“Human ears are better than measurements!” / “Humans can hear things we can’t even measure!”

[Pauli]

The fact that we don’t know all links between subjective impressions and objective data yet also does not mean that two devices which measure identical in every way might still sound different under otherwise identical boundary conditions: They wont.

The weak point of this statement is the »identical boundary conditions« phrase.

The term »boundary« needs to be defined in order to give any argumentative value to that phrase. But by doing so (or trying to do so) you won’t be able to keep clear of notoriously debatable areas. Does »boundary« mean the listening room – or the whole floor that room is part of? Or the whole house that room and floor are part of? Does »boundary« includes the physical conditions of the persons present? Their body temperature? Their bowel activities? Their transpiration? Does it include the current weather and climate? The current time of day? Does it cover the astronomical positions of planet earth over time, between several occasions of listening sessions? Etc etc etc ...

You know what I mean. Drawing a clear and unquestionable line here about what »boundary« should mean would be as necessary as it would be impossible.

You could wail »Oh – come on! Simply use common sense here!« But that won’t help at all ...

 

[krabapple]

IMO dismissing consistent feedback from experienced reviewers and listeners on flagship products as mere bias is an oversimplification. Measurements are essential for understanding performance, but they don't ultimately determine whether a component is enjoyable to listen to. Many of the industry's leading manufacturers deliberately voice their products for the most satisfying listening experience, even when that means deviating slightly from what appears "perfect" on a measurement graph.

So.......do they all aim for the same 'most satisfying listening experience' or do they all define it differently?

And whose listening experience are they satisfying? By what process is this determined?

The claims in your post are incoherent. You say that carefully controlled listening tests are vital yet you give credence to 'independent reviewers' who, apparently listening sighted, converge on what you say is a common perception of midrange 'leanness' in an amplifier.

Show us the measurements.

 

[sergeauckland]

IMO dismissing consistent feedback from experienced reviewers and listeners on flagship products as mere bias is an oversimplification. Measurements are essential for understanding performance, but they don't ultimately determine whether a component is enjoyable to listen to. Many of the industry's leading manufacturers deliberately voice their products for the most satisfying listening experience, even when that means deviating slightly from what appears "perfect" on a measurement graph.

I appreciate ASR's emphasis on objective measurements, but the new Marantz Model 10 is an interesting example. It's an uncompromising, ultra high end, Class D amplifier developed by some of the industry's most accomplished engineers, backed by an enormous R&D budget and world-class component sourcing. Yet a number of independent reviewers have consistently described its midrange as leaner than the traditional Marantz Class AB sound.

That doesn't necessarily mean those impressions are correct, but it does suggest that today's standard measurements - frequency response, THD, SINAD, and similar metrics may not fully describe every aspect of perceived sound quality. Characteristics such as harmonic structure, dynamic intermodulation distortion, transient behavior, or complex loudspeaker load interactions may contribute to qualities like midrange body, tonal richness, or emotional engagement in ways that current measurement suites don't yet fully quantify.

This doesn't diminish the value of objective measurements - they remain indispensable. It simply acknowledges that, at least for now, carefully controlled listening tests are still the ultimate benchmark for assessing how enjoyable a product is in real-world use.

But that has nothing to do with technical performance. It has everything to do with the emotional response to looks, feel, reputation and cost. Nobody has EVER shown that in a blind listening test when the emotional responses are removed, there's any difference between components that can't be easily explained by the measurements.

I happen to like valves and vinyl, not because of their technical performance, which is woeful, but because they look nice, and I enjoy the technical challenge of getting a pleasing sound out of such flawed technology. Much in the same way of steam locomotives or vintage cars.

S.

 

[Mart68]

Yet a number of independent reviewers have consistently described its midrange as leaner than the traditional Marantz Class AB sound

Just link to where they did a back to back blind comparison. Except you can't because they didn't. In such reviews it's commonplace for the reviewer to refer back to an amplifier he reviewed months, or even years back. He may genuinaly believe he recalls its 'sound'. He doesn't.

People assume that because the reviewers can seemingly do this, that they can too. Moreover detecting these 'subtle but important differences' has become a flag of expertise and discernment and so the result is lots of tedious, misleading nonsense.

 

[OCA]

But that has nothing to do with technical performance. It has everything to do with the emotional response to looks, feel, reputation and cost. Nobody has EVER shown that in a blind listening test when the emotional responses are removed, there's any difference between components that can't be easily explained by the measurements.

I happen to like valves and vinyl, not because of their technical performance, which is woeful, but because they look nice, and I enjoy the technical challenge of getting a pleasing sound out of such flawed technology. Much in the same way of steam locomotives or vintage cars.

S.

True, but I’m not buying audio gear solely for technical performance beyond what I can actually hear. When two components measure similarly, I choose the one that consistently delivers more listening pleasure and musical engagement. Sure, it could involve bias but if that bias makes me enjoy the music more every single time, it’s a useful one. The gear’s purpose is to enhance the listening experience, not win measurement contests. For some of us, cost, looks, or reputation aren’t driving factors but listening experience is.

Besides, who says the measurements themselves are infallible? ASR relies heavily on one person, his very specific setup and unique abilities and experience for most reviews. Replicating accurate SINAD or distortion measurements at home is extremely difficult, most people would introduce far more errors than the subtle listening biases they criticize. In that sense, over reliance on measurements carries its own risks.

 

[RandomEar]

IMO dismissing consistent feedback from experienced reviewers and listeners on flagship products as mere bias is an oversimplification. Measurements are essential for understanding performance, but they don't ultimately determine whether a component is enjoyable to listen to. Many of the industry's leading manufacturers deliberately voice their products for the most satisfying listening experience, even when that means deviating slightly from what appears "perfect" on a measurement graph.

It's fine to prefer speakers with a non-flat FR or tube amps with audible distortion. Nobody is taking that away from you by measuring those things. With speakers in particular, their performance depends on the interaction with your room. That does not mean that things which measure identical sound different. These are two totally different discussions to be had.

I appreciate ASR's emphasis on objective measurements, but the new Marantz Model 10 is an interesting example. It's an uncompromising, ultra high end, Class D amplifier developed by some of the industry's most accomplished engineers, backed by an enormous R&D budget and world-class component sourcing. Yet a number of independent reviewers have consistently described its midrange as leaner than the traditional Marantz Class AB sound.

Those reviewers, did they test correctly and in a controlled manner? If not, this is just the same useless bubble again which reviewers and manufacturers have been pushing for decades. This is like relying on a colorblind person wo doesn't know they are vision-impaired in that way to tell you which colors are the same.

That doesn't necessarily mean those impressions are correct, but it does suggest that today's standard measurements - frequency response, THD, SINAD, and similar metrics may not fully describe every aspect of perceived sound quality. Characteristics such as harmonic structure, dynamic intermodulation distortion, transient behavior, or complex loudspeaker load interactions may contribute to qualities like midrange body, tonal richness, or emotional engagement in ways that current measurement suites don't yet fully quantify.

It does not suggest that. That's the whole point. They tested in an uncontrolled manner and got uncontrolled, biased subjective results. The fact that they agree on something is no argument either, because group think is strong amongst subjectivists.

This doesn't diminish the value of objective measurements - they remain indispensable. It simply acknowledges that, at least for now, carefully controlled listening tests are still the ultimate benchmark for assessing how enjoyable a product is in real-world use.

Devices can lead to different levels of enjoyment but still sound 100% identical. Nicer knobs, higher price, better UI? All that influences sighted listening perceptions - as dumb as it is, the sound is not the thing these people really compare, even though they think they do.

 

[OCA]

Just link to where they did a back to back blind comparison. Except you can't because they didn't. In such reviews it's commonplace for the reviewer to refer back to an amplifier he reviewed months, or even years back. He may genuinaly believe he recalls its 'sound'. He doesn't.

People assume that because the reviewers can seemingly do this, that they can too. Moreover detecting these 'subtle but important differences' has become a flag of expertise and discernment and so the result is lots of tedious, misleading nonsense.

FWIW It was a back to back test with a Rotel Michi Class AB known for its no signature sound added honest sound production and he preferred the Marantz throughout, only mentioned thinner midrange compared to the Michi.

 

[DoctorRobert]

IMO dismissing consistent feedback from experienced reviewers and listeners on flagship products as mere bias is an oversimplification. Measurements are essential for understanding performance, but they don't ultimately determine whether a component is enjoyable to listen to. Many of the industry's leading manufacturers deliberately voice their products for the most satisfying listening experience, even when that means deviating slightly from what appears "perfect" on a measurement graph.

I appreciate ASR's emphasis on objective measurements, but the new Marantz Model 10 is an interesting example. It's an uncompromising, ultra high end, Class D amplifier developed by some of the industry's most accomplished engineers, backed by an enormous R&D budget and world-class component sourcing. Yet a number of independent reviewers have consistently described its midrange as leaner than the traditional Marantz Class AB sound.

That doesn't necessarily mean those impressions are correct, but it does suggest that today's standard measurements - frequency response, THD, SINAD, and similar metrics may not fully describe every aspect of perceived sound quality. Characteristics such as harmonic structure, dynamic intermodulation distortion, transient behavior, or complex loudspeaker load interactions may contribute to qualities like midrange body, tonal richness, or emotional engagement in ways that current measurement suites don't yet fully quantify.

This doesn't diminish the value of objective measurements - they remain indispensable. It simply acknowledges that, at least for now, carefully controlled listening tests are still the ultimate benchmark for assessing how enjoyable a product is in real-world use.

This is puzlling, coming from you. I have stated in a number of other threads that I really like what you have done for room correction. My home cinema system is more enjoyable because of that. But for me, that also proofs that room integration / room correction is where the magic is. The focus on mysterious not-understood immeasurable aspects of audio components is useless. I just don't believe that there is anything there (see the original post), but even if you do, it is dwarfed by the room and its interactions with speakers and sound waves. Discussions about metallic sounding metal tweeters, brittle sounding digital amps, most DAC-discussions, it's all about listening with your eyes. If you want to improve your sound quality, get components that measure right (check ASR) and improve/optimize the interaction with the room.

There is still enough there to keep this forum going.

 

[RandomEar]

The weak point of this statement is the »identical boundary conditions« phrase.

The term »boundary« needs to be defined in order to give any argumentative value to that phrase. But by doing so (or trying to do so) you won’t be able to keep clear of notoriously debatable areas. Does »boundary« mean the listening room – or the whole floor that room is part of? Or the whole house that room and floor are part of? Does »boundary« includes the physical conditions of the persons present? Their body temperature? Their bowel activities? Their transpiration? Does it include the current weather and climate? The current time of day? Does it cover the astronomical positions of planet earth over time, between several occasions of listening sessions? Etc etc etc ...

You know what I mean. Drawing a clear and unquestionable line here about what »boundary« should mean would be as necessary as it would be impossible.

You could wail »Oh – come on! Simply use common sense here!« But that won’t help at all ...

The listener is not part of the chain here when talking about "sounds identical". Of course, people can have vastly different listening impressions of two devices under otherwise identical boundary conditions. Those are subjective impressions and they are only loosely correlated with the actual sound waves reaching their ears. They may prefer the "sound" of the more expensive amp or DAC or speaker, or of the one with the red chassis, even though they perform identical from a technical perspective. That's the whole problem with sighted, uncontrolled listening. It does not give you results which are trustworthy. They can change from one hour to the next, or if you learn that one device has specific properties you did not know about before.

I agree, though, that the defning the exact boundary conditions for a type of device get's tricky real fast. For electronics, I think it wil be doable. For speakers it might be very difficult today. There's at least one good thread about it on ASR already - probably many more

 

[Mart68]

FWIW It was a back to back test with a Rotel Michi Class AB known for its no signature sound added honest sound production and he preferred the Marantz throughout, only mentioned thinner midrange compared to the Michi.

Is there a link to this review, out of interest? I thought it was relating to a difference from the alleged Marantz 'house sound' not an amp from another manufacturer - not that it matters.

 

[OCA]

Is there a link to this review, out of interest? I thought it was relating to a difference from the alleged Marantz 'house sound' not an amp from another manufacturer - not that it matters.

https://darko.audio/2026/03/the-dramatic-sound-forgiveness-rotel-michi-x3-series-2-review/

 

[Mart68]

Sure, it could involve bias but if that bias makes me enjoy the music more every single time, it’s a useful one.

This can't be argued with - but when that bias is costing extra thousands don't you think twice about how worthwhile it is?

FWIW I work on a similar principle - If I like the look of the equipment then I know I'm likely to enjoy it more so I will pay a premium for that, if necessary. But this is all psychological and nothing to do with the O/P.

 

[OCA]

This is puzlling, coming from you. I have stated in a number of other threads that I really like what you have done for room correction. My home cinema system is more enjoyable because of that. But for me, that also proofs that room integration / room correction is where the magic is. The focus on mysterious not-understood immeasurable aspects of audio components is useless. I just don't believe that there is anything there (see the original post), but even if you do, it is dwarfed by the room and its interactions with speakers and sound waves. Discussions about metallic sounding metal tweeters, brittle sounding digital amps, most DAC-discussions, it's all about listening with your eyes. If you want to improve your sound quality, get components that measure right (check ASR) and improve/optimize the interaction with the room.

There is still enough there to keep this forum going.

The DSP methods you enjoy along with many others were developed through extensive listening tests, not measurements alone. That's the point I'm trying to make: quantifying what the human brain ultimately perceives as "better" is far more complex than many people assume and should not just be discarded as bias.

I can produce ten different DSP solutions with nearly identical frequency and phase responses, yet they won't all sound the same. If I dismissed those audible differences as mere listener bias instead of investigating them, I would never have arrived at a robust processing method.

Measurements are indispensable, but they don't always tell us which solution people will consistently prefer. In practice, careful listening remains an essential part of developing DSP algorithms.

 

[sergeauckland]

True, but I’m not buying audio gear solely for technical performance beyond what I can actually hear. When two components measure similarly, I choose the one that consistently delivers more listening pleasure and musical engagement. Sure, it could involve bias but if that bias makes me enjoy the music more every single time, it’s a useful one. The gear’s purpose is to enhance the listening experience, not win measurement contests. For some of us, cost, looks, or reputation aren’t driving factors but listening experience is.

Besides, who says the measurements themselves are infallible? ASR relies heavily on one person, his very specific setup and unique abilities and experience for most reviews. Replicating accurate SINAD or distortion measurements at home is extremely difficult, most people would introduce far more errors than the subtle listening biases they criticize. In that sense, over reliance on measurements carries its own risks.

But listening pleasure and musical engagement are what YOU generate. Someone else might hate what you've bought, even though on blind listening, they are both identical.

The gear's purpose is to reproduce music. YOU get listening pleasure from using it, I might not, for exactly the same equipment. So it's not the listening experience you like, as that's identical. It's YOUR emotional response to the equipment, which gives you pleasure.

If we accept that one piece of equipment is only 'better' because WE prefer it, nothing to do with the equipment's performance, then all this nonsense about how one item is better than another goes away. Personal preference, not equipment performance.

As to measurements not being infallible, provided they are done correctly, they pretty much are. It's trivially easy to do SINAD measurements at home. If people just invested the same amount in measuring equipment as they do in useless cables, we'd all have decent measuring equipment at home.

S.

 

[krabapple]

True, but I’m not buying audio gear solely for technical performance beyond what I can actually hear. When two components measure similarly, I choose the one that consistently delivers more listening pleasure and musical engagement.

You call them that, but if the two measure the same then it's probably got nothing to do with what you 'can actually hear' -- the sound waves or the music.

This is the whole basis of blinding as a control to ensure that only what is actually heard matters.

Sure, it could involve bias but if that bias makes me enjoy the music more every single time, it’s a useful one.

OK , but the bias comes from non-audible factors. It's a mistake/delusion to believe otherwise.

 

[Mart68]

https://darko.audio/2026/03/the-dramatic-sound-forgiveness-rotel-michi-x3-series-2-review/

Doesn't level match, doesn't have fast switching comparison, not blind. Most likely - he is just mistaken.

Also he uses vinyl for some of the comparison without taking into account that the two phono stages may have different RIAA EQ curves.

 

[OCA]

As to measurements not being infallible, provided they are done correctly, they pretty much are. It's trivially easy to do SINAD measurements at home. If people just invested the same amount in measuring equipment as they do in useless cables, we'd all have decent measuring equipment at home.

I agree with you about the cable part. That said, for many people, listening is still much easier than making high-quality measurements.

Even Amir occasionally revises measurements after discovering an issue with the test setup or methodology. I could probably measure the SINAD of my own equipment and eventually arrive at a value that's reasonably close, but reproducing the accuracy and repeatability of a well-equipped lab is another matter.

Small variables can influence the result: the quality and noise on the AC mains (which fluctuates throughout the day where I live), environmental electrical noise, the operating temperature of the device, the measurement setup itself, calibration, grounding, and many other factors. None of these make measurements unreliable. They simply illustrate that obtaining laboratory-grade measurements consistently requires controlled conditions, experience, and appropriate equipment.

Listening, by comparison, has a much lower barrier to entry. The challenge is not whether to measure or listen, but understanding what each method can and cannot tell us.

 

[ahofer]

Listening, by comparison, has a much lower barrier to entry.

I notice that “it smells like a public pool” always wins over “I tested the water and the chlorine is high”. People would rather participate in a judgement than accept it from a third party.

 

[Redacted]

Can we have a new thread dedicated to "stuff I like to make up, 'cos my ears are more special than yours"?

And "Tales of mystery and imagination from Planet Audiophile"?

"The X-files - the truth about ear wax is out there"

 

[DoctorRobert]

The DSP methods you enjoy along with many others were developed through extensive listening tests, not measurements alone. That's the point I'm trying to make: quantifying what the human brain ultimately perceives as "better" is far more complex than many people assume and should not just be discarded as bias.

I can produce ten different DSP solutions with nearly identical frequency and phase responses, yet they won't all sound the same. If I dismissed those audible differences as mere listener bias instead of investigating them, I would never have arrived at a robust processing method.

Measurements are indispensable, but they don't always tell us which solution people will consistently prefer. In practice, careful listening remains an essential part of developing DSP algorithms.

I understand that the correction target is validated by listening tests. But you correct based on measurements. If it cannot be measured, you could not correct/optimize it. What you have created with your listening tests, is a model of how an ideal listening room should sound, I think. And you are correcting with the DSP for what the actual room does 'wrong'. The only thing is, that when it comes to acoustics, the simple 'straight wire' model that works for audio electronics, does not work any more. But it is still measurable.