The myth of "The Absolute Sound"

[crenca]

8 minutes ago, Jud said:

 

What the experiment shows is that attack and release are critical to instrument identification, tone surprisingly little. If differences in tone cannot even enable college music students playing and hearing these instruments every day to hear differences between oboe and violin, or between flute, saxophone and trombone, then exactly what audible differences are you relying on to distinguish the quality of two high end systems from each other? Attack and release? Even mid-fi and some lo-fi systems ought to be able to reproduce that. So again, what qualities are you evaluating to compare two high end systems if reproduction of the tone of an instrument (main tone and harmonics) is very plainly insufficient?

 

So you take a waveform and apply DSP so that your left with some inner "tone" quality, and this is then extrapolated to be some kind of golden indicator of reproduction?!  I can take the most intimate sound you know, let's for argument sake say the sound of your wifes voice, apply a little DSP, and make her sound like Daffy Duck or Donald Trump...and this means something significant to high fidelity?!?

 

Your making this up as you go along arent you...what are you smoking/drinking/popping, I want some 😋

 

If anything, the conclusion drawn from this test is the opposite from what you have it:  it shows the importance of non-distorted high fidelity playback of the entire waveform in the audio band...

[Jud]

Just now, crenca said:

If anything, the conclusion drawn from this test is the opposite from what you have it:  it shows the importance of non-distorted high fidelity playback of the entire waveform in the audio band...

 

Why no, that's exactly what it does not show. What it shows, again, is very precisely this: If you thought the tonality of an instrument was important to you being able to identify it, you were quite wrong. What is actually (and very surprisingly) critical are the attack and release of the instrument.

 

Thus if you thought accuracy in frequency enabling proper reproduction of fundamental and harmonics would enable you to distinguish between a violin and oboe, let alone a violin and viola, in your audio system, you would be incorrect.

[semente]

6 minutes ago, Jud said:

 

What the experiment shows is that attack and release are critical to instrument identification, tone surprisingly little. If differences in tone cannot even enable college music students playing and hearing these instruments every day to hear differences between oboe and violin, or between flute, saxophone and trombone, then exactly what audible differences are you relying on to distinguish the quality of two high end systems from each other? Attack and release? Even mid-fi and some lo-fi systems ought to be able to reproduce that. So again, what qualities are you evaluating to compare two high end systems if reproduction of the tone of an instrument (main tone and harmonics) is very plainly insufficient?

 

I think that ARQuint expressed in much better terms what I have been trying to say:

 

37 minutes ago, ARQuint said:

Simply, I play acoustic guitar and hear the sound of that instrument virtually every day. I also regularly attend piano concerts. I know what an acoustic guitar and a piano should sound like. I can therefore tell whether the output of an audio system, over a sample of recordings, corresponds to what an acoustic guitar or a piano sounds like. This is not intended as an absolute judgement, for example, to the extent that one should be able to readily distinguish a Martin from a Gibson acoustic guitar or a Steinway from a Yamaha concert piano. It refers to the unique timbre of the instrument, generally, as opposed to that of any particular brand, and whether that timbre is reproduced faithfully.

 

There are many attributes that one has to focus on to determine how realistic a playback system reproduces music, and by realistic I mean lifelike or what one would hear from the audience, the attributes that make up timbre and the spatial cues.

This of course requires a recording with a particular technique and a flat mic response and adequate positioning which aims at recreating a sonic perspective similar to what you would be able to listen if you were present at the original event.

 

Orchestral music can be recorded to sound like what you would listen live and has immense complexity and richness of attributes (many different timbres, many instruments playing simultaneoulsy, different groups playing different voices, pps and ƒƒƒƒs, real space, etc.).

You can evaluate anything from transient response (percusive) to tonal balance (legato) to dynamic range (swings) to low-level resolution (pp) to loudness threshold (ƒƒƒƒ) to intermodulation distortion (tutti) to hardness in the treble to etc...

 

One of my test tracks is the first movement of Rachmaninoff's Symphonic Dances. The score, which has a rich variety of instruments, features a piano, a harp and also a saxophone which often goes unnoticed in lesser systems or by less attentive listeners.

[crenca]

5 minutes ago, Jud said:

 

Why no, that's exactly what it does not show. What it shows, again, is very precisely this: If you thought the tonality of an instrument was important to you being able to identify it, you were quite wrong. What is actually (and very surprisingly) critical are the attack and release of the instrument.

 

Thus if you thought accuracy in frequency enabling proper reproduction of fundamental and harmonics would enable you to distinguish between a violin and oboe, let alone a violin and viola, in your audio system, you would be incorrect.

 

 

Nope.  You are extrapolating, speculating, and asserting way beyond the evidence.  All it may show (further testing is required) is that some parts of the wave form are necessary but not sufficient in themselves for identification (within the confines of the test).  Anything beyond that is not in evidence.

 

I like to muse on the significance of this test for high fidelity as well so I like how it underscores the real nature of sound, in that there is but one waveform to rule them all on the one hand, and how easy it is to tinker with and distort on the other hand.  It reveals how arbitrary "audiophile" language such as "transients" can be (granting their usefulness at the same time), etc. etc...   

 

This whole discussion is but a proxy for a radical subjectivism, which states that if a tree falls in a forest it does not really make a sound and even if it does, it is an impossible ideal that never really exists in the universe as it is.  The truth in @barrowsand your subjectivism is that yes there is preferences and taste, even in the art of recording/playback engineering.  Yet you take it too far when you push the difficulties to an absolute subjectivism.  TAS as a guide and method is not only useful and real, it is necessary...

[semente]

23 minutes ago, Jud said:

 

Why no, that's exactly what it does not show. What it shows, again, is very precisely this: If you thought the tonality of an instrument was important to you being able to identify it, you were quite wrong. What is actually (and very surprisingly) critical are the attack and release of the instrument.

 

Thus if you thought accuracy in frequency enabling proper reproduction of fundamental and harmonics would enable you to distinguish between a violin and oboe, let alone a violin and viola, in your audio system, you would be incorrect.

 

[gmgraves]

38 minutes ago, barrows said:

Wait, what?  Hahaha, I am assuming this is a joke, and it illustrates my point.  I listen for pleasure, so changes to my system which increase my pleasure are changes for the better.  I would rather evaluate changes on this basis, than on the basis of some (virtually impossible to have a true reference for) theoretical "absolute".

 

I would love it if you could provide some examples of such "adequately mic'ed" recordings?  I have one example which might pass muster:  Charlie Haden and Antonio Forcione's "Heartplay" from Naim recordings...  Still, not being present at the sessions, I admit to really not knowing what this recording "should" sound like, despite my experience with guitars and basses.

Since commercial recordings are not made to be accurate (not even so-called audiophile recordings), the best one can hope for are “passably mic’ed” recordings. That’s why I decided years ago to make my own. I have recordings that truly are “state of the art”, and guess what? They aren’t difficult to make, don’t require mega-buck equipment,  and they could have been made 50 - 60 years ago. All that’s required is that the recording company to WANT to make them! That they so rarely do, tells me all I need to know about the real world of commercial recording.

In the late 1950’s there was a small company called Everest recordings that more than occasionally got classical recordings correct. They were basically two people: Harry Belock, the president of Belock Instruments and his recording engineer, Bert Whyte (along with a couple of assistants). Try Bert Whyte’s recording of Copland’s “Appalachian Spring” with Walter Suskind conducting the London Symphony. Recorded with just two microphones, even today, 60 years later, it still sounds great. Tidal has it as a “Master” title, and the recording, while not immune to the physical ravages of time, has survived fairly well. I’m not saying that the recording is perfect, mind you, but it does go to show what can be done with a pair of decent (Neumann U-47, cardioid, IIRC) condenser microphones, simple mike preamps and a half track 1/4 inch recorder (I seem to recall that Bert favored a big, pro Magnecord in those days) running at 15 inches per second!

[Jud]

1 minute ago, semente said:

There are many attributes that one has to focus on to determine how realistic a playback system reproduces music, and by realistic I mean lifelike or what one would hear from the audience, the attributes that make up timbre and the spatial cues.

 

This is generalizing. The experiment reduced it to specifics.

 

Do you suppose, as I do, that any reasonable high end system ought to allow you to distinguish between a flute and a trombone? What has been shown by scientific experiment is that the part of the sound that enables you to make that distinction is the attack and release of the notes. This difference in attack and release between the two instruments is a very large scale difference, so large that any reasonable high end system can reproduce it and thus enable you to distinguish the two.

 

There is virtually no difference in attack and release between a viola and violin, let alone one model of violin and another, let alone playback of the same violin (or acoustic guitar) recording on two different high end systems. So what are we left with? Subtleties of tonality, perhaps. Yet the experiment shows even large differences in tonality, such as between a flute and trombone, cannot be reliably identified even by experienced musicians, so what hope do we have of reliably determining which is the "correct" tonality between two different systems playing back the same violin/acoustic guitar/flute/trombone?

[semente]

3 minutes ago, gmgraves said:

Since commercial recordings are not made to be accurate (not even so-called audiophile recordings), the best one can hope for are “passably mic’ed” recordings. That’s why I decided years ago to make my own. I have recordings that truly are “state of the art”, and guess what? They aren’t difficult to make, don’t require mega-buck equipment,  and they could have been made 50 - 60 years ago. All that’s required is that the recording company to WANT to make them! That they so rarely do, tells me all I need to know about the real world of commercial recording.

In the late 1950’s there was a small company called Everest recordings that more than occasionally got classical recordings correct. They were basically two people: Harry Belock, the president of Belock Instruments and his recording engineer, Bert Whyte (along with a couple of assistants). Try Bert Whyte’s recording of Copland’s “Appalachian Spring” with Walter Suskind conducting the London Symphony. Recorded with just two microphones, even today, 60 years later, it still sounds great. Tidal has it as a “Master” title, and the recording, while not immune to the physical ravages of time, has survived fairly well. I’m not saying that the recording is perfect, mind you, but it does go to show what can be done with a pair of decent (Neumann U-47, cardioid, IIRC) condenser microphones, simple mike preamps and a half track 1/4 inch recorder (I seem to recall that Bert favored a big, pro Magnecord in those days) running at 15 inches per second!

 

I used to listen to recordings of concerts on the radio that had been broacasted live by different national radios from several places in Europe and many, probably due to the simple mic pair hanging from the canopy setups, sounded a lot better that many commercial recordings.

I made digital captures of some of them but burned them into a DVD that's ended up missing when I moved abroad...

[Jud]

5 minutes ago, semente said:

 

 

Yes - very large measured differences, which one would intuit would then reliably indicate to listeners, especially instrumentalists, which was which, eh? Just like one would intuit a heavier object would fall to earth faster. Too bad ugly little science experiments negate our beautiful intuitions.

[semente]

3 minutes ago, Jud said:

 

This is generalizing. The experiment reduced it to specifics.

 

Do you suppose, as I do, that any reasonable high end system ought to allow you to distinguish between a flute and a trombone? What has been shown by scientific experiment is that the part of the sound that enables you to make that distinction is the attack and release of the notes. This difference in attack and release between the two instruments is a very large scale difference, so large that any reasonable high end system can reproduce it and thus enable you to distinguish the two.

 

There is virtually no difference in attack and release between a viola and violin, let alone one model of violin and another, let alone playback of the same violin (or acoustic guitar) recording on two different high end systems. So what are we left with? Subtleties of tonality, perhaps. Yet the experiment shows even large differences in tonality, such as between a flute and trombone, cannot be reliably identified even by experienced musicians, so what hope do we have of reliably determining which is the "correct" tonality between two different systems playing back the same violin/acoustic guitar/flute/trombone?

 

Either those experiments were made with very low-fi systems or I have very good ears, which I actually do have. At least according to the choirmistress of the choir I used to be a part of. 😁

 

I love the sound of acoustic instruments, it's part of the pleasure of listening to music.

[Jud]

1 minute ago, semente said:

 

Either those experiments were made with very low-fi systems or I have very good ears, which I actually do have. At least according to the choirmistress of the choir I used to be a part of. 😁

 

I love the sound of acoustic instruments, it's part of the pleasure of listening to music.

 

There are sites on the web where one can reproduce these experiments. You might find one and try it.

[semente]

Another one of my test tracks:

 

 

[gmgraves]

1 hour ago, Fitzcaraldo215 said:

Who cares about the original sound field?  It betters the stereo paradigm by a fair bit, at least for classical music.

 

By that I mean two things.  Objectively, by having sound sources better able to reproduce the enveloping surround and a center channel to improve imaging.  Subjectively, on an “absolute sound” scale, it delivers clearly more realism.  

I do, for one. In classical music, it is part and parcel of the concert experience. To be able to even partially recreate it on one’s home system is to be magically transported to the place where the performance is taking place. I.E it is an inseparable part of High-Fidelity and the reason why stereophonic sound was invented in the first place!

Of course, if all one listens to is pop and rock, then I can understand the indifference because, by definition, the original “sound field” doesn’t exist anyway. That’s fine. That music is not trying to recreate a performance, but rather to create one that hasn’t existed before. There are lots of recordings of Beethoven’s 5th Symphony recorded by lots of orchestras, but The Rolling Stones original recording of “I Can’t Get No Satisfaction” was the first and before they recorded it, it didn’t exist. Do you get what I’m saying?

[Ralf11]

How about putting real, acoustic instruments in a live space, like on a stage in a concert hall - but behind a visually opaque curtain.

 

Then, substitute similarly concealed recording & playback equipment.

 

What would the audience think?

 

or what did they think...

 

 

[semente]

4 minutes ago, Ralf11 said:

How about putting real, acoustic instruments in a live space, like on a stage in a concert hall - but behind a visually opaque curtain.

 

Then, substitute similarly concealed recording & playback equipment.

 

What would the audience think?

 

or what did they think...

 

 

 

Why Live-versus-Recorded Listening Tests Don't Work

 

http://seanolive.blogspot.com/2010/07/why-live-versus-recorded-listening.html

[semente]

20 minutes ago, Jud said:

 

Yes - very large measured differences, which one would intuit would then reliably indicate to listeners, especially instrumentalists, which was which, eh? Just like one would intuit a heavier object would fall to earth faster. Too bad ugly little science experiments negate our beautiful intuitions.

 

Good question. Perhaps instrumentalists are not trained or tuned into such aspects of sound.

 

Some audiophiles report being able to listen to differences in cables or capacitors which show no measurable difference. Or look at the @manisandher experience.

[crenca]

10 minutes ago, semente said:

 

Good question. Perhaps instrumentalists are not trained or tuned into such aspects of sound...

 

Or perhaps the hypothesis is being poorly tested, or perhaps the wrong conclusions are being drawn.  Beyond that, Jud's conclusions are several degrees removed/speculated further.  The counselor is leading you by the nose... 😉

[fas42]

7 hours ago, Confused said:

The recording is fine.  If it sounds horrible it is your over ambitions rig that needs sorting.  Get your rig sorted and you will be able to get to the essence of music.

 

Or...  maybe terrible recordings sound bad, terrible systems sound bad, terrible recordings sound bad on decent systems, decent recordings sound good on decent systems.  You decide.... 

 

Yes, well ... a recording of a poor system playing an aggressive track is certainly pushing the friendship ...

 

A simple rule of thumb - any recording which involves the microphone capture of some sort of audio system outputting through speakers as one of the sound elements needs to consider that sound source as being 'processed' - and all bets are off. At one end of the spectrum, Hendrix 'torturing' his Marshall amp, while singing along - the vocals should sound like a real person; the guitar element is only very vaguely connected with what the 'live' sound of his guitar, unplugged, is like.

[Fitzcaraldo215]

6 minutes ago, gmgraves said:

I do, for one. In classical music, it is part and parcel of the concert experience. To be able to even partially recreate it on one’s home system is to be magically transported to the place where the performance is taking place. I.E it is an inseparable part of High-Fidelity and the reason why stereophonic sound was invented in the first place!

Of course, if all one listens to is pop and rock, then I can understand the indifference because, by definition, the original “sound field” doesn’t exist anyway. That’s fine. That music is not trying to recreate a performance, but rather to create one that hasn’t existed before. There are lots of recordings of Beethoven’s 5th Symphony recorded by lots of orchestras, but The Rolling Stones original recording of “I Can’t Get No Satisfaction” was the first and before they recorded it, it didn’t exist. Do you get what I’m saying?

Facetiously, maybe,  I dismissed the sound field from consideration, for a reason.  

 

But, that plays into @barrows hands.  There is an infinite number of sound fields anywhere, slightly or drastically different  in the hall.  And, on reproduction, you do not know which sound field was which.  All of which is beside the point. 

 

But, it’s the plausible event of a symphony orchestra in question.  How can we not know, with whatever sound field, how he damn thing sounds, whatever the sound field to our ears was measured.

[barrows]

1 hour ago, gmgraves said:

That you, or others might not want “accurate” is another issue; one with which I have no complaint

I never said, or asserted, any such thing!  Of course I want "accuracy", my point is that: accurate to what reference?  everything I am getting at here is that virtually no one has a true reference of what the playback "should" sound like.

 

Indeed George, you are close, as having made your own recordings, and presuming that you did hear the music recorded live as well from a certain perspective, you have a better reference than most.  But even your own recordings as a reference are flawed, because of the inaccuracy of aural memory, so it is still not "absolute".

 

But, make no mistake, I do pursue accuracy as one of the qualities I prefer in my playback system (although I tend to prefer a slight amount of a forgiving quality, just slight mind you, in order to allow for enjoyment of less than perfect recordings-for example, I tend to prefer Vivid Audio loudspeakers to something like Magicos, for those who have experience with both).

 

[gmgraves]

29 minutes ago, Ralf11 said:

How about putting real, acoustic instruments in a live space, like on a stage in a concert hall - but behind a visually opaque curtain.

 

Then, substitute similarly concealed recording & playback equipment.

 

What would the audience think?

 

or what did they think...

 

 

Acoustic Research used to do that at their  Broadway Showroom in NYC in the middle '60's. They called it their "Live vs. Recorded"  demonstration. I heard it as a teen a number of times. They had a pair of AR3a speakers on a slightly raised stage with a scrim across the stage behind which was a string quartet. You could see the quartet's shadows against the scrim, but you couldn't actually see the musicians. They would either be actually playing or, just going through the motions while a recording of their playing played through the speakers. The idea was that one couldn't tell the difference between the sound of the AR3a's and the real thing. It might have worked for the middle-aged Wall Street types who seemed to frequent the place, But I could tell which was which with my back to the stage. You see, I was 16 years old and easily hear the tape hiss... Today? Not so much.  But the music they chose, was chosen because it didn't really stress the speakers and the real and the reproduced did sound remarkably alike. I still say Kudos to Villchur et al for having the cahones to publicly run a demonstration like that!

[semente]

23 minutes ago, crenca said:

 

Or perhaps the hypothesis is being poorly tested, or perhaps the wrong conclusions are being drawn.  Beyond that, Jud's conclusions are several degrees removed/speculated further.  The counselor is leading you by the nose... 😉

 

I could be wrong but it looks to me like @Jud could be trying to downplay the importance of tonal acuracy in a recording or playback system.

But if a system's response isn't flat then the recording and more importantly the music (it's really the same thing) will not be reproduced as it was meant to.

This is of the test tracks I use to check how the room affects the low end. If some of the bowed double-bass notes play significantly louder or softer than expected then you've found problematic frequencies. It won't sound good and it definitely won't sound like the artist intended. This is true for the whole of the audible range.

[gmgraves]

6 minutes ago, barrows said:

I never said, or asserted, any such thing!

And I never intimated that YOU did. I just said that there are some audio types who do not want accurate sound. Many just like big bass and bright, splashy highs that sound spectacular as opposed to accurate. 

 

By the way, when making a recording, I don't have to have an accurate aural memory. I'm right there and slip off the headphones and listen to the real thing WHILE  I'm recording it!

But the reality is you don't need a great aural memory to remember that real and reproduced sound different. And we, as the audio community, are "there" when they no longer do!

[fas42]

16 hours ago, marce said:

You have to have a good recording and master before you can start playing the music through your high fidelity system!

 

Nope ... that's the great thing about the music recording archive that we have - there's enough information on every one of them for most of us to be able to enjoy a subjectively captivating listening experience.

 

The real myth is that "your high fidelity system" is in fact, accurate - it's not, and not by a long margin - less than pristine recordings make all the shortcomings of the playback chain extremely obvious - unfortunately, that's where that dreaded 'sorting' is required ... 😜.

[semente]

14 minutes ago, barrows said:

I never said, or asserted, any such thing!  Of course I want "accuracy", my point is that: accurate to what reference?  everything I am getting at here is that virtually no one has a true reference of what the playback "should" sound like.

 

As @ARQuint said you have to use different recordings. Or you can use a microphone which is even better.

 

The only accuracy possible is to the recorded signal. You can use music instead of measurements but it won't be as effective; I think some people like @John_Atkinson also listen to pink noise.