The Fallacy of Accuracy

[amirm]

I am aware that the person who presents these opinions designed the Nautilus loudspeaker, that I have listened playing great in a four amplifier active system that is still one of my audiophile references. I am aware that he has a long CV creating speakers I have listened and can listen anytime. I am aware of tens of opinions of others in the forum, magazines and real life about his speakers. I am aware that this is just an example and that I have read from many high-end designers sharing similar thoughts. I am aware that other members are not fools and can interpret facts and opinions without the need of constant proof or your blessing .

And nothing in that CV included published research in which you quoted his opinion. There have been formal methods and research in determining user preference. You can ignore that but don't pass it as some kind of blessing and appeal to authority in that manner.

Concerning your last sentence, I am also aware that IMHO your arrogant and impolite style is not suited to a proper debate in audio. Do not expect me to go in your usual insult fights.

I did not address you personally whatsoever whereas you are doing that with me. Take your own advice and stay on the technical topic. I won't caution you again.

 

[amirm]

There is no choice. We have preferences at live performances. I have played bad sounding violins, good sounding violins and great sounding violins. I've used bad sounding bows, good sounding bows and great sounding bows. I've played in bad sounding auditoriums, good sounding halls and a brilliant sounding concert hall. And all of that is before there is any sound preproduction going on. It's live music!

Creation of art is not my forte so I can't comment on that. If you tell me that 10 great violinists will pick 10 different violins as the best, I believe you . Sound reproduction is my domain and there, the research has yielded significant insight into user preferences in loudspeakers and it is not remotely chaotic.

In the OP, Paul Barton of PSB says this:

"When I design a new product I apply the experience I've had measuring and then interpreting the measurements in terms of natural reproduction. Once the new design meets or exceeds the performance expectations by measurements, then the "listening step" fine-tunes the final speaker sound. There is no substitute for the "listening step"—maybe someday a computer can do the listening, but I'm not counting on this in my lifetime. Besides, I love to listen to music; it's a huge perk of my job."

Paul participated in the same research at NRC and absolutely follows what was learned there with regards to what measurements will result in high likelihood of listener preference.

Andrew Jones states the same:

"From a technical-measurement point of view, we mostly think we want a flat on-axis response measured anechoically, with smooth off-axis response, although even this concept has been contested."

That is why he can design a $500 speaker and still have it sound good.

Now, speaker *business* is not about the best performance. It is to survive and thrive in a highly fragmented and congested market. As a result, what you do as far as brand, marketing, distribution, etc. almost entirely trumps fidelity. So we have a lot of flavors, many of which are broken but still sell because these other factors are in their favors.

 

[JackD201]

Amar Bose had tons of research. How about them apples?

 

[amirm]

Amar Bose had tons of research. How about them apples?

From the Wiki:

• Fellow, IEEE, 1972 - for contributions to loudspeaker design, two-state amplifier-modulators, and nonlinear systems.
• Honorary member, Audio Engineering Society, 1985.
• The 2010 IEEE/RSE Wolfson James Clerk Maxwell Award, for "outstanding contributions to consumer electronics in sound reproduction, industrial leadership, and engineering education".[18]
• In 2011, he was listed at #9 on the MIT150 list of the top 150 innovators and ideas from MIT.
• Beryllium Lifetime Achievement Award, Association of Loudspeaker Manufacturing & Acoustics International, 2014.
• Founders Award at The Asian Awards 2015.

His net worth was as high as 1.8 billion dollars. He was a rare combination of researcher, inventor and master marketer.

Seems like pretty delicious apples to me.

 

[amirm]

With speakers I have two rules:

1. How likely is that if someone bought them sight unseen, they would like them?

2. How variable is their performance based on room and content being played?

Design methodologies that incorporate the answer to these questions are the winners in my book.

 

[Rodney Gold]

It's all up to the room/speaker interface and user preference
You cannot design for an unknown room and an unknown listeners preference so you as designer design for flat etc in an anechoic chamber ..and let the chips fall where they may.

Luckily , with modern DSP and a bit of room treatment , we can shape "accuracy" or system sound as to how you want it to be
I can make my system sound profoundly different with PEQ/DIRAC/Subs etc
We all do it , simple upgrading is making the system sound different..why upgrade anything if it doesnt change the sonics

 

[JackD201]

From the Wiki:

• Fellow, IEEE, 1972 - for contributions to loudspeaker design, two-state amplifier-modulators, and nonlinear systems.
• Honorary member, Audio Engineering Society, 1985.
• The 2010 IEEE/RSE Wolfson James Clerk Maxwell Award, for "outstanding contributions to consumer electronics in sound reproduction, industrial leadership, and engineering education".[18]
• In 2011, he was listed at #9 on the MIT150 list of the top 150 innovators and ideas from MIT.
• Beryllium Lifetime Achievement Award, Association of Loudspeaker Manufacturing & Acoustics International, 2014.
• Founders Award at The Asian Awards 2015.

His net worth was as high as 1.8 billion dollars. He was a rare combination of researcher, inventor and master marketer.

Seems like pretty delicious apples to me.

Indeed. The proceeds from his far from flat and quite room dependent model funded quite a few things we take for granted. Turned the mass market audio paradigm on its head too with those tiny sub/sats. What is curious is that another famous and well respected designer with probably even more research into the hearing mechanism under his belt, Dr. Geddes, takes a very different approach from Mr. Bose. Dr. Toole and Sean also differ in approach from Earl championing wider comparative dispersion. So who are right? I think they all are quite frankly. They just might have different performance targets.

While there is a baselines out there for the human population, it is possible that these may still need refinement. Stratification even. A challenge faced by the researchers is that while people generally hear and process sound in the same way biologically, the manner in which people listen do differ so we have sets, subsets and subsets within subsets. Tools for different jobs so to speak. If this is indeed the case then accuracy becomes just one of many boundaries for acceptable utility and not necessarily the ultimate goal each and every time. Add external cultural influences to the list of challenges. Tastes change over time in both individuals and populations.

I don't see a convergence anytime in the near future that is unless the machines take over and we're born into and are plugged into the Matrix. LOL. We may hear the same way but I'm pretty sure we don't all listen for the same things. It's one of the beauties of music after all. You can key in to something different every time you listen to the same song. The gear is just there to deliver the song in a way that you'll actually enjoy enough to keep on going.

 

[PeterA]

Down to earth short read...bravo.

Bob, do you mean the length of the OP or the length of the article?

 

[NorthStar]

Peter I don't have the kind of humor in reference to Al our OP. My comment was in reference to the post's content, the short article by Steve Guttenberg.
...A well respected audio reviewer/writer and much much more. And same for Al, highly always great to read.

I'm surprised that you asked.
___________

* I'll add this: I read occasionally comments referring to a poster's remarks as insulting. Nobody ever here is saying such remarks; humor yes, but never insults, never ever. And that too, when occasionally I am reading, it baffled me that such thought can cross some people's mind.
Yes, I am referring to Amir right now. ...Last guy on earth who would insult anyone.

We are not only highly respectable human beings, but also highly intelligent, sociable and humorous and serious @ the right times. ...All of us here.
...Over 5,000 members, including Mike (Blizzard) who should be reinstated.
And that, is also related to accuracy, human accuracy.

In audio as in humanity among ourselves, our friends, our families, our societies ? http://www.thefreedictionary.com/fallacy

 

[microstrip]

Indeed. The proceeds from his far from flat and quite room dependent model funded quite a few things we take for granted. Turned the mass market audio paradigm on its head too with those tiny sub/sats. What is curious is that another famous and well respected designer with probably even more research into the hearing mechanism under his belt, Dr. Geddes, takes a very different approach from Mr. Bose. Dr. Toole and Sean also differ in approach from Earl championing wider comparative dispersion. So who are right? I think they all are quite frankly. They just might have different performance targets.

While there is a baselines out there for the human population, it is possible that these may still need refinement. Stratification even. A challenge faced by the researchers is that while people generally hear and process sound in the same way biologically, the manner in which people listen do differ so we have sets, subsets and subsets within subsets. Tools for different jobs so to speak. If this is indeed the case then accuracy becomes just one of many boundaries for acceptable utility and not necessarily the ultimate goal each and every time. Add external cultural influences to the list of challenges. Tastes change over time in both individuals and populations.

I don't see a convergence anytime in the near future that is unless the machines take over and we're born into and are plugged into the Matrix. LOL. We may hear the same way but I'm pretty sure we don't all listen for the same things. It's one of the beauties of music after all. You can key in to something different every time you listen to the same song. The gear is just there to deliver the song in a way that you'll actually enjoy enough to keep on going.

Well said, Jack.

 

[amirm]

While there is a baselines out there for the human population, it is possible that these may still need refinement. Stratification even. A challenge faced by the researchers is that while people generally hear and process sound in the same way biologically, the manner in which people listen do differ so we have sets, subsets and subsets within subsets. Tools for different jobs so to speak. If this is indeed the case then accuracy becomes just one of many boundaries for acceptable utility and not necessarily the ultimate goal each and every time. Add external cultural influences to the list of challenges. Tastes change over time in both individuals and populations.

I hear what you are saying Jack. But these are ultimately words in dire need of some kind of verification. Last time I was at Harman, Dr. Olive talked about a story of a major Japanese car manufacturer saying that the Japanese have different taste in sound and therefore the Harman car audio product would not be to their liking. Dr. Olive made binaural recordings of the one car system and tested it in Japan, china and US. All the results agreed with each other. Contrary to popular belief, good sound was good sound regardless of what genes you have, or what upbringing you have.

Now I grew up with the same assumption of Japanese having different preference. "It just made sense." A lot of things in audiophile world "make sense to us." So we say them to each other with confidence. But that doesn't make them true. Harman went on to win the multi-billion dollar business from that Japanese company for their car infotainment business despite their system costing more than their competitors which were Japanese car audio companies.

So personally I have stopped believing what I read or makes sense to my lay belly. There is research that guides us and speaker manufacturers in regards to preference in speaker sound. The research is highly compelling. It was validated for me when I sat there and hear a group of high-end dealers take the same test and their outcome agreeing with the research.

Now there is still room for variation even in this mode. Speaker designers still listen. Their marketing departments still chime in for changes, etc. But for the 80%, there is a formula and research that guides us.

I don't see a convergence anytime in the near future that is unless the machines take over and we're born into and are plugged into the Matrix. LOL. We may hear the same way but I'm pretty sure we don't all listen for the same things. It's one of the beauties of music after all. You can key in to something different every time you listen to the same song. The gear is just there to deliver the song in a way that you'll actually enjoy enough to keep on going.

As I say routinely what we deny here, is routinely accepted by speaker designers as solid research and what many of them follow. I spoke to Paul Barton quoted in the article and that is what he does. I quoted Andrew Jones. Now, not every company is religious about that. But the research is out there and is quite strong. It says that if you make a well behaved speaker, people like it better. A great coincident of excellence in engineering resulting in our preferences!

As to Amar Bose, as I said, he is an example of how important business is to speaker manufacturing. I didn't put him forward as an example of this type of research.

 

[NorthStar]

Indeed. The proceeds from his far from flat and quite room dependent model funded quite a few things we take for granted. Turned the mass market audio paradigm on its head too with those tiny sub/sats. What is curious is that another famous and well respected designer with probably even more research into the hearing mechanism under his belt, Dr. Geddes, takes a very different approach from Mr. Bose. Dr. Toole and Sean also differ in approach from Earl championing wider comparative dispersion. So who are right? I think they all are quite frankly. They just might have different performance targets.

While there is a baselines out there for the human population, it is possible that these may still need refinement. Stratification even. A challenge faced by the researchers is that while people generally hear and process sound in the same way biologically, the manner in which people listen do differ so we have sets, subsets and subsets within subsets. Tools for different jobs so to speak. If this is indeed the case then accuracy becomes just one of many boundaries for acceptable utility and not necessarily the ultimate goal each and every time. Add external cultural influences to the list of challenges. Tastes change over time in both individuals and populations.

I don't see a convergence anytime in the near future that is unless the machines take over and we're born into and are plugged into the Matrix. LOL. We may hear the same way but I'm pretty sure we don't all listen for the same things. It's one of the beauties of music after all. You can key in to something different every time you listen to the same song. The gear is just there to deliver the song in a way that you'll actually enjoy enough to keep on going.

I have just read your post Jack, like right now. And I like it a lot; very good common logical sense.
And if I may add; my ears today are not the same as ten years ago. ...And neither all my audio gear. ...Speakers are the same though.
But the music sounds different, definitely, because of those ears. They don't sell those @ audio stores or amazon. They don't sell time either (no ears, no years).
So yeah, your above quote is ... right on.

 

[thedudeabides]

Agreed Bob. Jack's post is spot on IMHO.

One thing that has always interested me is the divergent path well known speaker manufacturers take regarding cabinet resonance.

Take Magico, YG and others who go to heroic lengths to minimize / eliminate.

And then you have a company like Harbeth, or other British speaker manufacturers with an entirely different view on this issue.

 

[DaveC]

Amir is 100% right but as usual oversimplified thinking on the topic... Loudspeaker design is juggling compromises, simple as that, and the old saying "there's more than one way to skin a cat" definitely applies.

For example take a list of well regarded speaker designers, say Linkwitz, Geddes, Danley, Sanders, A Jones, Toole... these guys are all skinning the cat in a different way and ending up with a result that closely matches the results of the preference testing done by Toole/Olive despite the fact the designs are sometimes VERY different from one another.

And to discuss this without also discussing the room and associated treatments/acoustics is meaningless. Some speaker designs work better in different acoustic environments, i.e. a dipole shouldn't be slammed into a corner while some speakers are designed ONLY to be placed in corners as part of their design. Wide dispersion direct radiator speakers require different treatments vs a directional designs like GedLee speakers.

So given the inherently compromised exercise of speaker design and their suitability for different acoustic environments, ALL of which may adhere to preference testing research, it's not surprising to me there are such a large variety of designs. And despite the fact that all these well-designed speakers perform well according to the preference testing people still have much different preferences within that envelope. What does preference testing say about paper vs aluminum coned woofers? Diamond vs beryllium tweeters? And on and on... all these things make a massive difference in the overall sound of the speaker while not effecting frequency response or dispersion, or any factors that might change it's ranking in how closely it conform to known preference testing.

So what might look like "chaos" to some is really not chaotic AT ALL when you understand the design intent and chosen compromises made by the designer. It is true there are many speakers that do not follow the preference testing, but even if you completely ignored these more compromised designs you'd STILL have a huge variety of different designs that do conform, and they get there in their own way and they all have advantages and disadvantages.

So no, there isn't going to be any "convergence" in speaker design anytime soon, but if there was going to be one it would be for a point source horn because that's the most ideal design.

 

[microstrip]

(...) So no, there isn't going to be any "convergence" in speaker design anytime soon, but if there was going to be one it would be for a point source horn because that's the most ideal design.

Dave,
We already had a point source panel speaker - the QUAD ESL63 electrostatic! If it wasn't for dynamics limitations it would still be my first choice.

 

[KeithR]

And to discuss this without also discussing the room and associated treatments/acoustics is meaningless. Some speaker designs work better in different acoustic environments, i.e. a dipole shouldn't be slammed into a corner while some speakers are designed ONLY to be placed in corners as part of their design. Wide dispersion direct radiator speakers require different treatments vs a directional designs like GedLee speakers.

So given the inherently compromised exercise of speaker design and their suitability for different acoustic environments, ALL of which may adhere to preference testing research, it's not surprising to me there are such a large variety of designs. And despite the fact that all these well-designed speakers perform well according to the preference testing people still have much different preferences within that envelope. What does preference testing say about paper vs aluminum coned woofers? Diamond vs beryllium tweeters? And on and on... all these things make a massive difference in the overall sound of the speaker while not effecting frequency response or dispersion, or any factors that might change it's ranking in how closely it conform to known preference testing.

Don't forget amplifiers either - many speaker designers make efficiency a high priority to use with simple, lower power amplifiers.

 

[JackD201]

I hear what you are saying Jack. But these are ultimately words in dire need of some kind of verification. Last time I was at Harman, Dr. Olive talked about a story of a major Japanese car manufacturer saying that the Japanese have different taste in sound and therefore the Harman car audio product would not be to their liking. Dr. Olive made binaural recordings of the one car system and tested it in Japan, china and US. All the results agreed with each other. Contrary to popular belief, good sound was good sound regardless of what genes you have, or what upbringing you have.

Now I grew up with the same assumption of Japanese having different preference. "It just made sense." A lot of things in audiophile world "make sense to us." So we say them to each other with confidence. But that doesn't make them true. Harman went on to win the multi-billion dollar business from that Japanese company for their car infotainment business despite their system costing more than their competitors which were Japanese car audio companies.

So personally I have stopped believing what I read or makes sense to my lay belly. There is research that guides us and speaker manufacturers in regards to preference in speaker sound. The research is highly compelling. It was validated for me when I sat there and hear a group of high-end dealers take the same test and their outcome agreeing with the research.

Now there is still room for variation even in this mode. Speaker designers still listen. Their marketing departments still chime in for changes, etc. But for the 80%, there is a formula and research that guides us.


As I say routinely what we deny here, is routinely accepted by speaker designers as solid research and what many of them follow. I spoke to Paul Barton quoted in the article and that is what he does. I quoted Andrew Jones. Now, not every company is religious about that. But the research is out there and is quite strong. It says that if you make a well behaved speaker, people like it better. A great coincident of excellence in engineering resulting in our preferences!

As to Amar Bose, as I said, he is an example of how important business is to speaker manufacturing. I didn't put him forward as an example of this type of research.

As far as cars go you're preaching to the choir. IMO Logic 7 (Harman) equipped cars are my favorite followed by Ford's in-house car audio group (Geddes).

As far as culture goes, and I did have the discussion with Sean himself right here in WBF's early days, it wasn't ethnicity but rather architecture. Of course architecture does follow cultural trends and we can trace some causality when it comes to voicing. There is a weaker indication but present nonetheless with speaker size and radiation patterns. Some things are so black and white that I don't think much study is needed. For example a big horn will work better than a free space full range speaker in a small room. The Japan effect. The thing is "well behaved" can and does take many forms and that is actually my point. The first question is always what will I use the speakers for and the rest follows. Granted that a good speaker should be consistent regardless of content, it's specs are only part of the overall list of consideration. Size, placement, required coverage, all stuff you deal with more than I do since you do balls to the wall theater systems. So for me it is a utility thing. If someone wants to do large scale at convincing levels then one has to abide by the laws of physics. That said, I always say, you can't cheat physics but you surely can use physics to cheat. You can treat your room to behave a certain way when using wide dispersion speakers or go with controlled directivity with need for less treatment at least from the midband up. Both speakers can be as you said "well behaved" but they will differ.

Don't get me wrong. I do think research is vital. What I don't fully agree with is that as Bruce's signature says, a blind squirrel getting the nut sometimes. Sure it is inefficient stumbling around by trial and error but it is still well within the realm of possibility. My opinion of the Bose sound is that while it lacks a lot it doesn't go out of its way to suck either. Sins of omission really. Perfectly fine for muzak which is how most people employ music anyway. Put these same people in a situation where they are made to focus and suddenly they aren't so great. That is certainly not surprising but it can be argued that in regular use, they are just fine. That's what Dr. Bose's research showed and he ended up with a billion dollar formula. He catered to the broader market. Now we audiophiles are a funny bunch. Every small difference down to the limits of audibility is given so much more weight but really that makes us no different from any other hobbyists be they bonzai guys obsessing over clippers, aquarists over gravel or shooters over springs. Hobbyists are by nature looking at things beyond the baseline and forming personal standards is at the center of any hobby. Diversification is inevitable in such a context IMO. The qualitative weighting differs from individual to individual beyond what is considered statistically as "sound" design. If in the Bose case, less is fine, on the other end of the spectrum there will always be people looking for more.

 

[slcaudiophile]

Before we get too far, we still have to build consensus on this inescapable truth from Andrew Jones:

"Accuracy in terms of closest approach to the original performance is not practical nor even possible. "

most accurate thing ive heard in a long long time.

 

[slcaudiophile]

Accuracy simply means lack of audible distortion being added by the playback chain - nothing more, nothing less. If two systems are accurate, using completely different styles of components, then they will sound identical, subjectively, on all recordings; yes, even using the infamous DBT!! And, that's exactly what I keep chasing ...

this doesn't make sense, at least to me. if 'accurate' means a lack of distortion, as you suggest, then there are lots of 'accurate' systems out there. lots of systems have little to no distortion (except tubes of course).

 

[rbbert]

this doesn't make sense, at least to me. if 'accurate' means a lack of distortion, as you suggest, then there are lots of 'accurate' systems out there. lots of systems have little to no distortion (except tubes of course).

I think by "lack of distortion" he means faithfulness to the original signal rather than simply measured harmonic and/or IM distortion. In other words, all kinds of "distortion" meaning any change to the original signal, whether easily measurable or not.