Master Thread: Are measurements Everything or Nothing?

[Thorsten Loesch]

you did research which you say breaks new ground and indicates that there is in fact something in the signal that is unknown but affects preference.

I wrote no such thing. In the process of commercial product development (as well as in "hifi" as hobby) I have come across situations where I have good reason to conclude that I was facing something that was not revealed by the standard set of classic objective tests.

There are many items, most of which I am happy to put down to the changes from the test bench to a more complex, less controlled setup with more chances of actualising problems (hum loops and missing earth based noise pick-up are extremely obvious examples of such.

I mentioned my experiences conversationally in a specific conversation. I am not trying to challenge anyone, change anyone's thinking etc. (I tried that when I was a lot younger and have long since given it up as a waste of time that becomes more precious the older I get.

My comments were dropped off here (I would have suggested to simply delete them if offensive - had I been asked) with results I would have hoped to be unpredictable but at the same predicted to be exactly what happened.

However you don't seem to accept the very high probability that the explanation is that your methodology was flawed, and you didn't even keep a copy of the data.

I responded to what people wrote. I called this thread an echo chamber.

But you still feel able to criticise the way science is approached on this forum?

Yes, absolutely. I feel able to point out where problems exist in the way things are approached.

No you may ignore what I write, get angry Greta Thunberg style "How dare you", you may reject my criticism or you may look and see if perhaps there is something here.

It actually makes no difference to me.

Thor

 

[manisandher]

Why do you refuse to see the difference between the validity of the theory and the practical application by a system with limited resolution?

I'm perfectly happy with the validity of the theory. I think it's great. But it has its limits... infinity being one of them.

My question clearly concerned the application of the theory on a finite music signal, which is what we do during measurements & analysis. In which case, I hope we're all agreed, that we're working with [very good] approximations only.

So, in practice, no, you cannot recreate the original [music] signal from its Fourier Transform... because you're always using "a system with limited resolution" to do so.

Everyone happy with this?

 

[ahofer]

No you may ignore what I write, get angry Greta Thunberg style "How dare you", you may reject my criticism or you may look and see if perhaps there is something here.

Many have taken the third option and looked. But you have given us bugger-all to go on. You could consider this criticism and re-design your approach. "Just asking questions" does not remove the evidence we have to-date.

 

[GXAlan]

I was facing something that was not revealed by the standard set of classic objective tests.

The classic objective tests are designed to be helpful proxies providing quick and useful information. Subjective differences not captured by “classic” tests should still be captured by specific objective tests intended for what you are looking at.

One easy comparison is ground loops. Some devices are more sensitive to it than others and while a good test setup will eliminate ground loops in both, it is possible that in real world homes the unit more resilient to ground loops will sound better.

 

[Vacceo]

isn't everything we listen to on the psychoacoustic level and therefore important to how we discern what we hear? What mood we are in? Whether we've had a few bourbons? Whether we have a preconceived opinion of the music we are listening to or the equipment listening on, etc.

The reaction to it, indeed. The phenomenon itself, no.

 

[pkane]

So, in practice, no, you cannot recreate the original [music] signal from its Fourier Transform... because you're always using "a system with limited resolution" to do so.

Everyone happy with this?

Not happy! Fixed it for you. The limitation has nothing to do with the Fourier transform. We simply don't have infinitely precise measurement devices or digital computers with infinite precision. These don't exist in nature and never will. But we can get close enough to where it really doesn't matter.

 

[pma]

you did research which you say breaks new ground

During the past 20 years on the various forums and since the beginning of seventies (I am 12 years senior to TL) I have met or read from a large set of “Big Names” who thought they broke “new ground”. Their findings were often contradictory and in the end you would see that the King is naked. Take it easy.

 

[pma]

Anyway, Thorsten stirs up stagnant water. It is good to have him here .

 

[ahofer]

If it makes no difference to you, then why are you posting here?

Jim

I suggest this bit of reading on "counter-expert" techniques:

Lifemanship, or How to Trip the Expert

In his book Gamesmanship and in his article “Golfmanship,”which the Atlantic published last September. STEPHEN POTTER disclosed the new technique by which the dub, or crock, can win various games of skill from a superior opponent. He now lays down a conversational method by which the full-blown...

www.theatlantic.com

 

[manisandher]

The limitation has nothing to do with the Fourier transform. We simply don't have infinitely precise measurement devices or digital computers with infinite precision.

Yes, that's what I meant.

But we can get close enough to where it really doesn't matter.

Dare I ask... how close is "close enough"?

Really, don't bother answering... lest I get accused of trolling again (as though I don't have 1001 better things to do with my time).

 

[ahofer]

Dare I ask... how close is "close enough"?

Below well-researched limitations of human hearing.

 

[Mulder]

In this case the "original event" is what was heard on the Studio's setup during the final mastering session.

But it's basically the same thing. The listener in front of his HiFi system cannot know how it sounded in the mastering studio.

Why do we need to distort the signal in some ways? Because microphones are not speakers in reverse, speakers are not microphones in reverse and microphones are not equivalent to ears. And because often the original dynamic range cannot be reasonably accommodated in most domestic listening.

Now is this harmonic distortion? No. But is still signal alteration (quite gross often enough).

Thor

Of course. But whatever you try to say, it makes no sense to me put in relation to what you previously claimed..

 

[pkane]

Dare I ask... how close is "close enough"?

Sure, as close as laws of nature and hardware would allow.

Physics imposes limitations on our recording/reproduction equipment, such as thermal noise, bandwidth limitations, noise made by air molecules colliding or hitting the eardrum, etc. Some of these cannot be overcome, but in most cases these are way below our ability to hear them, so not an issue for music reproduction. Computational precision is not a serious limitation compared to these.

For computational precision, use 64 bit floating point to get down to around -300dB accuracy. Use 128 bits if that floats your boat, or 512 bits or ... Computation libraries using variable size/precision exist and can be used, if extra precision is required.

 

[birdog1960]

Below well-researched limitations of human hearing.

which is another variable that should be considered. it is not static (pardon the pun)

 

[tomelex]

Science is about measurements too, like in Audio Science Review. Just love this place to bits.

Measurements are not done here to describe preferences, but they can totally describe all characteristics of a system you prefer.

 

[Mart68]

I wrote no such thing. In the process of commercial product development (as well as in "hifi" as hobby) I have come across situations where I have good reason to conclude that I was facing something that was not revealed by the standard set of classic objective tests.

There are many items, most of which I am happy to put down to the changes from the test bench to a more complex, less controlled setup with more chances of actualising problems (hum loops and missing earth based noise pick-up are extremely obvious examples of such.

I mentioned my experiences conversationally in a specific conversation. I am not trying to challenge anyone, change anyone's thinking etc. (I tried that when I was a lot younger and have long since given it up as a waste of time that becomes more precious the older I get.

My comments were dropped off here (I would have suggested to simply delete them if offensive - had I been asked) with results I would have hoped to be unpredictable but at the same predicted to be exactly what happened.



I responded to what people wrote. I called this thread an echo chamber.



Yes, absolutely. I feel able to point out where problems exist in the way things are approached.

No you may ignore what I write, get angry Greta Thunberg style "How dare you", you may reject my criticism or you may look and see if perhaps there is something here.

It actually makes no difference to me.

Thor

I think you are labouring under the misapprehension that there is a 'thinking' or a 'belief system' at work here rather than an acceptance of the current paradigm.

Everyone here would be delighted to be presented with evidence that the current paradigm is faulty. So far in hundreds of posts we have lots of assertions with no evidence.

That's why the thread is an echo chamber.

I have never become 'Angry Greta Thunberg style' in my entire life.

Like many intelligent men, you underestimate your audience.

 

[Ken Tajalli]

Ok, so have another go at answering this:

Can you reproduce the original [music] waveform from its Fourier Transform? And if not, why not?

I suspect your answer will now be different to your original answer .

Think about it, how do DACs work?
All they get, in the form of input, are mere instructions! There are no music signals sent or received.
Together within criteria that were accepted (bandwidth limit, sampling rate frequency, bitrate depth ...), they reproduce the original music from a mere powersupply voltage.
The answer is yes, and it is not an approximation!
When you specified Music , you specified:
- bandwidth
- dynamic range
Within those limitations, yes, we can reproduce the music from its FT components (how a DAC does it) perfectly, with negligible distortion.

 

[SIY]

which is another variable that should be considered. it is not static (pardon the pun)

Static. You keep using that word. I do not think it means what you think it means.

 

[pma]

Static.

Signal types [Randall]

 

[manisandher]

Think about it, how do DACs work?
All they get, in the form of input, are mere instructions! There are no music signals sent or received.
Together within criteria that were accepted (bandwidth limit, sampling rate frequency, bitrate depth ...), they reproduce the original music from a mere powersupply voltage.
The answer is yes, and it is not an approximation!
When you specified Music , you specified:
- bandwidth
- dynamic range
Within those limitations, yes, we can reproduce the music from its FT components (how a DAC does it) perfectly, with negligible distortion.

A DAC reproduces a music signal from [the music signal's] FT components? Well, I've learned something new today.