Master Thread: Are measurements Everything or Nothing?

[ahofer]

The SHL5+ presents a fairly conventional load.

Harbeth Super HL5plus loudspeaker Measurements

Sidebar 3: Measurements I used DRA Labs' MLSSA system and a calibrated DPA 4006 microphone to measure the Harbeth Super HL5plus's frequency response in the farfield; and an Earthworks QTC-40, with its small, ¼"-diameter capsule, for the nearfield responses.

www.stereophile.com

 

[ahofer]

Come on over….

 

[voodooless]

Come on over….

View attachment 411494

Lost cause:

Over the years I always appreciate upgrades for all other components. This makes me feel like I am losing my mind. Have any others experienced this regarding streamers? I want to try more.

He just wants to hear that he's wrong so he can go out and buy the next upgrade.

 

[ArthurDK]

I am an old guy, and a bloody amateur, but I have picked up certain things along the last 50 years in HiFi.
I might put my head right into a wasp nest now, but bear with me, and here we go:

This post is inspired by your ´Discrete OpAmp Review: Sonic Imagery vs. Sparkos´. The results are quite
impressive and very well made, but they do only tell one part of the whole history, in my opinion pr. see.

If the distortions are as small as stated, no humans in the world would be able to tell a difference in sound
between amplifiers in this quality class. An American University has proved that normal people cannot sense
distortion below 0.2%. I don’t know how they tested it, but there is a long way to 0.002% or lower that we see
in recent measurements. Be aware of the plural word ´distortions´ – It might not be correct, but for myself I
define ´static distortion´ and ´dynamic distortion´ to have some terms to relate to.

To me FFT analysis is a static analysis, in that it normally just consists of a single tone (typically 1 KHz) that is
analyzed for harmonic distortion, and that is an area that is quite easy for most amps. We could vary the
frequency, but it will still only reveal some of the distortion in an amp.

What we need is to measure the dynamic distortions too. They will reveal the overshoot, undershoot, ringing
and IM distortion. These things are in my opinion quite responsible for how the amp sounds.

It could be square waves, for instance 100 Hz, 1 KHz and 10 KHz or pulse waves to measure the ringing, and
2 tone FFT´s to measure IM distortion.

Somebody must have wondered why most of us are able to hear clear differences in the sound of 2 amplifiers
with a measured distortion of 0.00close-to-nothing.

Having done lots of listening tests over long periods I am a discrete op-amp believer, and my preferences are
Sparkos 3602 for standard single discrete op-amps, and Sonic Imagery 995 FET for Pro Discrete single op-amps
- and LM4567 (dual)/ LM47920 if it has to be chip op-amps. Long period testing will reveal ear fatigue better
than quick A-B tests.

BTW I have seen tests where it is necessary to bypass Sparkos opamps across the FB resistor to lower the
bandwidth and prevent ringing.

English is not my mother tongue, but I hope I got the message across.


Best regards
Arthur DK

 

[ahofer]

Cross-posting my new default response.

So I think we can distill the advice to newbies down to something like:

People at ASR tend to view good sound as (for electronics) fidelity to signal, and speaker output conforming to Toole and Olive's research. If you like the sound distorted, less accurate, or prefer some other speaker presentation, that's fine, but own it, don't pretend a) lesser fidelity is greater accuracy or b) there must be something wrong with these standards because of your personal preferences or c)you can hear something that can't be measured. There's no need to rationalize your tastes.

Most of us also believe that the way to test for *strictly audible* differences is by properly executed and level-matched double blind procedures, or through taking measurements and recording a result above audible thresholds. The fact that you noticed a difference outside of these conditions simply isn't evidence of a difference in signal quality at your ears. Even if it is a difference in the signal, as opposed to some sighted bias, it is likely to be a difference in amplitude rather than something more subtle.

Finally, all of the above mistakes are simply human. No human being is so "experienced" or "trained', or "sensitive" as to be able to make sighted comparisons objectively.

update: edited for completeness and links

Message to golden-eared audiophiles posting at ASR for the first time...

"Different" isn't better.

audiosciencereview.com

 

[BDWoody]

Welcome! I moved your post into this thread where we keep this running discussion.

To me FFT analysis is a static analysis, in that it normally just consists of a single tone (typically 1 KHz) that is
analyzed for harmonic distortion,

You're missing the multitone analysis.

If the distortions are as small as stated, no humans in the world would be able to tell a difference in sound
between amplifiers in this quality class.

Expanding that to amplifiers with comparable power, noise, distortion and frequency response into varying loads, so far I don't believe anyone has shown that they can. Lots of claims like yours, but no evidence as of yet.

Somebody must have wondered why most of us are able to hear clear differences in the sound of 2 amplifiers
with a measured distortion of 0.00close-to-nothing.

If they haven't explored psychoacoustics they will certainly wonder, and we hear from them all the time.

Control for sight and volume levels and see if these differences can still be identified.

 

[ahofer]

Somebody must have wondered why most of us are able to hear clear differences in the sound of 2 amplifiers
with a measured distortion of 0.00close-to-nothing.

See my post above:

Most of us also believe that the way to test for *strictly audible* differences is by properly executed and level-matched double blind procedures, or through taking measurements and recording a result above audible thresholds. The fact that you noticed a difference outside of these conditions simply isn't evidence of a difference in signal quality at your ears. Even if it is a difference in the signal, as opposed to some sighted bias, it is likely to be a difference in amplitude rather than something more subtle.

 

[ArthurDK]

OK, it seems here is no place for non-SINAD believers, so let us continue to measure unusable things like we always did - Things that does not show much difference between
different devices, and that does not contribute to anything. And let us keep believing that the best sounding device in the world is the device with the lowest SINAD.

 

[Blumlein 88]

OK, it seems here is no place for non-SINAD believers, so let us continue to measure unusable things like we always did - Things that does not show much difference between
different devices, and that does not contribute to anything. And let us keep believing that the best sounding device in the world is the device with the lowest SINAD.

What about null testing using music? What about controlling the listening conditions so the listening is done only using ears and What they are hearing?

 

[pkane]

OK, it seems here is no place for non-SINAD believers, so let us continue to measure unusable things like we always did - Things that does not show much difference between
different devices, and that does not contribute to anything. And let us keep believing that the best sounding device in the world is the device with the lowest SINAD.

Interesting, but unoriginal view. Doesn't seem to match what anyone here actually thinks. It does match well those who have formed an opinion about ASR with no actual reading or understanding.

 

[MatrixS2000]

And let us keep believing that the best sounding device in the world is the device with the lowest SINAD.

a - It’s the highest scoring SINAD.
b - It is not about “sounding” (sorry about the quotes), SINAD points to accuracy, the higher the SINAD scoring, the more accurate it is.
c - You may like a the sound of a DAC with a lower scoring SINAD, no harm or foul there, but just realize that the cost of that is a loss of accuracy when compared to a DAC with a higher scoring SINAD.

EDIT - corrected to help interpretation - added bolded words

 

[BDWoody]

OK, it seems here is no place for non-SINAD believers, so let us continue to measure unusable things like we always did - Things that does not show much difference between
different devices, and that does not contribute to anything.

It's an evidence based place. Come up with some and let's talk about it. Making the same unsupported anecdotal claims as everyone else makes is what doesn't contribute to moving anything forward. Are you unwilling to entertain the idea that you may have simply fallen prey to natural human bias?

 

[voodooless]

Ah, the good old "You guys only look at SINAD" strawman... It's getting really old

 

[ArthurDK]

What about null testing using music? What about controlling the listening conditions so the listening is done only using ears and What they are hearing?

Ears does not matter. It is all about SINAD and THD+N - Then you know exactly how an amp is sounding

 

[ArthurDK]

a - It’s the highest SINAD.
b - It is not about “sounding” (sorry about the quotes), SINAD points to accuracy, the higher the SINAD, the more accurate it is.
c - You may like a the sound of a DAC with a lower SINAD, no harm or foul there, but just realize that the cost of that is a loss of accuracy when compared to a DAC with a higher scoring SINAD.

It is the lowest, did you notice the minus in front of the figure??

 

[Smaestro]

Ears does not matter. It is all about SINAD and THD+N - Then you know exactly how an amp is sounding

But what if I'm at work in the studio - and I use your same amp to get that special sound flavor into my recording?

Then when you listen to it, you get that effect twice? That's too much. Hmm. I wish there was a solution for this.

 

[ArthurDK]

Interesting, but unoriginal view. Doesn't seem to match what anyone here actually thinks. It does match well those who have formed an opinion about ASR with no actual reading or understanding.

Your icon indicates that you may be a little old fashioned

 

[ArthurDK]

It's an evidence based place. Come up with some and let's talk about it. Making the same unsupported anecdotal claims as everyone else makes is what doesn't contribute to moving anything forward. Are you unwilling to entertain the idea that you may have simply fallen prey to natural human bias?

I just tried to point in a direction of using for instance square waves and dual tone test and such new fashion things, to try to indicate why some devices sound
different than others. Because SINAD and THD+N does certainly not.

 

[ArthurDK]

Ah, the good old "You guys only look at SINAD" strawman... It's getting really old

It is because it is REALLY OLD.

 

[Smaestro]

I just tried to point in a direction of using for instance square waves and dual tone test and such new fashion things, to try to indicate why some devices sound
different than others. Because SINAD and THD+N does certainly not.

You're forgetting frequency response.