McGurk can’t be applied to HiFi

[Simon 13th]

https://www.13thnote.net/2020/07/11/applying-bias-from-the-mcgurk-effect-to-audiophile-hifi/

what do you all think?

 

[Jimbob54]

I think you need to read this thread and if possible the linked actual studies https://audiophilestyle.com/forums/topic/59351-does-bias-affect-audio-test-results/#comments relating to the actual subject under discussion, being perception of audio equipment.

I also think referencing McGurk in an area that doesnt involve visual speech in any way, shape or form is just plain odd.

So, basically, I'm giving you an F. Possibly one the dog should have eaten.

 

[Jimbob54]

On the plus side, I would say you got the title of the thread absolutely correct.

 

[ahofer]

https://www.13thnote.net/2020/07/11/applying-bias-from-the-mcgurk-effect-to-audiophile-hifi/

what do you all think?

I think you are cherry-picking from the many ways bias operates in order to dismiss the effect of bias and noise on your decision-making. The McGurk effect is one small example of a huge constellation of bias issues. Another poster even gave you a map of those in your prior thread. For instance, your dismissal of the appearance of the amplifier in this article - that wouldn’t be McGurk, it would likely be confirmation bias. And you haven’t addressed “noise”, as decision-making science defines it, which needs to be controlled as well.

As I said in that other thread, researchers have repeatedly demonstrated that avoiding bias is about method and process, not being a special person, or self-appointed expert, and I quoted Danny Kahneman saying as much. You seem to keep using your listening experience as an excuse to dismiss precisely the process and method that could eliminate the noise and bias in your reviews.

I’m sorry, but I think you have barely scratched the surface of audio engineering and decision-making research, and your methods are proudly unscientific. That’s fine, but don’t expect applause when you bring it here.

Incidentally, since you seem to believe expertise-by-doing is a prophylactic against bad decision-making, I invite you to read some works by Philip Tetlock - https://www.theatlantic.com/magazine/archive/2019/06/how-to-predict-the-future/588040/

 

[ahofer]

I think you need to read this thread and if possible the linked actual studies https://audiophilestyle.com/forums/topic/59351-does-bias-affect-audio-test-results/#comments relating to the actual subject under discussion, being perception of audio equipment.

I also think referencing McGurk in an area that doesnt involve visual speech in any way, shape or form is just plain odd.

So, basically, I'm giving you an F. Possibly one the dog should have eaten.

I’ve seen people use McGurk - there was some objectivist panel at an AES forum where it appeared prominently, along with playing Stairway to Heaven. It’s a nice example of how our auditory system can be misled by queues from other senses, but I agree that it can be inapposite to how most people listen to hifi equipment. Our OP seems to think we view McGurk as the definitive dunk on subjectivism, as opposed to an enormous body of research that supports the commonly-held null hypotheses about audio equipment around here - particularly that measurements define all of what electronics can do to change the *audible* signal.

 

[Chrispy]

I think if its from your site it can usually be ignored, like your reviews....

 

[solderdude]

The McGurk effect shows our brain interprets the same auditory input differently depending on visual input.
It has nothing to do with enjoying music in so far there is no related visual input.
Just like the Laurel/Jenny thing but that does not use visual clues. It's the same audio input but brains make something different of it depending on some factors.

The moral of the story is... your brain is a poor measurement device but is really suited for enjoying music.

When watching music videos on crappy equipment the music does sound better than when the visual stimulus is switched off.
A music video sounds better when looking at the actual instruments being played (live performances and such).
Not looking at pop music with unrelated pictures for the sake of art.

 

[pozz]

@Simon 13th I think the point of your article was that certain cognitive biases can be defeated with enough training or ability. While true, the training is highly specific to each kind of activity. You were probably implying that audiophiles, because they have dedicated themselves to listening, are exempt from certain experiences of more casual listeners. In that case It would be really instructive to read a little into the background of psychological testing and how the validity of well-known methods like double blinding was established, and then move into reading the results of scientific listening tests.

It is pretty likely that there is a kind of assessment where audiophiles perform better. But looking at the serious published work it is not in recognizing small impairments or distortions, understanding tonal quality or assessing equipment (beyond matters like ease of use).

 

[Koeitje]

McGurk relates to the interaction between vision and hearing in speech perception. Of course it doesn't relate to hifi directly, but as solderdude points out it does show that the human brain is very easily fooled. When I was at university I followed some courses on linguistics and related subjects, and one of the things that becomes extremely clear is that while the human brain is capable of fascinating feats it also cannot be trusted.

Even if your argument that trained listeners are immune to biases is true (which I don't think it is), then it still makes a case for an objective approach to hifi. Because trained listeners are a tiny fraction of those interested in hifi. You aren't even a trained listener, because you spout so much nonsense in your reviews. Or maybe you are and are just spouting that crap to make an extra buck from advertisers.

 

[Simon 13th]

McGurk relates to the interaction between vision and hearing in speech perception. Of course it doesn't relate to hifi directly, but as solderdude points out it does show that the human brain is very easily fooled. When I was at university I followed some courses on linguistics and related subjects, and one of the things that becomes extremely clear is that while the human brain is capable of fascinating feats it also cannot be trusted.

Even if your argument that trained listeners are immune to biases is true (which I don't think it is), then it still makes a case for an objective approach to hifi. Because trained listeners are a tiny fraction of those interested in hifi. You aren't even a trained listener, because you spout so much nonsense in your reviews. Or maybe you are and are just spouting that crap to make an extra buck from advertisers.

i don’t think the human brain is fooled. The point of the article too was you can’t extrapolate one known mechanism carte Blanche to such thinking. Much like say the knee jerk reflex, which is a known pathway.

In saying common sense is such that if you are alive to what biases their might be - wanting it to be better for more money, brand name, price and so on, and look at it objectively.

 

[Simon 13th]

The McGurk effect shows our brain interprets the same auditory input differently depending on visual input.
It has nothing to do with enjoying music in so far there is no related visual input.
Just like the Laurel/Jenny thing but that does not use visual clues. It's the same audio input but brains make something different of it depending on some factors.

The moral of the story is... your brain is a poor measurement device but is really suited for enjoying music.

When watching music videos on crappy equipment the music does sound better than when the visual stimulus is switched off.
A music video sounds better when looking at the actual instruments being played (live performances and such).
Not looking at pop music with unrelated pictures for the sake of art.

that was exactly the point the article defeats. That the point is, at least on this effect alone, you can’t apply carte Blanche thinking of one well known phenomena to all.

if our brains interpreted auditory input differently that would make navigating ambulances with sirens a very tricky task.

to say the brains is a poor measurement advice. It’s all you’ve got!

 

[Jimbob54]

@Simon 13th I think the point of your article was that certain cognitive biases can be defeated with enough training or ability. While true, the training is highly specific to each kind of activity. You were probably implying that audiophiles, because they have dedicated themselves to listening, are exempt from certain experiences of more casual listeners. In that case It would be really instructive to read a little into the background of psychological testing and how the validity of well-known methods like double blinding was established, and then move into reading the results of scientific listening tests.

It is pretty likely that there is a kind of assessment where audiophiles perform better. But looking at the serious published work it is not in recognizing small impairments or distortions, understanding tonal quality or assessing equipment (beyond matters like ease of use).

I actually think he was saying cognitive bias can be avoided by simply being conscious of it. Which implies it is a trap to be avoided, rather than a factor to be designed out of the test.

 

[voodooless]

In saying common sense is such that if you are alive to what biases their might be - wanting it to be better for more money, brand name, price and so on, and look at it objectively.

You really can’t. You assume these processes are conscious. In reality they are mostly not. You actually have no idea what influences you, and in what way. Therefore you cannot unlearn everything. Also note that your auditory memory is only about 4 seconds. There is no way you can compare two devices days or even weeks apart. It just doesn’t make sense.

to say the brains is a poor measurement advice. It’s all you’ve got!

It’s the interpretation that is poor. The measurements themselves are probably quite good. And no, it’s not all we have. We’ve invented an AP device to help us make objective measurements.

 

[GGroch]

I’ve seen people use McGurk - there was some objectivist panel at an AES forum where it appeared prominently, along with playing Stairway to Heaven....

Here is the AES forum where at 5:20 McGurk is discussed. Simon, I agree that the McGurk effect is not a common issue when listening to HiRes audio in the home, (unless we are invoking it for fun by simultaneously reading hilariously erroneous lyrics). However, that McGurk is uncommon does not invalidate the impact of other bias on listening. You state: It it’s hard to make a case that the quality of the audio we hear is being adversely affected by our brain pathways. But our brain pathways always affect what we hear, there is no way around it. Cognitive bias cannot be eliminated by being aware of it.

Please view JJ Johnston's section of the above AES forum where he lays out an abbreviated description of how this works. J.J.'s far more detailed presentations on the topic are linked elsewhere on ASR. It is worth noting that J.J.'s background includes years at DBX where understanding how our brains interpret what we hear is a core competency. How we hear ties directly to how we think, and I will second Ahofer's recommendation of Daniel Kahneman on Thinking Fast and Slow. This outlines why bias cannot be thought or prayed away.

I am skeptical of whether measurement and blind testing accurately model the music listening experience. I question it not because I dispute bias. Bias is an inherent part of all normal music listening. When we seek to eliminate it, we create an artificial test environment that does not perfectly model the real world.

 

[Simon 13th]

You really can’t. You assume these processes are conscious. In reality they are mostly not. You actually have no idea what influences you, and in what way. Therefore you cannot unlearn everything. Also note that your auditory memory is only about 4 seconds. There is no way you can compare two devices days or even weeks apart. It just doesn’t make sense.


It’s the interpretation that is poor. The measurements themselves are probably quite good. And no, it’s not all we have. We’ve invented an AP device to help us make objective measurements.

they are very conscious. I know what influences me, but less so since reviewing as I have no motives to want it to be better at price or whatnot. I deconstructed that auditory memory point in the last post because if I put you in a concert hall blindfolded, then move you to home listening to hifi, you’d know Where you are. The question of extent to that point must by implication, be a more valid one.

 

[voodooless]

they are very conscious. I know what influences me, but less so since reviewing as I have no motives to want it to be better at price or whatnot.

You can believe that all you want.. that doesn’t make it true.

I deconstructed that auditory memory point in the last post because if I put you in a concert hall blindfolded, then move you to home listening to hifi, you’d know Where you are.

No, you just trawmened your way out it it..

 

[JeffS7444]

Well, you're right about the McGurk Effect not applying to our hifi listening unless one is watching videos of someone saying "Faa faa faa" (or is it "Baa baa baa"?). You seem to be falling into the trap of extrapolating things where you shouldn't, and applying a dose of rhetoric to make your point.

Perhaps we should start referring to double-blind ABX testing as "Pure listening tests". These require a bit of extra rigor to set up, but I'm sure it will be a trivial matter for the experienced listener/reviewer to perform and breeze through.

 

[voodooless]

I am skeptical of whether measurement and blind testing accurately model the music listening experience. I question it not because I dispute bias. Bias is an inherent part of all normal music listening. When we seek to eliminate it, we create an artificial test environment that does not perfectly model the real world.

You are 100% right. What is does however is save a lot of people a whole lot of money. I don’t think it is about the experience in the first place. It’s just about object data. That you eventually pay more for the product that looks nicer or because the brand name RME is on it.. that is personal perference. Then you might ask: why not leave everything to personal preference? Well, mostly because data shows that for some things preferences are figements of our imagination.

 

[Simon 13th]

Here is the AES forum where at 5:20 McGurk is discussed. Simon, I agree that the McGurk effect is not a common issue when listening to HiRes audio in the home, (unless we are invoking it for fun by simultaneously reading hilariously erroneous lyrics). However, that McGurk is uncommon does not invalidate the impact of other bias on listening. You state: It it’s hard to make a case that the quality of the audio we hear is being adversely affected by our brain pathways. But our brain pathways always affect what we hear, there is no way around it. Cognitive bias cannot be eliminated by being aware of it.

Please view JJ Johnston's section of the above AES forum where he lays out an abbreviated description of how this works. J.J.'s far more detailed presentations on the topic are linked elsewhere on ASR. It is worth noting that J.J.'s background includes years at DBX where understanding how our brains interpret what we hear is a core competency. How we hear ties directly to how we think, and I will second Ahofer's recommendation of Daniel Kahneman on Thinking Fast and Slow. This outlines why bias cannot be thought or prayed away.

I am skeptical of whether measurement and blind testing accurately model the music listening experience. I question it not because I dispute bias. Bias is an inherent part of all normal music listening. When we seek to eliminate it, we create an artificial test environment that does not perfectly model the real world.

Its not the brain pathways that affect what we hear but how the auditory cortex must relate to the other parts of the brain - the memory centres to recognise them and so on. So for instance if I’m really tired and walking up a hill and I see the hill as steep I will see it the same steepness but my interpretation will say that’s a really steep hill. Thats where the bias comes in.

I hear this debate in hifi a lot and I think it’s a mis understanding of the biology of how we work. So my brain pathway hasn't effected how steep it is. Can you imagine, if I started walking and it went from 30 degrees to 50 degrees because, as you say, it was inter-related, and I gave up walking on a easy scramble. Our environment would be very degraded if that were the case. Think how much it could apply to mess up your everyday life.

I watched a bit of JJ Johnston’s but gave up - this isn’t applying biology. He says initially that if you concentrate on one thing you will remember one thing and that’s why you need to do double blind. Total horsesh1t , sorry. what if you arent concentrating on one thing. What if when I eat a steak I concentrate on the sauce, how it’s cooked, the cut and so on. I fail to understand how people are taken in by this?

 

[Jimbob54]

Here is the AES forum where at 5:20 McGurk is discussed. Simon, I agree that the McGurk effect is not a common issue when listening to HiRes audio in the home, (unless we are invoking it for fun by simultaneously reading hilariously erroneous lyrics). However, that McGurk is uncommon does not invalidate the impact of other bias on listening. You state: It it’s hard to make a case that the quality of the audio we hear is being adversely affected by our brain pathways. But our brain pathways always affect what we hear, there is no way around it. Cognitive bias cannot be eliminated by being aware of it.

Please view JJ Johnston's section of the above AES forum where he lays out an abbreviated description of how this works. J.J.'s far more detailed presentations on the topic are linked elsewhere on ASR. It is worth noting that J.J.'s background includes years at DBX where understanding how our brains interpret what we hear is a core competency. How we hear ties directly to how we think, and I will second Ahofer's recommendation of Daniel Kahneman on Thinking Fast and Slow. This outlines why bias cannot be thought or prayed away.

I am skeptical of whether measurement and blind testing accurately model the music listening experience. I question it not because I dispute bias. Bias is an inherent part of all normal music listening. When we seek to eliminate it, we create an artificial test environment that does not perfectly model the real world.

The whole of that linked video should be mandatory viewing for all members, especially like me with no acoustic/ EE education or practical knowledge.