Who's afraid of DBTs

[trithio]

Somewhere earlier in the thread someone made the comment that not all auditory memory is "short term" using the example of a mothers/fathers/siblings/spouse/child's voice. Clearly we do retain enough memories of those voices that even after a few words over a poor phone connection we can recognize that voice. But ask us to describe why that voice is different from one that we recognize as not our mother etc., and we have a much harder time saying exactly why.

 

So if someone is in the audio business and, as such, is constantly listening to live music from all sorts of instruments, why couldn't that person form the same sort of memory about what a real violin, a guitar, a cymbal, a trumpet, etc. sounds like that persists as much as recognizing a mothers voice?

 

If you agree that is possible, then what Peter is in effect doing over a number of days is listening to all sorts of music and doing a mental check of how often he hears something that is "not right" when compared to his memory of the live instrument. At any given point, that might be the fault of the recording, not the playback, but, if after a number of days the number of "errors" is less than before it would seem that a legitimate improvement was made.

 

On the other side of the equation, if as you listen over those days, a particular recording keeps sounding "off," then you eliminate that recording from the test set. Again, over time, you begin to build the set of recordings that consistently sounds most like the real instrument.

 

Peter further acknowledges that a given tweak might, in an instantaneous ABX test set, make a violin sound better than before, but, until you have also heard that tweak's effect on the trumpet, the cymbal, the drums, etc. you don't really know if it was an overall improvement.

 

All of that makes intuitive sense to me. To prove that it cannot be so, I think you would have to show that we cannot ABX our family's voices against those of others and thus prove that there cannot be an extended memory of audio sounds.

 

You do raise many valid concerns. But there are a lot of potential apples and oranges there.

The differences between people's voices are actually very big. Not sure how much is measurable but the differences are in the same class as for example differences between detailed-to-shrill ss amps and very smooth tube amps. Both categories of differences are very easily heard by almost everyone. Just record two voices with the same mike and I bet you can almost instantly ABX between them.

 

The differences between two ss amps who measure the same are in a different league. That is IF they even exist or are audible. The sound of those two amps will have the same timbre and generally 'feel' the same.

 

The point about musicians and music experts. Yes a violin player would be very good at recognizing the sound of his violin. But his violin will sound quite different in his concert hall than in a recording of that. Go to a live acoustic concert and then buy the cd recording of that performance. You may be quite surprised. Will the musician be any better at identifying differences between two amps playing his violin? Maybe. Maybe not. For example some say that musicians are actually very bad at it. They are so used to the original live sound of their instrument that it doesnt matter what recording or equipment they hear. Their brain compensates for the differences and they actually always 'brainhear' the original live sound. Will be a potentially interesting experiment though. Anyone with more input on this?

 

What Peter & others do over X days may sound thoughtful and well planned at first. But as mentioned in another thread we hear with our brains more than with our ears. Your mood affects what you hear. Mightly. A song sounds very different when you are happy and when you are upset. And over 5 days you will go through thousands of moods. It's impossible to keep an objective and consistent hearing experience that long. You also have big variations in temperature, humidity and air pressure. All affect the soundwaves. A hearing test over days is pretty much the definition of an unmeasurable mess. Simply impossible to deal with the enormous amount of variables.

Those issues are very well adressed by short DBTs. Dont get me wrong, never said DBTs were perfect. They are just better than pretty much any other empirical method of comparison. At least to my knowledge. And you are welcome to expand that!

[Jud]

What Peter & others do over X days may sound thoughtful and well planned at first. But as mentioned in another thread we hear with our brains more than with our ears. Your mood affects what you hear. Mightly. A song sounds very different when you are happy and when you are upset. And over 5 days you will go through thousands of moods. It's impossible to keep an objective and consistent hearing experience that long. You also have big variations in temperature, humidity and air pressure. All affect the soundwaves. A hearing test over days is pretty much the definition of an unmeasurable mess. Simply impossible to deal with the enormous amount of variables.

 

Those issues are very well adressed by short DBTs. Dont get me wrong, never said DBTs were perfect. They are just better than pretty much any other empirical method of comparison. At least to my knowledge. And you are welcome to expand that!

 

Two things for you to consider:

 

(1) Is being a chef impossible? After all, over the many days you are a chef you will go through thousands of moods. The season of the year, your co-workers' clothing and perfumes, what else is being cooked at the same time, make it impossible to keep a consistent and objective smelling and tasting experience that long. It all affects the tastes and smells. It's pretty much the definition of an unmeasurable mess. Simply impossible to deal with the enormous amount of variables.

 

Or does knowledge, training and experience help to make it possible?

 

(2) Peter's explained that discerning short-term differences is not what he's after (and why not). Is it possible that because DBT is your "hammer," all tasks to be accomplished in audio look like "nails"?

[trithio]

Two things for you to consider:

 

(1) Is being a chef impossible? After all, over the many days you are a chef you will go through thousands of moods. The season of the year, your co-workers' clothing and perfumes, what else is being cooked at the same time, make it impossible to keep a consistent and objective smelling and tasting experience that long. It all affects the tastes and smells. It's pretty much the definition of an unmeasurable mess. Simply impossible to deal with the enormous amount of variables.

 

Or does knowledge, training and experience help to make it possible?

 

(2) Peter's explained that discerning short-term differences is not what he's after (and why not). Is it possible that because DBT is your "hammer," all tasks to be accomplished in audio look like "nails"?

Nice mock anecdote to my earlier mock anecdote. Be careful, I heard this stuff might get you banned

 

1. thats exactly why two meals by the same chef never taste exactly the same. A fact-of-life which seems to support my point 100% and 0% yours. Stoopid, annoying facts

 

2. that hammer. Or it might be the best known comparison tool. As usual you are welcome to name a better one.

And no, I do not agree with Peter's arguments about testing. Not at all. But that feels like trithio & jud having a gossip about peter. Not my idea of the five oclock tea

[goldsdad]

One issue I find aggravating it is nearly impossible to insert mathematical symbology here, so everything has to be explained pretty much in English. That takes a enormous quantity of words, when a half dozen symbols can sometimes tell the whole story exactly.

 

An obvious solution is to use a formula editor then post an image here in the forum.

 

An example taken from Wikipedia:

 

[Jud]

An obvious solution is to use a formula editor then post an image here in the forum.

 

An example taken from Wikipedia:

 

[ATTACH=CONFIG]16588[/ATTACH]

 

Owen, how would one create the image from the formula?

[Jud]

Nice mock anecdote to my earlier mock anecdote. Be careful, I heard this stuff might get you banned

 

1. thats exactly why two meals by the same chef never taste exactly the same.

 

And two listens to the same recording never or always sound the same?

[goldsdad]

Owen, how would one create the image from the formula?

 

The editor should have an image export function, but if not, just grab a screenshot.

[trithio]

And two listens to the same recording never or always sound the same?

 

According to pretty solid research into the echoic memory subject, if you wait more than about 5secs between 'listens' you'll never know the answer to that question. You cant remember sounds longer so you simply cannot compare.

Even worse. I dont remember right now how big that echoic memory storage was. But pretty sure it is smaller than one typical song. So, according to currently mainstream audio memory theory, your audiobrain is simply not capable of doing that kind of comparison with full songs. Tough luck I guess.

 

You should still be able to reliably compare small snippets in a fast switching environment. That would be the ABX methodology.

[Paul R]

Owen, how would one create the image from the formula?

The only way I could find to make that work was creating either JPEG or GIF images, and that sure was clumsy, compared to something more familiar like TEX.

-Paul

[sdolezalek]

The point about musicians and music experts. Yes a violin player would be very good at recognizing the sound of his violin. But his violin will sound quite different in his concert hall than in a recording of that. Go to a live acoustic concert and then buy the cd recording of that performance. You may be quite surprised. Will the musician be any better at identifying differences between two amps playing his violin? Maybe. Maybe not. For example some say that musicians are actually very bad at it. They are so used to the original live sound of their instrument that it doesnt matter what recording or equipment they hear. Their brain compensates for the differences and they actually always 'brainhear' the original live sound. Will be a potentially interesting experiment though. Anyone with more input on this?

 

Trithio: First, thanks for the thoughtful reply! Second, coming from a family of concert musicians (piano, violin and cello) I do wholeheartedly agree that they seem to have little appreciation for the differences in recorded sound (they must be listening for something else, but what that is, I don't know). But there are those of us who do spend lots of time listening to all sorts different equipment, and I do believe that the more you do of that, the more you can develop a sense for colorations and other distortions. Whether that "informed sense" can then be used effectively in an ABX test, I don't know, but that typically isn't what it was developed for, so I doubt that would be its best use.

 

There is a new review out in TAS by Anthony Cordesman on the $98,000 a pair Boulder 2150 Mono Power Amplifier.Boulder 2150 Mono Power Amplifier and 2110 Preamplifier | The Absolute Sound I found one of his quotes very interesting:

 

"Here’s the issue: With equipment this good—and this neutral and transparent—almost all of the colorations you hear come from the recording, the front end, the interconnects and speaker cables, the load presented by the speaker, and the complex room interactions that shape the sound of the speaker at a given listening position."

 

I think that further complicates the whole ABX issue because it implies that we have arrived at a level of sonic improvement where the interactions between equipment may be as or more important than the equipment itself. Put differently, even if I can hear a difference between Amp A and Amp B in an ABX test, is what I'm hearing the difference between the amps or the difference in how each amp interests with a) the signal it is fed and b) the speaker load it is driving and (maybe) c) the cables between them?

[goldsdad]

The only way I could find to make that work was creating either JPEG or GIF images, and that sure was clumsy, compared to something more familiar like TEX.

-Paul

 

 

My mistake. I thought exporting an image seemed less clumsy than typing an "enormous quantity of words, when a half dozen symbols can sometimes tell the whole story exactly."

[esldude]

I've got just a bit of time here, and wanted to make one other quick comment.

 

DBT is often seen (quite rightly) as a way to counter the effects of expectation bias. But what if expectation bias isn't necessarily the problem?

 

Let's look for example at a claim that's occasionally engendered some controversy, that different USB audio cables sound different, especially feeding a DAC with async USB input. I believe I've heard such differences. Here are the approximate prices, apparent construction quality, and my sound quality preferences (worst to best) for each of the different USB cables I've heard in my system:

 

Furutech GT2 - ~$110 - Excellent apparent construction quality (Worst SQ by far)

 

Wireworld Starlight - ~$80 - Excellent apparent construction quality

 

Audioquest Forest - ~$30 - Good apparent construction quality

 

Audioquest Carbon - ~$110 - Good apparent construction quality

 

Audioquest Coffee - ~$275 (I bought it for less) - Excellent apparent construction quality, very "technical looking"

 

Mapleshade Clearlink Plus - ~$130 - Poor to fair apparent construction quality, extremely thin conductors inside what looks like a piece of a woman's white mesh stocking (Best SQ by far) (Mapleshade has apparently tried to make the cable look a little better by using black rather than white mesh now)

 

There's no pattern of perceived sound quality here with regard to cost (except within the Audioquest line), nor with regard to apparent construction quality. The perceived sound quality did not match my expectations, nor others' observations I'd read. (From judgments I'd read by others, I expected the Furutech and Wireworld to sound better than the Audioquest Forest I had at the time. The Wireworld wasn't bad, but not as good as the less than half as expensive Forest. I did not at all expect the Furutech, with many good user reviews and at four times the price of the Forest, to sound purely awful. I hadn't read any user reviews of the Mapleshade cable; though I expected it to sound good for the price, I did not at all expect it to simply blow the AQ Coffee completely out of the water for perceived sound quality.)

 

So why am I posting these anecdotes - yes, that is absolutely all they are - here? After all, I've noted that my listening experience with the Wireworld and Furutech cables was completely at odds with other listeners' impressions, so there's no universal truth to be found through my sighted observations.

 

It's this: Whatever is the cause of this "disease" (anecdotal impressions by various listeners all over the lot), expectations isn't it. My listening impressions flouted my expectations multiple times, and particularly with regard to the cables I thought sounded worst and best by far. If expectations don't cause the "disease," then is DBT the "cure"?

 

Why yes DBT is the cure of exactly this. If there was some real quality to USB cable sound difference, you might expect more agreement. If there is not, the you might expect chaos. Your anecdotal results compared to others are chaotic or more precisely random. You did have an expectation bias. You were influenced to believe they could sound different. They did sound different. Your ranking was not in line with other people's. Quite possibly because there is no difference of the signal itself.

 

If the signal difference is really audible, you would expect to confirm that in blind testing, even if your preference differs from those of other people.

[esldude]

Somewhere earlier in the thread someone made the comment that not all auditory memory is "short term" using the example of a mothers/fathers/siblings/spouse/child's voice. Clearly we do retain enough memories of those voices that even after a few words over a poor phone connection we can recognize that voice. But ask us to describe why that voice is different from one that we recognize as not our mother etc., and we have a much harder time saying exactly why.

 

So if someone is in the audio business and, as such, is constantly listening to live music from all sorts of instruments, why couldn't that person form the same sort of memory about what a real violin, a guitar, a cymbal, a trumpet, etc. sounds like that persists as much as recognizing a mothers voice?

 

If you agree that is possible, then what Peter is in effect doing over a number of days is listening to all sorts of music and doing a mental check of how often he hears something that is "not right" when compared to his memory of the live instrument. At any given point, that might be the fault of the recording, not the playback, but, if after a number of days the number of "errors" is less than before it would seem that a legitimate improvement was made.

 

On the other side of the equation, if as you listen over those days, a particular recording keeps sounding "off," then you eliminate that recording from the test set. Again, over time, you begin to build the set of recordings that consistently sounds most like the real instrument.

 

Peter further acknowledges that a given tweak might, in an instantaneous ABX test set, make a violin sound better than before, but, until you have also heard that tweak's effect on the trumpet, the cymbal, the drums, etc. you don't really know if it was an overall improvement.

 

All of that makes intuitive sense to me. To prove that it cannot be so, I think you would have to show that we cannot ABX our family's voices against those of others and thus prove that there cannot be an extended memory of audio sounds.

 

I meant to comment about that earlier. You can recognize through aural memory the voice of your mother obviously. However, can you reliably under blind conditions, tell me which cable I am using to play a recording of your mother's voice? See those are two different things we are talking about here.

 

No one said you have zero aural memory after 4 seconds. The most precise memory goes to short term memory a few seconds later and then to longer term memory. That final long term memory is not as precise as the raw sensory perception was. Which in no way is indicating you can't recognize people's voices.

[esldude]

According to pretty solid research into the echoic memory subject, if you wait more than about 5secs between 'listens' you'll never know the answer to that question. You cant remember sounds longer so you simply cannot compare.

Even worse. I dont remember right now how big that echoic memory storage was. But pretty sure it is smaller than one typical song. So, according to currently mainstream audio memory theory, your audiobrain is simply not capable of doing that kind of comparison with full songs. Tough luck I guess.

 

You should still be able to reliably compare small snippets in a fast switching environment. That would be the ABX methodology.

 

Yes, that is a good description of how it works. Doing an ABX test with Foobar recently there was a short test signal I could reliably get (like 17 of 19 followed by 17 of 20 correct) if I kept the segment being auditioned to 2.5 seconds each. At 11 seconds my results were random. At 5 seconds my results were random. The difference in the two signals was quite far down in level and I didn't think it would be audible. But it was if the length of the sample listened to was short enough.

 

Again this isn't to say no memory is available beyond a few seconds. The longer memory however is a compressed lossy version. So differences have to be larger to be detected if your listened sample gets longer than the short few seconds of echoic memory.

[esldude]

Hi trithio. I'll try to be concise.

 

- Conceptually, I like the idea of DBT. As has been mentioned by at least one other commenter, I want to see that provision has been made for false negatives (not detecting a difference when one exists). When the question is precisely about whether a difference can be detected, or what level of difference can be detected by our senses, that question becomes a little more complex than it would be for something like pharmaceutical testing. So whatever the exact protocol for the DBT in a particular test is, I want to know whether that protocol has been checked against other tests of perception to determine a false negative rate.

 

- In practice, I want to see DBT that is scientific. After all, that's one of the primary advantages being claimed for it; if it's not done scientifically, then why go through any additional hassle versus good old *un*reliable sighted testing? I've seen almost no DBT in audio that could be called scientific. I've seen people doing careful, competently run tests on various audio systems, yes. But I wouldn't call that scientific any more than I'd rely on a bunch of people using home chemistry sets for scientific results in the field of chemistry.

 

- Also in practice, I would like to see what the results of training the subjects are. We are all pretty good at knowing what a difference in loudness sounds like. Many of us are quite good at knowing what a difference in frequency sounds like (those with "perfect pitch" are excellent at this). But how many of us know what jitter sounds like? What "smearing" of transients due to a ringing filter sounds like? Does training people to hear these things make a difference in DBT results?

 

Edit: Re training - There is of course a question (on which experiments may have been done, but I haven't seen them) as to whether, even if people aren't trained to hear the particular thing that makes A different from B, they can still say simply whether A sounds different than B.

 

Those are some of the questions that occur to me regarding DBT. Note I'm not saying this makes sighted testing results reliable, I'm purely talking about DBT.

 

Well for false negatives you need a history of positives to gauge that with. The better done blind tests will have a positive control known to generally be audible. If a test fails that the thing is suspect. Of course you want to know false negative rates for the actual thing being tested. About all you can do is use general principles to reduce the liklihood of that if you don't yet know the item under test is a positive. The most simple way is to increase sample size to reduce false negative rates.

 

As for training it does depend on what is being tested for. There is some work on that. Some things it hardly matters (loudness) others it does have an effect. The points about knowing what jitter sounds like etc. are good points.

 

In a perfect world all blind tests would meet ITU standards. Some simple things like basic distortion, loudness, frequency response are pretty easy in general. Even a minimal blinding will get good results. Obviously other areas may be more problematic. Requiring more careful testing to tease out the truth. That isn't so different in other areas of science really. You need CERN to find the Higgs boson. Other facilities at a lower level find other valid results. (interesting article about that here: How Three Guys With $10K and Decades-Old Data Almost Found the Higgs Boson First | WIRED)

 

That isn't a satisfactory answer to many. A simple blind test might miss something real that a better one would find. However even simple blinding is relatively powerful. So many have that gut feeling about their subjective experience, they have a hard time thinking simple blind testing is useful. They hold out they are really hearing something more marginal and difficult to find in blind tests. Of course you can't completely prove that is not the case with certainty. The reported subjective impressions often indicate a clear difference is experienced. All the other tests done on things makes that a long shot. Even fairly small things are detected blind, that a truly obvious difference gets missed unless the blind test is super persnickety precise seems very unlikely.

[mmerrill99]

According to pretty solid research into the echoic memory subject, if you wait more than about 5secs between 'listens' you'll never know the answer to that question. You cant remember sounds longer so you simply cannot compare.

Even worse. I dont remember right now how big that echoic memory storage was. But pretty sure it is smaller than one typical song. So, according to currently mainstream audio memory theory, your audiobrain is simply not capable of doing that kind of comparison with full songs. Tough luck I guess.

 

You should still be able to reliably compare small snippets in a fast switching environment. That would be the ABX methodology.

No, you are confusing the ability of using our perception of hearing as an instrument for picking differences between two audio samples (5secs limit on doing this). Obviously saying that we can't remember sounds after this period is patently wrong. We no longer have the level of detail stored to do the same trick. But that is not what our auditory perception is designed for. So all this talk of unreliability of our hearing is about how unreliable it is in performing a function that it was never designed for.

 

We do store our auditory impressions in some synthesisied, higher level format that we don't really know the details of. But there s no doubt that we have a stored auditory library of auditory objects which allows us to categorise different instruments & with some practise & expertise some can recognise exactly which instrument is playing - so we do remember specific signatures of auditory objects. In the same way as, with training, people can remember & recognise different types of distortions.

 

There's too much emphasis on A/B testing for differences, rather than what the auditory processing system can actually differentiate from memory - what audio playback sounds more like reality

[sbgk]

must say dbt discussions are extremely boring, is anything new actually being discovered or will the same debate be occurring if we come back in a 100 years time.

[mmerrill99]

must say dbt discussions are extremely boring, is anything new actually being discovered or will the same debate be occurring if we come back in a 100 years time.

Agreed & in the meantime advancement will be on-going without DBTs

[trithio]

Agreed & in the meantime advancement will be on-going without DBTs

 

Things advanced a lot even before anything resembling science was ever tried. They were however much slower and much more wrong. Mostly of the wrong direction kind. Of course science is not the only reason why things evolved so much faster during 20th than during 13th centuries. But it is one of the main reasons. If you ask me it was The main one by quite a margin.

Audio without science will mostly advance deeper into one's clueless pockets. Basically the con way.

 

 

P.S.

you dont seem to like me much and that surely doesnt help your interpretation of my words. Instead of trying to debunk my mere resumes of wiki articles and references like that echoic memory stuff, why dont you listen to esldude?! He's an older member with more credentials and makes a very good case for science. You may have less feelings about him as a messenger and that will surely help the message.

[trithio]

Trithio: First, thanks for the thoughtful reply! Second, coming from a family of concert musicians (piano, violin and cello) I do wholeheartedly agree that they seem to have little appreciation for the differences in recorded sound (they must be listening for something else, but what that is, I don't know). But there are those of us who do spend lots of time listening to all sorts different equipment, and I do believe that the more you do of that, the more you can develop a sense for colorations and other distortions. Whether that "informed sense" can then be used effectively in an ABX test, I don't know, but that typically isn't what it was developed for, so I doubt that would be its best use.

 

There is a new review out in TAS by Anthony Cordesman on the $98,000 a pair Boulder 2150 Mono Power Amplifier.Boulder 2150 Mono Power Amplifier and 2110 Preamplifier | The Absolute Sound I found one of his quotes very interesting:

 

"Here’s the issue: With equipment this good—and this neutral and transparent—almost all of the colorations you hear come from the recording, the front end, the interconnects and speaker cables, the load presented by the speaker, and the complex room interactions that shape the sound of the speaker at a given listening position."

 

I think that further complicates the whole ABX issue because it implies that we have arrived at a level of sonic improvement where the interactions between equipment may be as or more important than the equipment itself. Put differently, even if I can hear a difference between Amp A and Amp B in an ABX test, is what I'm hearing the difference between the amps or the difference in how each amp interests with a) the signal it is fed and b) the speaker load it is driving and (maybe) c) the cables between them?

 

Many thanks for the first-hand account about musicians. Especially since it fits some of my hunches

But I just don't trust a single word that TAS prints. I read it sometimes for various reasons but never as a source worth of any trust or second thoughts. For me they are one of the the poster-evils responsible for the current state of the hifi industry. Which as I said above I see as mostly 'advancing' into my pockets. A greedy hiene that disguises itself as the luscious 'woman in red'. Not saying that hifi is entirely like that. Just too much like that.

[trithio]

Guess I forgot to qualify most of my above post with a mandatory 'for my taste only'. I would like to think that my opinion on the hifi industry is a well informed one cause lately I do read lots of hifi press, reviews, comments and specs. But it is all still my personal anecdote.

[firedog]

I think lots of differences people hear are imaginary and blind testing would show that. I include myself in that.

 

There is nothing in the DBT protocol that forbids long listening periods.

 

I personally have trouble with quick A/B tests: One, I don't listen the same way when I'm in test conditions - it tenses me up. Two, the quick A-B switches and type of concentration required get me mixed up and tired very quickly and I can't listen properly to anything. So I don't know how useful they really are in determining what I hear when I'm not listening that way.

 

Some of the most touted audio DBTs are full of flaws when one actually examines them critically. So I don't accept them as worth much.

 

Trained listeners will certainly hear things that untrained listeners won't. I hear all sorts of things my friends don't - until I point them out. Then they do hear them. I've taught several people who said mp3s sound the same as Redbook to hear the difference - in blind testing. They just weren't aware of the differences before, so they didn't notice them. That's just how our perception works.

I have an artist friend who notices things about paintings that I "don't see" - until she points them out. Then my brain does start to notice them.

[trithio]

I think lots of differences people hear are imaginary and blind testing would show that. I include myself in that.

 

There is nothing in the DBT protocol that forbids long listening periods.

 

I personally have trouble with quick A/B tests: One, I don't listen the same way when I'm in test conditions - it tenses me up. Two, the quick A-B switches and type of concentration required get me mixed up and tired very quickly and I can't listen properly to anything. So I don't know how useful they really are in determining what I hear when I'm not listening that way.

I'd go further and say they are close to useless at that. But they are very good at telling if there are objective, measurable differences in sound. As you hear it during those precise few seconds. That's surely different than as you hear from the couch with a wine. But the absolute sound differences exist independently of your couch and wine

 

Some of the most touted audio DBTs are full of flaws when one actually examines them critically. So I don't accept them as worth much.

Most science experiments are flawed. I'd say the vast majority. Does not disqualify the methods. At all. And even some of the most clueless DBTs are better than 'my wife heard it too'

 

 

Trained listeners will certainly hear things that untrained listeners won't. I hear all sorts of things my friends don't - until I point them out. Then they do hear them. I've taught several people who said mp3s sound the same as Redbook to hear the difference - in blind testing. They just weren't aware of the differences before, so they didn't notice them. That's just how our perception works.

I have an artist friend who notices things about paintings that I "don't see" - until she points them out. Then my brain does start to notice them.

 

with you most of the way. But some of my artsy friends also see things I dont see even after they 'explain'. Things like 'that green shade makes me dreamy'. Sorry, it makes me hungry if anything

Also I wont put visual and auditory perception in the same pot. Not by a long shot. Huge physiological differences but I will point just one. Visual info is Huge. Look out the window and the total picture you see is bigger than 100GB. Much bigger. Your brain simply cant process all that and it selects bits and pieces according to your previous experience, mood, wishes and so on. Some people argue that it is hard to impossible to see things for the first time without being pointed out by others. Even if it's big as an elephant and sits in front of you . At first your brain will just filter it out cause it has no pattern for matching it. There is a funny anecdote about the american indians who at first did not see Columb's ships. They were looking at them but had no way to 'see' them. No matter how hard you try, you wont see all, even in a small painting. And you'll _never_ see the same as the guy next to you. Some interesting experiments into that, for example people at some conference were asked to describe their way in, in a visually detailed manner. Very different recollections about the same halway!

On the other hand, we are capable to process an incoming audio stream almost completely. At least when you concentrate, close eyes and so on. Your first level echoic memory will pick almost all audio info for a few secs, no need for much filtering. And assuming you both know how to concentrate on sound, the expert musician will not pick more. At most he will be able to 'see' some details faster, recognize notes faster/better and so on. But he wont hear more or anything that you simply dont.

But of course you are right when you say that experience can make big differences. Valid in every case. Everywhere. Everytime.

[Paul R]

Also I wont put visual and auditory perception in the same pot. Not by a long shot. Huge physiological differences but I will point just one. Visual info is Huge. Look out the window and the total picture you see is bigger than 100GB. Much bigger. Your brain simply cant process all that and it selects bits and pieces according to your previous experience, mood, wishes and so on. Some people argue that it is hard to impossible to see things for the first time without being pointed out by others. Even if it's big as an elephant and sits in front of you ..

 

Just to point out, the visual centers of the brain are very active while processing audio, and visual input can easily be overwhelmingly important when processing audio. (Which is of course, what the ABX testing proponents are so adamant about.)

 

Lots and lots of problems in that. Then you get to people like me, who tend to "see" sound as light and shapes. I can tell in just a second or two if there is something about a speaker I absolutely do not like, even if I can not "hear" it. It's like getting smacked upside my head and seeing stars sometime.

 

But it will take hours or days of listening to find what it is about the speaker I don't like. Like/Dislike is fast - identification of what is causing that input in the sound is *hard*. Unless of course, you want to accept that it sounds like a needle of light stabbing right between my eyes. (grin) I guess you can "see" why I don't listen long to components I don't like.

 

That apparently isn't the way most people perceive music, but the visual centers of the brain are active in sound processing for everyone. Music I really like is sort of an invisible picture show to me, and I always feel like it is a bit of a shame that not everyone shares that.

 

Conversely, and annoyingly to my art history minor wife, some paintings are like getting screamed at to me - literally. Others are like a symphony. She has a bit of trouble figuring out why I like some paintings and not others, or why the color of a room can be annoying enough to me I have to make a mental effort to ignore it.

 

It is also why, I tend to trust my own perception that a cable sounds different or high res / DSD sounds different on a particular system than Redbook. Impossible (or at least very very difficult) to explain, but it does't look the same to me.

 

Oh yeah, and I do agree that lots of what people hear is imaginary. My case is different, the stuff I "see" is 100% imaginary, but it is still real to me. Makes me more sympathetic I suppose.

 

 

-Paul

[trithio]

Just to point out, the visual centers of the brain are very active while processing audio, and visual input can easily be overwhelmingly important when processing audio. (Which is of course, what the ABX testing proponents are so adamant about.)

 

Lots and lots of problems in that. Then you get to people like me, who tend to "see" sound as light and shapes. I can tell in just a second or two if there is something about a speaker I absolutely do not like, even if I can not "hear" it. It's like getting smacked upside my head and seeing stars sometime.

 

But it will take hours or days of listening to find what it is about the speaker I don't like. Like/Dislike is fast - identification of what is causing that input in the sound is *hard*. Unless of course, you want to accept that it sounds like a needle of light stabbing right between my eyes. (grin) I guess you can "see" why I don't listen long to components I don't like.

 

That apparently isn't the way most people perceive music, but the visual centers of the brain are active in sound processing for everyone. Music I really like is sort of an invisible picture show to me, and I always feel like it is a bit of a shame that not everyone shares that.

 

Conversely, and annoyingly to my art history minor wife, some paintings are like getting screamed at to me - literally. Others are like a symphony. She has a bit of trouble figuring out why I like some paintings and not others, or why the color of a room can be annoying enough to me I have to make a mental effort to ignore it.

 

It is also why, I tend to trust my own perception that a cable sounds different or high res / DSD sounds different on a particular system than Redbook. Impossible (or at least very very difficult) to explain, but it does't look the same to me.

 

Oh yeah, and I do agree that lots of what people hear is imaginary. My case is different, the stuff I "see" is 100% imaginary, but it is still real to me. Makes me more sympathetic I suppose.

 

 

-Paul

 

It surely makes you more sympathetic

 

That's an interesting account of the visual-sound experience. That kind of stuff is not coming so often into my daily experience. And rarely when not explicitly cued. Maybe because I am the kind who enjoys keeping his mind on a very rational level. If I decide to tune-in though, all that sensorial fireworks you describe comes in spades. Colors, sound, smell and sometimes even almost tactile experiences. And I do enjoy night time music sessions with lights off and lots of induced visuals.

 

I rarely tried to use that in my comparisons though. Those I mostly kept on the rational, max focus level. Guess one may call that DBT-style. However, did try some of that feeling-based comparisons. For example between foobar and jriver. Cued one favourite album on one player. Did listen in relax mode one-two times for a few hours. Few times I even forgot what I was doing and just enjoyed music. Then switched the playe and repeated. No differences. Neither audio, nor visual, nor feelings. Not immediately after switching, not along the way. Nada. Same beautiful sound and music.

Tried that only a few times. But it kinda convinced me that if I dont hear differences in the focused DBT-style, I wont hear anything when relaxed either. Admitedly my 'proof' is based on a few trials only and is therefore quite thin. But I do not need extraordinary proof because the experience fits the current audio and hearing theories. Maybe I should try more of that when comparing say cables. But I dont see any sign to say that the extra test effort will bring any results. And that's a lot of effort. One full evening for just too cables. Sounds like too much for me.

Looks like it brings results for you. But you are walkin on a very thin ice there. The audio-visual interactions and your whole experience are so complex, I find it hard to put any of that in the 'proof' category. It's just just too much stuff hapening at once, things that you cannot or do not want do control. Things which are not exactly stable or predictable or corelated in any clear or consistent manner. At least in my case, I rarely see the same images when listening to same music. There is some similarity but vague, not even the same dominant color or theme or such.

Oh well. All the above is just a looong way of saying Different folks, different strokes

The only clear thing: in my book neither folks nor strokes are proof. Just anecdotes. And those are normally 100% neutral on the rational level. Unless they are particularly lame