Master Thread: Are measurements Everything or Nothing?

[Blumlein 88]

Pink noise has equal energy per octave of frequency. The energy of pink noise falls off in a range of one to three db per octave (pretty wide range in loudness there). Who determined that this is the correct slope of how we hear things? I can tell you that in the first twenty years of my audio career I did a LOT of live sound. I tried the whole pink noise thing with a special (and expensive) flat response microphone. The results were always underwhelming to say the least. When you had finished getting the system perfectly flat, you could speak into your perfectly flat mic and it sounded like someone speaking through a heavy curtain. Certainly not an accurate reproduction of the source.

Okay, let me try again. Using pink noise is a method of measuring frequency response. It has nothing......let me repeat......nothing to do with a correct slope of how humans hear. Zip, nada, no connection in this case.

It happens that much natural noise is close to pink in nature.

Pink noise falls off 3 db per octave, not 1-3 db. If you have an octave or fractional octave analyzer then perfect pink noise will give a flat response across your analyzer. To REPEAT, none of this is related to how anyone hears anything. It is an alternate method of measuring frequency response.

I don't know what was wrong with your approach in your live sound work. I've done only a little, and the results were not like you describe. Sounds like an error in your approach. Sounds like you reverse EQ'd for the pink noise slope or something. I don't know. I do know using pink noise can work rather well, and that your conception of why this is done is in error.

 

[Ricardus]

Pink noise has equal energy per octave of frequency. The energy of pink noise falls off in a range of one to three db per octave (pretty wide range in loudness there). Who determined that this is the correct slope of how we hear things? I can tell you that in the first twenty years of my audio career I did a LOT of live sound. I tried the whole pink noise thing with a special (and expensive) flat response microphone. The results were always underwhelming to say the least. When you had finished getting the system perfectly flat, you could speak into your perfectly flat mic and it sounded like someone speaking through a heavy curtain. Certainly not an accurate reproduction of the source. I'm not saying you can't measure the frequency response of the speaker in relation to pink noise, I'm saying that a perfectly flat response to pink noise is not equivalent to perfect reproduction of the source material.

What does how humans hear have to do with pink noise calibration? It's a calibrated mic and a spectrum analyzer on the other end. Not your ears. Dude. This is 101 stuff.

As for why you couldn't get pink noise to work, I won't speculate.

 

[Studio Guy]

You miss the point entirely. PROVE to me that a system with a perfectly flat pink noise response results in a perfect reproduction of the source material. As to your limited experience in live sound, I can only say that its like snakes, you don't really see the difference until you see your first rattlesnake. Most of the live bands I've heard (and I've heard a lot) have mediocre sound at best. I never settled for mediocre.

 

[Studio Guy]

What does how humans hear have to do with pink noise calibration? It's a calibrated mic and a spectrum analyzer on the other end. Not your ears. Dude. This is 101 stuff.

As for why you couldn't get pink noise to work, I won't speculate.

You are being obtuse. You can't use a ruler that's 20 inches long as a good approximation of what a human foot should be. I'm not saying you can't measure something with pink noise
, I'm questioning its validity as a scale in reference to perfect reproduction of the source material.

 

[Blumlein 88]

You miss the point entirely. PROVE to me that a system with a perfectly flat pink noise response results in a perfect reproduction of the source material. As to your limited experience in live sound, I can only say that its like snakes, you don't really see the difference until you see your first rattlesnake. Most of the live bands I've heard (and I've heard a lot) have mediocre sound at best. I never settled for mediocre.

False Premise: When Arguments Are Built on Bad Foundations – Effectiviology

effectiviology.com

 

[funnychap]

Yes, but our auditory system does not have neither infinite resolution nor infinite bandwidth.

Once you reach a certain point of performance, the audible differences become practically non existent. The reason why we experience one DAC as having a "character" that another DAC does not, despite both being well engineered, is because our brains are extremely good a fabricating such things. Do a blind test with proper controls, and I'll almost guarantee that those wonderful R2R flavor compounds will disappear like magic. At best there will be nothing but near infinitesimal differences left.

And no, DAC's do not create sound waves. They create a simple voltage that changes value over time.

The hard part is making sure that value is correct a all times. No magic or secret sauce though.

omg, forget it...

 

[Studio Guy]

False Premise: When Arguments Are Built on Bad Foundations – Effectiviology

effectiviology.com

Not a lot of proof there. Do you have science to back up your claims or not? To try to clarify this for those who don't see the point. We don't use white noise to measure frequency response, why? Because white noise has equal energy per frequency, so we use pink noise which has equal energy per octave. I am questioning how we know that this is the proper noise to use. How do you KNOW that this kind of energy slope is correct? How do you know that a noise with a decreasing amount of energy per frequency wouldn't be more natural and result in more accurate reproduction of the source material?

 

[Beave]

 

[Blumlein 88]

Not a lot of proof there. Do you have science to back up your claims or not? To try to clarify this for those who don't see the point. We don't use white noise to measure frequency response, why? Because white noise has equal energy per frequency, so we use pink noise which has equal energy per octave. I am questioning how we know that this is the proper noise to use. How do you KNOW that this kind of energy slope is correct? How do you know that a noise with a decreasing amount of energy per frequency wouldn't be more natural and result in more accurate reproduction of the source material?

You..................have...................many................misconceptions.

The more you write, the more it appears you would be starting on the steep part of the learning curve. It simply is not practical to bring you up to speed with a few forum posts. Plus you have shown no inclination to listen. Your statements indicate a lack of understanding of the basics regardless of any experience you have had.

And yes, we could use white noise to measure frequency response. One reason not to is it could stress tweeters. Nothing to do with human hearing parameters. Another reason is low frequencies might be obscured by environmental noise vs pink noise. Another is an octave type spectrum analyzer was more common. If FFT's are available then white noise will work okay (noting the previous issues in this paragraph). Nothing to do with human hearing parameters.

Or maybe I just lack the patience today, and someone else can offer you a few explanations.

 

[BDWoody]

Just moved another batch in from the denafrips thread.

 

[pkane]

You..................have...................many................misconceptions.

The more you write, the more it appears you would be starting on the steep part of the learning curve. It simply is not practical to bring you up to speed with a few forum posts. Plus you have shown no inclination to listen. Your statements indicate a lack of understanding of the basics regardless of any experience you have had.

And yes, we could use white noise to measure frequency response. One reason not to is it could stress tweeters. Nothing to do with human hearing parameters. Another reason is low frequencies might be obscured by noise vs pink noise. Another is an octave type spectrum analyzer was more common. Nothing to do with human hearing parameters.

Plus, we can measure using any signal at all, including music - the shape is irrelevant as long as it covers the frequency range of interest. Turns out frequency measurement results are the same, except you’re much more likely to fry your tweeters by using white/blue/purple/violet noise.

 

[solderdude]

I am questioning how we know that this is the proper noise to use. How do you KNOW that this kind of energy slope is correct?

As @Blumlein 88 already mentioned. The reason for using pink noise with speakers is a technical one.
When using pink noise from an analyzer to obtain the frequency response (flat line on the display) the analyzer simply applies the inverse slope of the pink noise so you get 'white noise' alike measurements on the display.
Those can be used at high SPL to ensure you only measure the noise from the speakers while not damaging the speakers.
Pink and white noise are basically the same except for the tonal shaping. In fact white noise is generated inside and filtered.
Pink noise is less 'unpleasant' to listen to during the measurement and ensures the speakers are not damaged during a minutes long 'session' where one adjust sliders (live) of a graphic EQ, which might take minutes with a 32 band equalizer.
High levels of white noise during that time might fry tweeters and is very unpleasant.

Narrow BW (terts) noise measurements (with a 3dB/oct slope ?) which kind of has a similar effect as a warble tone in that is smooths out narrow peaks and dips can give you a 'tonal balance' correction while disregarding sharp nulls and peaks.

Depending on the smooting/averaging over a certain frequency band any measurements can be more or less 'exact'.
When averaging noise long enough you can create very discriminate and accurate FR plots.

 

[Ricardus]

You miss the point entirely. PROVE to me that a system with a perfectly flat pink noise response results in a perfect reproduction of the source material. As to your limited experience in live sound, I can only say that its like snakes, you don't really see the difference until you see your first rattlesnake. Most of the live bands I've heard (and I've heard a lot) have mediocre sound at best. I never settled for mediocre.

Dude. Please change your name. You don't understand the most basic concepts here and clearly you're here to argue bad science. I can assure you, YOU ARE NOT A STUDIO GUY.

Also, to the Mods, I'm on more than a few audio/electronics/studio forums and I've noticed more new signups than I seem to recall over the past few years, and they all seem to be people who cause trouble and get banned in a month. They're all arguing bad science and audiophool crap. Even on my pro audio forums.

Just an observation.

 

[fpitas]

Dude. Please change your name. You don't understand the most basic concepts here and clearly you're here to argue bad science. I can assure you, YOU ARE NOT A STUDIO GUY.

Also, to the Mods, I'm on more than a few audio/electronics/studio forums and I've noticed more new signups that I seem to recall over the past few years, and they all seem to be people who cause trouble and get banned in a month. They're all arguing bad science and audiophool crap. Even on my pro audio forums.

Just an observation.

Luckily they quickly get shunted to this thread. It's the honey pot of ASR

 

[RayDunzl]

Why do people actually like some kinds of distortion but not others?

Some "distortions" are what make up the music itself.

They create much of the sound of various instruments.


The electrical signal of a single note plucked gently on a bass guitar, clean, no distortion intentionally added.

The fundamental is at 96.5Hz. All the other peaks are harmonics (a kind of distortion) of the fundamental frequency.

The image should be similarly applicable to other string instruments.

The rest of our musical instruments will have a similar plot, with different harmonic frequencies accentuated.

Without such "distortions", instruments would all sound like blowing on a beer bottle - no harmonics.


Grolsch 330ml bottle. Very low "distortion". Not often used as a musical instrument. Not much character there.

 

[fpitas]

When euphonic distortion is discussed, generally people think of harmonic content. But I wonder if some people like the IMD created, kind of fuzzing up the background. Similarly, I've always suspected that some people like vinyl for the background groove noise.

 

[NattyLumpkins]

Not a lot of proof there. Do you have science to back up your claims or not? To try to clarify this for those who don't see the point. We don't use white noise to measure frequency response, why? Because white noise has equal energy per frequency, so we use pink noise which has equal energy per octave. I am questioning how we know that this is the proper noise to use. How do you KNOW that this kind of energy slope is correct? How do you know that a noise with a decreasing amount of energy per frequency wouldn't be more natural and result in more accurate reproduction of the source material?

I'm not sure how much this would contribute to the discussion but a flat eq is not how we hear. In mixing the curve of human frequency sensitivity has a name I just can't remember it.

Hearing · Physics

 

[NattyLumpkins]

Some "distortions" are what make up the music itself.

They create much of the sound of various instruments.


The electrical signal from a bass guitar, clean, no distortion intentionally added.

The fundamental is at 96.5Hz. All the other peaks are harmonics (a kind of distortion) of the fundamental frequency.

The image should be similarly applicable to other string instruments.

The rest of our musical instruments will have a similar plot, with different harmonic frequencies accentuated.

Without such "distortions", instruments would all sound like blowing on a beer bottle - no harmonics.


Grolsch 330ml bottle. Very low "distortion". Not often used as a musical instrument. Not much character there.

Thanks for posting I wanted to ask, it looks like you have a setup for room equalization. How do you like the Behringer for doing that? I wanted to do the same thing with REW.

 

[ahofer]

I'm not sure how much this would contribute to the discussion but a flat eq is not how we hear. In mixing the curve of human frequency sensitivity has a name I just can't remember it.

Hearing · Physics

Fletcher-Munson? https://en.wikipedia.org/wiki/Equal-loudness_contour

We are reproducing a signal, not 'how we hear it'.

I'm a big fan of loudness compensation, but that's because the original music would be presented at much higher volume than I'm listening at. You can't play, for instance, a big band, at concert volume in your living room.

 

[NattyLumpkins]

Fletcher-Munson? https://en.wikipedia.org/wiki/Equal-loudness_contour

We are reproducing a signal, not 'how we hear it'.

I'm a big fan of loudness compensation, but that's because the original music would be presented at much higher volume than I'm listening at. You can't play, for instance, a big band, at concert volume in your living room.

Thanks yes that's it! and I agree with you.