Head-Fi.org › Forums › Equipment Forums › Headphones (full-size) › Do Objective Headphone Measurements Correlate to the Audiophile's Subjective Experience?
New Posts  All Forums:Forum Nav:

Do Objective Headphone Measurements Correlate to the Audiophile's Subjective Experience? - Page 12

post #166 of 193
Thread Starter 

LOL. You like headphones that measure well, at least with my measurements. The exception is the D5000. The immense bass and bass distortion of the D5000 tends to mask everything else that isn't so well.

 

Sony R10

r10r.txt.jpg

 

HD800

HD800MK L.txt.jpg

 

TH900

th900r.txt.jpg

(A clean floor considering it's a closed headphone)

 

index.php?action=dlattach;topic=351.0;attach=1373;image

And as you said, the TH900 does have notable frequency response irregularities. However I feel it has certain qualities (clarity, low non-linear distortion, articulation, fast decay, detail extraction) which allow me to overlook its non-flat response. I'm sure you agree.

 

 

The D5000 will not measure flat on my rig. Here is the D7000 (which from my subjective guess has about 5-6db less bass at 50Hz than the D5000)

index.php?action=dlattach;topic=29.0;attach=25;image


Edited by purrin - 6/20/12 at 9:02am
post #167 of 193
Quote:
Originally Posted by purrin View Post

LOL. You like headphones that measure well, at least with my measurements. The exception is the D5000. The immense bass and bass distortion of the D5000 tends to mask everything else that isn't so well.

 

 

And as you said, the TH900 does have notable frequency response irregularities. However I feel it has certain qualities (clarity, low non-linear distortion, articulation, fast decay, detail extraction) which allow me to overlook its non-flat response. I'm sure you agree.

 

 

The D5000 will not measure flat on my rig. Here is the D7000 (which from my subjective guess has about 5-6db less bass at 50Hz than the D5000)

index.php?action=dlattach;topic=29.0;attach=25;image

 

Ironically, the D-5000 is the one I bought based on measurements! If you dig Tyll's compensation curves, the resultant curve is pretty flat (as headphones go). Looking at your D-7000 curve, 50 hz is roughly 3 db higher than the 1 khz level, which is about the same as the D-5000 curves, again D-5000 w/Jmoney pads. Here are his measurements:

 

http://www.innerfidelity.com/images/DenonAHD5000JMoneyPads.pdf

 

I would like the bass THD to be lower, but still it's only gets to 1% at 40 hz for 90 db level, so it's not disasterous.

 

Interesting to see your measurements on the other phones I mentioned. I had a feeling the TH-900 would not measure spectacularly well, but, as you say, it has its compensations. And I would argue (but not to strenuously wink.gif ) that some of it's charm may be in it's strategic deviations from measuring objectively "flat".

 

To get back to the point of the thread, I think measurements can point out the general nature of a phones sound and help one to filter out phones from consideration that probably won't be to one's personal liking, but listening is still the final arbiter. Which I suspect is pretty close to your POV...

 

Kevin 

post #168 of 193

What's with all the talk about the Denon's low bass distortion?  They distort about as much as any other dynamic driver on the market.  It's a limitation of the technology for headphones it seems.  Even the HD-800, with its ring radiator to minimize distortion, still shows signs of this.

post #169 of 193
Quote:
Originally Posted by TMRaven View Post

What's with all the talk about the Denon's low bass distortion?  They distort about as much as any other dynamic driver on the market.  It's a limitation of the technology for headphones it seems.  Even the HD-800, with its ring radiator to minimize distortion, still shows signs of this.

 

Yes, LF distortion is not an uncommon occurance in headphones, but some others (such as the HD-800) do have less distortion and that's the point.

 

And there are plenty of phones that do worse (often significantly so), that's why I said it wasn't disasterous.

post #170 of 193
Thread Starter 
Quote:
Originally Posted by TMRaven View Post

What's with all the talk about the Denon's low bass distortion?  They distort about as much as any other dynamic driver on the market.  It's a limitation of the technology for headphones it seems.  Even the HD-800, with its ring radiator to minimize distortion, still shows signs of this.

 

Referring to non-linear distortion (harmonic, intermodulation, etc.) vs. linear distortion which is frequency response. Just conjecture of mine that the Denon D2000/5000 and to a lesser extent, the D7000, have higher non-linear distortion in the bass compared to the TH900. The Denons' bass sounds muddy in comparison to the TH900's. Lowest non-linear distortion (in the bass) headphones are the Audeze's. The HD800 also has fairly low non-linear distortion figures for the bass. The huge drivers of Audeze's and the HD800 helps alot.


Edited by purrin - 6/22/12 at 11:03am
post #171 of 193

Very interesting thread purrin.  I've been toying around with the idea of creating an experiment that would measure at just what point we can detect changes.

 

Off-topic to your thread, but semi relevant to the spirit behind it:

 

It is my opinion the methodology surrounding the double blind tests that have been done in the past have been really sloppy.  One should never conduct a test with a hypothesis such as "Can audiophiles tell the difference between x and y" without a proper measurement system analysis of what your testers (your measurement systems) are capable of accurately detecting in the first place.  

 

To my knowledge, no one has ever conducted a proper statistically valid measurement system analysis, prior to conducting an "audiophile" double blind a/b test.

 

I wish I didn't have a conflict for the Los Angeles meet in August.  I have a methodology in mind* and would love to have that many head-fi members in one place to conduct an experiment. 

 

*I am a statistical analyst by trade and my background is in creating design of experiments for medical device studies.  


Edited by TWIFOSP - 6/25/12 at 7:26am
post #172 of 193

That is a very important consideration TWIFOSP, I don't think I have ever read about a pre-test design being used before either. I can't imagine how difficult it will be to (a) operationalize what "qualified hearing acuity" is, and (b) how to ensure that all participants are assessed in the same manner, on the same measures, and in a very similar time-frame. That said, any attempts to improve the quality of these subjective measures will go a long way to allowing us to determine what if any correlation exists between the objective frequency response profiles of headphones, and the subjective experience of users. Man it would be nice to see some of the larger equipment manufacturers get behind such an initiative and provide funding. Not to likely I think as they have a great deal to lose if their products don't measure up to their expectations. Who knows, we need to have dreams after all.

post #173 of 193

Those are valid points, but you don't necessarily need to compare all listeners to each other.  That's the point of a measurement system's analysis.  Once you've determined a measurement systems precision and accuracy, you can measure the results against themselves.  Accounting for hawthorne and other psychological effects should be easy enough with false positives.  My hypothesis is that experienced listeners are far more capable than past a/b "tests" have let on.  

 

What I have in mind could easily be accomplished at a large head-fi meet with borrowed equipment.  

post #174 of 193
Quote:
Originally Posted by Sonic Defender View Post

That said, any attempts to improve the quality of these subjective measures will go a long way to allowing us to determine what if any correlation exists between the objective frequency response profiles of headphones, and the subjective experience of users. Man it would be nice to see some of the larger equipment manufacturers get behind such an initiative and provide funding. Not to likely I think as they have a great deal to lose if their products don't measure up to their expectations. Who knows, we need to have dreams after all.

 

I'd like to see what Sean Olive at Harman comes up with, if he gets the chance:

 

youtube.com/v/cV0S7BTMMV0?start=2426&end=2511

Wilkinson: Does Harman test headphones, like you do speakers?
Sean Olive: Yeah, we do. A lot of the testing, right now, is done over in Vienna Austria, where AKG is located. Here in Northridge (CA) we want to get more involved in the actual subjective testing; because compared to loudspeakers, headphones (in terms of scientific research) [are] about the same place where we were 15 years ago with loudspeakers.

 

Olive on subjective testing and the importance of meaningful measurements in headphones:

youtube.com/v/cV0S7BTMMV0?start=2580&end=2690

(note/apologies: couldn't get embedding to work with start/end times)

post #175 of 193

Going back for a moment to the OP:

Do Objective Headphone Measurements Correlate to the Audiophile's Subjective Experience?

in all honesty, I can only reply "sometimes".

 

 

Also - there's no known way of measuring the output of a set of headphones, and then interpreting the data received,

which will indicate exactly how those headphones sound.

 

There's simply no substitute (yet, anyway) for careful, attentive listening.

post #176 of 193

Here's my opinion on the subject.  Objective information cannot tell you what something will sound like, it can confirm what others hear though.  Subjective information can confirm other subjective information.  EG, if 10 people say it sounds like this, and 1 person says otherwise, it's more likely that the 10 are right. Objective can confirm subjective, if a few people say it sounds like x, and the graph confirms that it's possible for it to sound like x, then it is confirmed.  It is useful, and just creates more accuracy in people's ears.

post #177 of 193
That's true. The first time I listened to the ED10 I knew something was off, then I saw Tyll's charts and thought, "well, that'll do it".
post #178 of 193

I came across this interesting bit of info on Sean Olive's blog:  http://seanolive.blogspot.com/2012/05/more-evidence-that-kids-even-japanese.html

 

This is a slide from a presentation that he made.  I think the sample size of the study was a little small and the group was not really random.  But it's still telling.

 

Screen Shot 2012-06-29 at 12.55.59 AM.png

post #179 of 193
Thread Starter 

Very interesting and consistent with the results of my own informal study. They've obviously got more data points. Another difference is that I only allowed people who I considered "trained listeners" vs. random Head-Fi'ers. Similar to my own informal study, the least accurate headphone (A) had the most variance in opinion.

 

Individual+Listener+Loudspeaker+Preferences.png

 

May be interesting to conduct a more controlled test at the August meet.

 

P.S. Good to know that my UCI friends have good ears :-)


Edited by purrin - 6/29/12 at 8:43am
post #180 of 193
Originally Posted by k3oxkjo View Post
Originally Posted by purrin View Post

LOL. You like headphones that measure well, at least with my measurements. The exception is the D5000. The immense bass and bass distortion of the D5000 tends to mask everything else that isn't so well.

 

 

And as you said, the TH900 does have notable frequency response irregularities. However I feel it has certain qualities (clarity, low non-linear distortion, articulation, fast decay, detail extraction) which allow me to overlook its non-flat response. I'm sure you agree.

 

 

The D5000 will not measure flat on my rig. Here is the D7000 (which from my subjective guess has about 5-6db less bass at 50Hz than the D5000)

index.php?action=dlattach;topic=29.0;attach=25;image

 

Ironically, the D-5000 is the one I bought based on measurements!

 

http://www.innerfidelity.com/images/DenonAHD5000JMoneyPads.pdf

 

 

Yup, that PDF and the discontinuation is what propelled me to buy the D5000 with J-money pads as well, from 30Hz to 16kHz that's very flat.

 

HD800 D5000 Jmoney.jpg

 

 

Edit: D5000 left, HD800 right.

 

Like others are saying the bass isn't something to write home about, but it's a very musical headphone imho.


Edited by kiteki - 7/1/12 at 12:32pm
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Headphones (full-size)
Head-Fi.org › Forums › Equipment Forums › Headphones (full-size) › Do Objective Headphone Measurements Correlate to the Audiophile's Subjective Experience?