mike1127
Member of the Trade: Brilliant Zen Audio
- Joined
- Oct 16, 2005
- Posts
- 1,114
- Likes
- 25
I agree with the scientists---tests must be blind to overcome human bias. E.g., food testing scientists have demonstrated that people will prefer the pretty package over the plain one---doesn't even matter what's inside.
But I have reservations about many audio DBTs. Why? In a nutshell, because I doubt that listeners can be sensitive to the right things under the test conditions.
Now, a longer explanation.
The human mind is hierarchical. At the bottom are the sympathetic and parasympathetic nervous systems, our inheritance from early animals. Higher up are more complicated structures, like the reptilian brain and the mammalian brain. At the top is consciousness. Evolution threw that in last.
It's a bit like a military unit, headed up by a sergeant. The men under him take care of a lot of menial tasks automatically. Often they don't even talk to him, when there's no need. Sometimes they do communicate with the sergeant, either bringing things to his attention or answering his questions.
Let's say that sergeant is human consciousness. Under him are guys like Private Ear (he processes sounds and talks to the sergeant about them), Private Movement (he handles body movement), Private Heat Sensitive Nerves, etc.
What happens when you touch a hot stove? Private Heat Sensitive Nerves will detect that and dispatch an urgent message to Sergeant Consciousness. But something happens first. Private Movement gets that message! You jerk your hand back before you even feel the heat.
How do you know, consciously, that you touched something hot? You know it first through movement, and only a moment later through heat nerves.
So the Privates talk to each other as well as to Sergeant Consciousness. You can know things about the world through more than one Private.
I observe in myself, and others have agreed, that listening to music involves not just Private Ear, but also Private Movement and Private Emotion.
I always consciously notice the sound: Private Ear talks to Sergeant Consciousness. But it appears that Private Ear is also having side conversation with Private Movement and Private Emotion. Do I consciously decide what way of moving expresses the music? No. Do I analyze it prior to “deciding” what emotion it should have? Nope. That stuff happens unconsciously.
There are many DBT protocols. I am not arguing that valid DBT is impossible. Let's look at just one protocol first.
Let me repeat: I am not claiming valid DBT is impossible. I think that's the number one misunderstanding when I explain this stuff.
Let's look at just one protocol. It's the music-snippet, quick-switch style. You listen to a very short clip of music on device A, then instantly switch to device B.
As a musician, I'm aware that music has large-scale structure. Often, its effects are apparent only after listening to a decently long segment. As the music plays, Private Ear is communicating stuff to Sergeant Consciousness, but is also talking to Privates Movement and Emotion. What can Sergeant Consciousness pick up in a one-second clip? Certainly a bit of information. But for the test to be valid, we must assume that it doesn't change how Private Ear talks to Private Movements and Emotion. That seems like a shaky assumption.
Let me amplify this point. Let's say I claim that devices A and B sound different. What does that mean, exactly, about how Sergeant Consciousness experiences them? He might pick up the difference from what Sergeant Ear says, but he might only get the difference from what Privates Emotion and Movement say.
Here's the problem. If the test protocol encourages Sergeant Consciousness to talk only to Private Ear, then consciousness will miss the vital difference.
Often someone says to me: “Your point is not relevant---nobody said ABX has to be short-snippet---it can be long listening.”
Well, the vast majority of audio scientists' knowledge about the brain's experience of sound is based on short-snippet tests. Everything they tell you about mp3 codecs, for example. All that research was based on short-snippet tests. So implicating short-snippet tests has some relevance to the state of scientific knowledge about listening.
There is more to say about long-term listening tests, but I'll leave it here for now.
But I have reservations about many audio DBTs. Why? In a nutshell, because I doubt that listeners can be sensitive to the right things under the test conditions.
Now, a longer explanation.
The human mind is hierarchical. At the bottom are the sympathetic and parasympathetic nervous systems, our inheritance from early animals. Higher up are more complicated structures, like the reptilian brain and the mammalian brain. At the top is consciousness. Evolution threw that in last.
It's a bit like a military unit, headed up by a sergeant. The men under him take care of a lot of menial tasks automatically. Often they don't even talk to him, when there's no need. Sometimes they do communicate with the sergeant, either bringing things to his attention or answering his questions.
Let's say that sergeant is human consciousness. Under him are guys like Private Ear (he processes sounds and talks to the sergeant about them), Private Movement (he handles body movement), Private Heat Sensitive Nerves, etc.
What happens when you touch a hot stove? Private Heat Sensitive Nerves will detect that and dispatch an urgent message to Sergeant Consciousness. But something happens first. Private Movement gets that message! You jerk your hand back before you even feel the heat.
How do you know, consciously, that you touched something hot? You know it first through movement, and only a moment later through heat nerves.
So the Privates talk to each other as well as to Sergeant Consciousness. You can know things about the world through more than one Private.
I observe in myself, and others have agreed, that listening to music involves not just Private Ear, but also Private Movement and Private Emotion.
I always consciously notice the sound: Private Ear talks to Sergeant Consciousness. But it appears that Private Ear is also having side conversation with Private Movement and Private Emotion. Do I consciously decide what way of moving expresses the music? No. Do I analyze it prior to “deciding” what emotion it should have? Nope. That stuff happens unconsciously.
There are many DBT protocols. I am not arguing that valid DBT is impossible. Let's look at just one protocol first.
Let me repeat: I am not claiming valid DBT is impossible. I think that's the number one misunderstanding when I explain this stuff.
Let's look at just one protocol. It's the music-snippet, quick-switch style. You listen to a very short clip of music on device A, then instantly switch to device B.
As a musician, I'm aware that music has large-scale structure. Often, its effects are apparent only after listening to a decently long segment. As the music plays, Private Ear is communicating stuff to Sergeant Consciousness, but is also talking to Privates Movement and Emotion. What can Sergeant Consciousness pick up in a one-second clip? Certainly a bit of information. But for the test to be valid, we must assume that it doesn't change how Private Ear talks to Private Movements and Emotion. That seems like a shaky assumption.
Let me amplify this point. Let's say I claim that devices A and B sound different. What does that mean, exactly, about how Sergeant Consciousness experiences them? He might pick up the difference from what Sergeant Ear says, but he might only get the difference from what Privates Emotion and Movement say.
Here's the problem. If the test protocol encourages Sergeant Consciousness to talk only to Private Ear, then consciousness will miss the vital difference.
Often someone says to me: “Your point is not relevant---nobody said ABX has to be short-snippet---it can be long listening.”
Well, the vast majority of audio scientists' knowledge about the brain's experience of sound is based on short-snippet tests. Everything they tell you about mp3 codecs, for example. All that research was based on short-snippet tests. So implicating short-snippet tests has some relevance to the state of scientific knowledge about listening.
There is more to say about long-term listening tests, but I'll leave it here for now.