Measurements vs Listening

How much stock do you place in measurements & do they correlate to sound quality?

  • Completely. If I can't measure it, it's not there.

    Votes: 3 14.3%
  • I put a lot of stock in measurements, but they don't tell the whole story.

    Votes: 11 52.4%
  • I read the measurements, but I think there's things we can't measure.

    Votes: 6 28.6%
  • I rarely pay attention to measurements, just power output, etc.

    Votes: 1 4.8%
  • None. Measurements are irrelevant. The ear is the only judge.

    Votes: 0 0.0%

  • Total voters
    21
Rob Babcock

Rob Babcock

Moderator
How much credence do you put in measurements? Do you feel that a components sound & performance is entirely related to its measured specs? A recent discussion has got me curious just where the bulk of the AH crowd falls on the spectrum of objective measurements vs subjective listening.

I tried to keep the answers short, so maybe not everyone will see themselves in one of those responses. Please post an explanation of your opinions if you like.

This should be fun! :D
 
Last edited:
Mudcat

Mudcat

Senior Audioholic
Since I've been developing a method of evaluating speaker cables, I've become firmly entrenched in the measurements side. I feel that a company that sells a product for what many think is an outrageous amount of money should have a product that meets some minimum requirements. While those rminimum equirememts have not been determined as yet, we can compare them to minimal cost cables.

My evaluation method which Gene will be posting soon as part of the speaker cable face off (2), (which evaluates LCR parameters and cost against other cables tested), will show that very expensive cables don't have the measurememts to back up their cost. Though we are waiting, what will Gene's analysis will show.
 
Rob Babcock

Rob Babcock

Moderator
Hmmm....did you vote, Mudcat? This is the first poll I've run, and I'm not seeing a result.
 
gene

gene

Audioholics Master Chief
Administrator
Fortunately for cables, measurements do tell the whole story. For loudspeakers and other electronics it becomes far more involving. Making the right acoustical measurement of speakers and correlating that to sound is very tricky and often incomplete.
 
Rob Babcock

Rob Babcock

Moderator
I certainly don't mean just cables: I guess I'm really curious how the AH readership balances the objective view of audio as a science vs a view that it's an art. My gut feeling is that the guys here tend very strongly towards measurements over subjective listening, but we shall see.

BTW, I agree with respect to cables, Gene, and I agree that it's much more complex for other components. I haven't case my official vote yet- I'll let the others have their say before I start chiming in! :p
 
gene

gene

Audioholics Master Chief
Administrator
I guess I'm really curious how the AH readership balances the objective view of audio as a science vs a view that it's an art.
Rob, lets also not forget that not all measurements are truly objective. Often more the case than not, a reviewer (not from Audioholics of course ;) ) will misinterpret a measurement, or make it incorrectly (especially for speakers) and imply a certain characterstic for a product that may actually not be.

Measurements also need to be closely scrutinized as to their validity and interpretations. I for one value a subjective review on products to determine if the product has the feature set and/or looks interesting enough to consider and then take a look at the measurements to see how it performs within the boundaries of the tests.
 
Rip Van Woofer

Rip Van Woofer

Audioholic General
You know where I came down on that question!

Although I do concede that there is still a place for subjective evaluation when it comes to speakers. Not that measurements are incapable of telling nearly all where they are concerned, but that doing complete, valid measurements is very difficult, time-consuming, and beyond the means of all but a handful of laboratories (and probably techs). The Harman Int'l. facility described in the recent speaker articles is one.

And hey, some folks genuinely prefer a bit of boost or attenuation here and there in the frequency response. Perfecly legit even if it isn't my thang. But few (including me) have the training to say, "oh, I'll probably like the sound of that speaker because it has an attenuated response by 3dB between 2000Hz and 4000Hz and a 2dB/oct rolloff above 10kHz". So the ears still have a role there...at least for everyone who doesn't have their own anechoic chamber and calibrated test gear. (My wife keeps saying "no!")

Oh, and for the "is audio science or art" question: Science, period. In the olden days when hi-fi was aborning and technology and measurement were cruder, I'd say maybe there was some art in there as a matter of necessity, but not today. Ditto for the disciplines on which audio engineering is based, from the "hard" sciences like physics to the "soft" domains like psychoacoustics. Only the music we listen to is art. Audio, as Dr. Floyd Toole says, is science in the service of art.
 
Last edited:
Mudcat

Mudcat

Senior Audioholic
I know this is off topic, but Rips statement about art or science got me thinking about some amps I saw on eBay.

http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&category=39783&item=5714733526&rd=1


This one is just $10500 worth of ugly.


http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&category=39783&item=5714733553&rd=1

And this is, well you have to read the specs. 40 W/Ch, S/N or 82 dBa, and a whopping 1 - 3% THD. But it is Art (I mean it is either really pretty or someone is very adept at Adobe Photoshop 7).

Back to our regulary scheduled program.
 
gene

gene

Audioholics Master Chief
Administrator
82dB SNR is rated at 1watt so its actually not a bad spec. The distortion is high, but some people like the colored sound of tube distortion. Tubes disrtort much more gracefully than solid state amps. 40wpc is prob ok for high efficient speakers in a small to medium sized room, but not for hometheater. Its esoteric at its finest. It will probably really shine with battery powered cables ;)
 
WmAx

WmAx

Audioholic Samurai
gene said:
Fortunately for cables, measurements do tell the whole story. For loudspeakers and other electronics it becomes far more involving. Making the right acoustical measurement of speakers and correlating that to sound is very tricky and often incomplete.
The way a loudspeakers sounds is not due to any unmeasureable parameter. One only need to observe the delayed energy storage(relative to the source signal), frequency response(which is not as straighforward as may be typically considered), the polar response(frequency response at at all angles) in detail, distortion mechanisms in order to analyse how a speaker will 'sound'. The 'mystery' part that any people embrace, IMO, is coming from three things (1) the behaviour in the specific room it is used(this is another variable that may not be overlooked) (2) incomplete/flawed measurements (3) typical audiophile belief that we can hear things that can not be measured.

I notice you did not deny measurements, but wisely stated it was 'far more involving' to correctly measure and coorelate. :)

-Chris
 
Last edited:
C

cornelius

Full Audioholic
Regarding speaker measurement, I was always intrigued by the step response test. I think it's a very important and telling test of how a speaker will perform.
 
Rip Van Woofer

Rip Van Woofer

Audioholic General
I'm with ya totally, WmAx. I especially enjoyed point (3)! But in fact, in an article by Floyd Toole (based on a keynote speach he gave at an AES convention) that appeared in The Audio Critic, even a speaker's behavior in a typical room can be fairly well predicted given truly complete anechoic measurements (and presumably the application of some fancy math). Not 100%, but pretty good as I understand it.

Question: I'd always thought that energy storage problems would mainly manifest themselves as frequency response anomalies - so we're back to frequency response as the "ultimate" measurment (leaving aside distortion for now). Am I correct? And ditto for baffle step problems? (thanx Cornelius).
 
D

docferdie

Audioholic
I believe in measurements especially when listening to digital content. A CD represents music in a "measured form" so to speak so there really isn't anything subjective about its playback. Either the gear can represent the bitstream appropriately or it cant.
On the other hand amplifying and playing back live sound is a whole different ballgame. There may be attributes present in live sound that a certain piece of equipment can play back which is not necessarily known or yet unmeasurable by current technology--sort of how electrons and protons were the smallest components of matter until technology was developed that proved the existence of quarks.
 
Last edited:
WmAx

WmAx

Audioholic Samurai
Rip Van Woofer said:
I'm with ya totally, WmAx. I especially enjoyed point (3)! But in fact, in an article by Floyd Toole (based on a keynote speach he gave at an AES convention) that appeared in The Audio Critic, even a speaker's behavior in a typical room can be fairly well predicted given truly complete anechoic measurements (and presumably the application of some fancy math). Not 100%, but pretty good as I understand it.
You are absolutely correct. I am familiar with Dr. Toole. However, I meant that to apply to the understanging by reviewers and general listeners.

Question: I'd always thought that energy storage problems would mainly manifest themselves as frequency response anomalies
No. Their are two forms of stored energy (1) non-delayed resonance: manifested by the frequency response(the Q of fr corner slope, for exampe, will exibit different degrees of damping/enegy storate) or a frequency response irregularity. [and] (2) delayed resonances: resonances that can be hidden in some freqency response measurments, appearing as a linear frequency response(, but persist to decay non linearly relative to the surrounding spectrum. Such examples are the cabinet walls, which can harbor such resonances, the cabinet volume(if sufficient acoustic damping materials are not used and propogation time is sufficient relative to frequency of interest).

Delayed or non-delayed resonance will cause modifcations to timbre[1] if sufficiently present.

[1]

The Modifications of Timbre by Resonances: Perception
Floyd E. Toole, Sean E. Olive
JAES, Volume 36, No. 3, 1988, March

-Chris
 
Last edited:
Rip Van Woofer

Rip Van Woofer

Audioholic General
Excellent, thanks! So to bring it down to my level, scenario (1) does indeed screw up the frequency response, while scenario (2) doesn't, but is instead manifest by what I'll call "ringing" - the speaker and cabinet simply don't "shut up" when they're supposed to even though they still measure flat. I suppose that, subjectively described, it would "muddy" the sound and affect the timbre (add a coloration) much as an over-reverberant room does.

I admit that I am still a bit confused with my limited understanding to envision how any coloration of sound would not somehow show up in the frequency response measurements; but I am not about to argue with Mssrs. Toole and Olive and am inclined to take their (and your) word for it!

Thanks, therefore, for the AES preprint citation. Perhaps I'll order it and, with luck and study, understand the concept better.

Re: the pricey tube amps mentioned above: the cliches "more money than brains" and "style over substance" come naturally to mind!
 
Last edited:
Rob Babcock

Rob Babcock

Moderator
gene said:
Rob, lets also not forget that not all measurements are truly objective. Often more the case than not, a reviewer (not from Audioholics of course ;) ) will misinterpret a measurement, or make it incorrectly (especially for speakers) and imply a certain characterstic for a product that may actually not be.

Measurements also need to be closely scrutinized as to their validity and interpretations. I for one value a subjective review on products to determine if the product has the feature set and/or looks interesting enough to consider and then take a look at the measurements to see how it performs within the boundaries of the tests.
Okay, you're splitting hairs there, Gene! :p I mean this in the philosophical sense. Let's assume for the sake of the poll that the measurements are being conducted correctly. What I want to know is just how much weight any metric has in the minds of the AH readership.

BTW- a big "DUH!" from me. I couldn't figure out why I couldn't see the poll results, but I imagine it's not visible 'til I cast my vote!
 
WmAx

WmAx

Audioholic Samurai
Rip Van Woofer said:
Excellent, thanks! So to bring it down to my level, scenario (1) does indeed screw up the frequency response, while scenario (2) doesn't, but is instead manifest by what I'll call "ringing" - the speaker and cabinet simply don't "shut up" when they're supposed to even though they still measure flat. I suppose that, subjectively described, it would "muddy" the sound and affect the timbre (add a coloration) much as an over-reverberant room does.

I admit that I am still a bit confused with my limited understanding to envision how any coloration of sound would not somehow show up in the frequency response measurements; but I am not about to argue with Mssrs. Toole and Olive and am inclined to take their (and your) word for it!
1st, let me slightly revise(edit) my last reply. You are indeed correct, that for the most part resonances(even delayed resonances) will show up on frequency response(IF-and here's the catch)the frequency response measurement is high enough resolution and it can capture the delayed signal(dependant on issues to be explained). A typical impulse/FFT calculated FR may not reveal all of the resoances due to various issues(which are described in detail in the Toole/Olive paper). I was IN ERROR to classify diaphragm related resonances in as delayed resoances, btw. They are typically grouped with non delayed resoances. I sometimes have a brain fart. :) Cabinet resonances can qualify as either, depending on the situation. Let's say that a resonance was realtivey short time dealy relative to teh frequency, but the resonance occured on a part of the panel at suitable distance from the loudspeaeker element and mic placment, to not show up or signficantly in the amplitude response. However, the radiation of this resoance woudl reflect off room boundaries and contribute to the overall amplitude response off axis into the room. This would be a delayed resonance in this manner, but perhaps not if the room boundary was taken out of the equation(dependant on the frequency and propogation time to the resonant object). The same can be said for the inernal volume. If the propogation delay is large enough relative to a given frequency(which depends on cabinet size), the resoance can be classified as dealyed. But this is entirely depeandatn on the actual delay times vs. frequency.

Delayed resonances are simple; related to the overal time propogation(the delayed engery storage) vs. the direct source of the signal(transducer).

Thanks, therefore, for the AES preprint citation. Perhaps I'll order it and, with luck and study, understand the concept better.
I can email you some scans of the pages if you wish; though I have no idea what the copyright issue is on these papers. If you want a copy, email me(wmax@linaeum.com).

-Chris
 
Last edited:
Mudcat

Mudcat

Senior Audioholic
Rob Babcock said:
What I want to know is just how much weight any metric has in the minds of the AH readership.

It has been evaluated, chewed up, regurgitated, and settled upon, at least until further review that in an evaluation of speaker cables; resistance will be weighted 41%, inductance weighted 32%, and capacitance weighted 27%. These will be combined into weighted technical eval number which will then be weighted at 60% with cost weighted 40% for a final weighted average.

I know this does not cover the opinions of the entire AH readership, but from the minds of people much smarter than I when it comes to cable metrics.

Hope this helps. Maybe Gene, Steve or Clint can do something so that excel files can be attached like *.doc files.
 
D

djoxygen

Full Audioholic
Rip Van Woofer said:
And hey, some folks genuinely prefer a bit of boost or attenuation here and there in the frequency response. Perfecly legit even if it isn't my thang. But few (including me) have the training to say, "oh, I'll probably like the sound of that speaker because it has an attenuated response by 3dB between 2000Hz and 4000Hz and a 2dB/oct rolloff above 10kHz". So the ears still have a role there...
This pretty much sums it up for me. The measurements will tell you how the gear in question performs, and I know that I like things to be accurate while others may bask in the 3rd order harmonic glow of their tube amps. I want to hear and see what the artists, producers, and directors want me to hear and see, and my gear shouldn't get in the way of that. But for those that don't know how the measurements relate to their preferences, a little look and listen will always be necessary.
 
newsletter

  • RBHsound.com
  • BlueJeansCable.com
  • SVS Sound Subwoofers
  • Experience the Martin Logan Montis
Top