment that measured identically but sounded different? If the answer is yes, then 'measurements' are clearly not sufficient to describe how something will sound. Of course the argument is dependent on the measurements being incomplete, which they always are.

Hello Groucho. Another example of what microstrip mentioned is the Carver Challenge from many years ago. Bob Carver had a challenge to mimic the sound exactly from one amp to another. After nulling them out and making sure that the sound signatures measured exactly the same, he had what was two amplifiers that sounded the same. This, of course, turned out NOT to be the case but it fooled some rather good ears at the time right off the bat.

There has also been a speaker manufacturer (the name alludes me at the moment) that had two sets of speakers. Both measured the same but both did not sound the same. Measurements IMO can only reveal so much about what we hear.

Tom

 

[PeterA]

Fortunately Dieter Burmester answered your good question.

SOUND AND MUSICALITY

As former professional musician Dieter Burmester is still driven by his passion for music which he lets run free in his home studio as often as possible. He is convinced that musicality cannot be described adequately with the means of measuring data.

Measurements are able to make a statement about the technical and mechanic quality of the piece of equipment. They cannot, however, predict the actual sound. Proof is given by the fact that it is possible to build two devices, which have exactly the same technical data but a completely different sound.

Nice post. It reminds me of the story, though not completely analogous, of a guy who was auditioning three Lyra cartridges, an Olympos, a Titan i, and an Atlas. They were mounted on three identical arms playing on the same turntable. A Lyra rep was doing the demonstration. When asked which cartridge the customer wanted to buy, he pointed to the Titan. He wanted to buy that specific cartridge and not a different brand new Titan.

He knew full well that even the same model of a particular cartridge can sound different. I don't know if the cartridges' measurements were ever compared or whether they would even show a difference, but they had the same specifications, obviously. His experience over the years told him that two samples of the same cartridge can sound different.

 

[jkeny]

I am not sure that hearing aid design is an apposite analogy. The hearing aid discussion makes me think of my tertiary hobby, amateur radio, where we talk a lot about audio, but strictly in the clinical context of communications quality and intelligibility. I assume that hearing aid design also is focused on voice intelligibility.

It is not obvious to me that our sonic goals in high-end audio, such as dynamics and soundstaging, have much to do with hearing aid design and measurements.

Maybe you're right but I was trying to example an industry that has the incentive & the money to do the research & do whatever measurements are necessary to fully characterise perfect sound reproduction. Ok maybe they have a more limited goal then our audio reproduction systems but even with these curtailed objectives they haven't produce a system without audible issues.

Does this not say something is seriously wrong with the claim that "all measurements necessary can be done"? Does it not highlight the fact that there are still some aspects that are noticeable by our auditory processing mechanism that elude current measurements? I'm not saying that they are impossible but until these new measurements are conceived, they're unknown & therefore impossible to do, currently.

 

[Bill Hart]

I rarely wade into threads such as these because they are often more about voicing entrenched views than advancing knowledge. If good science says, by changing the value of X, you achieve Y as a measurable outcome, fine. How that translates into listening to music is more difficult. We elevate the recording to a vaunted place in determining audio system performance, knowing full well that it is at best a representation of a performance--if there ever was a performance (as opposed to a studio pastiche). Jules Coleman wrote an interesting think piece several years ago about striving for 'accurate' reproduction of the recording when the recording itself is not 'accurate' to the performance. Is it all 'buy what you like'? Probably not. But, despite my dyed in the wool, analog orientation, I've been spending time listening to how archivists 'restore' old acetates and transcriptions using digital technology. These are hardly 'purist' and are unquestionably 'fiddled with' but I have been blown away by how 'real' some of this ancient stuff can sound. Look at Steve's visit to DK's place- positively antediluvian technology, delivering the goods musically.

 

[CGabriel]

I do not think any amount of science will prove anything to a pure subjectivist. The argument will always be that science does not know it all, does not test the right stuff, and is too easily manipulated.

I do not think any amount of subjective testing will prove anything to a pure objectivist. There will always be bias, a lack of controls in the testing, and one more experiment/one more test condition to try.

I do think both sides could learn a lot from the other, but learning takes significant effort.

Very well said. Fortunately the majority of our members are somewhere in between the extremes. I don't think it matters what your bias may be as long as you treat other people respect. I have a lot of respect for many of the members even though some of them I may not agree with. I look for intellectual honesty when talking with people here. I don't mind asking or answering tough questions. However, when someone, in their behavior, displays an underlying hidden agenda or they don't contribute to the community other than be antagonistic - I will usually just stop talking to them.

There is is too much to do in real life than waste valuable energy pissing into the wind.

 

[Tony Lauck]

My question and answer were direct, and specific. I will repeat my answer again. People are raising objections to the value of measurements when what is presented to them is results of listening tests. When we perform listening test of one MP3 encoder versus another, we don't say "the measurements show encoder A to sound better than encoder B." We say that listening tests show the encoder A to sound better than B. This is 100% subjective evaluation and occurs all the time in acoustic research.

A measurement on the other hand, is objective. It does not include a human being and hence, is quite robust and reliable.

One cannot mix and match these two concepts and have the statements have any meaning.

I am not imagining anything about you so not sure why you say that. See my answer above regarding the rest.


My focus right now is to be sure that people don't take any graph and think it is a measurement. What is being refused to be accepted as proper audio science is listening tests results first and foremost in the recent arguments. The measurements that correlated with them are only relevant once we get past the listening test data.

You can be damn sure I won't take any graph as a measurement. When I see graphs the first step I take is to try and figure out who/how/why these graphs were made and my starting assumption is that the purpose of the graphs at best is to convince and at worst to deceive. This is especially true when the graphs have engineering and commercial significance. I've seen research papers made by competitors in my field use all the well known tricks to lie with statistics.

However, in the case of these graphs, there is, presumably* objective data behind these curves. Each datum is probably in a computer file as a result of a mouse click made by a subject in a blind listening test. This data is objective. It is not fundamentally different that the subject used his ears listening to sound waves, or used his eyes looking at oil drops descend in a fluid. It does not matter whether the sense perceptions were recorded manually on paper or by computer. It does not matter whether the perception was by the experimenter himself or by another person. Even if measurements are automatic their validity depends on correct experimental setup by a person whose subjective judgement is integral to their value.

*Psycho-acoustics is not my field. If it were I would probably be up on the details of various political controversies surrounding curves of this sort. (I've picked up my cynicism from doing hard time at standards committee meetings.)

TL;DR: There is nothing that we humans deal with that is objective. It is all subjective, like it or not.

 

[steve williams]

'at best to convince worst to deceive' ,that's exactly how I consider subjective opinion.
Keith.

Keith

do you really believe that in this hobby that there isn't room for both camps. I do agree however that all too often it is the tone and style with which both camps make their position known and this is what I sense gets under people's skins. I say to them, get a thicker skin, develop brass cajones or just don't read the threads from these people or ignore them. There is room for both sides. It is just a matter of developing a big boy style about one's own impressions

 

[Atmasphere]

What about if you don't know? Would you reliably detect the absence? Further would most people? Even those who consider lossy compresion a crime... For the recordthe audiophile in me even when I can't fully recognize mp3 from lossless prefer lossless
Don't laugh people do yourselves a favor and compare with knowledge removed mp3 320 to full CD and see how good you are in determining whihc is which without prior training.

As I said, this is why people don't want to pay for a nice stereo and for that matter, music itself. They think they don't care when what is really happening is that the music, what's left of it, fails to inspire.

They can measure everything in terms of accuracy and signal transfer beyond levels of accuracy the ear is able to do. I would say they are able to measure everything needed to specify the perfect hearing aid. They don't know the processing needed by the ear however. So they lack parameters to fully specify the perfect hearing aid. But they do pretty good, and are getting better all the time. The advancements aren't in measuring basic things.

Calling BS on this one. The ear can here very suble amounts of higher-ordered harmonic distortion that is very difficult to measure. The reason for this is that the ear uses those harmonics to gauge how loud a sound is. This fact is one that is ignored by the audio industry. We can't and don't measure harmonic distortion to the degree needed; if we did you could look at an amplifier spec sheet and know immediately how it sounds. Since that isn't the case, logic dictates that there must be a problem with the measurement. It has been this way for 60 years. That's why we continue to have these inane discussions which will continue as long as the industry ignores human hearing/perceptual rules.

A good question to ask is: as a designer, could I make two pieces of equipment that measured identically but sounded different? If the answer is yes, then 'measurements' are clearly not sufficient to describe how something will sound. Of course the argument is dependent on the measurements being incomplete, which they always are.

As a case in point you can have two amps on the bench with identical frequency response and one will sound bright while the other does not. The problem is not a frequency response error, its caused by distortion- and oddly and also counter-intuitively by the one with the much lower distortion figure. It is because the ear translates distortion into tonality, and the amp with the lower distortion imparts the coloration because it has more higher-ordered harmonic distortion than the one that has higher distortion overall. Until the industry starts paying attention to how we perceive sound we will continue with these silly discussions. Although I am by no means the first to say things like this the industry seems to have a lot of inertia about measurement and is still stuck in the 1950s. I say that in the face of computer aids to measurement as they continue to ignore human physiology. If the spec is not meaningful to the human ear it is the Emperor's New Clothes.

I guarantee that if we did not have ears we would not play around with high end audio gear. In the world of audio the ear is the beggar king. The ear is the most important aspect of any audiophile's stereo. We can't do anything to change them except damage them- they are the result of millions of years of development. How about we make specs that pay attention to how they work rather than coming up with specs that don't mean anything to the ear? Logical huh, but we humans are short in that department...

 

[ddk]

Steve yes of course, but your ( or anyone else's ) opinion of one component ,in an unfamiliar system in an unfamiliar room, tells me nothing.
Now if you quote the specification or the measurements ,I have something upon which to base at least an initial opinion.
Keith.

Are you serious? Steve please link Lamm & Wilson sites for Keith, the specs are all there, I'd like to know what the specs tell him. While you're at it we need Shunyata's specs too, then there's the tape deck with Nick's head amp, and you already have an in depth record of your listening room construction, that should reveal all. Keith, all the specs and measurements you need can be had please tell us what to expect from Steve's system. I'd really like to know before I make the trip to his place since even with owning the same electronics I have no idea what to expect in his room. I'd love to know how to read the measurements the way you can, please share.

david

 

[steve williams]

Are you serious? Steve please link Lamm & Wilson sites for Keith, the specs are all there, I'd like to know what the specs tell him. While you're at it we need Shunyata's specs too, then there's the tape deck with Nick's head amp, and you already have an in depth record of your listening room construction, that should reveal all. Keith, all the specs and measurements you need can be had please tell us what to expect from Steve's system. I'd really like to know before I make the trip to his place since even with owning the same electronics I have no idea what to expect in his room. I'd love to know how read the measurements the way you can, please share.

david

That is certainly an interesting point as we do have the same electronics.

The point for me Keith and as David makes, is that there are room for both camps

I am not sure if David has measured his room but I would make a prediction based on the size of the room in relation to the size of the speakers as well as the irregular shape of the room that it doesn't measure all that flat (but I am guessing). Sadly if this is the case Keith you might pass up a chance to hear what I now state to be my new reference. The sound that comes from David's system is pure unadulterated enjoyable music which allowed me to sit for 4 days in front of his speakers listening.

Keith are you suggesting that you only buy equipment based on measurements rather than listening as well and as a corollary if something doesn't measure right but sounds great does this mean you won't buy it or are you suggesting that if it sounds good it must measure good.

I have no idea how David's room measures but I can tell you that it was the finest sound that I have ever heard. Period. End of story

 

[Johnny Vinyl]

Can a room/system sound great when only using one's ears and listening experience?

 

[andromedaaudio]

Ah its all relative and personal , what got me into this hobby was a presentation by avalon ascendants driven by Ml 32 and 432 amps , as i for example read david ddk doesnt have a high opinion of levinson electronics of the past ,and went for another direction .
Its proves to me its all relative and personal .
Regarding audioscience , measuring tools are a TOOL to reach a design goal , nothing more but nontheless can be very important and satisfactory , i measured /designed my system and in important ways it sounds correct ....and good .

 

[steve williams]

I think what David was suggesting was that for such a highly efficient speaker as his the need for power isn't an issue as he has found the sound of an SET tube amp produced what he likes to call natural.

 

[NorthStar]

Fortunately Dieter Burmester answered your good question.

SOUND AND MUSICALITY

As former professional musician Dieter Burmester is still driven by his passion for music which he lets run free in his home studio as often as possible. He is convinced that musicality cannot be described adequately with the means of measuring data.

Measurements are able to make a statement about the technical and mechanic quality of the piece of equipment. They cannot, however, predict the actual sound.
Proof is given by the fact that it is possible to build two devices, which have exactly the same technical data but a completely different sound.

Then we live in a free audio world where science and the art of listening are two different entities.

But one can be developed further in improving the other. ...With all permutations/combinations of variables entering into primary consideration.

______

* PeterA earlier mentioned "imaging"; as how to get best imaging...having our two stereo loudspeakers sounding the same across the entire audio spectrum.
First they have to be designed that way, within the closest tolerances possible...within a quarter decibel, or better, between each other. ...With the less margin error possible.
We cannot measure one pair and expect all the pairs to be sounding the same...simply too many variables...from manufacturing production facilities...tools replacement/modification...variations in batches...part's values...crossovers' derivations...material's composition...the room where they were measured...the mics used...their positioning...the room where they'll be finally resting and playing and re-positioning @ various times and then replaced. ...Brief, one speaker can be anywhere from a fraction of a decibel to more than one decibel different in loudness.

We know that for best imaging a healthy pair of ears, that are tuned with each other, without giving an emphasis to one over the other one (no two ears are identical...left and right), ...then each side wall has to be equidistant to their respective loudspeaker, proper amount of toe-in, proper distance from the loudspeakers to register all the music coherently, in phase with the room and our ears, etc.

Amir mentioned @ several times that the measurements (graphs) often posted by him are not measurements but psychoacoustical analyses from the art of listening (science).
...From trained listeners with their sound preferences of the loudspeakers reproducing specific music selections. ...So that they can design their loudspeakers according to them listener's preferences, with that reference audio frequency curve from 20 to 20.
But we know that above and below 20 have an influence between. ...So the variation in amplifiers and preamplifiers and the sources and cables enters the equation.
...And the various quality music recordings, of course. ...From the original master to the last replication/generation/modification/transfer/upsampling ...etc.

What is audio science? It's all of it...expanded audio knowledge from the recorded music to the manufacturing of audio components, loudspeakers, using measurements, latest researches in advanced material composition...acoustical properties...from the reference/preference frequency curve of trained music listeners...to the acoustical room treatments, to the fine tuning in exploring a vast assortment of physical tools and tweaks...to the connections between audio components and loudspeakers...to the main electrical grid with constant power flow voltage in providing all the current to all the components well regulated. ...And to the music reproduction in our own listening rooms @ home. ...And how we are involved in our own emotional psychoanalysis/interpretation. This sounds like a serious personal music love affair, a dedicated passion on a personal level between us and everything else surrounding us...the music playing in our room from the gear we picked and tuned to our preferred taste...and based on our own personal experience, knowledge and all that jazz...including our own dévouement ... dedicated resources ... time, sweat, pain, love, money and all. ...War and peace. :b

I'm just sharing my ideas, from my very minimal perspective in knowledge and in financial freedom.
This science, the audio science, the knowledge of acoustics, music recordings, reproduction, ...is constantly evolving...with the contribution of everything from everyone.
...Mike Lavigne, Gary from Genesis, ...them...and we can learn much more from many many more..and from them. Audio Science is the people's experiences and researches, and questions, and discussions, and challenges, and more questions, and measurements, and interpretations, and listening sessions, and the art of being emotionally involved in the music listening we love...each and everyone. ...Audio science is the sharing of that deep involvement in the art of music reproduction.

Can we explain all that emotional involvement, can we measure it? Of course we can...by always improving upon it. :b

 

[andromedaaudio]

For the midbass/highs that is , iirc for the woofers he needs more power , ive played a couple of years with a top quality 300 B set , but to say it was the holy grail no , it all comes down to different human beings / preferences

I think what David was suggesting was that for such a highly efficient speaker as his the need for power isn't an issue as he has found the sound of an SET tube amp produced what he likes to call natural.