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 9

post #121 of 193

What LFF said....

 

That's more of a knock against headphone manufacturer's unit-to-unit consistency than the measurement system. The same exact headphone will measure the same at different times (I believe purrin already posted measurements of the same headphone taken months apart). 

post #122 of 193
Quote:
Originally Posted by kiteki View Post

 

In blue, all that proves is that we can't fully trust single measurements, since another model may have less distortion, less noise, and more extended FR, or the adverse, in theory.

 

Looking at Sonove, Goldenears, innerfidelity, and custom IEM manufacturers, the measurements deviate quite a lot too.

A lot of these deviations can also come from how they treat the raw data. I'm pretty sure almost every major measurement that you see typically have some type of a loudness curve applied to it. I know Tyll does this and I'm pretty sure Purrin does this. If you're using a different curve, the presentation of the data comes out differently, but the raw data could be very similar. Most manufacturers choose curves which make their headphones look razor flat in measurements (see BT's provided DT48 measurements, Superlux's provided measurements, etc, etc). Yes, I do think this is a factor, but it isn't a major one in most cases. As long as you know how to compare the relative data, it's fine. It's more of a couple of headphone manufacturers having poor QC that are getting revealed by measurements. 

post #123 of 193
Quote:
Originally Posted by SanjiWatsuki View Post

A lot of these deviations can also come from how they treat the raw data. I'm pretty sure almost every major measurement that you see typically have some type of a loudness curve applied to it. I know Tyll does this and I'm pretty sure Purrin does this. If you're using a different curve, the presentation of the data comes out differently, but the raw data could be very similar. Most manufacturers choose curves which make their headphones look razor flat in measurements (see BT's provided DT48 measurements, Superlux's provided measurements, etc, etc). Yes, I do think this is a factor, but it isn't a major one in most cases. As long as you know how to compare the relative data, it's fine. It's more of a couple of headphone manufacturers having poor QC that are getting revealed by measurements. 


The compensation curve Tyll applies is to remove the pinnae effects from the microphone measurements. Other manufacturers do the same because they measure with IEC ears as well.

post #124 of 193
Quote:
Originally Posted by LFF View Post


The compensation curve Tyll applies is to remove the pinnae effects from the microphone measurements. Other manufacturers do the same because they measure with IEC ears as well.

 

The thing is, the compensations seem to be quite different when the manufacturer puts them out, more than just removing how the ear hears. I've always wondered how in the world Beyerdynamic came up with things like this. 

post #125 of 193
Quote:
Originally Posted by kiteki View Post

In blue, all that proves is that we can't fully trust single measurements, since another model may have less distortion, less noise, and more extended FR, or the adverse, in theory.

Looking at Sonove, Goldenears, innerfidelity, and custom IEM manufacturers, the measurements deviate quite a lot too.

*sigh* Yes, you can twist and manipulate your interpretation to serve whatever master you like. I think it's kind of in poor form to use measurements and data as a data point for dismissing all measurements and data though (tell me you see the irony) - LFF's argument is more plausible and likely the case. Especially since purrin isn't the only one who's noted the massive discrepancies between Beyer products. Really, what's easier to believe:

- That a hobbyist with no (as far as we know) vested interest in the outcome of any of his non-commerical measurements has screwed up every time he's tried different Beyer headphones, but can consistently measure other products.

or

- That a huge multinational company that has engaged in other tomfoolery (like harassing people for selling their products on eBay, used) and that only cares about making huge profits in the advent of headphones being a growth segment has absolutely stopped caring about QA or customers in general and just wants to increase their bottom line by another few %.

IF this wasn't Beyerdynamic, but another manufacturer (lets pick on Bose or Monster because they're crowd favorites around here), I doubt there'd be any objection to discrepencies - it'd be "yet another piece of proof that they make bad equipment." But because it's a "friend" it becomes a shoot the messenger kind of thing. Toole said it best a number of years ago (and I'll butcher the pharaphrase) - in science, when you measure something and the results are different, you try to figure out why. In religion, you just dismiss it and go along saying "this is how it is."
post #126 of 193
Quote:
Originally Posted by SanjiWatsuki View Post

 

The thing is, the compensations seem to be quite different when the manufacturer puts them out, more than just removing how the ear hears. I've always wondered how in the world Beyerdynamic came up with things like this. 


That my friend...is called marketing.

post #127 of 193
Quote:
Originally Posted by SanjiWatsuki View Post

 

The thing is, the compensations seem to be quite different when the manufacturer puts them out, more than just removing how the ear hears. I've always wondered how in the world Beyerdynamic came up with things like this. 

 

Same way Tobacco companies found researchers who released study after study showing tobacco was in fact good for your health?  It aint difficult to fudge this kind of thing.  Purrin has posted some nice explanations for how graphs can be manipulated to look nice.  You can't trust manufacturer released graphs for the same reason you can't trust research done by the company making the product in question. 

 

As for Tyll's graphs, they are made in a very different way, and use a different compensation method.  With Tyll's graphs, the ideal FR is not a flat line, so the most neutral headphone in the world would still not measure flat, or even close to flat.  You have to interpret them. with the ideal curve in mind  But AFAIK, Purrin's FR graphs are the first headphone graphs I've seen where the ideal is flat, just like it is with speaker FR measurements.  So what you see is what they sound like to the human ear.  Much easier to read.  But you can't compare the two. 

post #128 of 193

I know I've asked before, but both Tyll and Purrin have raw uncompensated versions of their graphs.  What is the difference?

post #129 of 193
Quote:
Originally Posted by TMRaven View Post

I know I've asked before, but both Tyll and Purrin have raw uncompensated versions of their graphs.  What is the difference?
I know that typically headphone measurements tend to be smoothed out a little bit to hide the sharper peaks and troughs which would make it look confusing.
post #130 of 193
The raw uncompensated measurements are exactly what they are, raw measurements where the microphone's response and other artifacts introduced by the measurement rig are not corrected.
post #131 of 193

I know what raw uncompensated graphs are, I meant what is the difference in the two's raw graphs- since rhythmdevils keeps saying Purrin's graphs are perfectly flat when his graphs are flat.

post #132 of 193
Quote:
Originally Posted by TMRaven View Post

I know what raw uncompensated graphs are, I meant what is the difference in the two's raw graphs- since rhythmdevils keeps saying Purrin's graphs are perfectly flat when his graphs are flat.

I think Tyll is the only one who publishes the raw data alongside the compensated/averaged data. I know GE specifically says they don't, same for Ryumatsuba, and I've only ever seen single lines (or L/R channel lines) from purrin - which leads me to believe they're the graphs *after* whatever compensation is applied. There's a Stereophile article on the whole matter of these kinds of measurements and how raw graphs can different wildly depending on how they were done and how the compensation can also differ wildly; that's why we don't want to compare GE to IF to Purrin et al.
post #133 of 193

There have been multiples of measures of various units of the same models.  By far, Beyerdynamic as a manufacturer has had the widest discrepancies unit to unit whereas a company like Sennheiser has had the most consistent measurements unit to unit.  Not even Ultrasone or Monster have come close to Beyer in inconsistency.  The closest were the pre RMA'd/updates LCD3's.  

 

We might have to consider publishing overlays of various models where we have multiple measures.

 

Edit - What's odd about this Beyer inconsistency is that it's most noticeable in it's newer, high-end line including 'Tesla' drivers.  Their older lineup is much more consistent and in some cases perform better technically depending.  Some of us hear this difference, others don't seem to notice or mind.


Edited by Anaxilus - 6/18/12 at 4:16pm
post #134 of 193
Thread Starter 

Related: http://www.head-fi.org/t/533508/new-beyer-dt1350/1740#post_8467838

 

Admittedly, I haven't really cared that much to formally track headphone variance and consistency because the super vast majority have been fine (from my subjective experiences and verified via objective measurements.) Formal tracking of this (and precision testing) is probably something I should do.

 

BTW, Tyll did find consistency issues with the early LCD3s, just as I did. Tyll even wrote an investigative report about it.

 

 

P.S. Somehow, I feel that even I had formally tracked the DT1350, performed multiple precision tests on these pairs and of the measurement rig, and found two more "defective" Teslas, some people still wouldn't be satisfied and continue to give the 5th degree rather than asking Beyer WTFIGO?

 

Also, I hate to say it, but if anyone's taken apart a T1 and compared to the internals of the DT880... Yes, I'm biased. But I have this image in my mind that bean-counters at Beyer are rolling around in a pile of Euros laughing at us.

 

rolling.jpg


Edited by purrin - 6/18/12 at 5:11pm
post #135 of 193
Would have to say +1 to what you said Purrin, and I haven't even tried the T1. The T70 was bad enough.
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?