In Defense of Audiophiles, Bose, Pass, Toole and Science

I don’t know why I look at Audio Science Reviews equipment reviews, they usually make me bang my head against my desk. The claims they make of being scientific is pretty half-baked. Don’t get me wrong, I appreciate measurements, and the time it takes to conduct them, along with insights into the causes, but judging all electronics based on 40+ year old measurements which have not really become closer to explaining human perception and enjoyment, they claim to be objective scientists. They are not. Let me tell you some of the people who are:

  1. Bose
  2. Harman
  3. Nelson Pass
  4. Floyd Toole

This may look like a weird list, but here is what all these have in common: They strive to link together human perception and enjoyment of a product to measurements. Each have taken a decidedly different, but very successful approach. They’ve each asked the question differently. I don’t always agree with the resulting products, but I can’t deny that their approach is market based and scientific.

Floyd Toole’s writing on room tuning, frequency response and EQ combines exact measurements with human perception, and as big a scientist as he is he remains skeptical of measurements, and with good reasons.

The process Nelson Pass uses is exactly right. His hypothesis is that a certain type of distortion, along with other important qualities, are what make for a great sounding amp, and lets face it, the process, and his effectiveness cannot be denied as not being scientific or financially successful. Far more scientific than designing or buying an amp based on THD% at 1 watt alone.

Bose is also very very scientific, but they come at the problem differently. Their question is: What is the least expensive to manufacture product we can make given what most consumers actually want to hear?" Does it work? They have 8,000 employees and approximately $4B in sales per Forbes:

Honestly, I don’t know how your average Bose product would measure, but you don’t get to these numbers without science. Assuming they measure poorly, doesn’t that mean measurements are all wrong?

The work Harman has done in getting listening panels together, and trying out different prototypes while adhering to previous science is also noteworthy. Most notably and recently with their testing of speaker dispersion which has resulted in the tweeter wave guides in the latest Revel speakers. They move science forward with each experiment, and then put that out into their products.

Regardless of the camp you fall into, crusty old measurements, perception measurements or individual iconoclast, we also must account for person to person variability. It’s been shown for instance that most people have poor sensitivity to phase shifts in speakers (like me), but if you are THAT person who has severe sensitivity to it, then all those studies don’t mean a thing.

My point is, let’s not define science as being purely in the domain of an oscilloscope. Science is defined by those who push the boundaries forward, and add to our understanding of human perception as well as electron behavior through a semi-conductor and air pressure in a room. If it’s frozen in 40 year old measurements, it’s not science, it's the worship of a dead icon.



Good analogy whoopycat.   We should continue to investigate the nature of audio reproduction, perception, and appreciation.  Certainly, at this point in time all of the meaningful measures have yet to be defined and curiosity about what is missing is a primary driver for scientific investigation.  The bottom line, to me at least, is that we can draw some conclusions from the measurements we can make, but we certainly can't completely describe the behavior of any piece of gear nor how we will perceive it.  At this point, the quote below certainly obtains.

"Not everything that counts can counted and not everything that can be counted counts." (commonly attributed to Albert Einstein, but probably coined by William Bruce Cameron).


I'm sorry but I don't see any correlation between studying random seasonal basketball stats and studying precise electrical measurements.

None whatsoever.
cd, I didn't say anybody should study basketball stats, but I tell you what, let's trade.  You send me the precise electrical measurements for midrange transparency, and I'll send you the NBA players who were the best this year at defending the pick & roll. 

Oh heck, I know you're not interested, but here they are anyway:!?sort=SCORE_POSS_PCT&dir=-1&SeasonType=Regular%20Season&TypeGrouping=defensive

The point of my analogy is if you were to ask for those stats 20 years ago you would've just gotten laughed at.  Now they are at anyone's fingertips.  Meanwhile in home audio, we're still looking at the same old measurements of frequency response and distortion and saying this is what something sounds like.  As Erik said, that's not science.  Precise or otherwise, that's the equivalent of heading to a sports analytics conference with a pack of baseball cards.


I understand the emergence of sporting stats. Modern day football (soccer) is now riddled with terms like assists, blocks, tackles etc which were not deemed necessary some 20/30 years ago.

I still believe there’s only a dubious relationship between sporting stats and sporting success, (as it is with any other form of social science). One that you can only attempt to establish if you highlight the ones which add credence to your argument whilst ignoring the ones that don’t. Sport is a one off event with many, many variables. You’d need a computer like the one Douglas Adams wrote about in The Hitchhiker’s Guide to the Galaxy.
["You know nothing of future time," pronounced Deep Thought, "and yet in my teeming circuitry I can navigate the
infinite delta streams of future probability and see that there must one day come a computer whose merest
operational parameters I am not worthy to calculate, but which it will be my fate eventually to design."]

However, loudspeaker testing and analysis has come a very long way with the advent of computer software and can be deemed a true science. Nowadays everyone uses this software and testing is as rigourous as its ever been. They know what to look for: eg

• On-axis frequency response
• Impulse response
• Cumulative spectral decay
• Polar response
• Step response
• Impedance
• Efficiency/Sensitivity
• Distortion
• Dynamics

Heck, there’s even at least one website devoted to this kind of thing.

The true test is whether the loudspeakers are actually getting better. By and large, I’d say they were. But I acknowledge there will always remain a subjective element.

Some will always prefer cassette tape to CD, a 1960 Ferrari to one from 2020, a Technics turntable from 1970 to one from now etc. That’s human nature.

Fair enough, but let’s not kid ourselves about which ones measure better.
However, loudspeaker testing and analysis has come a very long way with the advent of computer software and can be deemed a true science.

My point was never that science does not occur. My point was that judging products by an oscilloscope alone, when you have not done the work of measuring user preference, or listening for yourself, is _not_ science. It’s a mechanical process of measurement, or quality assurance at best.

When you are designing a piece of equipment to go in a car, air conditioner or computer, yes, that's all you need. Set the parameters, set your tolerances, and attempt to achieve those values for the least amount of money.

Judging a product for human consumption on that basis alone is where the term science is misapplied, unless you have previously worked to establish those were desirable.