NORMS AND STANDARDS FOR DISCOURSE ON ASR

[Absolute]

My man @svart-hvitt, I wish you had this much drive and enthusiasm for something substantial and meaningful instead of insisting on these legendary endeavors into nonsense-land.

Serious question; do you actually read and remember all of your various sources and quote them as you come across relevant stuff or do you google keywords that suit your next argument and copy-paste the highlights where applicable?

 

[svart-hvitt]

My man @svart-hvitt, I wish you had this much drive and enthusiasm for something substantial and meaningful instead of insisting on these legendary endeavors into nonsense-land.

Serious question; do you actually read and remember all of your various sources and quote them as you come across relevant stuff or do you google keywords that suit your next argument and copy-paste the highlights where applicable?

I have always had a fascination for epistemology, from I was a young boy, long before I knew the word and the term. About 15-20 years ago I started reading more formal texts about it.

I have about 20 years of experience with vox populi based designs, not least a database of millions of datapoints for real-life «polling», which also enables me to create virtual «persons» to test ideas and hypotheses. Some of the insights from real-life vox populi mechanisms can be duplicated using simulations or math. In some ways, the insights are astonishingly simple, but very few people realize and act on these general «rules».

 

[Wombat]

'e pissed 'em off, ably, seems to the outcome re ongoing epistemology focus.

Seriously, your enthusiasm for epistemology is unlikely to be shared by most forum members. One of the first considerations for general communication is to 'know' your audience in order to kindle their interest. Another is not to overload them with too many points and too much information, in order to maintain their attention.

 

[Blumlein 88]

To be clear speaking for myself, I don't feel pissed off. EDIT:It isn't a topic for me unless spoon fed. So I might not have much to contribute anyway.

 

[watchnerd]

I gave up on epistemology when I looked at the salaries of epistemologists.

 

[watchnerd]

To be clear speaking for myself, I don't feel pissed off. But it itsn't a topic for me unless spoon. So I might not have much to contribute anyway.

 

[Cosmik]

We're planning to have a garden party in a few weeks' time. In case it rains, the wife and I have discussed whether we need to order a marquee to be on the safe side, and if so, could we get away with not ordering it until we know we need it. The wife said that we could consult the weather forecast a week before the party. I replied that the weather forecast is not reliable enough to work that far in advance. Thinking like a scientist she proposed an experiment: keep monitoring the weather forecast for the next few weeks and see whether it really does work a week in advance.

I believe this is what science is for many people. From the outside, it is not possible to see the 'dimensionality' of a system. An experiment addresses 2 dimensions, say, and if it 'works' repeatably it is declared that the problem is solved. But the problem isn't solved, and like the infamous millennium bridge in London, another dimension comes along to ruin things. So you solve that problem with some more ad hoc testing and patching up, but without some pure thought - as opposed to empirical testing - you don't know that there isn't another dimension waiting in the wings. Nor do you know that you don't know ((c) Donald Rumsfeld).

It has been proposed that as long as an experiment has the so-called 'minimum hypothesis' attached to it ("If I do this, something interesting will happen.") it is still science. If so, I think science has got problems.

 

[Wombat]

We're planning to have a garden party in a few weeks' time. In case it rains, the wife and I have discussed whether we need to order a marquee to be on the safe side, and if so, could we get away with not ordering it until we know we need it. The wife said that we could consult the weather forecast a week before the party. I replied that the weather forecast is not reliable enough to work that far in advance. Thinking like a scientist she proposed an experiment: keep monitoring the weather forecast for the next few weeks and see whether it really does work a week in advance.

I believe this is what science is for many people. From the outside, it is not possible to see the 'dimensionality' of a system. An experiment addresses 2 dimensions, say, and if it 'works' repeatably it is declared that the problem is solved. But the problem isn't solved, and like the infamous millennium bridge in London, another dimension comes along to ruin things. So you solve that problem with some more ad hoc testing and patching up, but without some pure thought - as opposed to empirical testing - you don't know that there isn't another dimension waiting in the wings. Nor do you know that you don't know ((c) Donald Rumsfeld).

It has been proposed that as long as an experiment has the so-called 'minimum hypothesis' attached to it ("If I do this, something interesting will happen.") it is still science. If so, I think science has got problems.

Weather forecasting includes probability. The weather is not a controlled entity although it does lend itself to uncertain predictability.

 

[Cosmik]

Weather forecasting includes probability.

So does coin tossing prediction.

 

[Wombat]

So does coin tossing prediction.

For an outdoor function I'd go with the professional forecasts. Feel free to toss a coin.

 

[Absolute]

I have always had a fascination for epistemology, from I was a young boy, long before I knew the word and the term. About 15-20 years ago I started reading more formal texts about it.

Interesting fascination that, but do you read and remember the sources or do you google them up to find papers and quotes you find relevant to your point?

The reason I'm asking is because many of those papers seems so far off in relevancy that it comes across as a form of debating technique where you bombard readers with "authorities" wherever there's a semi-fitting quote or remark to strengthen your case.

I have about 20 years of experience with vox populi based designs, not least a database of millions of datapoints for real-life «polling», which also enables me to create virtual «persons» to test ideas and hypotheses. Some of the insights from real-life vox populi mechanisms can be duplicated using simulations or math. In some ways, the insights are astonishingly simple, but very few people realize and act on these general «rules».

In what form do you have experience with such designs? I could always say I have experience with vox populi based designs working as a seller of Volkswagen or Apple laptops without that giving me much authoritative weight.

What do you mean by "creating virtual persons to test hypotheses?" Creating internet troll personas to study reactions?
It all sounds a bit vague when you formulate yourself so casually and non-specific

Forgive me if this comes across as passive-aggressive, I genuinely don't intend it that way, but I'm sure you can see how the thought process may lead people to this line of thinking.

 

[svart-hvitt]

Interesting fascination that, but do you read and remember the sources or do you google them up to find papers and quotes you find relevant to your point?

The reason I'm asking is because many of those papers seems so far off in relevancy that it comes across as a form of debating technique where you bombard readers with "authorities" wherever there's a semi-fitting quote or remark to strengthen your case.

In what form do you have experience with such designs? I could always say I have experience with vox populi based designs working as a seller of Volkswagen or Apple laptops without that giving me much authoritative weight.

What do you mean by "creating virtual persons to test hypotheses?" Creating internet troll personas to study reactions?
It all sounds a bit vague when you formulate yourself so casually and non-specific

Forgive me if this comes across as passive-aggressive, I genuinely don't intend it that way, but I'm sure you can see how the thought process may lead people to this line of thinking.

Would my answering of your questions in any way alter your existing opinion?

Some of the references I used, are things I already read. The references used in the opening post are something that came my way this summer. Some - but not all - of the other references I’ve used are things that I find to illustrate how others answered similar questions. I am open on my sources of ideas instead of pretending it all came from my own head. I have never heard anyone getting formal critique for being open on references used.

Readers of this thread complain about my writing style. That’s fair. I believe it comes from people not being used to discussing epistemologic questions and a need for instant gratification, plus downright poor writing - which is a sin I hardly do on my own on ASR. If I showed you an SNR measurement of a dac, a more familar theme providing you with an instant gratification through a data point, you’d be more happy, I guess.

You say you don’t understand what I’m writing because of messy composition; but isn’t your hostility a sign that some of the content in what I wrote provoked you? So are you complaining about substance in content or a feeling of cognitive dissonance?

You ask me about my experience in designs of say optimal behaviour in decision making processes. So let me give you an exercise. Genelec’s R&D director Aki Mäkivirta recently wrote that «Neutrality is the safe choice». And he added: «In many walks of life, it is wise to stay neutral. This saves you from a lot of trouble...».
Source: https://www.genelec.com/blog/neutrality-safe-choice

How would you go on to prove Dr. Mäkivirta right or wrong? You can use math, simulations, empirical data/observations or prose to prove your position.

 

[svart-hvitt]

No. I used the term generically as people do in conversation. For @svart-hvitt to run with that word is beyond pedantic. I have explained the reason for not doing DAC listening tests countless times. It is #1 in the FAQ for measurements, linked to from the home page: https://www.audiosciencereview.com/...rstanding-audio-measurements.2351/#post-65101

View attachment 30358

If he believes DACs should be double blind tested, he should take the initiative to conduct them, together with resources to run the tests and publish them. Until then, I appreciate not hearing yet again "why do you measure." I do because hardly anyone in the industry does. Many of you can't do these measurements but can conduct blind tests. So if it is important to you, go and do it. You don't need a $30,000 instrument.

If all of this is cover by him to complain for the sake of complaining (which is what it reads to me), then stop. None of this is constructive and is wasting the forum resources.

@amirm , you wrote (my underlining):

«None of this is constructive and is wasting the forum resources».

I just checked the review of the Totaldac d1-six dac, which measures poorly and costs €13.500:

https://www.audiosciencereview.com/...and-measurements-of-totaldac-d1-six-dac.8192/

Since July 25th, there are 957 posts in that thread

Talk about waste and the efficient use of resources...

 

[Blumlein 88]

Rightly or wrongly:

 

[Cosmik]

Genelec’s R&D director Aki Mäkivirta recently wrote that «Neutrality is the safe choice». And he added: «In many walks of life, it is wise to stay neutral. This saves you from a lot of trouble...».
Source: https://www.genelec.com/blog/neutrality-safe-choice

But he stresses that it is flatness at the listener's ears as measured in a certain way, using an automated gizmo. But should we really be aiming for flat frequency response at the listener's ears? Or just a flat speaker, allowing the room to do what it will do?

The same company gives us the answer without realising it:

Once GLM has served you a flat in-room frequency response on a silver plate, the result might be brighter than you like. While a flat frequency response is necessary as a reference, it may not be perceived as the best option with all types of content and environments, particularly when working long hours and monitoring above 80 dB SPL. I therefore typically roll off gently by around 3 dB above 10 kHz, but that's a question of personal taste, and may be influenced by room and listening distance also.

I really, really love that paragraph. It sums up the issue perfectly: for all the talk of neutrality measured at the listener's ears, it is perceived as "too bright" (as predicted by we neutral speaker people), so feel free to fiddle with the EQ manually until it's not so painful! It is the perfect encapsulation of the debate, and shows why the "neutral at the listener's ears" is wrong. The listener reads the room automatically, and adjusts their hearing to perceive a neutral sound from a neutral speaker.

 

[PierreV]

You say you don’t understand what I’m writing because of messy composition; but isn’t your hostility a sign that some of the content in what I wrote provoked you? So are you complaining about substance in content or a feeling of cognitive dissonance?

I am willing to bet that quite a few of us just go "oh, no, not again" and skip most of what you have to say because of the semi-related citations bombing.

Source: https://www.genelec.com/blog/neutrality-safe-choice
How would you go on to prove Dr. Mäkivirta right or wrong? You can use math, simulations, empirical data/observations or prose to prove your position.

You don't. That Dr. Mäkivirta makes a fuzzy analogy to push his brand marketing message. Whether he is right in some kind of absolute framework upon which humans have not agreed yet, whether he is right in his narrow field of experience, whether you agree with him or do not is irrelevant.

There's no way a fuzzy rhetorical analogy can be proven wrong or right.

 

[mitchco]

One of Sean Olive's conclusions in his study on: The Subjective and Objective Evaluation of Room Correction Products

"A flat in-room target response is clearly not the optimal target curve for room equalization. The preferred room corrections have a target response that has a smooth downward slope with increasing frequency. This tells us that listeners prefer a certain amount of natural room gain. Removing the room gain, makes the reproduced music sound unnatural, and too thin, according to these listeners. This also makes perfect sense since the recording was likely mixed in room where the room gain was also not removed; therefore, to remove it from the consumers' listening room would destroy spectral balance of the music as intended by the artist."

A well done evaluation and certainly has been my experience with so called room correction products. A couple of slides worth pointing out from the study. From Sean’s slide deck is a preferred subjective ranking of average magnitude responses, objectively measured at the primary listening position:

The top preference (red trace) is a flat, but tilted measured response. If 0 dB is 20 Hz, then it would be a straight line to -10 dB at 20 kHz.

Note that this tilted measured in-room response is perceived by our ear/brain, as subjectively flat or a neutral response according to Sean’s research:

See how an objectively measured response of 20 Hz and straight line to -10 dB at 20 kHz (in-room) is subjectively perceived as a neutral or flat response to our ears/brain (red trace overlaid in the above chart). Most participants in the study preferred a frequency response from 20 Hz with a straight line to -10 dB at 20 kHz. A measured “flat” in-room frequency response is not the preferred target, as it sounds too thin or lacking bass.

Link to PDF slide show.

 

[Cosmik]

One of Sean Olive's conclusions in his study on: The Subjective and Objective Evaluation of Room Correction Products

"A flat in-room target response is clearly not the optimal target curve for room equalization. The preferred room corrections have a target response that has a smooth downward slope with increasing frequency. This tells us that listeners prefer a certain amount of natural room gain. Removing the room gain, makes the reproduced music sound unnatural, and too thin, according to these listeners. This also makes perfect sense since the recording was likely mixed in room where the room gain was also not removed; therefore, to remove it from the consumers' listening room would destroy spectral balance of the music as intended by the artist."

A well done evaluation and certainly has been my experience with so called room correction products. A couple of slides worth pointing out from the study. From Sean’s slide deck is a preferred subjective ranking of average magnitude responses, objectively measured at the primary listening position:

View attachment 30468

The top preference (red trace) is a flat, but tilted measured response. If 0 dB is 20 Hz, then it would be a straight line to -10 dB at 20 kHz.

Note that this tilted measured in-room response is perceived by our ear/brain, as subjectively flat or a neutral response according to Sean’s research:

View attachment 30469

See how an objectively measured response of 20 Hz and straight line to -10 dB at 20 kHz (in-room) is subjectively perceived as a neutral or flat response to our ears/brain (red trace overlaid in the above chart). Most participants in the study preferred a frequency response from 20 Hz with a straight line to -10 dB at 20 kHz. A measured “flat” in-room frequency response is not the preferred target, as it sounds too thin or lacking bass.

Link to PDF slide show.

"This tells us that listeners prefer a certain amount of natural room gain."

No it doesn't. It's an observation of a correlation that has no explanatory power.

This is actually the perfect example of the debate. A correlation between preference and frequency response is not an explanation of it, even if the experimenter decided at the outset that this was his hypothesis. It's like a weather forecaster noticing that pine cones close up in wet weather and deciding that it is the pine cones that are causing it to rain.

This is where 'science' loses its philosophical moorings and simply observes correlations and doesn't think about it any further. A neutral speaker in a room will give the sloping response, but it isn't the in-room frequency response that is the cause of perceived neutrality. Many different in-room frequency responses will give the same perceived neutrality if the experiment is repeated in many different rooms with different speakers. The listener is hearing the neutral speaker, not the in-room frequency response. But there's a 'blob' of average-ish speakers in average-ish rooms that all produce a more-or-less similar response when set up something like correctly (i.e. with some compensation for their non-neutral dispersion). This is all the experiment shows.

But different types of speakers vary in terms of dispersion. That's a whole layer of confusion that a purely empirical approach will fall foul of: "I don't understand it. The box speaker sounded fine with the Harperson 37B target response, but the panel speaker with an identical measured response at the listener's ear sounds different."

 

[pkane]

This is where 'science' loses its philosophical moorings and simply observes correlations and doesn't think about it any further.

Right. It’s as far as science can go in the absence of a better model of human perception and cognition. It’s not impossible to imagine that in the future we’ll be able to do better by constructing a more detailed mathematical model derived from base principles instead of statistics. But for now, we observe and correlate. This data may yet become a test case for a more complete/comprehensive model sometime in the future.

 

[svart-hvitt]

"This tells us that listeners prefer a certain amount of natural room gain."

No it doesn't. It's an observation of a correlation that has no explanatory power.

This is actually the perfect example of the debate. A correlation between preference and frequency response is not an explanation of it, even if the experimenter decided at the outset that this was his hypothesis. It's like a weather forecaster noticing that pine cones close up in wet weather and deciding that it is the pine cones that are causing it to rain.

This is where 'science' loses its philosophical moorings and simply observes correlations and doesn't think about it any further. A neutral speaker in a room will give the sloping response, but it isn't the in-room frequency response that is the cause of perceived neutrality. Many different in-room frequency responses will give the same perceived neutrality if the experiment is repeated in many different rooms with different speakers. The listener is hearing the neutral speaker, not the in-room frequency response. But there's a 'blob' of average-ish speakers in average-ish rooms that all produce a more-or-less similar response when set up something like correctly (i.e. with some compensation for their non-neutral dispersion). This is all the experiment shows.

But different types of speakers vary in terms of dispersion. That's a whole layer of confusion that a purely empirical approach will fall foul of: "I don't understand it. The box speaker sounded fine with the Harperson 37B target response, but the panel speaker with an identical measured response at the listener's ear sounds different."

I think you’re circling in on the main issue with the «gold standard», the vox populi method of research.

In vox populi one assumes that vox=veritas, to keep it in latin, or preference=truth in English.

In «Galton’s ox» we saw that vox was a very good indicator of veritas, but Galton had the advantage of a reference of truth, i.e. the ox weight is indicated by a calibrated weight. In audio related vox populi research, the calibrated reference a.k.a. truth isn’t available. Now, Freeman Dyson comes to mind again:

«It’s a very dangerous game. If you work with a computer model for years and years and years, always improving the model, in the end you end up believing it. So it’s difficult to remain objective».

Even if vox populi isn’t Dyson’s «computer model», it’s still a model where full calibration to truth is not possible. Now, let’s manipulate Dyson a bit:

«If you work with [a vox populi model] for years and years, always improving the model, in the end you end up believing it. So it’s difficult to remain objective».

Since the «gold standard» is so heavily associated with vox populi, why not consult the masters of vox populi research on the matter, i.e. the economist. Somewhat surprising, maybe for believers in the «gold standard» of audio science, that the economist is he who knows the most about vox populi, including its epistemology.

I already referred to Akerlof (2019) and Rayner (2012) - in which a failed research program was discussed. Aren’t all the dots becoming visible now for everyone to connect?