Everything sounds the same

[fas42]

Just for the hell of it, I'll throw in a, Gasp!!, computing analogy - current audio is like Windows, where more and more "features" were added which just got in the way of people being able to do what they wanted to do - it was full of 'bugs' and bad design choices, and was severely compromised by having a very "old ideas thinking" past.

 

Ummm, so what's the best thing to do with, err, Windows ...

[sandyk]

20 minutes ago, Sonicularity said:

We certainly cannot account for every individual's perception of a given measured sonic input, but we can verify that the sonic input measures precisely the same.   

 That's fine for an analogue input, but not so easy to measure when the input and outputs are digital, before the digital output is converted to analogue by a D to A conversion. (Signal Integrity)

The digital output may also have other crap riding along with the binary data that could influence the quality of the conversion to analogue. USB Audio is particularly susceptible to this problem.

[Superdad]

1 hour ago, mansr said:

The idea that humans, of all animals, can hear things that elude the finest scientific instruments is nothing short of preposterous.

 

So if someone presented you with two complete sets of measurements of two DACs--say the full suite that John Atkinson runs on every DAC they review--but did not tell you which set came from which, you would confidently be able to tell us which one performs better subjectively?  Every time? One would be from a $200 cheap-as-chips DAC, the other from say dCS or Meitner.

[lmitche]

1 hour ago, sandyk said:

 

Which cheek ?

Which tongue? 

[Superdad]

[fas42]

The battle here is between those who believe they've got it under control, and those who know they haven't got the necessary knowledge ...

[mansr]

2 minutes ago, Superdad said:

So if someone presented you with two complete sets of measurements of two DACs--say the full suite that John Atkinson runs on ever DAC they review--but did not tell you which set came from which, you would confidently be able to tell us which one performs better subjectively?  Every time? One would be from a $200 cheap-as-chips DAC, the other from say dCS or Meitner.

The one with the most distortion is likely the dCS or other hyper-expensive product. It is the only way they can be substantially different from a decent design with a reasonable price.

[mansr]

1 minute ago, fas42 said:

The battle here is between those who believe they've got it under control, and those who know they haven't got the necessary knowledge ...

Which category do you place yourself in?

[Jud]

25 minutes ago, mansr said:

In order to identify a pattern, one must first register a difference. If there is no discernible difference, no detection of a pattern is possible.

 

Sorry, the brain's pattern identification mechanisms do not work in this nice, neat way.  Otherwise instruments, which are far more sensitive to small differences, would inevitably be better at pattern identification.  For example, instruments can detect things far smaller or further away than our eyes can see, but we are better at recognizing faces. 

 

Ignoring rather than detecting differences may also be key to pattern recognition, as with recognizing a disguised person or recognizing that two different instruments in two different recordings are both violins, or both Telecasters.

[lmitche]

1 minute ago, mansr said:

The one with the most distortion is likely the dCS or other hyper-expensive product. It is the only way they can be substantially different from a decent design with a reasonable price.

Man, you are one cynical dude. How do you get out of the bed in rhe morning?

[Sonicularity]

10 minutes ago, sandyk said:

 That's fine for an analogue input, but not so easy to measure when the input and outputs are digital, before the digital output is converted to analogue by a D to A conversion. (Signal Integrity)

The digital output may also have other crap riding along with the binary data that could influence the quality of the conversion to analogue. USB Audio is particularly susceptible to this problem.

 

Why can't we measure an analog output?  We can certainly record it, make copies of it, and send it to a myriad of devices in many formats over several different protocols.

[rickca]

2 minutes ago, lmitche said:

Man, you are one cynical dude. How do you get out of the bed in rhe morning?

The alarm on a $5 clock radio?

[fas42]

2 minutes ago, mansr said:

Which category do you place yourself in?

 

Still definitely the latter ... I can always organise a compromise workaround, for the moment - but in terms of being able to create a deliverable product that ticks every box that I consider important, no.

[lmitche]

Just now, rickca said:

A $5 clock radio?

Rick, you are the best straightman ever!

[gmgraves]

1 hour ago, fas42 said:

 

Religion is an early form of science - it attempts to explain why things are as they are ... such as, why is the universe there. As far as I'm aware, current science still hasn't quite got a secure hold on the latter ...

Religion has always been what it still is, an attempt by one group of people (priests) to hold dominion over the hoi-polloi. It has nothing to do with science. Using religion to explain natural events like earthquakes, storms, thunder and lightening, floods, and the stars in the sky are merely tools to reinforce the power of the gods and further ingrain in the populace, by association, the power of the priestly class. IOW, one ignores or crosses the priests at one's peril because they say that they can summon the wrath of the gods.

 

OTOH, science is Deus ex machina. Early scientists like Archimedes, Hippocrates, Pythagoras and Socrates looked for logical explanations for natural phenomena and how to use the natural world to improve life. They did not sit around invoking the gods to their bidding. Problems like measuring the volume of irregular shapes, moving water uphill, mechanical advantage, the shape of the earth and it's circumference are all the work of early science and with nary a god in sight. The Hippocratic Oath invokes no gods in order to heal, it relies solely on the expertise of the physician to prescribe remedies found in nature through diet, exercise, and herbalism.

[lmitche]

27 minutes ago, Superdad said:

I should have said, "which forked tongue?"

 

[danadam]

30 minutes ago, Jud said:

 

Sorry, the brain's pattern identification mechanisms do not work in this nice, neat way.  Otherwise instruments, which are far more sensitive to small differences, would inevitably be better at pattern identification.  For example, instruments can detect things far smaller or further away than our eyes can see, but we are better at recognizing faces. 

 

Ignoring rather than detecting differences may also be key to pattern recognition, as with recognizing a disguised person or recognizing that two different instruments in two different recordings are both violins, or both Telecasters.

Do you have any example of a case where some measurement says that 2 images should be indistinguishable to human eye and yet people can distinguish them due to their better pattern recognition?

[pkane2001]

28 minutes ago, Jud said:

 

Sorry, the brain's pattern identification mechanisms do not work in this nice, neat way.  Otherwise instruments, which are far more sensitive to small differences, would inevitably be better at pattern identification. 

 

Physical property detection and pattern identification are very different things. One is limited by the sensitivity of the instrument, the other by the complexity of an algorithm.

 

Instrument sensitivity has nothing to do with its ability to do pattern detection, except if it’s unable to measure a difference. If there is no measured difference then no pattern detection is possible.

 

[gmgraves]

1 hour ago, mansr said:

What's the difference? The idea that humans, of all animals, can hear things that elude the finest scientific instruments is nothing short of preposterous.

 Do you actually believe that? or are you just waxing contrariness in order to create controversy? While I fundamentally agree that concrete, measurable data is needed to explain why certain things sound the way that they do, I also know that some things that can be measured make no difference to the sound of an audio component, and some things that can be heard are simply not represented by the measurements. For instance, you have two amplifiers that measure, identically. They have the same power, the same frequency response and both have vanishing levels of distortion. On paper, there should be no discernible difference between the two amps, yet listening shows that they sound distinctly different from one another. How do you measure that? The differences, it turns out, might be in the power supplies or the types of coupling and bypass capacitors used in each, Those things won't show up in the standard measurement suite, yet they are easily heard. 

I'm all for measurements, they are crucial, but just as crucial is a discerning ear. Measurements might tell you "why" but you ears will almost always tell you "what". I'm not forgetting the role that expectational and confirmational bias plays in all this, that's why when the measurements tell you one thing and your ears tell you something completely different, caution must be taken to avoid being fooled. But clearly to me, both measuring and listening are important.  John Atkinson's technical review of the Schiit Yggdrasil said that the thing didn't perform very well. Had I heeded that review, I wouldn't own one now. But, I have compared it to DACs such as the MSB Diamond DAC IV, the Benchmark DAC3 DX, The dCS Vivaldi, the AudioGD R2R7, etc. And NO DAC that I have tried has come within a country mile of the Yggdrasil in terms of overall musicality. 

[elcorso]

 

"If dogs bark, it's because we ride" Said Don Quixote to Sancho Panza, his humble squire.

But he should have added "and even the deaf dogs ..."

 

Have a good night and goodbye ?

 

Roch

 

PS / By the way, nobody answered my simple questions ...

[AudioDoctor]

On 9/18/2018 at 9:52 AM, wgscott said:

Sounds like the first element on the periodic table.

 

Hydrogen has been replaced with Bullshit?  Is it's Atomic Symbol Bu or Pu?

I am so fucking clever...

[jabbr]

1 hour ago, Jud said:

 

Sorry, the brain's pattern identification mechanisms do not work in this nice, neat way.  Otherwise instruments, which are far more sensitive to small differences, would inevitably be better at pattern identification.  For example, instruments can detect things far smaller or further away than our eyes can see, but we are better at recognizing faces. 

 

These are two different levels of the "stack"

 

Consider:

1) an image sensor

2) a software program that operates on an image file

 

Facial recognition is typically implemented in a software program e.g. facebook, which has no access to the original image sensor.

 

In any case that's how it works in the brain. The retina is the sensor, the cortex performs high level pattern recognition.

 

Quote

 

Ignoring rather than detecting differences may also be key to pattern recognition, as with recognizing a disguised person or recognizing that two different instruments in two different recordings are both violins, or both Telecasters.

 

Right. The pattern recognizer is a classifier, which groups disparate signals together into classes aka patterns.

 

Regarding electronic signals, an oscilloscope allows you to visualize an electronic signal as a graph. Although the scope may also contain software to manipulate, transform and apply statistics to the signal, such software is not exhaustive.

 

Consider a 10 Gsps scope which realistically might be require to capture a high speed digital audio signal (and 50 Gsps+ scopes exist). Such a scope can product an overwhelming amount of data for an audio track. Such data indeed captures any reasonable difference in a digital audio signal yet the analysis/pattern recognition to analyze such data just isn't there compared with the brain.

 

Now also consider the amount of raw data produced by e.g. the Hubble Space Telescope and then consider the application of truly "big data" statistics and analysis software in modern science... yes Astrostatistics is a thing! https://en.wikipedia.org/wiki/Astrostatistics which goes waay beyond the telescope itself.

 

[fas42]

19 minutes ago, gmgraves said:

Religion has always been what it still is, an attempt by one group of people (priests) to hold dominion over the hoi-polloi. It has nothing to do with science. Using religion to explain natural events like earthquakes, storms, thunder and lightening, floods, and the stars in the sky are merely tools to reinforce the power of the gods and further ingrain in the populace, by association, the power of the priestly class. IOW, one ignores or crosses the priests at one's peril because they say that they can summon the wrath of the gods.

 

You're confusing the https://en.wikipedia.org/wiki/Iron_law_of_oligarchy

with https://en.wikipedia.org/wiki/Theories_about_religions. Pecking order behaviour doesn't correspond with the desire to understand, but if the latter becomes institutionalised then it forms excellent starting blocks for the ambitious.

 

[mansr]

24 minutes ago, gmgraves said:

in terms of overall musicality

What does that even mean?

[Ralf11]

musicality is like euphonicity, but in drag