Auditory Scene Analysis (ASA) explained

[jkeny]

Actually the more meaningful part comes in the article itself. Let's review that and see if we a) already know about it and b) agree with it.

This should be fascinating insight to people who say they "trust their ears." The process as correctly stated is bi-directional. What you think, influence what you think you are hearing. And more so can even change what you are hearing!

Let's remember the simple exercise I mentioned in another thread where you play the same file over and over again. And how doing so results in "hearing" differences in detail even though nothing has changed as far as what the ear is picking up.

Yes, it is a fascinating insight but not for the reasons you claim.

You are reading this with your usual bias. What it is saying is that strong top-down influences (notice it doesn't say biases) - such as attention (which isn't a bias unless you have a negative bias towards hearing something & so don't give proper attention when listening) ; goals of the individual in the specific situation - this is the bit you are basing your interpretation on, right? - you are interpreting goals to mean a desire for a particular outcome but this shows that you haven't read or don't understand the text - goals means listening listening to a conversation in a noisy room or listening for an announcement over Tannoy or listening for the creak of a floorboard if you suspect an intruder, etc.
"expectations" you have interpreted to suit your own agenda but it's the wrong interpretation if you read the literature - I explain it below.

So let's look some more into it, Amir. The top-down aspect of auditory processing is not just limited to the situation you are presenting - a case for blind testing - in fact it argues against this. Top-down processing is part & parcel of all our auditory processing, not just sighted listening - so it argues against blind testing as between each trial of an ABX test for instance we will hear it differently & therefore any differences will be masked/confused by this variation.

But that's not the case that I am making, I'm just stating that your conclusions are superficial & one-sided

Top-down process actually happens all the time at a much more granular level than you understand. For instance when we hear a bell, the attack portion of the sound immediately allows us to categorise the sound as a bell & we anticipate (predict) that the bell will ring for an extended period based on our knowledge of how bells sound in the world. So we are subconsciously anticipating this extended fade & if it doesn't happen a mismatch might arise in our consciousness or in more subtle examples we adjust the model because a bottom-up signal contravened what our expectation was.

So yes, we focus our attention either consciously, on what interests us, or automatically on what startles us.

All of what I say is already in the introduction which I quoted & you chose to ignore, instead picking out isolated text which you don't understand & which you try to bend into a meaning which supports your agenda

Rather, top-down
processes such as attention, expectations and prior knowledge facilitate
scene perception. Thus, scene analysis is linked not only with the extraction
of stimulus features and formation and selection of perceptual objects, but
also with selective attention, perceptual binding and awareness.

This focus is part of why training is needed for blind testing - we need to find the audible spot that is different & then use it in the blind trials to focus our attention - without this training only gross differences can be ascertained.

Of course this describes the placebo effect where our expectation of difference in an outcome results in us concluding the same. The "goals of the individual" in hearing a difference is met. "Prior knowledge" that we have changed something so therefore the sound must be different, comes into play.

Not in as slavish a way as you portray.This is why listening to a device over a long time negates all you say because most of us are not slaves to this 'bias' as you portray it. Many have experienced initial enthusiasm about the sound of a new device only to find during longer term extended listening that the initial sound which was liked was a distortion. If expectation bias was such a force then there would be no release from this bind as you seem to maintain. In fact again it argues against typical blind testing as what is happening is a test, on a day, of a particular sound or music segment, & seldom repeated for more than the 12 trials necessary. I don't know of anyone who has done repeated blind testing of the same device using different music (or segments) many times over the course of a fortnight - a typical period of time that we use to accommodate to the sound character of a new device.

Does audio science already know this? Of course. This is why we perform controlled experiments where we attempt to remove the influence of these factors. We don't tell you if something has changed. And with it, we neuter the brain's influence over your hearing perception to conclude that something "must be different."

Is it important to know why such bias and variation exists? No. We have empirical data that has told us this for decades. It is why we have controls and rigor in our testing for audio differences.

If you look at Diana Deutsch's site here you will see her research into listening to tones. She shows that we can't compare two tones when separated by 6 other tones in the same piece of music never mind between pieces of music. So again your view of all things to do with blind testing is simplistic.

It is shown that our auditory sense works mainly by pattern matching, not by being able to accurately compare one piece of music to another & so setting up tests to specifically test what our auditory sense is not designed for is disingenuous. Deriving conclusions from such tests is a further sin.

The paper goes on:
...snip

All confirming what I just explained. Notice the last line: that we perceive differences even when nothing ("the environment") has changed.

In summary, audio science and what is investigated here are completely consistent. We can investigate the "why" in this type of research but the "what" is already known. That humans are full of preconceptions and multiple-sensory inputs and willing imagination. If you want to make an audio assessment that is durable across listeners and represents what the device is doing, you need to follow what is already part and parcel of audio science. That is, reduce as much as you can the reverse connection between the brain and your ears. What is presented in this paper (and book I reference) simply reinforce what we know as a matter of science, but unfortunately choose to ignore in our audio hobby.

Sure, Amir - all is already known (by you) & you have no room for further learning.
We get it.
It's pretty similar to Ethan Winer's statements that all audio was sorted out 50 years ago - nothing left to advance (except room treatments, of course)

 

[jkeny]

Except that (outside of room acoustics) this research in no way, no shape supports your assumption. Indeed all the research involves sounds that are measureable and then we examine what is audible.

ASA is a fledgling research area in the scheme of things & like all early attempts at grappling with tough questions, it uses simple sound patterns with obvious differences to test hypotheses - more complex & subtle differences don't usually come until the field is mature & theories confirmed & generally agreed. If you read the latest ASA research, a major theme of it is to move away from using these simplistic sounds & sound patterns into more realistic audioscapes that are encountered in the real world.

You are mixing up the techniques used for scientific testing with how such tests relate to real world signals encountered in everyday life. Your simplistic notion is that because they haven't used such real world signals then the research doesn't support my analysis.

In acoustics where we ignore that we have two ears and proceed to use one microphone to measure things this does come true. But for upstream products, nothing remotely in the research touches on this whatsoever.

You mean that you disagree with Harmon's loudspeaker testing where they used only one speaker for the tests? Interesting!!

What it does say as I have repeated quoted from your own references is that what we think we "hear" is actually a product of multiple sense and a creative brain. That is the part you want to learn and with it, go back to your experiments and correct their major flaws. Once you do that, you will see that measurements predict why there is or is not an audible difference. There is no conflict there.

I'm happy for you to live in ignorance of ASA & it's profound implications - it doesn't bother me in the slightest.

 

[jkeny]

I'll give a very recent example of this lack of bias slavery.

I recently started a tour of my USB isolator device & gave it first to Pearse (Sigolad) who I know & is local (he has a Lampizator DAC)
His first report here

Straight away yesterday I could hear the difference, the 2 words that come to mind are a fuller and natural sounding presentation.
There is definitely more body to music and a very natural sound, dare I say it very analogue but with all the detail and more.
Not sure how many hours you have on the units but it sounds even better today, maybe the batteries are now fully charged but even with them charged it sounds like it is still improving.

It is definitely a step forward from the prototype we have heard on several occasions over the last year.

And his next report where he isn't hearing the audible improvement :

Another update and this time I am a little confused.

I have not had much time for listening this week but I did do comparisons between the battery powered ISO Hub and the Intona and on both occasions I struggled to find any difference which has me confused.

When I put the ISO Hub in last Sunday I heard an improvement on tracks I am well used to.
Now when I go back to the Intona I still hear that improvement !!!

And after I discovered what the problem was - a power cable issue:

OK normal service resumed....finally found the issue, there was a poor connection on the umbilical between the Battery Pack and ISO Hub.
All sounding great again and the Intona goes to the back of the class.

Very happy now with the sweetness and the way notes hang that much better in the air with the ISO Hub so I will be handing it back to John at the weekend for next in line.

And someone even posted this:

Fascinating, and a strong blow against the "expectation bias" explanation for any differences heard in these things. You heard a difference first, no difference in a later run, had the balls to state it, and then found the problem and got the sound back. Good ears, Pearse, and nice work,

His final report:

Had a glorious afternoon listening to some great music and have grown really fond of this added hardware in my system, so much so I am at pains to part with it now.
Anyway a special request John, would you build a one off in advance of release for me at full one off build costs, you know I mean that?

Looks like it will be a while longer before I see the ISO Regen as I did not make the group for first shipping, seems a strong track record of past purchases (i did buy a Regen off Uptone) added to number of posts on the forum trumped me being in the first 10 to commit to a purchase, sad days but that seems to be the way of the world these days.

Anyway back to happier things, really enjoying the USB Isolator now and it has gelled really well with my system.
Recent changes with new WTFPlay and some tweaking with blinds on the Raidho tweeters has allowed me an almost transcendental afternoon of listening, I even immensely enjoyed listening to Sade greatest hits!

For those who shun Sade listen to the drums and percussion as there is an Irish guy now long gone called Dave Early playing drums and percussion on many of her tracks who was a master and a lunatic at the same time, spent some mad illicitly fueled times with the man some years back at gigs in Sligo where he played and in after hours parties. He had some stories to tell of life on the road with Van Morrison and the sessions in London with Sade.

So now are we going to twist the expectation bias into contortions to handle these different listening reports & say:
- there is no difference to be heard here & all he is reporting is his expectations
- first he heard it sounded good because he knows me & wants to please me
- then he reported that he couldn't hear any difference between it & the Intona USB isolator - why? because he was pissed off with me for some reason so he was biased to not hear a difference
- then he heard a difference & is glowingly basking in the sound. Why? Because I talked him into liking me & he again wants to please me

Come on. This is all such a crock of sh1t. We are not slaves to our expectations & most of us have relatively reasonable control particularly when long term listening is engaged in
We all know this - it's the pseudo-scientists that insist we do blind tests & wake up to the 'fact' that we hear no difference!! Really?

 

[amirm]

Not in as slavish a way as you portray.This is why listening to a device over a long time negates all you say because most of us are not slaves to this 'bias' as you portray it. Many have experienced initial enthusiasm about the sound of a new device only to find during longer term extended listening that the initial sound which was liked was a distortion.

Well, let's examine this using one of your own references, Bergman's listening test examples: http://webpages.mcgill.ca/staff/Group2/abregm1/web/downloadstoc.htm

There are tons of tests which you can take. None, none have any "take home" component for "long term listening." The effect is available and immediate to your perception.

So based on this topic of your own choosing, i.e. ASA, you just put words in the mouth of science that is not remotely applicable or ethical. This is evident by lack of any quotation to back what you say above. I don't know of any branch of auditory science that justifies your position in the slightest.

This demonstrate what I find very objectionable in your posts: working backwards from audiophiles beliefs to shoehorn them into science when that science does not in any form of fashion sanction or support such. It is all a corrupt disregard for the facts to sell hardware that has no way of showing its efficacy.

Why do you attempt it then? Because you hope that the reader is not familiar with the terminology or science being discussed and takes your argument at face value.

The whole long-term bit has been tested anyway and unfortunately for you, it had the opposite outcome. Listeners using quick switching could detect distortions at far lower levels than those who did that at their leisure, in their own home and system. Here is Clark in his AES paper:

I have taken countless tests of small impairments and passed ABX tests. All of that was possible at times to be able to make comparisons in as short a period of 0.1 to 0.2 seconds. Any switching longer would remove my ability to find distortions that were objectively there. You have no friends in audio science for that argument.

 

[jkeny]

Yes, Amir, you prove again & again that you don't read & understand what I write (or indeed the scientific method in general) - as I said "ASA is a fledgling research area in the scheme of things & like all early attempts at grappling with tough questions, it uses simple sound patterns with obvious differences to test hypotheses - more complex & subtle differences don't usually come until the field is mature & theories confirmed & generally agreed. If you read the latest ASA research, a major theme of it is to move away from using these simplistic sounds & sound patterns into more realistic audioscapes that are encountered in the real world."

Bergman is the father of ASA & his pages reflect the early & ground breaking work he did - as a result they use standard listening tests to probe the obvious differences & test hypothesis - the field is moving towards more realistic & more complex soundscapes for listening tests.

The ABX test (when designed & conducted by someone who knows about perceptual testing) is fine for revealing the type of distortions that you have trained yourself in - it's not that suitable for revealing the type of perceptual differences being spoken of here - realism of the illusion, soundstage depth & solidity.

I gave links to ABX tests conducted by Ultmusicsnob on Gearslutz & head-fi which demonstrates this sufficiently well. In these tests he is able to A/B differences in these types of differences but his example is rare as he is a recording/mixing engineer who has a lot of experience in this sort of analytic listening & he still found it very difficult as he couldn't use his usual audible differences that he usually looks for. He was also far more dedicated than most in his efforts & spent a number of days chasing down differences until he found what he could readily identify in A/B testing. The thing is that each piece of music required him to listen for a totally different audible difference in order to 'pass' an A/B test. The overarching point is that he had already identified in sighted listening that he had a preference for the high-res version of the files through his work & continued exposure to the two different versions of the files.

Now it may be that he could just have done a blind preference test & not gone to all this bother & it would have been just as valid but the blind test of choice suggested on forums is the Foobar ABX test
2013 Gearslutz thread - Foobar 2000 ABX Test - Redbook vs 192/24

I gave another example before of a staunch (but open-minded) objectivist on Pinkfishmedia forum who couldn't hear any differences between DACs & organised a number of sighted & blind listening tests in his home over a number of months. It wasn't until the 4th such get together when someone arrived who could hear the difference & pointed out what to listen for - they could all then hear the differences after 3 months of tests not hearing any difference.

So let's get onto to training.
The home listening test you mention is by the late, Tom Nousaine - do you notice anything wrong with what he did in that test?
I'll leave you to try to answer that for the readers & maybe you can give them a link to this test so that they can read it for themselves?

 

[amirm]

I gave another example before of a staunch (but open-minded) objectivist on Pinkfishmedia forum who couldn't hear any differences between DACs & organised a number of sighted & blind listening tests in his home over a number of months. It wasn't until the 4th such get together when someone arrived who could hear the difference & pointed out what to listen for - they could all then hear the differences after 3 months of tests not hearing any difference.

Would you please post the measurements of those DACs?

Meanwhile, can you explain why the ASA effects are easily audible on any computer with any DAC? Heck those soundstreams may even be lossy compressed. So in the context of the topic you started, i.e. ASA, can you say why the experimenters do not care whatsoever about using high-end audio gear? Why is it that they don't think the distortions you imagine to be there, interfering with ASA outcomes?

The home listening test you mention is by the late, Tom Nousaine - do you notice anything wrong with what he did in that test?
I'll leave you to try to answer that for the readers & maybe you can give them a link to this test so that they can read it for themselves?

The test used hardware that was purpose built to test this theory. Have you ever bothered to build hardware to test your theories?

 

[amirm]

Yes, Amir, you prove again & again that you don't read & understand what I write (or indeed the scientific method in general) - as I said "ASA is a fledgling research area in the scheme of things & like all early attempts at grappling with tough questions, it uses simple sound patterns with obvious differences to test hypotheses - more complex & subtle differences don't usually come until the field is mature & theories confirmed & generally agreed. If you read the latest ASA research, a major theme of it is to move away from using these simplistic sounds & sound patterns into more realistic audioscapes that are encountered in the real world."

Bergman is the father of ASA & his pages reflect the early & ground breaking work he did - as a result they use standard listening tests to probe the obvious differences & test hypothesis - the field is moving towards more realistic & more complex soundscapes for listening tests.

The stream of consciousness you are dumping on us is hard to follow John. Did you not say this?

How I relate ASA to this hobby is that for me, it explains a lot of the rift between what is heard & what is currently measured - it suggests to me that our auditory perception is a pattern matching analytic technique which we use to make sense of the ever changing sounds from what surrounds us. That might be our playback systems just as much as it might be the bird singing outside my window, at the moment.

That's pretty specific statement yet nothing in the research even directionally supports that. Indeed ASA is about the opposite.

Let's examine the "cocktail party" effect which is the posterchild of this research. That says that in a room full of people talking, you are able to tune into a specific conversation and hear and understand that. The brain is able to filter the surrounding sounds that are coming from all directions, in favor of specific source of sound from the one voice you want to hear. In this regard, ASA is about unhearing things, not the other way around. In other words, it is a form of noise filtering.

So if you want to infer anything, you can deduce that ASA says the brain is able to unhear anything outside of the main messaging it wants to hear. In the case of music+distortion, we obviously want to hear music and not distortion and hence, ASA would directionally tell us that we have the ability to ignore such distortions.

Do we have evidence of this in our audiophile life? Yes. Ever read a review of someone saying they heard a speaker and it didn't sound good to them. But a few weeks later they did? I am sure you have. This is called "adaptation." The brain notices that together with what you want to hear, there is this auxiliary distortion that comes along with it that conveys no new information so it adapts to its signature and unhears/tunes it out.

We have this in room acoustics in spades. Take a concert hall. We have massive reflections coming from all directions. Taken literally would drive us mad as every sound is accompanied with thousands of "echos." Since moving into caves, the brain has adapted to know that reflections are a) harmless and b) convey to more meaningful information if they have characteristics that says they are within a room (i.e. certain delay and amplitude).

Not only does the brain ignore the echo aspects of the reflections, it actually takes advantage of them by summing their energy to improve intelligibility. This is why research shows that comprehension improves when reflections are there.

Ask any home theater owner who has a projector in their room with the fan, whether they hear it once a movie starts. They say most of the time they do not because again, the brain adapts and dials out what is clearly noise and unwanted signal.

All of this comes with loads of research data and references don't even dream of arguing against them. Instead, explain why you have picked ASA to advance your argument I have quoted above when its role is the opposite: i.e., unhearing unwanted parts of the music.

The answer is simple and is what I stated before: you picked it in the hope that no one will know the term or bother to go and research more than you have so they just buy your manifestation on face value. Because any learned reading of the topic clearly neuters your case.

 

[amirm]

Oh, another example I forgot to mention above. Namely, LP playback.

We all know that LP has verifiably more track noise and pops. We hear it. We can measure it. No question about it. Yet the people who love LP don't mind those clear and obvious deviations from the source. Why? Because of adaptation. They have learned to "unhear" the track noise and pop. I don't have that ability. So at shows when LP is being played back and a loud pop comes, I am the only one who jumps! I usually look around and feel foolish because nobody else even blinks when that happens.

This is supported by "ASA" because the listener is segregating the noise stream from music and latches onto hearing the latter.

It is much harder for me because I have spent years trying to hear distortions and noise so unconsciously, I am actually searching for the noise and sources of impairments.

Note that there is non-ASA explanation of these effects to but we don't need to go there.

 

[jkeny]

Would you please post the measurements of those DACs?

Meanwhile, can you explain why the ASA effects are easily audible on any computer with any DAC? Heck those soundstreams may even be lossy compressed. So in the context of the topic you started, i.e. ASA, can you say why the experimenters do not care whatsoever about using high-end audio gear? Why is it that they don't think the distortions you imagine to be there, interfering with ASA outcomes?


The test used hardware that was purpose built to test this theory. Have you ever bothered to build hardware to test your theories?

So I ask you what's wrong with Nousaine's test & you deflect with another question!
Care to answer?
Once we hear from you on this the reader might learn something - at the moment they are just learning more about you & your motivation in this thread.

While we are waiting for your answer you should ponder what 853guy already quoted to you ""The (zealot) pathologizes others for doing things he doesn’t understand without ever realizing it is his understanding that may be limited (...) their main skill is capacity to pass exams written by people like them". Nassim Nicholas Taleb

 

[jkeny]

At this stage it's obvious that no amount of explanation will inform Amir of what ASA is about - he is impervious to new learning based on his posts here.
Everything I post about ASA is filtered & bent into his myopic world view - I mean it's truly incredible that anyone can post such misinformation in bold "In this regard, ASA is about unhearing things, not the other way around"

I'm not interested in & would be a fool, wasting my time on such a fruitless exercise.

So far we have had both Frantz & Amir posting on this thread, not because they are interested nor inquisitive about the subject but for their own nefarious reasons.

I will gladly try to answer any posts which show a genuine interest in ASA.

Just to summarise:
- ASA is the study of auditory perception & central to how & why we hear what we hear
- it's central to fuller understanding of this hobby
- IMO, it begins to explain some of the more contested reports of audibility that we see cropping up regularly.
- it's a difficult area to explain as it requires an understanding of the findings about how auditory perception works & some understanding about the technology involved in audio recording & playback. I focus my attention on playback as it's the only one I have any control over
- AFAIK, what has not emerged yet is a direct application of ASA to audio technology although there are a number of companies who know about this branch of research & have paid credence to it's findings - Dolby, Bob Stuart of Meridien, Siegfried Linkwitz & probably others I'm unaware of

I hope that I have at least outlined what ASA is about & some of my thinking in how I see it applying to specific areas in audio playback?