Better explanation (with pictures)

[mike1127]

Let me explain more about this idea of "listening to sound as sound," and why I think it's flawed as an audio test.

(EDIT: if you quote this post, please edit out the URLs to jpegs so we don't have to look at them several times on one page. Thank you.)

First I will show a very rough model of the brain during listening. Please note: (1) This is rough. (2) I am not a biologist but inferring certain facts from my teachers who are more familiar with the brain. (3) There is a certain degree of concrete non-speculation to this---it represents my experience from introspection and meditation.

Here's a picture:

To explain:

1. Sound waves arrive at the ear.
2. Information travels along the auditory nerve.
3. The brain has different stages of modeling internal sound. At the most primitive level, sounds are "abstract." At a higher level of integration, different frequency ranges are integrated to become distinct sources of sound. In the abstract form, you are hearing some highs, some grumbling lows, etc. In the integrated form, you are hearing trumpets and tubas (and can separate them in your attention).
4. The motor cortex is involved in our experience of sound. That's why music is toe-tapping or dancy.
5. There is a high degree of up-and-down and cross-connectivity in the brain. You can see that each part can influence other parts, or be influenced by other parts.
6. We have an executive control to decide where to put our attention (or decide to "just enjoy the sound").
7. Ultimately we can only be aware of what becomes conscious, and things only reach consciousness through layers of pre-processing and pre-organization.

Now to address what I feel is a misconception.

I am challenging the idea that if two systems are noticeably different in sound, then they are always different while listening to "sound as sound."

This idea is at the basis of many quick-switch tests.. in fact in testing and comparing in general.

The reason I am challenging this idea is we can become aware of sound through different pathways. For example, the following picture might be listening to "sound as sound". The green represents the path by which we mainly become aware of sound.

Now, let's say we are listening to "sound as dance"; that is, we focus on the movement of our toe, or whole body, in response to sound.

Is the same information is available both through the green and purple channels? I propose this is extremely unlikely. That is because music has a wealth of information, and we only become aware of it by organizing this information into some higher level of abstraction. It is highly likely that the green and purple paths have different ways of organizing the available information. To illustrate in a visual analogy:

The 2D picture represents information: this is information about the objects that are present, their shape, size and color. Notice that what we can see depends on the direction we are looking. From some directions information is obscured that is available in other directions.

From my introspection, meditation, and drawing on ideas about how we perceive and organize detailed sensory information into a model of the world, I find it likely that the green and purple channels represent two different views on the available information.

To recap: Listening to "sound as sound" would be adequate if

1. The same information flows through both green and purple channels.
2. We can actually be aware of all that information, choosing at will what things to notice.

I find that this proposal is unlikely and in fact speculative. It is the unexamined assumption which is at the basis of quick-switch and other types of testing.

 

[Pio2001]

Hello Mike,

Shouldn't this topic have been posted into the "The experience of music, and comparisons" thread (which could itself have been part of the "Difficulty of Blind Testing" thread) ?

Anyway, my opinion is that it is often possible to divide problems into smaller problems. It is not always practical, but here, I think that most of the time, it is.
In other words, I think that it is possible to conciously translate the purple path into the green one.
It may not be easy, it may take some time and require some mental work, but there are no absolute boundaries between the different levels of concious analysis.

From a practical point of view, taking your example of the "sound as dance", what are the characteristics that make the sound a "dance" ?

Mostly the musical content, but also, in a lesser extend, the sonic quality.
What characteristics of the sonic quality will have an effect on the "dancing" quality of the sound ? Let's say, the listening level, the dynamics, and, in some cases, the frequency response.

So, since we are dealing with sound reproduction, what characteristics are relevant ? Listening level can be controlled with the volume knob. We must accept the dynamics that the studio has put into the recording. What's left ?

Mostly the bass response, that can make the rhythm more or less present. Maybe the amount of room reverberation can have an indirect effet (it is the case for me on some tracks, like Blue planet Corporation - Intrigue, or Juno Reactor - Mars).

These factors can be analyzed for themselves, independantly of the "dancing" feeling. That's what I mean with "translating the purple path into the green one".

In the end, for me, in blind tests, it is essential that the test methodology is monitored by the listener. Only him or her knows what is to be heard, and consequently how the listening must be done in order to hear it.
The test director only deals with blinding and evaluating the score, unless he or she wants to test the listening method itself.

And again, this only applies to "unknown differences". Once a difference is heard, but the blind test fails, if the blind test methodology is into question, the hard, but unquestionable way to deal with it is to absolutely reproduce the conditions where the difference is heard, parameter after parameter, beginning with the musical choice, the listening environment, then the listening duration, up to the listener's mood. Once all parameters have been set to the configuration in which the difference was heard to begin with, the only variable that can't be set in the same way is the fact that the listening is blind.
If all other causes are eliminated (which can be very hard indeed in practice), the only explanation left for the failure will be that the difference comes from the knowledge of the source identity.

 

[mike1127]

Quote:

Originally Posted by Pio2001 /img/forum/go_quote.gif Hello Mike, Shouldn't this topic have been posted into the "The experience of music, and comparisons" thread (which could itself have been part of the "Difficulty of Blind Testing" thread) ?

Hi Pio2001,

By the way, I'm about to listen to your mp3/flac samples. I'll report back to you.

Regarding this new thread, I don't like how my threads get derailed and filled with posts off the point entirely. This particular post is the most explicit I've made my ideas, so my hope is that this new thread will stay closer to the topic.

Quote:

Anyway, my opinion is that it is often possible to divide problems into smaller problems. It is not always practical, but here, I think that most of the time, it is.

It seems to me that we have different "paradigms." One difference between us appears in your statement above. Your description of the division of problems refers to a analysis by decomposition. We could also say you are producing a particular model of the problem, in which the problem is well-understood by dividing it into parts, analyzing those separately, "encapsulating" many details within the parts, and thereby simplifying our understanding of the whole system. You may be aware that some philosophers believe that some systems cannot be analyzed by decomposition. Behavior which is not predicted by decomposition is referred to as emergent behavior.

In my paradigm, the brain and especially consciousness is likely an emergent behavior.

Computer software is well-understood by decomposition. Conscious experiences, in my view, are not.

Quote:

In other words, I think that it is possible to conciously translate the purple path into the green one. It may not be easy, it may take some time and require some mental work, but there are no absolute boundaries between the different levels of concious analysis.

I disagree that the paths can be made equivalent. I would agree with you there are no absolute boundaries. There is no reason to impose a restriction on the capabilities of our brains. But the evidence from introspection on my listening process is that the green and purple pathways are so radically different that they cannot be made equivalent, and analysis doesn't help much.

Quote:

From a practical point of view, taking your example of the "sound as dance", what are the characteristics that make the sound a "dance" ? Mostly the musical content, but also, in a lesser extend, the sonic quality. What characteristics of the sonic quality will have an effect on the "dancing" quality of the sound ? Let's say, the listening level, the dynamics, and, in some cases, the frequency response. ...... What's left ? Mostly the bass response, that can make the rhythm more or less present. Maybe the amount of room reverberation can have an indirect effet (it is the case for me on some tracks, like Blue planet Corporation - Intrigue, or Juno Reactor - Mars). These factors can be analyzed for themselves, independantly of the "dancing" feeling. That's what I mean with "translating the purple path into the green one".

What you are doing is modeling the behavior of our brains along the purple path as a system by composition. In my paradigm, such models do a very poor job of predicting any aspect of aesthetics. For example, we have no measurements that will predict danciness. Decomposing the problem, as you have done, in my view, gets us nowhere.

Quote:

...... Once all parameters have been set to the configuration in which the difference was heard to begin with, the only variable that can't be set in the same way is the fact that the listening is blind. .....

My view is that we don't have a good handle on what those "parameters" are, so to set them in the same way is a difficult problem that no one understands well.

 

[Pio2001]

Quote:

Originally Posted by mike1127 /img/forum/go_quote.gif My view is that we don't have a good handle on what those "parameters" are, so to set them in the same way is a difficult problem that no one understands well.

The one that understands them the best is still the listener

At least he/she knows what parameters are important to him/her.

 

[Pio2001]

Quote from the "Dificulty of Blind Testing" thread :

Quote:

Originally Posted by mike1127 /img/forum/go_quote.gif Therefore, I feel a valid comparison would be one in which the musical effects are evident. Because I can only "get a handle" on the music when I listen to a fairly substantial section of the piece from the beginning, and with a bit of time before the next listen, I find that quick-switching is not a valid test.

That's exactly the kind of parameters I'm talking about

 

[mike1127]

Quote:

Originally Posted by Pio2001 /img/forum/go_quote.gif Quote from the "Dificulty of Blind Testing" thread : That's exactly the kind of parameters I'm talking about

Okay, I agree I can make some progress understanding the conditions that are valid for me. I continue to attempt that. However, does that mean that I can control my own listening process well enough that any blind test is absolutely conclusive? Not even close. And I don't think quick-switching is valid, per the original post.

 

[Pio2001]

Quote:

Originally Posted by mike1127 /img/forum/go_quote.gif However, does that mean that I can control my own listening process well enough that any blind test is absolutely conclusive? Not even close.

Positive blind tests can easily be conclusive. Negative blind test are another story. The rule is that failed tests prove nothing in themselves.

However, every controlled parameter in a failed blind test can bring a bit of interesting information : that a given difference is not obvious under these controlled conditions.

 

[mike1127]

Quote:

Originally Posted by Pio2001 /img/forum/go_quote.gif Positive blind tests can easily be conclusive. Negative blind test are another story. The rule is that failed tests prove nothing in themselves. However, every controlled parameter in a failed blind test can bring a bit of interesting information : that a given difference is not obvious under these controlled conditions.

To get to back to my brain model per the diagram, I don't think it's easy to control the pathways one uses to listen. In fact I don't know how to do it, and no one I've talked to knows (and most don't care). Don North was telling me about research that shows listening for enjoyment uses a different part of our brain (i.e. different pathway) than critical listening for comparing.

I don't think that means it's impossible... I do harbor the hope that I can train myself to simultaneously "listen naturally" while keeping one part of my brain glancing sideways at what's going on.

There's a related experience in meditation. Zen meditators discover through repeated introspection that they have a habit of "conceptualizing" or "doing" things, rather than discovering them or letting them emerge.

Once I was sitting in my car, waiting for the red light to change. I suddenly had the feeling the light was about to change. I guessed that I had seen the opposing light turn yellow, but I could not figure out where or how. The opposing light was not visible directly, although it was a tube-shaded fixture that permitted one to see reflections off the inner side of the tube at night. This was broad daylight, though, and looking directly at the tube shade, I couldn't make out any yellow. But, the light changed. So I believe had seen something.

At the next light was a similar configuration. I looked directly at the place where I might be able to see reflected yellow. Nothing was apparent. Maybe I thought I saw something once or twice, but it was hard to tell. Then the light turned. So I had missed it.

This got me thinking. In the first experience, noticing the yellow just happened. In the second, I consciously sought it out. Notice that we might use the same words for each experience: "I noticed the yellow change."

Yet it appears I was more sensitive in the first situation, where I was probably using my peripheral vision (known to be more sensitive to some things). And there's probably also an important factor of "not expecting anything in particular"---I was more open to discovering things.

A better example from music. Let's say I'm listening to rock music and noticing the tight bass. Pehaps it just "comes to my attention" while I'm paying attention to something else or resting my attention broadly immersed in everything. Or, perhaps I deliberately seek it out. In either case, we say in English, "I noticed the tight bass." And like Zen meditators learn, we tend to regard those as the same experience because one set of words describes them. But they are different experiences.

For example, let's say I'm doing a first-time listen to a new recording, and I "happen to notice the tight bass." I will have a particular pathway for experiencing that, a pathway related to its novelty and surprise. Now let's say I listen again, and I'm expecting it. I even try to notice some of the same things about it. That's a different pathway, will create a different conscious experience, and there's no reason to believe it's sensitive to the same things! As Zen meditors learn, though, humans are vulnerable to the illusion that any two experiences described (named, or pointed to) by the same words are really the same experience.


-Mike

 

[Pio2001]

Since the two pathways are symbolic and I have no clue about the experiment that allowed to discover them (and that we would just have to reproduce in order to get the same results in the brain), all I can say is the following.

For the first path, before every playback, think about something else. For example, multiply two numbers mentally, and just think about hitting the play button at a given moment of your calculus.
Sound should hit directly your conciousness without being processed, at least for a third of a second, while you're still calculating.

For the second path, imagine yourself sitting near a dancefloor. You are hesitating about getting up and dance, and tell yourself that it will depend on what the DJ is going to play. Hit the button, and just see what you feel about.

 

[Pio2001]

BTW, if you practice meditation, you may be able to willingly trigger the first pathway with deep meditation : separate completely your attention from the sound until your attention is focused on nothing.
However, according to some scientific experiment, only monks with many many years of practice can willingly and repeatably trigger different kind of brain activity, as measured with functionnal magnetic resonance imagery.

 

[mike1127]

Quote:

Originally Posted by Pio2001 /img/forum/go_quote.gif Since the two pathways are symbolic and I have no clue about the experiment that allowed to discover them (and that we would just have to reproduce in order to get the same results in the brain), all I can say is the following. For the first path, before every playback, think about something else. For example, multiply two numbers mentally, and just think about hitting the play button at a given moment of your calculus. Sound should hit directly your conciousness without being processed, at least for a third of a second, while you're still calculating. For the second path, imagine yourself sitting near a dancefloor. You are hesitating about getting up and dance, and tell yourself that it will depend on what the DJ is going to play. Hit the button, and just see what you feel about.

I very much appreciate your ideas. I like how you don't dismiss my model, but continue to work with it. That's a valuable attitude in life. I will take some of ideas during the blind testing I'm working on today.



Quote:

BTW, if you practice meditation, you may be able to willingly trigger the first pathway with deep meditation : separate completely your attention from the sound until your attention is focused on nothing. However, according to some scientific experiment, only monks with many many years of practice can willingly and repeatably trigger different kind of brain activity, as measured with functionnal magnetic resonance imagery.

I think it may take a lot of practice listening to things blind. I don't think I necessarily have to reach the level of mastery of a monk. There are cross-relationships between blind listening and the kind of meditation a monk does, but they also have different specific foci.