Art vs Science of Audio Reproduction: Which Team Are You In?

[Irvrobinson]

I believe it is a matter of education. In the beginning, each of us can listen to different systems and decide which sounds best. But at this point, "best" is invariably just personal preference, often colored by brighter highs and enhanced bass.

Then we begin to learn about accuracy. And if we're lucky, we have someone to teach us. Now we begin to listen for accuracy, and our preference always starts to sway in that direction. The more we learn, the more we sway.

Finally comes the epiphany. It is our money, our system, our ears, and accuracy doesn't necessarily trump personal preference. Now we begin to moderate our bias toward accuracy w/ a portion of personal preference. And this is where the debates heat up.

So what are you saying? That accuracy may not be so pleasing, so the ultimate evolution of system design is to tweak accuracy to preference? Not me, at least not for acoustic music. For movies and tv accuracy chasing does appear to be a waste of time, but, for me, acoustic music sounds better on an accurate system.

 

[herbu]

So what are you saying? That accuracy may not be so pleasing, so the ultimate evolution of system design is to tweak accuracy to preference?

No sir. I'm saying that someone who is just beginning their audio journey is likely uneducated in the art or science of audio, but can still tell what sounds good to them. If they choose to learn about the subject, they will begin to learn about "accuracy", and their preference will invariably shift in that direction.

If they continue to learn, and invest, they will ultimately reach the conclusion that the most accurate on paper may not be their top preference. (Perhaps they decide that actively bi-amped speakers get top priority. ) Now they will begin to temper pure accuracy with personal preference.

Not saying they abandon accuracy, just temper it with the realization that an audiophile doesn't buy a system solely because it has the best specs. Rather they construct their system with the specs and attributes that are most pleasing to them. Of course, accuracy remains prominent among those attributes, just not always the final word. (Do your components have the absolutely best s/n ratio available?)

 

[KEW]

I think for people with an acoustic reference (such as many musicians), accuracy is never trumped. Anything that doesn't sound accurate is wrong.

I might prefer a certain inaccurate sound for a synthesizer solo (which has no reference for what is the true sound of it ), but if it is a sax, trumpet, violin, etc, it has to sound real.

 

[Johnny2Bad]

The purpose of any Audio (or A/V gear) is to reproduce stored data representing a musical performance via an analog interface ending at our brain's interpretation of sound. So the ultimate judge of success must be "how does it sound".

We can attempt to correlate a given conclusion about sound quality with measurements. But if there is a conflict, by definition, the measurement must be wrong somehow. It's also worth mentioning that the most sophisticated audio measurement cannot equal the operating conditions of even simple real-time musical conditions.

With regard to accuracy, the brain not only *can* be fooled, in the process of "hearing" the sub-conscious actively seeks to fool the conscious. So it is not absolutely necessary for accuracy when a subjectively preferential inaccuracy is present. This complicates correlation of measured criteria to sound quality.

 

[Irvrobinson]

It's also worth mentioning that the most sophisticated audio measurement cannot equal the operating conditions of even simple real-time musical conditions.

"Operating conditions"? Do you have any idea what you're talking about?

 

[Johnny2Bad]

Why, now that you ask, I do. Do you?

Assuming your answer would be "no", I'll rephrase:

When an amplifier is asked to play even a simple musical note, say the middle C key of a piano, you are asking the amp to reproduce about a dozen notes that comprise the fundamental plus harmonics inherent in the sound of a piano (It's those harmonics that make a piano and a guitar sound different when playing the same fundamental).

Yet we take an amplifier that tests well for THD, then re-test with 2 simultaneous sine wave signals and look for intermodulation distortion. Typically the IMD is higher than the THD at the same output level/load impedance.

Well, where is your intermodulation test for a dozen simultaneous sine wave signals that represent one key played on one instrument, the piano? And would the resulting distortion % be the same as your two signal IMD test? We can only guess, because we cannot test that condition; its too complex for our test gear. A normally skeptical person would assume higher distortion than the THD value and higher than the 2-signal IMD value, but we don't know the value, there is not test for it.

Yet we listen to music under that condition, and typically much more complex than that condition (ie more complex than a single piano key being played).

 

[ski2xblack]

When an amplifier is asked to play even a simple musical note, say the middle C key of a piano, you are asking the amp to reproduce about a dozen notes that comprise the fundamental plus harmonics inherent in the sound of a piano (It's those harmonics that make a piano and a guitar sound different when playing the same fundamental).

Yet we take an amplifier that tests well for THD, then re-test with 2 simultaneous sine wave signals and look for intermodulation distortion. Typically the IMD is higher than the THD at the same output level/load impedance.

Well, where is your intermodulation test for a dozen simultaneous sine wave signals that represent one key played on one instrument, the piano? And would the resulting distortion % be the same as your two signal IMD test? We can only guess, because we cannot test that condition; its too complex for our test gear. A normally skeptical person would assume higher distortion than the THD value and higher than the 2-signal IMD value, but we don't know the value, there is not test for it.

Yet we listen to music under that condition, and typically much more complex than that condition (ie more complex than a single piano key being played).

Two sine waves of equal magnitude vs. the fundamental and typically lower amplitude harmonics of musical notes from specific instruments will not result in IM products of equal magnitude. Most harmonics are but a small fraction of the amplitude of the fundamental, and any IM products resulting would be but a tiny fraction of that.

Here is middle c on a steinway piano:

 

[Irvrobinson]

Why, now that you ask, I do. Do you?

Assuming your answer would be "no", I'll rephrase:

When an amplifier is asked to play even a simple musical note, say the middle C key of a piano, you are asking the amp to reproduce about a dozen notes that comprise the fundamental plus harmonics inherent in the sound of a piano (It's those harmonics that make a piano and a guitar sound different when playing the same fundamental).

Yet we take an amplifier that tests well for THD, then re-test with 2 simultaneous sine wave signals and look for intermodulation distortion. Typically the IMD is higher than the THD at the same output level/load impedance.

Well, where is your intermodulation test for a dozen simultaneous sine wave signals that represent one key played on one instrument, the piano? And would the resulting distortion % be the same as your two signal IMD test? We can only guess, because we cannot test that condition; its too complex for our test gear. A normally skeptical person would assume higher distortion than the THD value and higher than the 2-signal IMD value, but we don't know the value, there is not test for it.

Yet we listen to music under that condition, and typically much more complex than that condition (ie more complex than a single piano key being played).

This is what I thought you were talking about, but I'm still not sure you really know what you're talking about.

The big picture you want has been available for years. See this link:

http://www.libinst.com/Audio DiffMaker.htm

 

[gene]

The purpose of any Audio (or A/V gear) is to reproduce stored data representing a musical performance via an analog interface ending at our brain's interpretation of sound. So the ultimate judge of success must be "how does it sound".

We can attempt to correlate a given conclusion about sound quality with measurements. But if there is a conflict, by definition, the measurement must be wrong somehow. It's also worth mentioning that the most sophisticated audio measurement cannot equal the operating conditions of even simple real-time musical conditions.

With regard to accuracy, the brain not only *can* be fooled, in the process of "hearing" the sub-conscious actively seeks to fool the conscious. So it is not absolutely necessary for accuracy when a subjectively preferential inaccuracy is present. This complicates correlation of measured criteria to sound quality.

You make good points and so far nobody has successfully been able to characterize human preferences with distortion to any degree of certainty. However one point I'd like to make, if an amplifier does well distortion wise with a single tone FFT and sine-sweep, it will also do well with IMD too. This is why I rarely test them separately. You can find misbehavior quite easily with traditional testing.

Audible preferences are a bit more understood with amplifier distortion tests than loudspeaker distortion tests however. We still have a long way to go.

 

[Ampdog]

The title is Art vs. Science of Audio REproduction. I then also presume that some exactitude is part of the theme. Thus, one way of determining this would be to determine whether there is any difference between the input and output of a system. Has anything been added/removed from the original, whatever that was. (Whether we know what the original was becomes immaterial.)

Further, when using the term "Audio", one presumes that such differences must be audible to denote an applicable difference. Scientifically this has become easy given modern analysing capabilities. Sadly hearing is a very inaccurate tool to measure this with; hearing characteristics have by and large been quantified well enough over decades, to enable measurements (the right ones!) to give an outcome one way or the other.

Art is often confused with whether the listener (trained or untrained) likes what he hears. His verdict will be the more accurate the more experience he has of the live product, but I doubt whether that can eliminate taste.

To the camps of 'subjectivists' and 'objectivists' I would like to add 'realists'. I would like to believe I belong there.

Bottom line: To me there is no room for art in "REproduction" per the above view-point.

 

[Tao1]

For now I am leaning more towards team science because I am still learning the ins and outs of audio. Even though good technical reviews of products are still a bit lacking in this day and age, I wish we had what we have now 15 years ago when I bought my Bose on ear headphones. $250Can for those, and I felt buyers remorse right away, since they didn't sound that spectacular. Unfortunately I didn't have much to go on back then.

Anyway, now that I have access to the science part, I will be able to compare different products I come across over the years, and then be able to reintegrate the 'art' portion of the equation once I have some reference of where the science lays.

 

[KEW]

At $3500 for a pair of 8' cables, these seem like a pretty good deal for anyone who believes science has no place in audio!

Early on, I concluded that in many respects (perhaps in most), the HT Armour cables rivaled the very best wires that I’ve experienced, including in-house favorites like Synergistic Research’s latest Atmosphere and Wireworld’s Platinum Eclipse Series 7, the latter significantly more expensive. In this face-off, the HT narrowly missed the top rung, but only in a couple of areas. It came up a bit short trying to match Synergistic Research’s 3-D imaging and soundstage holography. In the lowest octaves, Wireworld Eclipse produced a more satisfying solo cello from artist Pieter Wispelwey, with more contrast and bloom, a broader image, and a greater sense of the musician aggressively laying into the strings of his instrument (plus a deeper well of resonance rising up from the entire doublebass/cello section).

The Harmonic Technology Armour cables are statement products, pure and simple; they join the ranks of some of the most elite wires I’ve heard. While every set of cables has a unique signature, I have found that at these exalted levels (and depending on associated equipment) the margins of sonic differences begin to narrow. As I said at the outset, HT’s latest offerings are battle-ready. They enter the fray with all the virtues critical to the attainment of a great high-end system. Highly recommended.

http://www.theabsolutesound.com/articles/harmonic-technology-pro-7-reference-armour-loudspeaker-cable-and-armour-link-iii-interconnects/

 

[Tao1]

At $3500 for a pair of 8' cables, these seem like a pretty good deal for anyone who believes science has no place in audio!



http://www.theabsolutesound.com/articles/harmonic-technology-pro-7-reference-armour-loudspeaker-cable-and-armour-link-iii-interconnects/

The trick to selling snake oil is you keep the price at a place where it is not economically viable to be sued for misrepresentation, let alone create a fraud case that is worth the attention of law enforcement.

 

[killdozzer]

I think most of you are going the wrong way about this. First of all, not only that personal opinion is under no threat, it is being promoted as the measure of everything. So you shouldn’t worry yourself with hurting someone’s feelings or shutting someone up with testing. I mean, you didn't do it so far.

Personal opinions, emotional responses and, unfortunately, snake oil are always going to be there. So you shouldn’t preoccupy yourself with those. Whenever in doubt look at your slogan above and remember most of us came here to escape mysticism, snake oil or plain old selling snow to Eskimos.

It is not about art and science. Not at all. The worst thing you can do is name “snake oil vendors” artists. This would render you clueless on the topic of art. Con is no art.

Let’s deal with the first one: you said you can hear something instruments are not measuring. OK, build an instrument that can measure it. Or simply ask the producer to provide such an instrument. Remember, quantifying a speaker is not a question whether there is such a thing as human soul. There is no magic or alchemy; producers are building equipment not by spell, but by means of technology. So, perhaps there is something in there that you don’t understand how it came to be, but rather than selling it as mystical, find a way to detect what it is and by doing so further improve the overall production.

Second one is how the "unknown" being used. There is no other reason for someone to run away from testing (especially when technology is in question) than hoping he could earn more by selling YOU YOUR OWN EMOTIONAL RESPONSE back as though it is his own product. It is a placebo being used for the sole purpose of taking more money from you. So we’re NOT talking about two options that are simply out there for everyone to choose and they are not the same. One of those is fundamentally unethical. It is plain wrong and thus couldn’t be further from art. You seem to forget, art is not a mystical property of a conductor or transistor, art is the actual song coming out of your speaker.

Ask yourself; when you’re reviewing your own emotional response to a speaker, what are you really doing?? Who is it for and what use is that to anyone? I can’t show you a picture of my dog and demand or expect you to have the same emotional response to it. So the most useless thing by far is saying “I had an emotional response to something so it’s good”. Your emotional response is not withheld inside of the box you’re buying!! Why would you agree or let someone charge you for it? For Christ’s sake! You can have an emotional response and a placebo effect EVEN AFTER THE TESTING IS DONE. Also, you can have a large portion of a possible placebo even if the price is low.

Now, with the “big price placebo” the whole topic goes down the drain. This is simply below human. It is paying for it. It is a prostitution of your senses. And it is the “fool me”, “lie to me”, “make me believe into something that’s not there” type of lifestyle. Now this cannot be promoted and is harmful to everyone. It is not something you can decide for yourself. This should never be supported. Tricking and coning cannot become a standard. And it couldn’t be further away from art.


In the end: one option is saying; look you are a human being and of course you will have your emotional responses, you will experience your music differently every day, but this is no reason to let someone con you into giving him money he doesn’t deserve. Giving him money for something that’s in you and a part of you. And if it’s not in you let’s find a way to measure it.

The other option is saying: look I gave 29,000.00$ for a cable and it is better than all the rest!! I can’t admit I’m an idiot and I have been fooled, I have my ego to think of. If you want me to take a DBT the awful truth might come to light and I would be faced with an undisputable fact that I’m an idiot, and this is no option for me, so, “the man” is not selling snake oil there must be something special about his cables.

The latter is disgusting.

 

[PENG]

Why, now that you ask, I do. Do you?

Assuming your answer would be "no", I'll rephrase:

When an amplifier is asked to play even a simple musical note, say the middle C key of a piano, you are asking the amp to reproduce about a dozen notes that comprise the fundamental plus harmonics inherent in the sound of a piano (It's those harmonics that make a piano and a guitar sound different when playing the same fundamental).

Yet we take an amplifier that tests well for THD, then re-test with 2 simultaneous sine wave signals and look for intermodulation distortion. Typically the IMD is higher than the THD at the same output level/load impedance.

Well, where is your intermodulation test for a dozen simultaneous sine wave signals that represent one key played on one instrument, the piano? And would the resulting distortion % be the same as your two signal IMD test? We can only guess, because we cannot test that condition; its too complex for our test gear. A normally skeptical person would assume higher distortion than the THD value and higher than the 2-signal IMD value, but we don't know the value, there is not test for it.

Yet we listen to music under that condition, and typically much more complex than that condition (ie more complex than a single piano key being played).

I won't argue your point about distortions, but then you need to consider audibility, by us human. Also, I assume you would that all good audio equipment design engineers should aim for high fidelity, not to create some special brand sound by adding intentional distortions of any sort. Given today's technological know how that is not a problem at all, this is not even considered "high tech".

Since you own a Mc gear albeit a tuner, you probably know who Roger Russel is and here's some interesting article he wrote:

http://www.roger-russell.com/truth/truth.htm#goodamplifiers

and more about audio distortions related topics:

http://www.roger-russell.com/truth/truth.htm

 

[fmw]

Sorry but I don't see anything in audio reproduction that can be termed art. Technique certainly. Craft OK. Science no doubt. But art? No. The art comes from those who create and perform the music.

 

[panzeroceania]

I think limiting yourself by listening tests and double blind scenarios is very risky. I'd rather have a speaker tested beyond the limits of human hearing by repeatable, measurable tests, then to have someone tell me if they think it sounds good or not. human ears and eyes compensate for what they perceive and aren't great instruments of measurement.

That being said, I do think there are aspects of audio science that we don't fully understand yet, or don't fully measure, or that we think is "good enough" but we haven't bothered to measure in full accuracy.

In that case I would rather create better tests and instruments.

The problem I often hear is it is "engineers that make things good enough" vs "incompetent snake oil salesmen"

I don't know about you, but neither one of those is an acceptable option for me. I'd rather have a speaker engineered well beyond my ability to appreciate it and tested to be guaranteed, that way it will perform as it should in various installation environments and sound correct to even the most discerning ear I may have as a guest.