Master Thread: Are measurements Everything or Nothing?

[nonnyno]

But they just haven't. Believing is an act of faith, we're discussing science.


I am a non-expert and I'm challenging these assertion because I haven't seen the proof. I cannot prove it or disprove it myself because I'm no expert, but perhaps, if a proof was given, I could understand it.


Science here means data-driven? Like what, sociology? That's not real science.

I think some humility is in order. Not speaking about you specifically here. Our host for example knows this (and I appreciate it), but plenty of members don't. They're the most dangerous kind of people: those who don't know but think they know.

be careful saying science is about being data driven. Much of science is entirely theory or model driven and its also about context. We can't prove or disprove the existence of god with science. We think the big bang happened but aren't certain. There's data that supports certain theories but we aren't sure that the theories are right. I prefer to think about science much as per this definition; the intellectual and practical activity encompassing the systematic study of the structure and behavior of the physical and natural world through observation and experiment.

That then leaves things open to further development and change.

 

[Pugsly]

What I don't understand about the subjectivity is that if this site has its desired effects the subjectivists will still be able to buy stupidly priced gear (with measurements beyond even what a bat with golden ears could pick up). It's just that it will be truer to the original recording - truer hifi. They can still carry out their hobby and play with components and let their subjectivity dictate choices. Even though all the kit will sound the same, in reality the biases we humans create will mean it doesn't for them unless they do a blind test and as they don't listen to music through blind tests all the time, the game doesn't have to end. They'll just get better quality gear out of it - not be bluffing themselves that they have better quality gear. What have they got against getting better designed gear?

 

[Robin L]

Had to get rid of 7 turntables and 2000 LPs about 3 years ago. I do not regret it.

 

[symphara]

be careful saying science is about being data driven. Much of science is entirely theory or model driven and its also about context. We can't prove or disprove the existence of god with science. We think the big bang happened but aren't certain. There's data that supports certain theories but we aren't sure that the theories are right. I prefer to think about science much as per this definition; the intellectual and practical activity encompassing the systematic study of the structure and behavior of the physical and natural world through observation and experiment.

That then leaves things open to further development and change.

I actually fully agree with everything you wrote. There’s a difference though between “we know X because we can prove it“ and “we think X is true because we have some data showing it’s true, we do some modeling etc”. The latter approach, at least in software, means bugs. Sometimes extraordinarily bad ones.

One of the biggest issues I have, for example, is the “in-room predicted response” graphs. Who’s room? My room applies a completely different envelope function over sound compared to your room, and I have no way (insufficient data and modeling) to accurately obtain the said response in my room. On top of that, I don’t know what other characteristics about a speaker makes it pleasant to me (like, say, step response), and if they’re measured, how to exactly interpret the measurements and what to look for.

If anything, some measurements can show that electronics or speakers can be bad. In this respect they are useful. More than that, I’m in doubt.

 

[Suffolkhifinut]

be careful saying science is about being data driven. Much of science is entirely theory or model driven and its also about context. We can't prove or disprove the existence of god with science. We think the big bang happened but aren't certain. There's data that supports certain theories but we aren't sure that the theories are right. I prefer to think about science much as per this definition; the intellectual and practical activity encompassing the systematic study of the structure and behavior of the physical and natural world through observation and experiment.

That then leaves things open to further development and change.

The thing about science is it follows a very narrow path. If science is true it will try to disprove what it’s trying to prove, true science is honest, no one should doubt that. Before you start a program must be set out while ignoring everything else.
When the Leakey family found the earliest known human remains in Kenya, Archeologists went to help them investigate. The Leakeys sent them all packing as their training led them down a predetermined path. The Leakeys believed their closed minds prevented them looking for the truth or as we would call it Fidelity.

 

[nonnyno]

Hey nonnyno
I think even I can welcome you to this site from my lofty coupla months here!
I will not be able to debate with you. However, are you saying you approve of the site and its brief, or not?

Yes absolutely I approve wholeheartedly of this site - I have become a veritable fanboi (I've always wanted to say that). I am as I said a relaxed objectivist. I sympathize with and understand how the subjectivists come to their way of thought, even if ultimately I don't agree with them. I don't think either side helps meet in the middle because I think the way we communicate with each other (language used) has been badly thought through but that is a wider issue than just on this site. All the subjectivists hear is criticism of their chosen widget (and thus they switch off and stop hearing us and criticise us) because we have stated very bluntly that it measures badly and thus must be cr*p as a result. All we hear is that they think we're ideological idiots and we must be wrong as their widget sounds awesome they hear it and everyone else tells them so. Ultimately the communication ensures that both sides thinks the other is stupid. I mostly focus on the measurements to define the pool of equipment from what I will pick and from that pool I use my other choice criteria. I accept that subjectivists will be happy with their lot too. Ultimately if they insist on calling components, hi fidelity, when they patently aren't then they're bluffing themselves. If they simply want to create a sound they enjoy more (which is largely about perception) and don't much care about measurements or sonic engineering then I hope they enjoy themselves - they're not doing anyone any harm. They're buying perception and as long as they can afford them who am I to quibble.

 

[nonnyno]

The thing about science is it follows a very narrow path. If science is true it will try to disprove what it’s trying to prove, true science is honest, no one should doubt that. Before you start a program must be set out while ignoring everything else.
When the Leakey family found the earliest known human remains in Kenya, Archeologists went to help them investigate. The Leakeys sent them all packing as their training led them down a predetermined path. The Leakeys believed their closed minds prevented them looking for the truth or as we would call it Fidelity.

I think I only partly agree with you. There are current scientific developments and approaches that are realising the limits of single and narrow discipline approaches to science - climate change research for example. Economics has chosen to go the opposite route as it tried to gain credibility by placing itself in the class as a pure science - its not. its a social science and that is not a bad thing. There are a growing number of eminent economists who believe that this approach has been a mistake for economics.

 

[nonnyno]

LOL, yes I agree but if they're happy with their sunk costs then so be it. I have a friend who has lots of money and spends a fortune on hifi. We argue endlessly over beer, sound clips and good food whilst his wife rolls her eyes at us and says boys and their toys and then says but are you happy - Great times. He doesn't deny that my kit sounds very good but he's convinced that his has the edge at orders of magnitude more money sunk. It does but not by that much.

 

[Suffolkhifinut]

I think I only partly agree with you. There are current scientific developments and approaches that are realising the limits of single and narrow discipline approaches to science - climate change research for example. Economics has chosen to go the opposite route as it tried to gain credibility by placing itself in the class as a pure science - its not. its a social science and that is not a bad thing. There are a growing number of eminent economists who believe that this approach has been a mistake for economics.

Economics isn‘t a science, it is anything but! Keynesian economics promote increasing public spending to stimulate the economy. This is only true when you have the means of production. Applied now it would do nothing to help the UK economy. While science is always true regardless of circumstances.

 

[alitomr1979]

@BDWoody I just ordered the pre90 to extend the inputs of my preamp, and will buy locally the mixer I need to use my microphone in order to tests the DACs I have here. I know data not only suggests that I wont be able to notice differences, they practically predict I wont, but I will perform the tests to provide MORE evidence to the subjectivists in the thread.

Question to the knowledgeable people in these matters: Is it acceptable to level match using pink noise?

 

[alitomr1979]

Economics isn‘t a science, it is anything but! Keynesian economics promote increasing public spending to stimulate the economy. This is only true when you have the means of production. Applied now it would do nothing to help the UK economy. While science is always true regardless of circumstances.

NO.

 

[Robin L]

if you quote me, please quote everything and not just half a sentence you can use to twist it

I quoted a whole thought that turns out to be a sentence onto itself. There's no requirement to include the rest of your post to point out a major fallacy in that post. The "golden ear" is a high-end myth that gets disproved by double blind testing.

 

[nonnyno]

I actually fully agree with everything you wrote. There’s a difference though between “we know X because we can prove it“ and “we think X is true because we have some data showing it’s true, we do some modeling etc”. The latter approach, at least in software, means bugs. Sometimes extraordinarily bad ones.

One of the biggest issues I have, for example, is the “in-room predicted response” graphs. Who’s room? My room applies a completely different envelope function over sound compared to your room, and I have no way (insufficient data and modeling) to accurately obtain the said response in my room. On top of that, I don’t know what other characteristics about a speaker makes it pleasant to me (like, say, step response), and if they’re measured, how to exactly interpret the measurements and what to look for.

If anything, some measurements can show that electronics or speakers can be bad. In this respect they are useful. More than that, I’m in doubt.

I can see where your scepticism comes from. I have the same thoughts. I reframed thus. I need some objective basis for measurement. That basis needs to be as theoretically clean as possible (e.g. introduce as few factors or variations on each measure as possible). The test bench approach does this. You then say well if it tests well on the bench it is likely but not guaranteed to test well and sound ok somewhere else. It may not sound pleasing to a user in their environment but we can guarantee that it will produce near as dammit the signal that was recorded. You might have to tweak stuff to get something that sound pleasing in your environment and for your ears (assuming of course the signal produced is something that you would have found pleasing in the first place ;-> no amount of tweaking will ever make death core sound good to me - sorry to all those death core fans). That is possible through peq (or other stuff I don't understand yet) or by adding something else to the mix. What is true is that it doesn't matter what you do to an item which measures badly (assuming that the bad measurements are audible). You might be able to make it sound pleasing but it won't play you what was recorded. Measuring well simply says its capable of reproducing what was recorded without bias. If you don't mind that you wont ever be able to hear the music you attempted to play as it was intended by the producer no matter what you do to the equipment then that's fine. If you do mind then that's an issue. That's what I am using the measurements as the basis for. I want to know that true reproduction is possible. that doesn't mean I wont want to color it afterwards. Its possibly a different attitude.

 

[nonnyno]

Economics isn‘t a science, it is anything but! Keynesian economics promote increasing public spending to stimulate the economy. This is only true when you have the means of production. Applied now it would do nothing to help the UK economy. While science is always true regardless of circumstances.

Actually that is entirely incorrect. Science doesn't say that something has to apply always. It says it must apply predictably within a set of criteria (some of which may be circumstantial). The example you've given is simply talking about applicability and not the definition of science. Keynesian stimulus works where there are no blockages to the productive consumption of the money. If there are such blockages then you get inflation (hugely simplified explanation I know - don't shoot me). Depending on how and where such stimulus was applied the UK could be very nicely set up to thrive in the future. Economics is not a science for many reasons but certainly not the one stated in the quote above.

I don't think Keynes ever claimed economics was a science.

 

[dc655321]

One of the biggest issues I have, for example, is the “in-room predicted response” graphs. Who’s room? My room applies a completely different envelope function over sound compared to your room, and I have no way (insufficient data and modeling) to accurately obtain the said response in my room.

The "room" in PIR is just a statistical representation, of course. Most domestic listening environments are of scales with a similar order of magnitude (~< 10m in any direction, etc), so amenable to such treatment.

You may already be aware of this, but Amir's first speaker measurement/review also provides a reasonable overview of the many components that go into a "spinorama" set of measurements. One thing that @hardisj often does with his reviews (and lacking here) is plot the his actual in-room response against the PIR (Kef R3 review here):

On top of that, I don’t know what other characteristics about a speaker makes it pleasant to me (like, say, step response), and if they’re measured, how to exactly interpret the measurements and what to look for.

So a major part of your dissonance here is that you don't understand what you're looking at (using step response as a particular example)?
What makes a "good" step response, etc? Is that correct?

 

[nonnyno]

The "room" in PIR is just a statistical representation, of course. Most domestic listening environments are of scales with a similar order of magnitude (~< 10m in any direction, etc), so amenable to such treatment.

You may already be aware of this, but Amir's first speaker measurement/review also provides a reasonable overview of the many components that go into a "spinorama" set of measurements. One thing that @hardisj often does with his reviews (and lacking here) is plot the actual in-room response against the PIR (Kef R3 review here):

So a major part of your dissonance here is that you don't understand what you're looking at (using step response as a particular example)?
What makes a "good" step response, etc? Is that correct?

Oh cool thank you for that explanation and plot.

 

[voodooless]

@BDWoody I just ordered the pre90 to extend the inputs of my preamp, and will buy locally the mixer I need to use my microphone in order to tests the DACs I have here. I know data not only suggests that I wont be able to notice differences, they practically predict I wont, but I will perform the tests to provide MORE evidence to the subjectivists in the thread.

What? A microphone to test DAC’s? That’s not a good way.

Question to the knowledgeable people in these matters: Is it acceptable to level match using pink noise?

No! Thinks like SPL meters are not good enough.

 

[Suffolkhifinut]

Actually that is entirely incorrect. Science doesn't say that something has to apply always. It says it must apply predictably within a set of criteria (some of which may be circumstantial). The example you've given is simply talking about applicability and not the definition of science. Keynesian stimulus works where there are no blockages to the productive consumption of the money. If there are such blockages then you get inflation (hugely simplified explanation I know - don't shoot me). Depending on how and where such stimulus was applied the UK could be very nicely set up to thrive in the future. Economics is not a science for many reasons but certainly not the one stated in the quote above.

I don't think Keynes ever claimed economics was a science.

Think I did say in a previous post Science is narrow, it does function under defined criteria. On Keynesian economics a few years ago the German Government had a subsidised car trade in scheme. If a car was traded in for a small more economical model the Government put (I think) €2000 towards the cost. The result Korean car imports surged and the sale of German cars fell. So much for productive consumption of the money?

 

[SIY]

"Science says..." is a key indicator that someone doesn't know what science is.

Ditto the use of "theory" to mean "hypothesis."

 

[nonnyno]

Think I did say in a previous post Science is narrow, it does function under defined criteria. On Keynesian economics a few years ago the German Government had a subsidised car trade in scheme. If a car was traded in for a small more economical model the Government put (I think) €2000 towards the cost. The result Korean car imports surged and the sale of German cars fell. So much for productive consumption of the money?

The fact that they hadn't correctly identified where consumption could be productive is a different issue and does not disprove Keynesian economics. Furthermore what was the goal of the German government? The driving of more economical, newer cars, or stimulus of the German car manufacturing industry? I assume dealers in Korean cars in Germany did very well out of it so it probably stimulated some areas of the economy quite well!