Your hearing is not that good… Sorry, it’s just not

[JSmith]

But can we actually hear what is being measured

It depends on many variables... sometimes yes, sometimes no. That doesn't mean anything though when referring to device design... personally I want to know my devices are transparent with plenty of headroom in all areas and wish to see the test measurements that show same.


JSmith

 

[Keith_W]

On the flip side, sometimes we can't measure what is easily perceived by the ear. For example, if I toe my speakers in or out by 5 degrees, my ears easily tell the difference. I will bet any amount of money I can detect the change blindfolded. But REW can't. REW may measure a tiny, almost imperceptible, change in dB at higher frequencies, all the while my ears are screaming "are these even the same speakers?"

It is true that microphones do not hear the same as ears, but a 5 degree toe in will definitely show up in REW measurements. You just need to take the right measurement and know where to look. Even if there is no change in the frequency response, you will see the peaks move in the ETC.

 

[GoFaster]

It is true that microphones do not hear the same as ears, but a 5 degree toe in will definitely show up in REW measurements. You just need to take the right measurement and know where to look. Even if there is no change in the frequency response, you will see the peaks move in the ETC.

I'l do it again. REW doesn't see 5 degree toe in. Maybe i'll remeasure tomorrow

 

[Ken Tajalli]

It depends on many variables... sometimes yes, sometimes no. That doesn't mean anything though when referring to device design... personally I want to know my devices are transparent with plenty of headroom in all areas and wish to see the test measurements that show same.

+1

 

[Speedskater]

On the flip side, sometimes we can't measure what is easily perceived by the ear. For example, if I toe my speakers in or out by 5 degrees, my ears easily tell the difference. I will bet any amount of money I can detect the change blindfolded. But REW can't. REW may measure a tiny, almost imperceptible, change in dB at higher frequencies, all the while my ears are screaming "are these even the same speakers?"

Measuring extremely small differences in electronics is easy. Orders of magnitude better than the human ear.
But with transducers (loudspeakers) and rooms, it's an entirely different ballgame.

 

[Cbdb2]

But can we actually hear what is being measured, when all modern equipment is so well designed and built?

The problem is that not all modern equipment is well designed and built. Some audiophool ideas take hold with manufacturers and can "ruin" there designs, like no negative feedback or tuning an amp by ear. And some of these faults are deliberate to make there amps sound different so they stand out when auditioning (a lot of audiophiles like a bit of added distortion). Like the $60k tube amp with 1% distortion at 1watt output and a 10db boost at 100hz. And of course rave reviews in the audio rag.

 

[Henry53]

The problem is that not all modern equipment is well designed and built. Some audiophool ideas take hold with manufacturers and can "ruin" there designs, like no negative feedback or tuning an amp by ear. And some of these faults are deliberate to make there amps sound different so they stand out when auditioning (a lot of audiophiles like a bit of added distortion). Like the $60k tube amp with 1% distortion at 1watt output and a 10db boost at 100hz. And of course rave reviews in the audio rag.

Again when distortion figures are measured in 0.001 increments what relevance is one more 0 on sound?

 

[Henry53]

It is true that microphones do not hear the same as ears, but a 5 degree toe in will definitely show up in REW measurements. You just need to take the right measurement and know where to look. Even if there is no change in the frequency response, you will see the peaks move in the ETC.

But do any of these measurements mean it sounds 'better' to the listener?

 

[Cbdb2]

Again when distortion figures are measured in 0.001 increments what relevance is one more 0 on sound?

You missed my point. I look at measurements, these days, for gross deficiencies which are still out there in both cheap and uber expensive equipment. Especially when your forced to buy online with out auditioning. And you can't believe the reviews in the audio rags.

 

[Keith_W]

But do any of these measurements mean it sounds 'better' to the listener?

As Toole keeps saying, an omnidirectional microphone does not hear the same as a two ears and a brain. What might look horrible in REW might be inaudible or subtly audible in reality. Conversely, what might look subtle might sound very different. Lots of audible phenomena does not even show up on a simple on-axis sine wave sweep. You need special measurement procedures to uncover what is going on. Then once you have the measurements, you have to interpret them with understanding of psychoacoustics - i.e. knowledge of what is perceptible and what is not, what is important and what is not.

And that is only the beginning. You asked if it sounds "better". That is a matter of subjective preference and taste, and that is not universal. You can view preference as a cluster around a target - most of us prefer the same thing, but not all of us do.

 

[Henry53]

As Toole keeps saying, an omnidirectional microphone does not hear the same as a two ears and a brain. What might look horrible in REW might be inaudible or subtly audible in reality. Conversely, what might look subtle might sound very different. Lots of audible phenomena does not even show up on a simple on-axis sine wave sweep. You need special measurement procedures to uncover what is going on. Then once you have the measurements, you have to interpret them with understanding of psychoacoustics - i.e. knowledge of what is perceptible and what is not, what is important and what is not.

And that is only the beginning. You asked if it sounds "better". That is a matter of subjective preference and taste, and that is not universal. You can view preference as a cluster around a target - most of us prefer the same thing, but not all of us do.

There is also perception of what a test, number or measurement means. I can weigh the food on a plate, I can measure its temperature, it tells me nothing about what it tastes like. Adding 5 more zeros to a number has no relevance what so ever if we cannot actually hear the effect of the first 2. I loved a story from uni over 40 years ago. A professor was conducting a long term experiment. It involved a piece of 12v battery powered equipment. He set it up every evening and in the morning found it switched off. He wrote several notes asking politely for it to be left on, to no avail. Finally he wrote a sign, Danger 12,000 millivolts, no one ever switched it off again.

 

[Henry53]

It depends on many variables... sometimes yes, sometimes no. That doesn't mean anything though when referring to device design... personally I want to know my devices are transparent with plenty of headroom in all areas and wish to see the test measurements that show same.


JSmith

And what measurement accurately demonstrates 'transparency'?

 

[Keith_W]

And what measurement accurately demonstrates 'transparency'?

Transparency is easy to demonstrate on electronic equipment. That's why I get bored of electronics reviews, Amir may as well be measuring a hundred apples from the same orchard and saying "this one has 0.05% more sugar than that one". The fun only comes when Amir finds something which is outstandingly bad ... then it's worth reading for the lulz.

Loudspeakers OTOH are more interesting. Not only are they much harder to measure, they are by nature more compromised. Then "transparency" becomes a question of how closely it approximates an ideal loudspeaker.

 

[JSmith]

And what measurement accurately demonstrates 'transparency'?

The general goal is to reduce noise and all forms of distortion, whilst maintaining a flat frequency response in the audible range.


JSmith

 

[Blumlein 88]

Again when distortion figures are measured in 0.001 increments what relevance is one more 0 on sound?

There is no relevance. It has become below audibility at that point.

 

[Cbdb2]

The general goal is to reduce noise and all forms of distortion, whilst maintaining a flat frequency response in the audible range.


JSmith

Yes and I would add lack of dynamic compression, stability (no ringing), and RFI rejection. I guess these could be seen as distortion and noise.

 

[Blumlein 88]

And what measurement accurately demonstrates 'transparency'?

We have had whole threads about where to draw the line. In terms of everything other than the speakers. Flat frequency response +0/-.1 db from 20 hz to 20,000 hz. Distortion below -80 db from max level. SNR of at least 80 db probably like a little more, with quiescent noise not audible. Lots of quibbles and blanks to fill in, but that list is more or less what you need.

 

[solderdude]

While that is correct there is also a thing like different listening SPL and people getting used to a sound signature.
The 'getting used to' is particular annoying (when evaluating over a longer time period) and a blessing at the same time (for owners and sellers).

As an example, when one listens to a head-/ear-phone (or speakers in a room) exclusively for several days, one that does not do anything objectionable but is tonally not conforming to some found target, then the brain adapts and may find the speaker/headphone very convincing and pleasant to listen to. It is only on direct comparison with something that has a 'better' tonal balance.

The weird thing is when listening to the 'skewed' one directly after listening to a 'conforming' one the familiar one sounds 'colored' when using that.... but after a while finds that skewed one to sound 'preferred' again and not colored and finds the 'reference' to sound 'off' again for a short while.
Humans are weird that way and don't act like measuring equipment.

That does not mean one cannot asses coloration or performance of sound with direct comparisons when one has a known reference at hand. Those should apply to the 'rules of transparency' but people may get as much, or even more, enjoyment from one that does not conform to the 'tonal balance' part of 'audible transparency'.

IMO there is leniency for only the 'tonal accuracy' part of 'audible transparency' when it comes to home listening/music enjoyment for the above mentioned reason (the brain able to adapt)

For that reason it is important, when doing evaluations/reviews of transducers to reset the brain with 'known' good performing transducers.

So, agreed that getting the basics right (acoustic 'transparency' for humans) requires minimal technical requirements there should be room for preference (tonal balance/perceived enjoyment of reproduced music).
Science shows there is a range where a tonal balance may fall in rather than a single (averaged) target that must be met.
At least for transducers. For the electronics side the 'transparency' part, obviously, is much better defined/measured.