Measured Improvement or Non-measured Improvement?

[dallasjustice]

This is a poll with only 2 choices.

"Improvement" is defined as a subjective audible improvement at seated position. The improvement is equally valuable, whether measured or not.

"Measured" is defined as an actual measurement that confirms the improvement before and after the improvement is made. This definition does not include a theoretical possibility for such a confirmatory measurement. For example, a microphone frequency response measurement at seated position confirming the improvement.

 

[Speedskater]

Measurements (at the listening position) just don't work that way!

 

[Kal Rubinson]

Is this a trick?

Measured improvement? Reassuring.
Non-measured improvement (assuming this means measured results are null and not that no measurements were done)? Provocative.

 

[dallasjustice]

If you feel that measurements are totally meaningless and can never be used to demonstrate any audible improvement, this may not be your favorite poll.

 

[JackD201]

I use measurements but probably nowhere near as in depth as you do dallas. Put me in what Kal calls the reassuring camp but that's it. I'm fully aware of my measuring and measurement assessment limitations so that makes me very careful to not read too much into what I'm seeing. Confirmation bias can be every bit as strong as expectation bias after all. I don't know how to even vote on this poll since I do need to rely on both to different degrees in different instances.

 

[Kal Rubinson]

If you feel that measurements are totally meaningless and can never be used to demonstrate any audible improvement, this may not be your favorite poll.

I am not in that camp. I am also not in the camp that feels that measurements and/or double-blind auditions are essential.

 

[dallasjustice]

I am not in that camp. I am also not in the camp that feels that measurements and/or double-blind auditions are essential.

Agreed. I don't discount experiences not supported with measurements. However, if you had an equally satisfying improvement available to your listening experience, would you prefer to also have a confirmation from measurements taken before and after the improvement? If given a choice, wouldn't you want to have some knowledge as to why something is an improvement?

 

[andy_c]

This is a poll with only 2 choices.

"Improvement" is defined as a subjective audible improvement at seated position. The improvement is equally valuable, whether measured or not.

"Measured" is defined as an actual measurement that confirms the improvement before and after the improvement is made. This definition does not include a theoretical possibility for such a confirmatory measurement. For example, a microphone frequency response measurement at seated position confirming the improvement.

It's a messy question. If one measures smoothed steady-state on-axis frequency response in the statistical region (above Schroeder frequency) and EQs the result to flat, the subjective result will be horribly bright and harsh. This will be confirmed by quasi-anechoic on-axis measurements which will show a response that rises with increasing frequency. So this metric is very poor for predicting subjective performance.

OTOH, if one EQs the steady-state response below the Schroeder frequency to be flat, and even better, with multiple subs, as flat as possible at multiple seating positions (without boosting nulls), this will almost surely result in a subjective improvement, unless one prefers bass suckouts or one-note bass. I'm assuming here that multiple subs would be used to fill in minor nulls, rather that trying to boost them with EQ.

In short, the metric must be known to correlate with improved sound quality to make sense, and many acoustical measurements, such as the aforementioned steady-state high-frequency response, have been shown to be poor choices as a metric.

 

[Kal Rubinson]

Agreed. I don't discount experiences not supported with measurements. However, if you had an equally satisfying improvement available to your listening experience, would you prefer to also have a confirmation from measurements taken before and after the improvement? If given a choice, wouldn't you want to have some knowledge as to why something is an improvement?

But of course! Even a fervent subjectivist would.

 

[Kal Rubinson]

It's a messy question.

Yes.

In short, the metric must be known to correlate with improved sound quality to make sense, and many acoustical measurements, such as the aforementioned steady-state high-frequency response, have been shown to be poor choices as a metric.

I was assuming that the measurements (whatever they are) were appropriate to the matter at hand.

 

[dallasjustice]

But of course! Even a fervent subjectivist would.

Feel free to vote for the first option.

 

[dallasjustice]

I included the word confirm in my definition. Therefore, the hypothetical assumes you have a choice between a confirmed measured improvement and a non-measured improvement. I know some folks don't think it's ever possible to use a microphone to confirm through measurement any subjective improvement. I am not interested in changing minds. I just want to know whether folks use a form of the precautionary principle in their decision making process.

If equal choices are competing for my dollar, one is measured and the other is not, I always choose the measured one. This is the precautionary principle. The idea is that I choose externally demonstrated improvements for 2 reasons. First, it helps deal with confirmation bias if the measurement is a standard and well accepted one. Second, the measurements help refine future decisions. I can create a trail of measurements correlated to improvements. Such a trail helps me make future decisions.
Michael.

It's a messy question. If one measures smoothed steady-state on-axis frequency response in the statistical region (above Schroeder frequency) and EQs the result to flat, the subjective result will be horribly bright and harsh. This will be confirmed by quasi-anechoic on-axis measurements which will show a response that rises with increasing frequency. So this metric is very poor for predicting subjective performance.

OTOH, if one EQs the steady-state response below the Schroeder frequency to be flat, and even better, with multiple subs, as flat as possible at multiple seating positions (without boosting nulls), this will almost surely result in a subjective improvement, unless one prefers bass suckouts or one-note bass. I'm assuming here that multiple subs would be used to fill in minor nulls, rather that trying to boost them with EQ.

In short, the metric must be known to correlate with improved sound quality to make sense, and many acoustical measurements, such as the aforementioned steady-state high-frequency response, have been shown to be poor choices as a metric.

 

[mep]

I'm trying to ascertain the purpose of you starting this thread Dallas. I almost feel like it's one of those questions where there is no 'right' answer. Sort of like asking someone if they have stopped beating their wife yet.

 

[still-one]

Measurements are not required. I either prefer A or B. I am looking for sound not specs.

 

[Ethan Winer]

If you feel that measurements are totally meaningless and can never be used to demonstrate any audible improvement, this may not be your favorite poll.

LOL.

Are you aware of the research Sean Olive from Harmon did? He found that listeners universally prefer speakers that measure flat:

Listeners prefer loudspeakers that are flat

So I think it's a mistake to discount measurements. Another problem with assessing stuff by ear alone is the key of the music interacts with peaks and nulls in the room. So going by ear alone to position speakers, for example, might sound great with one piece of music but terrible with another. Versus measuring the room which shows all frequencies at once.

--Ethan

 

[rockitman]

180 views and two votes. Not a very popular poll I guess...

 

[mep]

180 views and two votes. Not a very popular poll I guess...

Meaningless in my opinion.

 

[ddk]

This is a poll with only 2 choices.

"Improvement" is defined as a subjective audible improvement at seated position. The improvement is equally valuable, whether measured or not.

"Measured" is defined as an actual measurement that confirms the improvement before and after the improvement is made. This definition does not include a theoretical possibility for such a confirmatory measurement. For example, a microphone frequency response measurement at seated position confirming the improvement.

Why don't you have a button for both? There are things that can be measure and many more that can't, both are important. Measurements are only a tool and worthless without the right interpretation and implementation, irrespective of a theoretical improvement or decline.

david

 

[dallasjustice]

Why don't you have a button for both? There are things that can be measure and many more that can't, both are important. Measurements are only a tool and worthless without the right interpretation and implementation, irrespective of a theoretical improvement or decline.

david

How can one have a confirmatory measurement and no measurement at the same time? Sorry, I never studied quantum mechanics. Maybe it's possible.

 

[ddk]

How can one have a confirmatory measurement and no measurement at the same time? Sorry, I never studied quantum mechanics. Maybe it's possible.

I said negative or positive measurement not no measurement. The problem with measurements and DSP, you can take a measurement that shows a peak or a dip and get the DSP to fix it and get the better measurements. Many times the better measurement can sound worse. You can also start with great, flat measurements but you have a dull sound and you might need to introduce peak somewhere to bring life to the room, in this case the worse measurement will sound better. Then there's the situation with electricity, you can get better measurements in certain parameters with a high quality transformer installed but sonically there are penalties. We've gone over this in another thread where I find the noisier, worse measuring mains preferable to the cleaner, quieter sound of the transformer.There's value in the measurements but at the same time you can't measure perception and very important things like tone, tonal range, timbre, solidity and naturalness. IMO both have their place.

david