Head-Fi.org › Forums › Equipment Forums › Sound Science › Testing audiophile claims and myths
New Posts  All Forums:Forum Nav:

Testing audiophile claims and myths - Page 93

post #1381 of 2862
Quote:
Originally Posted by kiteki View Post

- 99.99% is filtered (source?)

 

From the work of James "JJ" Johnston.  You could check out some of the stuff he posts on his blog occasionally.

post #1382 of 2862

Oh yeah, I remember him talking about "steered" in that video, that you hear what you focus on.

 

Sighted listening is flawed from an evidence POV, I don't think anyone is contesting that.  In order to have evidence that capacitors sound different you need to blind test them.  Sighted listening is flawed but that doesn't somehow = everything sounds the same.

 

Correlation != causation etc.

post #1383 of 2862

Okay, I read the links now

 

"This can be demonstrated by the old “backward masking” demonstration that was originally brought up in a different context, where it was alleged that there was satanistic content to part of the song “Stairway to Heaven” when it was played in reverse. Interestingly enough, when the reversed song is played to an unsuspecting audience (this has happened quite a few times in lectures), the audience hears nothing, or a very few random syllables. When, however, the “words” are presented, the audience hears the words clearly and correctly, even though the actual sound presented to the ear is not changed. This is a direct result of how we understand speech, when expecting speech with some particular content, we guide our feature extraction and object resolution to find the parts of our audio surroundings in order to extract the information, and in this case can do so even when there is no such information."

 

 

Exactly.

 

This is why so many blind tests fail with random audiences.

 

If you don't know what to listen for in a complex ocean of sound, you won't hear it.  In blind testing, you have to be familiar with the equipment, the content, and know what to listen for, if you don't there's little chance you'll hear anything, just like the audience didn't when played stairway to heaven backwards until they were told what to look for.

 

Yet another example defending the subjective sighted listening position.

 


 

Sortof joking again.  Just pointing out you can twist pseudo-scientific implication facts however you want.

 


 

Edit:

 

"The partial loudnesses from the cochlea are integrated somewhere at the very edge of the CNS such that some memory of the past is maintained for up to 200 milliseconds. Level Roving Experiments show that when delays approaching 200 milliseconds exist between two sources, the ability to discern fine differences in loudness or timbre is reduced. It is well established you need very quick, click-less switching between signals when trying to detect very small differences between signals, otherwise you lose part of your ability to distinguish loudness differences."

 

Ok now I finally know where all this 0.1 second ABX switch parroting is coming from.

 

Sorry but the ABX switch time interval is completely up to the participant, they should be able to choose between 0.1 seconds or 20 seconds, it's entirely up to the participant.

 

As soon as you say stuff like it has to be under 0.2 second switching, the blind test procedure has confined itself to that variable which is dependant on seperate findings.

 

It's like saying I'll blind test two cars for you but only in a car parking lot.

 

If the participant can blind identify which sports car is which, they should be able to drive the cars however they want, hm?


Edited by kiteki - 6/7/12 at 8:37pm
post #1384 of 2862
Quote:
Originally Posted by kiteki View Post

Okay, I read the links now

 

"This can be demonstrated by the old “backward masking” demonstration that was originally brought up in a different context, where it was alleged that there was satanistic content to part of the song “Stairway to Heaven” when it was played in reverse. Interestingly enough, when the reversed song is played to an unsuspecting audience (this has happened quite a few times in lectures), the audience hears nothing, or a very few random syllables. When, however, the “words” are presented, the audience hears the words clearly and correctly, even though the actual sound presented to the ear is not changed. This is a direct result of how we understand speech, when expecting speech with some particular content, we guide our feature extraction and object resolution to find the parts of our audio surroundings in order to extract the information, and in this case can do so even when there is no such information."

 

 

Exactly.

 

This is why so many blind tests fail with random audiences.

 

If you don't know what to listen for in a complex ocean of sound, you won't hear it.  In blind testing, you have to be familiar with the equipment, the content, and know what to listen for, if you don't there's little chance you'll hear anything, just like the audience didn't when played stairway to heaven backwards until they were told what to look for.

 

Yet another example defending the subjective sighted listening position.

 

 

See the bolded text. The point of this example is that, once the "words" were presented, the audience immediately heard what they were told to look for, regardless of whether it was in there or not. It is not "defending" the subjective position - it is damning it. 

post #1385 of 2862

If the difference between two systems is slightly enhanced soundstage between 50Hz-100Hz on one system and normal soundstage on the other, no one will ever find it unless you tell them exactly what to look for.

 

If you could hear it in 0.1s ABX, it'd have to be the right piece of music switched at the right time.  Even then, your mind could easily filter it out in the switch, since it's "locked on" to a certain kind of sound.  So, an XXXX versus YXYX test could easily be more effective in some cases, and should be up to the participant.

post #1386 of 2862
Quote:
Originally Posted by kiteki View Post

If the difference between two systems is slightly enhanced soundstage between 50Hz-100Hz on one system and normal soundstage on the other, no one will ever find it unless you tell them exactly what to look for.

 

If you could hear it in 0.1s ABX, it'd have to be the right piece of music switched at the right time.  Even then, your mind could easily filter it out in the switch, since it's "locked on" to a certain kind of sound.  So, an XXXX versus YXYX test could easily be more effective in some cases, and should be up to the participant.

 

I don't recall ever specifying a specific timeframe for ABX. Or even that ABX is the only method to use. I'm fine with participant directed changes.

 

However, the point of the above was that if you tell someone to listen for an expanded soundstage at a certain point, or in a certain range, the odds are good that they'll report hearing it - even if you are playing the same thing both times, with no actual change for them to hear. If you direct them to hear something, you immediately undermine the test - any data whether positive or negative is now unreliable.

post #1387 of 2862
Originally Posted by liamstrain View Post

 

However, the point of the above was that if you tell someone to listen for an expanded soundstage at a certain point, or in a certain range, the odds are good that they'll report hearing it - even if you are playing the same thing both times, with no actual change for them to hear. [/]

 

Using schizophrenics in examples over and over isn't very useful.  I'm sure they see detail in blu-ray which isn't there too, while someone else needs their prescription glasses to see any difference in blu-ray at all.

 

 

Originally Posted by liamstrain View Post

If you direct them to hear something, you immediately undermine the test - any data whether positive or negative is now unreliable.

 

If you say so.

post #1388 of 2862
Quote:
Originally Posted by Head Injury View Post

You do that in every post.

 

It's true we filter out information all the time. Human memory performs at different levels. Only the information that survives beyond sensory memory and into short term memory is of any use beyond reaction. The rest is filtered out and we don't make use of it.

 

There's a psychological aspect as well. We only tend to remember what matches our preconceptions or expectations, our schemas. So in the case of high end audio, we're more likely to remember everything that matches what we expect to hear from a component. So with an expensive cable, we'll remember the positive, and with a cheap cable we'll remember the negative. That's if we have a preconceived notion that the high end cable will sound better. It ties into expectation bias.

 

I don't know where he got the specific number, but it's most likely he didn't mean it as factual and instead as a generalization. Why don't you try to message him yourself?

 

This is a real problem which cannot be escaped 100% - but from my experience expectations this work both ways - if you are expecting apiece of equipment to sound better in a certain way and it does not, then what you hear will contrast with what you are expecting to hear - we need our schemas to make sense of what we perceive but I strongly disagree that they override any possiblity of objective evaluation within a particular category.  For example you drink a cup of coffee and you are expecting it to have sugar in it - you can still notice the absence of sweetness relative to what you are expecting.  

 

The mechanism you mention of sensory memory being filtered before being retained into short term memory is particularly interresting, but as Kiteki posted it works both ways in that if we are not sure what we are listening for it can it very difficult to extract any useful information from what we perceive, especially to contrast very similar audio cues which must be noticed and also evaluated.  For this reason, along with others, I think that blind testing is of limited use in high end audio - and that measuring performance with lab eequipment is a much more useful tool for evaluating equipment performance.

 

As for sighted testing, I think it is of some personal value and can always be compared against measured performance - but I definitely don't think blind testing is reliable enough evidence to serve as a blanket body of knowledge with which to challenge anecdotal observatoins.  Is however useful in demonstrating that under blind test conditions we cannot reliably discern differences between said equipment.  In addition it is not reasonable to expect people to be supermen who ardently follow the results of their experimentation.  For example if you move into a house and see ghosts, and you don't see ghosts when you are not in this house, then apart from taking anti-psychotic medication you may also consider moving houses?  What if you decide to take the meds but still see ghosts?  What if someone else moves into the house independently and also claims to see ghosts?  You can call them crazy or idiots , or send them to a series of lectures but this probably won't help.  The underlying causes existed long before the diagnosis, and will persist long after.

 

My point is if there is some kind of subconscious agent that is causing us to perceive something on a regular basis - this is a real problem from a psychological standpoint, and the best course of action is not necessarily to put everyone on medication or send them to psychotherapy.  Of course I use the example of ghosts, because many here consider certain differences in audio equipment to be just that - phantoms - but this example unfairly presupposes that this is in fact the case.

 

To clarify this - intangible things can have tangible value, and cannot be dismissed out of hand as a disease which can be cured.  There is also the possibility that behind the intangible, there is tangible material that has yet to be measured.


Edited by drez - 6/8/12 at 12:02am
post #1389 of 2862

the 200ms limit is from the processing in the dedicated audio neural structures - very specific to fine loudness, timbre distinctions

 

the question of short term/long term/working memory limits, decay, capacity are for after the features have been encoded, reach the brain itself

 

this short summary article looks plausible to this non specialist, has good provenance: http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2657600/ - very little hard info, only times I could find were 5-10s as possible "short term" memory limit - additionally there is the concept of limited capacity for different "items/features" in working memory - as few a 4, possibly a few more with training

 

this suggests training, and deliberate focus are useful for discrimination, <10 seconds clips, switching may use more of the brain's "short term" capacity

 

this doesn't preclude longer term listening where hypothetical "more complex" judgments, perceptions, but possibly along fewer "aural dimensions" can transfer to "long term" memory

 

but there appears to be a case for relatively short clips, switching times being useful for discriminating some features of "the sound"


Edited by jcx - 6/7/12 at 11:38pm
post #1390 of 2862

Originally Posted by drez View Post

 

For example you drink a cup of coffee and you are expecting it to have sugar in it - you can still notice the absence of sweetness relative to what you are expecting.

 

[/]

 

Exactly.  If you usually expect coffee with normal sugar and one day it has sweetener, you notice it.  If you ordered a Coke and you are served Diet Coke in a Coke glass you notice it.

 

If an identical twin calls you pretending to be the other one, you notice it.  If you buy a $1000 fake watch in Thailand and think it's real, you think it's real.  Are these four cases related at all?

 

Likewise, if a schizophrenic saw extra detail in blu-ray, which wasn't there, it doesn't mean I'm going to stop reading blu-ray reviews all of a sudden, or think it's the same as DVD.

 

??

 

Originally Posted by drez View Post

 

For this reason, along with others, I think that blind testing is of limited use in high end audio - and that measuring performance with lab eequipment is a much more useful tool for evaluating equipment performance.

 

Lab equipment is entirely on a case by case basis.  In acoustic audio $10,000+ lab equipment is still completely useless compared to human perception in some areas.  It's very dependant on which specific case you're looking at.

 

Anyway, you don't need visuals to listen to audio / music.  There is nothing visual in audio.  Blind evaluation should be used a hell of a lot more than it is.  Blind testing is OK if performed correctly.

post #1391 of 2862
Originally Posted by jcx View Post

 

this short summary article looks plausible to this non specialist, has good provenance: http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2657600/ - very little hard info, only times I could find were 5-10s as possible "short term" memory limit - additionally there is the concept of limited capacity for different "items/features" in working memory - as few a 4, possibly a few more with training

 

[/]

 

but there appears to be a case for relatively short clips, switching times being useful for discriminating some features of "the sound"

 

I believe 0.1 second switches will be useful for faint volume differences, if you took a break for 10 seconds, it must be impossible to hear faint differences in volume, at least that's how it seems to me.

 

Volume is directly intertwined with frequency response.

 

There are countless other sounds to look for though, which could hypothetically only link to long term memory, in which case the rapid switch could only be confusing, or elicit an illusion of sameness.

 

When people are trying two different perfumes, they don't rapid switch, they take a break, etc.

 

We can discuss it until the horses come home.  In the end if someone thinks X device sounds different than Y, they can blind test it in any way they like.

 

 

If you're testing vast numbers of random people, you need a very revealing testing setup, and you need to tell them what to look for in the differences (if you can), and you need to take aside the positive results and re-test them.

 

If you want to test for something like different fluroescent light hertz rate flicker, or UVA versus UVC light, it would be impossible in an ABX, but you may find it in a blood test, there is always that too.

post #1392 of 2862

Much of this has to do with motivation, but some sounds get dampened, but.. when getting into the elusive realm of those warm euphoric harmonies of high end all hell breaks loose. To go about it in a flowery way is not very productive, so I'll point to some interesting research.

 

 

 

Quote: http://www.neuroscience-of-music.se/eng2.htm
 
Before 1992 it was unknown that one of the major functions of the mammalian inner ear is to selectively damp the loud frequency components of sound. Without this function we would not be able to follow normal speech, and music would be a torture rather than a pleasure.

That our ears are frequency selective amplifiers of weak sound was first suggested by Thomas Gold in 1948 and was finally proved in the 1980s. We now also know that there is a complement of low-level amplification: high-level damping.

The border between amplification and damping lies at ca 60 dB SPL, which is the typical mean sound level for normal speech. This means that weak spectral components in normal speech are amplified already in the inner ear, and strong spectral components in normal speech are damped already here.

The mechanism of frequency selective damping consists of a flexible membrane, which runs along the full length of the coiled inner ear and is set into vibrations by sound levels above 60 dB. The vibrations absorb high-level sound energy and take it out as thermal energy, which then can no longer distort or mask the sensation of weak sound components or even act as a stressor for the delicate sensory cells.

Due to the specific anatomy of this so-called basilar membrane, each sound frequency has its own place of maximal vibration. This enables the inner ear to amplify some frequencies while at the same time, but at different places, damping others.

An example: A complex sound has 70 dB at 300 Hz, 50 dB at 500 Hz, and 80 dB at 800 Hz. The weak second component is amplified, whereas the other two components are damped. Without their damping we would hear the second component much worse or not at all.

 

 

 

Quote: http://www.neuroscience-of-music.se/eng6.htm
 
Before 2000 it was unknown that a precognitive version of absolute pitch is general in humans. This ability is subconsciously applied in speech and it can be considered as a further means of speaker identification, in addition to voice timbre.

Application of the precognitive, absolute pitch-memory increases with speech emphasis. This is neuroanatomically plausible, because the major precognitive command center for speech (in the anterior cingulate cortex) is not only linked to a pitch memory but also to centers for arousal and emotion.

The motivation dependent use of the absolute pitch-memory can be assumed as one of the evolutionary foundations of song, and thus of music in general.

 

 

 

Quote: http://www.neuroscience-of-music.se/eng7.htm
 
Before 2004 it was unknown if the human brain is hard-wired for the perception of octave circularity. In all advanced musical cultures the names for tones in scales are repeated, when the octave interval (frequency ratio 2:1) is reached. For example, in the European music system, the octave-spaced tones of 110, 220, and 440 Hz are all called A. Only in technical descriptions an additional octave number is added, like A2, A3, and A4. The universal practice of tone name circularity indicates an equally universal circularity in pitch perception. Octave circularity in pitch perception was also observed in the monkey.

The fiber network between the neuron layers is such that all signals triggered by octave-spaced tones, such as A2, A3, and A4, are likely to be pooled. They could then be trans-coded into an additional signal for a general A. This would explain why all tones called A, regardless from which octave, have a common pitch quality, a so-called pitch chroma. Thus, the auditory thalamus can be considered as the anatomical basis of our internal chroma map.

Evolutionary advantage: Complex harmonic sounds, such as animal calls, vowels, or violin tones, easily become octave ambiguous. If one of partials 1, 3, or 5 has a low amplitude in the sound spectrum, partials 2, 4, 6 and 8 can be mistaken by the pitch extracting mechanism as partials 1, 2, 3, and 4 of a different complex sound, whose pitch is one octave higher. This is the reason why even experienced musicians easily make octave errors. Apparently octave errors were irrelevant during the evolution of hearing. However, animals gained in pitch stability within a one-octave frame by simply collapsing all simultaneous pitch candidates, as determined in the auditory midbrain, onto one single and general octave frame at a subsequent stage of processing, i.e., in the auditory thalamus.

Data on pitch shift caused by the medical drug carbamazepine and on statistics of inner-ear tones (SOAEs) now revealed the first physiological evidence of hard-wired octaves in the human brain.

 

Edit: Typo.


Edited by Albedo - 6/8/12 at 12:34am
post #1393 of 2862
Originally Posted by Albedo View Post

Before 1992

 

in 1948

 

in the 1980s.

Before 2000 
Before 2004

 

in 2007

 

in 2014

 

in 2025

 

in 2070

 


 

All we can say with any certainty is that audio science is very limited and incomplete in 2012.  Statements to any other effect are only self-induced illusion.

 

There is no 100% perfect audio playback system in 2012 either, that's pretty certain too.


Edited by kiteki - 6/8/12 at 12:51am
post #1394 of 2862
Quote:
Originally Posted by drez View Post
The mechanism you mention of sensory memory being filtered before being retained into short term memory is particularly interresting, but as Kiteki posted it works both ways in that if we are not sure what we are listening for it can it very difficult to extract any useful information from what we perceive, especially to contrast very similar audio cues which must be noticed and also evaluated.

 

Blind ABX testing can work well for verifying the existence of an alleged concrete difference, which can be previously discovered in any way, sighted or not. There is usually no shortage of those, I see many claims of gear having "better sound stage", "more impactful bass", "improved details", "less sibilance", and more being made with great confidence. Statements like these would imply already knowing exactly what to listen for ?

 

Quote:
Originally Posted by drez View Post
As for sighted testing, I think it is of some personal value and can always be compared against measured performance - but I definitely don't think blind testing is reliable enough evidence to serve as a blanket body of knowledge with which to challenge anecdotal observatoins.

 

What reliability advantage does sighted listening have over an otherwise identical, but blind test (assuming that the "not knowing what to listen for" issue is already dealt with), other than being more likely to produce the - more desirable to audiophiles - positive result due to its own flaws ? Two wrongs do not make a right.

post #1395 of 2862
Quote:
Originally Posted by Albedo View Post

Much of this has to do with motivation, but some sounds get dampened, but.. when getting into the elusive realm of those warm euphoric harmonies of high end all hell breaks loose. To go about it in a flowery way is not very productive, so I'll point to some interesting research.

 

These are interesting, but do not have much to do with how "golden" the hearing is, rather than how "smart" it is. The effects described above still work on audio that is downmixed to mono, dithered to 8 bits, and downsampled to 22050 Hz. They show how intelligently the brain can process complex high level information, but audiophilia is about obsessive micro-optimization of the low level performance of audio playback.

New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Sound Science
Head-Fi.org › Forums › Equipment Forums › Sound Science › Testing audiophile claims and myths