accuracy is subjective

[Yuri Korzunov]

  If we define accuracy in audio playback as "how close something is to the original," then we have to ask "what is the original?" and "what are we comparing it to?" The original is a sound field experienced by a human being. The reproduction is a sound field experienced by a human being.
 
These two sound fields are never the same, as current audio technology does not reproduce 3-D sound fields. So suppose we have an original sound field A, and two different reproductions B & C. How do we decide which is closer to A? 
 
I claim that there is no way to do that without involving an individual human decision. For some people B is closer. For some people C is closer.
 
Therefore accuracy is subjective.

 
It is exact understanding of future of the audio branch - capturing and restorting sound field.
 
Need separate accuracy of sound field reproduction and accuracy of transmitting/storing asudio signal.
 
Both kinds of accuracy may be estimated by measurement tools. Of course, estimation of sound field demands recording/transmitting audio signal and will have lesser precision than estimation of recording/transmitting, I suppose.
 
Subjectivism begin after appearing term "threshold of audibility".
 
This threshold may be partitially ignored by double blind test. But for sound field accuracy I don't imagine how do it correctly.

 

[old tech]

   
Re piano, wonder what you think of these tracks. There has been some bleed into classical from the latest vinyl craze, but it sure doesn't have anything to do with accurate sound reproduction.

Different...

 

[Iron-Buddha]

Interesting conversation.
 
From my own experience, I've found bass boosted headphones (like the Campfire Vega's) to be more accurate, as compared to the same artist in concert.   Some people would "disparage" this type of tuning as "fun but less accurate".  
 
My suspicion is that this type of mixing http://www.soundonsound.com/techniques/mixing-bass, which is a requirement for stereo mix-down but also an art and subject to the engineer's discretion, robs some bass energy.  Now, re-injecting it on the playback end is a "one-size fits all" attempt at fixing an extremely variable problem.  Hence, everyone has different preferences as to how much bass boost they want (if any).  An audio sweep, as commonly used to generate frequency response graphs, doesn't undergo a similar mixing process.   
 
In theory, I think you could line the entire system up to reproduce the same bass as the original note...... but since the black box of mixing is variable and the recording/playback system is not a closed system, in practice, I think it's pretty futile, making the net effect that accuracy becomes subjective and inaccurate.  

 

[gregorio]

 
[1] From my own experience, I've found bass boosted headphones (like the Campfire Vega's) to be more accurate, as compared to the same artist in concert.  
[2] Some people would "disparage" this type of tuning as "fun but less accurate".  
 
[3] My suspicion is that this type of mixing http://www.soundonsound.com/techniques/mixing-bass, which is a requirement for stereo mix-down but also an art and subject to the engineer's discretion, robs some bass energy.  Now, re-injecting it on the playback end is a "one-size fits all" attempt at fixing an extremely variable problem.
 
[4] In theory, I think you could line the entire system up to reproduce the same bass as the original note......
[4a] but since the black box of mixing is variable and the recording/playback system is not a closed system, in practice, I think it's pretty futile, making the net effect that accuracy becomes subjective and inaccurate.  

 
1. As so often seems the case, audiophile statements are based on inappropriate comparisons and then followed by what is effectively faulty/irrational reasoning and conclusions to explain the results of those comparisons!
 
2. That type of tuning is indeed "fun but less accurate". You and others obviously don't like the "disparaging" connotations of that fact, so you are apparently trying to get around it by in effect redefining the meaning of word "accurate".
 
3. And here we have the false statements/explanations derived from reasoning which is erroneous because it's all based on an inappropriate comparison to start with. That type of mixing (which covers pretty much ALL modern mixing) does NOT "rob some bass energy", in fact, the exact opposite is typically the case! The "problem" you are trying to fix has absolutely nothing to do with accuracy!
 
4. Which original bass note, the one you experienced where you were standing at a live gig or the one which is actually on your recording? That question goes to the heart of the problem! "Accuracy", as far as audio equipment is concerned, has a simple meaning/determination: How closely does the output signal match the input signal? This is a purely objective determination as it can be precisely measured. You on the other hand are trying to match the output signal not with the input signal but with something almost completely different, an experience you had at a live gig?! With this comparison all you can do is form a subjective opinion of which you personally like more, you cannot make any determination of "accuracy".
4a. The solution to this "problem" is to understand what "accuracy" actually means and to realise that comparing apples to oranges doesn't tell you anything about accuracy. Simply redefining "accuracy" as subjective is not the solution though! Or rather, it's not the solution if you care about avoiding irrational/erroneous reasoning and making false statements.
 
G

 

[Yuri Korzunov]

   
"Accuracy", as far as audio equipment is concerned, has a simple meaning/determination: How closely does the output signal match the input signal? This is a purely objective determination as it can be precisely measured.

 
I agree, can be easured.
 
Main issues there:
 
1. What difference is audible?
 
2. What each kind of differences have degree of impact to sound perception.
 
These points create subjective aspect of interpretation of measurements results.

 

[gregorio]

  Main issues there:
1. What difference is audible?
2. What each kind of differences have degree of impact to sound perception.
These points create subjective aspect of interpretation of measurements results.

 
No! I'm not saying those issues are unimportant or not worth asking, I'm saying that they are different issues. Accuracy is an objective measurement, which by definition means that any individual's perception of it is a separate issue. For example, we can measure the input to DAC chip, the output of the chip, then make an objective measurement of the difference, this is an objective measurement of it's accuracy. Let's say there is a measured difference of 0.1dB. We can argue till the cows come home about our perception/subjective opinion of this chip's accuracy, anything from audibly transparent to a "night and day" difference but that's an argument about perception not about accuracy. Even if we all absolutely agree that it's a "night and day" difference, that still doesn't affect the accuracy of the DAC chip in any way, it's still accurate to 0.1dB!
 
G

 

[Yuri Korzunov]

  but that's an argument about perception not about accuracy.

 
For my purpoces I don't consider accuracy as abstract value. For me, as developer, need practical using of measured accuracy.
 
Otherwise, these measurements have no sense.
 
From my experience and experience of other people I try found correlation between measured values and subjective perception.
 
When you develop audio processing tools you improve some parameters. As example - noise. It demands efforts. Efforts is time/money.
 
You need to choose what you will improve first: noise or phase, as example. Sometimes need founde balance between several features, that impact to accuracy.
 
Need permanently search directions of improvement what give maximal subjective sound improvement in current situation.
 
Accuracy don't improved for accuracy. Accuracy improved for people.

 

[gregorio]

  For my purpoces I don't consider accuracy as abstract value. For me, as developer, need practical using of measured accuracy. ...

 
That's the thing about an objective measurement, it's immune to what you consider! And accuracy is not an abstract value, it's an objective comparative value.
 
I understand that as a product developer you have engineering practicalities to balance and marketing considerations which are largely about how your target demographic perceive accuracy. Let's be honest though, many of the audiophile products out there have nothing to do with accuracy, most digital products, cables, etc., have no more accuracy than equivalent products costing a hundred times less and sometimes they're even less accurate! What these audiophile products create instead, is the perception of (greater) accuracy.
 
It's not surprising that consumers are confused between accuracy and perception, the audiophile product industry has been deliberately blurring this line for decades, their existence often depends entirely on this, because in many cases greater accuracy at levels anywhere near audibility became an impossibility years ago. This is the science forum though, where we try to separate fact from marketing obfuscation.
 
G

 

[Yuri Korzunov]

Yes. Accuracy is objective. Perception is subjective.
 
What is objective measure that accuracy is enough?

 

[old tech]

  Yes. Accuracy is objective. Perception is subjective.
 
What is objective measure that accuracy is enough?

It should be enough, indeed it is a requirement for playback if true hi fidelity is the goal.
 
Accuracy in playback means that the listener hears exactly what the artist/label/engineer wants the audience to hear.  So accuracy in the high fidelity context means having a playback chain that is a silent witness to the recording, neither adding or subtracting to the sound of what is originally on the CD or LP.  If one uses their stereo to change the sound because of subjective preferences then that is fine, but it is moving away from fidelity to what the artist/label/producer intended.
 
I think where people often get lost on this subject is confusing production with playback on the listening experience. Some compare it to paintings where beauty is in the eye of the beholder.  While that is true, whatever beauty one finds in the painting, if they wanted a copy of that painting to hang on their own wall then the reproduction should be as accurate as possible to the original, without additional colouring, faze or distortion.
 
The main issue with accurate reproduction is that it reveals both the good and bad of a recording.  Good recordings can sound brilliant but poor ones can actually sound worse on an accurate stereo because the flaws are more readily apparent.  What I don't understand though is why be lumbered with a coloured playback chain which then colours anything being played on it when it can be touched up for individual tastes, or to compensate for poor recordings, with a good equaliser?

 

[Yuri Korzunov]

  It should be enough, indeed it is a requirement for playback if true hi fidelity is the goal.

 
Edge beween "enough"/"no enough" is subjective.
Double blind test may add objectivity in this matter.
But proper double blind test demands: methodics, laboratory, big number of measurements and participants, careful performing and protocol of test for doing of conclusions by results of the tests.
 
Sometimes detect the edge is technically impossible. As example, 16 vs. 24 bit or 44 vs. 96 kHz. Because, as rule, it is different modes (used curcuits, algorithms, etc.) of apparatus.
 

What I don't understand though is why be lumbered with a coloured playback chain which then colours anything being played on it when it can be touched up for individual tastes, or to compensate for poor recordings, with a good equaliser?

 
Me seems rather coloration may be liked.
For music producer (mixing, post-production) vinyl scratches may be considered as element of art for impact to mood of listener.

 

[Iron-Buddha]

   
1. As so often seems the case, audiophile statements are based on inappropriate comparisons and then followed by what is effectively faulty/irrational reasoning and conclusions to explain the results of those comparisons!
 
2. That type of tuning is indeed "fun but less accurate". You and others obviously don't like the "disparaging" connotations of that fact, so you are apparently trying to get around it by in effect redefining the meaning of word "accurate".
 
3. And here we have the false statements/explanations derived from reasoning which is erroneous because it's all based on an inappropriate comparison to start with. That type of mixing (which covers pretty much ALL modern mixing) does NOT "rob some bass energy", in fact, the exact opposite is typically the case! The "problem" you are trying to fix has absolutely nothing to do with accuracy!
 
4. Which original bass note, the one you experienced where you were standing at a live gig or the one which is actually on your recording? That question goes to the heart of the problem! "Accuracy", as far as audio equipment is concerned, has a simple meaning/determination: How closely does the output signal match the input signal? This is a purely objective determination as it can be precisely measured. You on the other hand are trying to match the output signal not with the input signal but with something almost completely different, an experience you had at a live gig?! With this comparison all you can do is form a subjective opinion of which you personally like more, you cannot make any determination of "accuracy".
4a. The solution to this "problem" is to understand what "accuracy" actually means and to realise that comparing apples to oranges doesn't tell you anything about accuracy. Simply redefining "accuracy" as subjective is not the solution though! Or rather, it's not the solution if you care about avoiding irrational/erroneous reasoning and making false statements.
 
G

 
 
If you want to insist that the recording is gospel, and play it back ruler flat, no EQ, then sure, that is a measurable form of accuracy compared to the recording.  I just don't believe that is the true input signal.   
 
A musician goes into a studio, lays down a track (the TRUE input signal).  The engineer applies a host of EQ, compression, etc. (based on what factors, I don't know) cuts a recording.   I play it back on my headphones.   What do I want to hear?  I want to hear what the musician played. I cannot objectively know what I am hearing is accurate unless either (i) I took independent measurements in the studio to compare to, or (ii) I know the entire chain of modifications that were made in the process of mixing.   At any rate, since these changes are not "zero", I need to undo them on playback to get to the original input.   The "undoing" process, while theoretically possible in a closed system with full information, is a pipe dream in practice.  Accuracy is objective and can be measured, but no lab test on my headphones is going to give me insight into how accurate the output matches the performance in real life, only to the recording. 
 
As for what is on the recording, Jude and Jerry Harvey (and assign as much weight to this as you want, I'm not saying their word is gospel) all seem to agree that a bit of a bass hump sounds closer to being in the recording studio.   Both of them having actually had the benefit of being in the studio.   I never had the privilege.  All I can do is go to concerts.  I agree with you 100% that this is a terrible reference point, but this is all I got man.  Maybe you have access to your favourite musicians in studio, then more power to you.  If you want to insist that the recording is gospel, then I disagree but respect your right to an opinion.  At least you have a reference point to cling to.  If you jump down this other rabbit hole, it's like being atheist, you just have to admit you don't ******* know and just eyeball it. 

 

[old tech]

   
Me seems rather coloration may be liked.
For music producer (mixing, post-production) vinyl scratches may be considered as element of art for impact to mood of listener.

 
Yes that is the point.  Putting the vinyl scratches aside as that is not production, if the music producer wants to colour the original recording through mixing or final mastering that is part of the music production chain.  The playback chain should be accurate if the listener wants to accurately hear what production effort/artistic expression etc was put into the final product.

 

[Yuri Korzunov]

   
 
If you want to insist that the recording is gospel, and play it back ruler flat, no EQ, then sure, that is a measurable form of accuracy compared to the recording.  I just don't believe that is the true input signal.  

 
Practical measurement of accuracy is technical impossible. Here important matter, how we get reference original? Capturing system also have accuracy limitation.
 
And what is "true" sound in studio/concert hall? Each sitting place have own "true" sound.
 
We can estimate accuracy of recording/playback apparatus only. Via test signals and measurement tools.

 

[Don Hills]

... A musician goes into a studio, lays down a track (the TRUE input signal).  The engineer applies a host of EQ, compression, etc. (based on what factors, I don't know) cuts a recording.   I play it back on my headphones.   What do I want to hear?  I want to hear what the musician played. ...

 
The musician doesn't want you to hear what they played. They want you to hear it after the engineer has made it sound the way the musician wants it to sound.