Measurements and the Correlation of What We Hear

[RogerD]

Good point.

How'd those BGs work out?

Hi,

That project has been put on hold temporarily. But Rich's Ampex with BG's that I spec'd will be running next week,stay tuned.

 

[RBFC]

The very same thing happened to me in an audio club situation: 2 very different systems were played, the first which created/retrieved a very airy, realistic ambience of the recording space, nicely holographic in character, the second the classic hifi sound, in your face, intense, sharp, the sound obviously coming straight at you from the drivers. I turned to one of the leading lights of the club, who does reviews for a well known audio website, and said, "Remarkable how the second failed to get anywhere near the first in SQ", and he said, "What do you mean? Obviously the second is the better system!".

After recovering from falling off my chair, I pondered this. In hindsight, this was all about expectation: the second system behaved exactly like a conventional setup was supposed to behave, and therefore was the better system: it delivered an anticipated, desired result. The first system was disturbing, it was being "unnatural", therefore was wrong.

So we may still have quite some way to go ...

Frank

Frank,

There have been many discussions over the years in many different venues that consider this topic. Some audiophiles prize tone quality, realistic dynamics, and overall volume capability, while others value the sense of space and ambient cues more highly. I know, a proper system does all those things correctly, but the real world interferes and compromises are the rule. Thus, the "camps" are born.

Lee

 

[microstrip]

I've heard all of that many, many times before, micro. I used to believe it myself. Then I tested it, and it simply vanished.

Tim

Yes, but you tested it in conditions and with a methodology that would lead to that conclusion and naturally compromise the results. I am sure that you also have heard (read) this many times ...

 

[microstrip]

I do enjoy reading these discussions. On one hand you have folks who don't understand electronics claiming that we are not measuring everything and the other those who do understand electronics claiming we do. (...)

I also enjoy them, but I am very glad my teachers in electronics did not apply your evaluation criteria - I would have never graduated, as electronics was a key subject!

 

[audioguy]

Science cannot be denied,.

But it is --- regularly, frequently and consistently on these threads.

 

[Phelonious Ponk]

Yes, but you tested it in conditions and with a methodology that would lead to that conclusion and naturally compromise the results. I am sure that you also have heard (read) this many times ...

Yes, though usually after I had actually described the conditions and methodology . A critical difference is I believe in expectation bias, and I know I am not immune. And so I have tested and re-tested my own beliefs, numerous times, under different conditions, using different methodologies. This includes extensive listening over time to a wide variety of recordings, finding myself hearing things that, upon returning to quick switching, vanished.

So have I concluded that all electronics sound the same? No. Have I concluded that all well-designed amplifiers operating within their tolerances all sound the same (boy is "within their tolerances" a fungible concept!)? No. But I have concluded that many audiophiles, some of whom deny the most basic principles of audiology to hold on to their illusion, most of whom deny their own vulnerability to bias and the efficacy of any test that might reveal it, are quite skilled at hearing what they want to hear. And so I take what most of them say with a grain of salt. Some? A bagful. The fringe? I find that there's really not much point in trusting in the reality of anything they hear or say. If, every once in awhile, a glimpse of something real peaks through, I figure I can afford t miss it.

Tim

 

[MylesBAstor]

But it is --- regularly, frequently and consistently on these threads.

Well the only relationship between real science and what is discussed here is purely coincidental.

 

[Orb]

Well the only relationship between real science and what is discussed here is purely coincidental.

Very true IMO.

 

[fas42]

There have been many discussions over the years in many different venues that consider this topic. Some audiophiles prize tone quality, realistic dynamics, and overall volume capability, while others value the sense of space and ambient cues more highly. I know, a proper system does all those things correctly, but the real world interferes and compromises are the rule. Thus, the "camps" are born.

Lee

Thanks, Lee, I appreciate those comments: I probably didn't mingle enough with a variety of audio people to appreciate that there was such a divide ...

Frank

 

[Phelonious Ponk]

Well the only relationship between real science and what is discussed here is purely coincidental.

http://scholar.google.com/scholar?q=auditory+memory&hl=en&as_sdt=0&as_vis=1&oi=scholart

I'm sure if you bother to read any of that you'll be able to find things to object to. It was a quick Google search and I haven't read it myself. The point is that the science being discussed here right now -- that volume effects quality perception and that auditory memory fades very quickly over time -- is very well established and well-supported. There is a lot more out there if you care to look. Google is your friend. Unless what you really want is to avoid any science that might challenge your beliefs.

Tim

 

[fas42]

This includes extensive listening over time to a wide variety of recordings, finding myself hearing things that, upon returning to quick switching, vanished.

Sorry to possibly disturb you again, Tim, but you're touching upon a key element of audio here that I would emphasise is extremely important. You're surmising that when you quickly switched between two different circumstances that a seeming difference disappeared, and that that was evidence that there was no real difference in the first place.

This is going to upset some, but I would propose that in fact there WAS a difference, in least in some of these situations, and that the difference DID disappear, and this was a real occurrence. Huh?! Well, what I talking about is that the ACTUAL PROCESS OF SWITCHING itself altered the nature of the electronic environment that the system was operating in, which then altered the SQ. Hence you weren't only testing the difference between two situations but also the effect of switching between them.

This is all too ridiculous, I can hear people saying, BUT it is a very accepted aspect of the process of science research to eliminate the impact of the way the research is done. A scientist wants to understand what's happening in some box, but in the very action of looking inside that box, or attaching some measuring device to it, he can very easily interfere with what's happening in the box. In the ultimate form of this, it's called Heisenberg's Uncertainty Principle.

But we only talking about simple old audio, you say! Well, all my investigations have made me very aware that the slightest element not correctly in order, or the actual process of changing an element can make a huge difference, and it only gets worse the better the system becomes. Only if you're an audiophile, that is; normal listeners would be oblivious.

It's an "unfortunate" fact of life that our ear/brains can become very sensitive to these sort of variations, and I fear most audio enthusiasts always end up attuned to this degree. So, either decondition yourself, or accept that you will have to keep chasing after that elusive "correct" sound.

Frank

 

[fas42]

The point is that the science being discussed here right now -- that volume effects quality perception and that auditory memory fades very quickly over time -- is very well established and well-supported.

Perhaps part of the problem is the use of the word "quality": that to me is directly associated with a lack of audible distortion, in the conventional electronics sense. With real sounds, there is normally no such factor in operation, unless perhaps one has defective or impaired hearing. If I move closer to or further away from a real instrument, equivalent to changing its volume, the general impression for me is that the sound it makes becomes more or less intense, rather than change in "quality".

Which is how it should be for an audio system ...

Frank

 

[MylesBAstor]

http://scholar.google.com/scholar?q=auditory+memory&hl=en&as_sdt=0&as_vis=1&oi=scholart

I'm sure if you bother to read any of that you'll be able to find things to object to. It was a quick Google search and I haven't read it myself. The point is that the science being discussed here right now -- that volume effects quality perception and that auditory memory fades very quickly over time -- is very well established and well-supported. There is a lot more out there if you care to look. Google is your friend. Unless what you really want is to avoid any science that might challenge your beliefs.

Tim

Pray tell Tim what you ever did that was scientific? Did you ever do any scientific research in your life? I did for close to 20 years. You can Google FYI, and see some of my publications, about five to ten of the thirty or so I published.

You can read and quote all the books til the cows come home but until you've actually done something, I suggest listening. All you sound like are engineers who graduate and think because they've read a EE book, they know there is to know about circuitry. Most of them are in for a rude awakening when they have real engineering job.

And btw, if you've read my posts, you would know I talked ad nauseum about the many issues of short vs. long term memory, serial vs. parallel processing in the brain, how memories are stored, etc. Not news and the foundation for motor learning theory.

 

[Gregadd]

The question is are we slaves to our ignorance. Don't we learn. When we first got smoke alarms and we heard an alarm we jumped expecting a fire. That was expectation bias. A false positive. We were told that if the alarm went off there was smoke somewhere. More often than not this is not t he case. Used with common sense the smoke detector remains an essential tool. No one refers to those who rely on it as smoke detector-fools. No one calls the designers snake oil salesman.

How do we verify the smoke detector alarm? With our senses. if you perceived that the alarm sounded and it did not, that's an hallucination. You ought to see your doctor. Or curb your substance abuse.

I keep asking this question. The purpose of a stereo is to create an illusion. While that illusion is created outside our head. The effect is in our brain. Our head can never be wrong. The only thing the person who attempted to create the illusion can do is scratch his head and say well that was not the illusion as was striving for. It's the illusion maker who must change.

 

[fas42]

The purpose of a stereo is to create an illusion. While that illusion is created outside our head. The effect is in our brain. Our head can never be wrong. The only thing the person who attempted to create the illusion can do is scratch his head and say well that was not the illusion as was striving for. It's the illusion maker who must change.

I will say, hear, hear! And add, the illusion has a much better chance of effecting that reaction in our brains if as much information is fed to our ear/brain sense as possible, with minimal distortion. Easy to say, damned hard to do well enough!

If we do enough on our side of the fence, the illusion maker gets off the hook, i.e., even "crappy" recordings sound good ...

Frank

 

[Phelonious Ponk]

Pray tell Tim what you ever did that was scientific?

Pray tell, Myles, what that has to do with anything? I can understand that evolution is widely accepted to be a proven scientifc theory and link creationists to collections of studies supporting it without being Darwin.

I listen every day.

Tim

 

[Phelonious Ponk]

Sorry to possibly disturb you again, Tim, but you're touching upon a key element of audio here that I would emphasise is extremely important. You're surmising that when you quickly switched between two different circumstances that a seeming difference disappeared, and that that was evidence that there was no real difference in the first place.

You're not disturbing me, and that's not quite the case. What I'm assuming is that the difference was at least negligible and at most non-existent.

This is going to upset some, but I would propose that in fact there WAS a difference, in least in some of these situations, and that the difference DID disappear, and this was a real occurrence. Huh?! Well, what I talking about is that the ACTUAL PROCESS OF SWITCHING itself altered the nature of the electronic environment that the system was operating in, which then altered the SQ. Hence you weren't only testing the difference between two situations but also the effect of switching between them.

I'm never sure with you, Frank, if you're talking about the introduction of a physical switching mechanism and its impact on the signal chain, the idea that the physical act of throwing a switch might have an effect that could change the signal quality, or the metaphysical impact that my personal switching vibe might have on my time/space/ear/brain continuum.

This is all too ridiculous, I can hear people saying

Except for the time/space/ear/brain continuum thing, I don't think it is ridiculous, actually. But I don't think it accounts for what I thought I heard, or what audiophiles consistently report hearing, either. And I don't think it explains the fact that they, even I, almost always heard something obviously better in sighted, long-term listening that almost always becomes something questionable to inaudible when subjected to volume-balanced, rapidly-switched and, particularly, blind listening. Is there something in Heisenberg's Uncertainty Principle that would explain why, in high-end audio, the principle never adds, it never removes a negative, it only seems to subtract something the participant wanted to hear?

Tim

 

[fas42]

Except for the time/space/ear/brain continuum thing, I don't think it is ridiculous, actually. But I don't think it accounts for what I thought I heard, or what audiophiles consistently report hearing, either. And I don't think it explains the fact that they, even I, almost always heard something obviously better in sighted, long-term listening that almost always becomes something questionable to inaudible when subjected to volume-balanced, rapidly-switched and, particularly, blind listening. Is there something in Heisenberg's Uncertainty Principle that would explain why, in high-end audio, the principle never adds, it never removes a negative, it only seems to subtract something the participant wanted to hear?

Tim

As regards the switch mechanism, both its inclusion AND the impact of the switching operation could be relevant; I have certainly heard the effect of both instances at different times. I'll let the comment about the metaphysical thingy go through to the keeper, as we say down here

Your comparison of long term listening highlighting positives vs. rapidly switching dispersing differences I could see as being caused by the latter not allowing the system time to restabilise electronically: even though signals and DC levels may be matched at volt levels, at millivolt or one thousandth of a volt, 60dB down, there may be major differences in measurable voltages, and the electronics all have to stabilise to a new operating state. Capacitors are very much part of the equation here, even theoretically such a part never reaches stability, if you do the maths you find that even in the simplest situation the capacitor is always trying to reach an equilibrium which it never achieves.

Hence the long term listening allows the best chance of maximum stabilisation to occur, and hopefully best sound. Though I unfortunately have the other problem: long term the system starts to degrade again because other influences come into play, the properties of materials like plastic become relevant. Here I am talking about pretty subtle stuff, but it makes all the difference between the system sounding real, or just being hifi.

Frank

 

[Gregadd]

I do enjoy reading these discussions. On one hand you have folks who don't understand electronics claiming that we are not measuring everything and the other those who do understand electronics claiming we do.

First, you don't hear electronics (you don't hear current and voltage, but the results of those things as they wiggle about a speaker, or the results of how they wiggle because of the way air moves across a microphones transducer.

If you change the electronics, of course you change what comes out the speaker.

The goal of an amplifer, be it one stage or many, is to put out an amplified replica of what went in. We do that routinely and at far lower levels of distortion than you can hear.........than any one of you can hear. Note I said hear, as what your brain does with the information is another story unique to you and me.

If you hear a difference when you swap out an amp, then it is because you did not make a replica:

obvious, but you changed:

amplitude.....look at it on a scope..easy
phase.....look at on a scope as well
harmonic spread.....look at it on a spectrum analyzer (linear or non linear distortions)

do it with music, do it with a pure tone (that is a sinewave by the way), square waves, triangular waves, I don't care.

The issue why you hear differnces in amps and preamps and all such stuff is that they affect these things differently. They do not generate an amplified replica but they add or subtract from it. There is no magic.

So, as an example of phase effect, if the upper frequencies are slowed down a bit and the lower are allowed to pass earlier, you get that "phat" tube sound....but you can get that "phat tube sound" with solid state...if you use a transformer as well....no magic...but surely that "phat" sound does sound nice with many types of rock and jazz and small scale instrumental stuff.


What measurment affects what.....I will tell you, the recording is the biggest effector of what you hear on any given system, one can record in ambience, width , depth, whatever you want, even a dog barking 50 feet to the right of your speaker if you want. But electronics has NO CLUE itself of what the recording means when it pushes the air at the speaker air interface....NONE. Electronics job is to simply replicate the input signal and its the transducers job to simply replicate the electronic command...but we all know thats where things fall down fast.




If I perform just a sinewave THD test on a piece of gear, and not look at phase, then one gear will shift phase say 50 degrees and the other 120 degrees but the sinewave test will not reveal that because it is not looking for phase change.

You can't take any one test, except for a null test, and declare that no other parameters are different. In any case, at some point, the differences in amplitude, phase, and spectrum are so small that none of us can hear them, and that seems to be around 85 or 90 db in a normal home environment.

Tom

What Tom is saying then is the illusion maker (amp)is perfect. If you are not getting the perfect illusion its because your illusion maker is defective or you have an imperfect idea of what the illusion should be.
Moreover any attempt to go beyond what current measurements deem to be perfect is foolish and futile.
Of course once you hook up the perfect illusion maker to a speaker and play music through it, we can easily show it's not so perfect at all.

 

[Orb]

Do we agree on the following?

1. That there are measurements that tell us extremely well what we *could* hear. Take the frequency response measurements. If it has a dip of 2 db at 5 Khz, surely we all agree that is audible and the thing that told us that was the FR measurement.

2. That measuring is not the same as understanding. If I ran the above measurement point by a random person on the street they would surely not be able to make the determination I just did. So we need to separate knowledge of interpreting data from ability to gather it.

3. We all want single numbers of simple concepts of performance. Take miles per gallon. There are two numbers that were just revised in US because they were too misleading. Even with revised number is likely to be wrong for my commute to work. So we should also accept that in audio, coming up with single numbers or graphs should not be a prerequisite for understanding audio performance.

Just to expand:

I am coming back to this because I feel one area not considered in this debate about measurements and correlation to what we hear, is that there are fundamentally two very different approaches in operation and they do not mix.
I feel the discussion needs to appreciate that we have the engineering measurements that provide excellent information on the performance-behaviour-trait-characteristics of an audio product (including speakers), and then we have the scientific physics-psychoacoustics of waveforms/sounds and measurements relating to how we perceive sounds-music-rythm.

The 1st rely more upon sinewaves,squarewaves, distortion,FR,etc.
While the 2nd is more focused on complex waves found in nature and from musical instruments/voice, where it is critical to look at the sound in both frequency and time domain (envelope of the sound) and the fundamental/partials-harmonics (that help to define tone and timbre,etc).

To correlate what we hear we need to consider the scientific approach and this means we can only truly look to match an audio products performance to what we hear by using a musical complex wave that is ideally a major chord; this can be a recorded real instrument or derived by waveform synthesis.

However as I mentioned in the past the scientific complex waveform approach is meaningless from an engineering perspective as it would be impossible to relate this to an electrical components performance-behaviour-traits, hence why sinewave tones-squarewaves-etc are critical.

This means we cannot solely use existing engineering related measurements in such debates where we want to put forward the case of correlating an audio products performance measurements to actual specifics such as timbre-sibilance-etc and affect on an instruments reproduced sound.
This fits in with my recent posts and links that I have provided about complex waveforms and instruments (and one scientific paper that matches these to perception), and why they would be critical in assisting to conclude such debates as these with actual usable information.
Otherwise we will have a debate that will reiterate endlessly, but I am happy to join in repeatedly like others as it passes the time

Thanks
Orb