What to trust ear or measurement?

[Sal1950]

Though to play devil's advocate: someone may find a slight coloration in their system amounts to making most recording generally more pleasing, and thus never feel the need to reach for any tone controls or EQ.

And if the reply is "but introducing coloration is a bad thing," then whoever sometimes uses EQ to alter recordings to his taste has to remember that thing about glass houses.

My only thought is,
Don't paint yourselve into a corner with a coloration that can't be shut-off if or when desired.

 

[dualazmak]

The way I'm now approaching it is that I look to instrumented testing to point me towards components that more accurately reproduce the source signal. For example, if I were buying an AVR, the first one I'd consider would be the Denon 3600. Now, with speakers (and possibly amplifiers), I would take the same approach, but considering that you can often hear differences between speakers that measure reasonably well, I would then listen to a bunch of them and evaluate them subjectively, as to how I enjoy listening to them.

Yes, I know the room is critical to how they will sound. The best option would be to evaluate speakers in my own room. At some point, I could see going with a set of speakers that measure worse than another set if I simply enjoy spending time listening to them more than the better measuring speaker. IMO, the point of listening to music is enjoyment, and I'm not sure if I could just convince myself that the other set of speakers measures better and thus should sound better, even though I don't perceive them to sound better.

Hello MarcT,

I like and agree with your approach. I also always like to have my "reference sound" from which I would like to achieve some improvement in step-by-step approach by changing single or minimum factor(s) for possible improvement. If we change multiple factors at once, we cannot identify which would be the deciding factor for the improvement, or if we have plus and minus factors, there should be no improvement at all. I believe this step-by-step approach should be important both for measurements and ear-listening evaluations.

 

[tomelex]

I agree and would add, that is because the term "coloured" has a bad connotation to it. Had you said "euphonic", "pleasant & musical", etc, the response may have been different.

Coming to the original question, my response is contingent upon the fact that few measurements correlate directly and predictably to what we hear.
Hence, the tendency is to "trust one's ears".
The reason here is two-fold: a) measurements; b) personal preferences in sound reproduction, which give further credence to the "trust what my ears tell me" approach.

My personal take is to refer to both measurements (such as are available), and auditioning.
If I had meaningful measurements, i.e. such that can lead me to predict the sonic result, it would help me to make better choices of what to audition. But audition I would, of course, for music is what it's all about, is it not


Hello, this is my first post on these shores... Have a great and safe month, everyone!

"If I had meaningful measurements, i.e. such that can lead me to predict the sonic result, it would help me to make better choices of what to audition"

Hi, welcome from tomelex!

Lets change this around "the fact that few measurements correlate directly and predictably to what we hear" to I would say; if peoples ears can not correlate to measurements, then why trust ears?

1) Adequate (FULL measurements are meaningful and you can have them if you really want to) do describe the system at a technical level for reproduction of the recording, they take a lot of time, they include the room effects, etc,....
2) Would you agree that at your home system you love, and someone did full meaningful measurements as you say, then at another place, if that system also agreed with the measurements from your home, you "should" then be able to say that you have meaningful measurements?
3) Measurements do correlate to what you like, to your preferences!
4) Measurements d e s c r i b e your preferences!

 

[snapsc]

As a manufacturer of audio products... and knowing that any 10 people have different hearing curves, different rooms and different preferences, would you base your business model on designing a product with the best numbers or on a product that may not have the best numbers but that you could sell to the most people?

Most manufacturers say they use engineering and measurements as the first part of their design process and then tweak to what they or their hand picked focus group thinks sounds best.

Assuming they tweak and choose (their sound) wisely.. they should sell a lot of product because they already know it will appeal to people.

So, again, unless the best measurements result in a sound which a lot of buyers agree is a sound they will buy, why would a manufacturer make such a product?

 

[MattHooper]

My only thought is,
Don't paint yourselve into a corner with a coloration that can't be shut-off if or when desired.

Yup, makes sense!

 

[Sal1950]

As a manufacturer of audio products... and knowing that any 10 people have different hearing curves, different rooms and different preferences, would you base your business model on designing a product with the best numbers or on a product that may not have the best numbers but that you could sell to the most people?

Most manufacturers say they use engineering and measurements as the first part of their design process and then tweak to what they or their hand picked focus group thinks sounds best.

Assuming they tweak and choose (their sound) wisely.. they should sell a lot of product because they already know it will appeal to people.

So, again, unless the best measurements result in a sound which a lot of buyers agree is a sound they will buy, why would a manufacturer make such a product?

Many do design some peoples preferences and not transparency.
After the tech design tweaking, that's a good marketing spiel but that's all it is.
Let's look at that another way. If a manufacturer designs a product (DAC AMP, whatever) to be accurate to it's source, but then goes back and modify s it to sound pleasant, he just BROKE it, since it will then never be accurate with any source.
Some of them do this, others don't. That's the reason for doing review measurements like's done here. We need to find out if our windows are clear or tinted.
Then you can buy whatever you chose once you know what your really getting.

 

[solderdude]

As a manufacturer of audio products... and knowing that any 10 people have different hearing curves, different rooms and different preferences, would you base your business model on designing a product with the best numbers or on a product that may not have the best numbers but that you could sell to the most people?

Are you talking about transducers or electronics ? There is a rather large difference between them. The only thing they have in common is that hey both need to be there to produce sound. One is easy to measure and doesn't need to have 'special sound qualities' the other actually determines sound. That's where engineering, circumstances and acoustics make meaningful differences.

 

[Gregm]

"If I had meaningful measurements, i.e. such that can lead me to predict the sonic result, it would help me to make better choices of what to audition"

Hi, welcome from tomelex!

Lets change this around "the fact that few measurements correlate directly and predictably to what we hear" to I would say; if peoples ears can not correlate to measurements, then why trust ears?

I would trust ears because I did not have at hand the measurement that relates to the specific effect I heard. If I had a measurement of that, the measurement would highlight the effect -- OR, if the measurement showed an artifact that I had not detected, then I was not hearing well.

1) Adequate (FULL measurements are meaningful and you can have them if you really want to) do describe the system at a technical level for reproduction of the recording, they take a lot of time, they include the room effects, etc,....
2) Would you agree that at your home system you love, and someone did full meaningful measurements as you say, then at another place, if that system also agreed with the measurements from your home, you "should" then be able to say that you have meaningful measurements?
3) Measurements do correlate to what you like, to your preferences!
4) Measurements d e s c r i b e your preferences!

2) I'm not sure I get your gist, but if you mean that given a set of measurements that fully defines my preferred sound -- then those are meaningful, absolutely!

3 +4 : of course, and like you above, I'd put it another way: there must be elements in the measured numbers that describe my preferences. For example: high dynamic range. Regards!

 

[Kal Rubinson]

So, again, unless the best measurements result in a sound which a lot of buyers agree is a sound they will buy, why would a manufacturer make such a product?

Are you not aware of the Harman test findings and the work of Toole/Olive? Most ("a lot of") buyers prefer speakers with the best measurements in blind tests. The bias of other, non acoustic, factors skews the real market.

 

[QMuse]

Are you not aware of the Harman test findings and the work of Toole/Olive? Most ("a lot of") buyers prefer speakers with the best measurements in blind tests. The bias of other, non acoustic, factors skews the real market.

Speaking of that, when you are reviewing speakers do you peek at the measurements first or not?

 

[snapsc]

Kal... absolutely understand that speakers that measure well are preferred. The implication is that many speaker manufacturers will work toward producing speakers that measure well.

Then... as some have said, they will get their team together and listen... and while they may sound good, the team may feel that tweaking is needed... all, with the ultimate goal of selling more speakers.

In recent years I’ve owned Tekton and Salk speakers. Both sounded great in the same room... but different... the Tekton's dynamic the Salk more refined.

Why... each is a talented designer. Each admits to caring about measurements. Each has an opinion of what sounds “best”. Each sells a lot of speakers.

Just my theory... the market for speakers is made up of many sub markets... each with slightly different owner sound preferences.

The manufacturers know this and try to make sure that their product will have as wide of an appeal as possible and/but at the same time they have a picture in their mind of what sounds right.

Back to the Harman focus groups and measurements... it would be very interesting to follow up with participants at the 1,5 and 10 year marks after their focus groups to see what they own.

If someone just getting into this hobby asked me for advice.... I would advise..

1. Start by listening to those speakers that measure well... if you actually live in an area where you can hear them... which is becoming more and more difficult.

2. If you like more than a few try to narrow your list to a few and try them at home for at least 2 weeks. Be a little analytic and try to identify what you like and don’t like about each.

3. Don’t be afraid of internet brands even though you can’t listen b4 you buy.

4. Remember that it is likely you will want to try something different in the future so take care of them.

So... yes... I think measurements are important and should matter to designers and customers.

 

[snapsc]

Sal... I agree that you should try to know what is designed to be colored ( although designers don’t often admit this) and what was designed to be uncolored and then pick what sounds best in your likely colored listening environment.

And, be aware that over time, tastes often change, rooms often change and then you sell the old and bring in the new.

 

[BDWoody]

Sal... I agree that you should try to know what is designed to be colored ( although designers don’t often admit this) and what was designed to be uncolored and then pick what sounds best in your likely colored listening environment.

And, be aware that over time, tastes often change, rooms often change and then you sell the old and bring in the new.

Or, simply adjust the tone controls if you started with a transparent chain...

 

[Kal Rubinson]

Speaking of that, when you are reviewing speakers do you peek at the measurements first or not?

Not possible. The measurements are not even performed until the review is submitted to the editor.

 

[Kal Rubinson]

Just my theory... the market for speakers is made up of many sub markets... each with slightly different owner sound preferences.

Yes. The Toole/Olive results indicate that the majority of listeners will prefer the flat/accurate speaker sound but there is a distribution, even under blind conditions, that indicates a minority of listeners have different preferences.

If someone just getting into this hobby asked me for advice.... I would advise..
1. Start by listening to those speakers that measure well... if you actually live in an area where you can hear them... which is becoming more and more difficult.
2. If you like more than a few try to narrow your list to a few and try them at home for at least 2 weeks. Be a little analytic and try to identify what you like and don’t like about each.
3. Don’t be afraid of internet brands even though you can’t listen b4 you buy.
4. Remember that it is likely you will want to try something different in the future so take care of them.

And try not to see/know which speakers you are listening to so that you are not influenced by appearance or preconceptions. Hah!

 

[QMuse]

Not possible. The measurements are not even performed until the review is submitted to the editor.

Maybe @amirm should follow the same routine..

 

[Vuki]

Maybe @amirm should follow the same routine..

It would be much more entertaining

 

[thefsb]

Seems to me you worked real hard to write a post that no one would really understand WTF your saying?

I admitted defeat above.

 

[thefsb]

Maybe @amirm should follow the same routine..

I asked about that once. Iiuc, he wants to present his subjective impression taking everything he knows into account. I think that's as valid as any approach to writing a subjective listening review. I compare it to Robert Parker's wine reviews. He is consistent and uses language I can relate to. So after a while I learned what kinds of thing Parker likes and the language he uses for it. He has bad taste but his are among the most useful reviews. Parker, I mean.

 

[ttimer]

Yes. The Toole/Olive results indicate that the majority of listeners will prefer the flat/accurate speaker sound but there is a distribution, even under blind conditions, that indicates a minority of listeners have different preferences.

From what i recall of Tooles research, they chose sets of speakers with wildly different measurements for their experiments. Which makes a lot of sense if you want to determine the basic building blocks of what makes a popular speaker.

However, nowadays, the speakers of mainstream brands in the popular price brackets are much more similar in measurements than what was used in the experiments. Often the modern speakers are quite flat over some parts of the frequency range and less so over others. Which parts are flat and which are not will vary between speakers but there is often no clear hierarchy of "this speaker is more neutral than the other".

My hunch would be that within this practical set of speakers there is more heterogeneity in listener preferences than among the test groups in Tooles experiments.