Golden ears vs super scopes

[jabbr]

There are two different approaches to obtain obtain objective information:

 

a) Double blinded tests of two components using so-called "golden eared" individuals.

b) Equipment based measurements, high resolution scopes which can measure picovolts and femtoseconds.

 

Do we believe that everything audible is measurable? Do we know the audibility limits? Which data sways subjectivists (if any)? Which data sways objectivists?

[Superdad]

The main issue--and the reason for much of the "objectivist"/"subjectivist" divide--may be that the traditional measurements being taken are not looking at or showing up many of the variances in gear that people readily hear.  Much as specialized tests and advanced imaging and genetic testing is required to identify certain diseases which will not be found in routine blood lab work, there are measurements and research to be done which go beyond what an Audio Precision rig or high-speed scope will see.

 

I need not recite to you the history of analog and digital measurement advances which revealed factors (e.g. TIM and jitter) that were for a long time not even considered. 

There is little reason to believe that current popular measurement regimens are telling us all there is to know about audible differences.

[lucretius]

1 hour ago, jabbr said:

There are two different approaches to obtain obtain objective information:

 

a) Double blinded tests of two components using so-called "golden eared" individuals.

b) Equipment based measurements, high resolution scopes which can measure picovolts and femtoseconds.

 

Do we believe that everything audible is measurable? Do we know the audibility limits? Which data sways subjectivists (if any)? Which data sways objectivists?

 

Why bias the dbx tests with "golden eared" individuals? ☺️

[John Dyson]

1 hour ago, lucretius said:

 

Why bias the dbx tests with "golden eared" individuals? ☺️

I didn't check into the test that was mentioned, but on any moderatly noisy tape deck & DBX, even  I could EASILY hear the noise modulation.  That was one of the big disadvantages of DBX.   There were other secondary issues like transient problems, but I don't truly know if they are a problem in reality.   I couldn't use DBX back in the past because my very good hearing at the time ,and the noise modulation was too obvious to be limited to 'golden ears'.

I am very very busy right now, but I have an interesting technical report on NR systems that I'll post later.   The report had objective information and used 'scopes a lot to show the dynamics behavior.  (The test I am thinking about was DBXI, DolbyA and TelcomC4.)   Give me a day or so -- super, mega busy!!!   Not that the paper is directly applicable to the specifc issue, but might have interesting background info on the matter.

 

John

 

[alfe]

7 hours ago, jabbr said:

There are two different approaches to obtain obtain objective information:

 

a) Double blinded tests of two components using so-called "golden eared" individuals.

b) Equipment based measurements, high resolution scopes which can measure picovolts and femtoseconds.

 

Do we believe that everything audible is measurable? Do we know the audibility limits? Which data sways subjectivists (if any)? Which data sways objectivists?

... 😂

[alfe]

7 hours ago, Superdad said:

The main issue--and the reason for much of the "objectivist"/"subjectivist" divide--may be that the traditional measurements being taken are not looking at or showing up many of the variances in gear that people readily hear.  Much as specialized tests and advanced imaging and genetic testing is required to identify certain diseases which will not be found in routine blood lab work, there are measurements and research to be done which go beyond what an Audio Precision rig or high-speed scope will see.

 

I need not recite to you the history of analog and digital measurement advances which revealed factors (e.g. TIM and jitter) that were for a long time not even considered. 

There is little reason to believe that current popular measurement regimens are telling us all there is to know about audible differences.

One Alex may hide another😜

[mfsoa]

If b) returned absolutely perfect measurements, why would one assume that this would provide the more desirable listening experience?

 

The best preamp I ever had had probably 1000 times more distortion than my current one, but sounded so much better.

 

 

 

[jabbr]

7 minutes ago, mfsoa said:

If b) returned absolutely perfect measurements, why would one assume that this would provide the more desirable listening experience?

 

The best preamp I ever had had probably 1000 times more distortion than my current one, but sounded so much better.

It is obvious that “distortion”, for you is not a good measure of listening pleasure. If your old preamp was so much better, why would you downgrade?

 

Perhaps you would like to know, in a detailed way, why your old preamp sounded so good, so that you could replace it, rather than suffer through your suboptimal listening experience as you are know?

[jabbr]

9 hours ago, Superdad said:

The main issue--and the reason for much of the "objectivist"/"subjectivist" divide--may be that the traditional measurements being taken are not looking at or showing up many of the variances in gear that people readily hear.  Much as specialized tests and advanced imaging and genetic testing is required to identify certain diseases which will not be found in routine blood lab work, there are measurements and research to be done which go beyond what an Audio Precision rig or high-speed scope will see.

 

You are not, by any means, limited to using an AP analyzer alone to measure UpTone’s  equipment.

 

As @Archimago does (IIRC) you could use an ADC to collect data, such as the RME ADI-2 FS and then use your laptop to analyze in any way you wish.

 

Similarly a 2Ghz Tektronix scope can be had for <$2k (Used) and these can do very sophisticated measurements. 
 

The idea that you need $100k equipment to do sophisticated measurements is false. 

Quote

 

I need not recite to you the history of analog and digital measurement advances which revealed factors (e.g. TIM and jitter) that were for a long time not even considered. 

There is little reason to believe that current popular measurement regimens are telling us all there is to know about audible differences.

Current popular measurement regimens like phase error plots? 1/f noise? 3D output impedance plots? EMI patterns? 
 

These are all well known measurements used in the electronics industry on a daily basis.

 

Yes! By all means let’s stop talking about THD and RMS jitter (terms from the 1970s) and get into modern measurements!

[jabbr]

9 hours ago, lucretius said:

 

Why bias the dbx tests with "golden eared" individuals? ☺️

😉

 

there is a certain range of response, just as different people have different baseline vision tests. 
 

When it comes to complex feature extraction, that rests more in the brain, and so-called “golden eared” folks mostly have more experience listening for subtle differences — these are the folks we would want tested 

[CG]

1 hour ago, jabbr said:

The idea that you need $100k equipment to do sophisticated measurements is false. 

 

I agree!  (Bet that completes your day! 😉)

 

The big challenge is moving beyond the established test regimes that were adopted more than a half century ago, because they were what the practitioners could do at the time.  

 

A very simple example, which I will continue to beat until the carcass is a pile of molecules, is making spectral and other plots without the use of averaging.  Averaging certainly lets you measure way into the noise, but it purposely ignores that very noise to do so.  In a broad sense, noise is not only that hissing sound we all know.  It's random or pseudo random disturbances to the signal content.  If you take 10000 samples of tones and average out the disturbances that rarely repeat or repeat in random ways, you definitely can detect tones that are consistent over those 10K samples.  But, think you can hear junk that goes on randomly for a quarter second?  (10K samples at 44 KHz sample rate)  Just how is that effectively captured using current test techniques?  There is a place for averaging and there is a place for capturing peak events and the integrated envelope.  

 

In addition, gear is measured under ideal conditions.  By that I mean, without other audio components attached to it.  The current loops and other limitations caused by connecting all these audio components through random cabling and attached to random electrical supplies certainly must cause disturbances in real world gear.  Think of how often John Atkinson reports in his test measurements how he's had to ground this in a funny way or float some other connection in order to get decent measurements.  What is that?  The old idea that "properly designed equipment is immune to these situations" is a load of twaddle.  If that was the case, why even test for anything? 

 

Using test gear to measure performance requires actually having a good idea of what to look for.  We usually fail to do that in a thorough way.

 

Using our ears is a far more integrated approach, but is flawed because the "test gear" is hardly uniform between instruments and from day to day.

 

Neither is ideal, so it makes sense to apply both.  However, thinking and reason is required in any case.

 

But... Very few people genuinely, really care.

[jabbr]

24 minutes ago, CG said:

In addition, gear is measured under ideal conditions.  By that I mean, without other audio components attached to it.  The current loops and other limitations caused by connecting all these audio components through random cabling and attached to random electrical supplies certainly must cause disturbances in real world gear.

Absolutely. That’s stress testing.

 

One example of mine: I set up a Raspberry Pi with a standard RPi SMPS and WiFi streaming to a a few DACs with and without the Uptone ISO Regen... Without doing detailed measurements I was trying to create a high EMI, high jitter situation where I expect the ISO Regen to shine...

 

a) With the iFi iMicro (+\- BL) I couldn’t hear an ISO Regen effect after months of listening.

 

b) The Topping D7x had  clicks and pops either way (AC powered)

 

c) The Pro-Ject S2D had clicks and hum when fed directly from the RPi but a great sound when fed from the Regen. (USB bus powered)

[Teresa]

17 hours ago, jabbr said:

...Do we believe that everything audible is measurable? Do we know the audibility limits? Which data sways subjectivists (if any)? Which data sways objectivists?

 

I don't know if everything that is audile can be measured or what audibility limits exist. I believe that there are many kinds of subjectivists and many kinds of objectivists.

 

I have saved tons of money as a poor audiophile subjectivist because I don't go by measurements alone. For example: if there are two audio components, one with better specifications than the other, and the one with the worst specifications is 1/3 the price I will try that one first, more times that not I am satisfied with the less expensive one.

 

I'm not really swayed by data as long as the specifications are decent. For examples: I would never pay over $1,000 for any audio component and I would prefer to pay $200 - $500. I would never spend more than $50 for any cable and would prefer to spend $20 instead.

 

So the way I see it there are budget subjectivists that only slightly compare specifications. If they are the same price I tend to try with the better specifications first. But if the specifications are close I would go with the cheapest one. And I want equipment to last many decades, and I don't replace anything until it breaks and cost too much to repair.

 

Then there are spend thrift subjectivists in which price is not a concern, many of them believe the higher the price the better the sound. While I have heard many super expensive excellent sounding components the two times I went CES. Once when I sold audio equipment at the Good Guys and later when I was a reviewer at Positive Feedback. I would never lay down that kind of cash.

 

I would guess that most subjectivists are somewhere in-between. I would also guess there is also a wide divide among objectivists.

 

Not worrying too much about specifications and auditioning affordable components by ear has saved me a ton of money.

 

BTW my ears are pink like the rest of my body. 😀

 

I believe no one hears music the same and everyone's ear / brain system only apply's to them and no one else, so I don't believe in the concept of golden ears.

[Archimago]

 

Nah guys... It's not "Golden Ears" vs. "Super Scopes"!

 

It's Golden Ears:

http://archimago.blogspot.com/2014/02/musings-golden-earism-philips-golden.html

 

AND super scopes!

 

I wish all golden eared subjectivist reviewers in magazines and online sites had a "standard" test and certification they must pass before accredited to be writing.

 

As for whether instruments capture "everything". I believe the answer is clearly a yes. There are anomalies that  measurement devices can easily pick up which ears will miss... So long as there is an ideal which the output of a device can be compared to, instruments can make that determination way better than ears. "High end" companies and pure subjectivists will not want to concede this of course; it will take away much of the mystique that these companies and writers need.

[Iving]

8 minutes ago, Archimago said:

As for whether instruments capture "everything". I believe the answer is clearly a yes. There are anomalies that  measurement devices can easily pick up which ears will miss...

 

[Iving]

8 minutes ago, Archimago said:

So long as there is an ideal which the output of a device can be compared to

 

What "ideal"? Output of another device/instrument?

[jabbr]

1 hour ago, Teresa said:

I have saved tons of money as a poor audiophile subjectivist because I don't go by measurements alone. For example: if there are two audio components, one with better specifications than the other, and the one with the worst specifications is 1/3 the price I will try that one first, more times that not I am satisfied with the less expensive one.

 

I'm not really swayed by data as long as the specifications are decent. For examples: I would never pay over $1,000 for any audio component and I would prefer to pay $200 - $500. I would never spend more than $50 for any cable and would prefer to spend $20 instead.

You are on the right track for sure!

 

I guess the point is that it’s hard to show that a >$50 cable has better specs than a $20 cable within reason. I have heard differences between RCA cables, I have no doubt that there are measurable electrical parameters, but the cable companies don’t want to publish these because everyone would see how simple it would be to replicate. So really I haven’t seen “better” measurements from audiophile cables.
 

I don’t really obsess about cables myself. I use Corning fiberoptic cables which are cheap and state of art. I use well made XLR cables with good connectors. I don’t obsess about USB cables because I’m interested in other things.

 

By all means go for the best made cables which are also reasonably cheap. 

[lucretius]

11 hours ago, jabbr said:

these are the folks we would want tested

 

First, you'd have the problem of identifying them.

[Audiophile Neuroscience]

1 hour ago, lucretius said:

13 hours ago, jabbr said:

these are the folks we would want tested

 

First, you'd have the problem of identifying them.

Nope, they tend to stick out !

This is a picture of me....

[jabbr]

6 minutes ago, Audiophile Neuroscience said:

Nope, they tend to stick out !

This is a picture of me....

That is most certainly not you ... @sandyk gets high levels of jitter around XX and communication bandwidth drops to 0 😂

[Audiophile Neuroscience]

7 hours ago, Archimago said:

As for whether instruments capture "everything". I believe the answer is clearly a yes.

Hi Archi

I believe believe is a good choice of words.

 

7 hours ago, Archimago said:

There are anomalies that  measurement devices can easily pick up which ears will miss...

If so, it doesn't follow that the converse is not true

 

7 hours ago, Archimago said:

So long as there is an ideal which the output of a device can be compared to, instruments can make that determination way better than ears.

 

How do you know. Evidence?

 

[Audiophile Neuroscience]

12 minutes ago, jabbr said:

That is most certainly not you ... @sandyk gets high levels of jitter around XX and communication bandwidth drops to 0 😂

 

I don't think @The Computer Audiophile permits @sandyk to post in objective-Fi otherwise, having met me, he could testify that this is my true likeness....down to the last measurement! 🤣

[Audiophile Neuroscience]

@The Computer Audiophile talking of measurements, is it just me or does the post count get stuck from time to time?

[Iving]

3 hours ago, lucretius said:

First, you'd have the problem of identifying them.

 

Perhaps - but it really depends what you want to demonstrate, and your hypothesis / research question is framed accordingly.

 

e.g. #1 - A sample of 100 self-identified "golden-eared" audiophiles were tested to establish whether they could discriminate (music-relevant stimulus) A from B ...

 

e.g. #2 - A random sample of 500 adults were given a musical nuance test to establish what proportion (if any) could be identified as "golden-eared" so defined ...

 

etc etc

 

As in the spiritual life the scientist may find sooner or later what she or he is looking for.

 

Subject of course to methodological rigour ...

 

and p (reliability or reproducibility) ...

[Audiophile Neuroscience]

19 minutes ago, Iving said:

it really depends what you want to demonstrate, and your hypothesis / research question is framed accordingly.

Yes I agree. Toole (I am not his greatest fan but) already compared experienced listeners with non experienced and found that they came to much the same conclusions. The experienced listeners were just better at doing it, more consistent and more definite.

 

The suggestion IIRC was that you needed fewer experienced/trained listeners to yield a statistically significant result.