Where is Audio Truth ?

[Audiophile Neuroscience]

It would appear there are many threads on many forums debating computer audio evidence vs hearing differences when listening. People have dug trenches, grenades have been lobbed, feelings (and maybe some egos) bruised, threats made or implied, ridicule slung,….and this has been interspersed with respectful discussion over differing viewpoints. So all in all its like most other so-called subjectivist vs objectivist debates I have stumbled upon, ending with people picking up their marbles and retiring to their corners.

 

I say "so-called" subjective vs objectivist as, ironically, I see both sides of the debate as necessarily subjective.

 

As some may already know I am sceptical about "data" and "measurements" presented as evidence. As a scientist in my own field I have seen how they can mislead or be misinterpreted or just be plain wrong. They become subjective in their interpretation as to what they mean. They become potentially dangerous in that they start to dictate behaviour, based on assumptions/conclusions on the available evidence of what we know about things here and now.

 

We still of course need measurements and still need to strive to improve both their accuracy as well as their relevance to the outcomes we are studying. I believe evidence should guide our opinions and behaviour, just not dictate it rigidly or have the need to slavishly stick to it. Evidence is only as good as the next batch of evidence, the new tool to measure it, or the new revelation we are not measuring what we thought we were. Evidence changes, and as such is a somewhat capricious commodity.

 

What is more disturbing, for some, than the 'attack' on physical evidence being relatively unreliable is the whole unsavoury notion of psychological influences. A rejection that they have any valid role to play in interpreting or changing the 'real' world. Surely things psychological are an affront to the physical evidence, 'not real', 'imaginary' and even pejorative ? If physical evidence can't explain something the assertion is that you are nuts, unworthy, stupid in believing otherwise.

 

I think most of us are aware that psychological things can alter the way we perceive things and seemingly at times with out any accompanying change in the real world or sensory processing apparatus. The question is does it matter? If it comes down to enhancing the enjoyment of an entirely subjective experience like music on our HiFi, itself a basically 'illusory phenomenon', then I would say no. I say illusory because we are relying on known pycho-acoustical neurological processing to recreate something that isn't there, not just a recreation of a past event, but an actual acoustic image of instruments and people an a soundstage.

 

I think there is a deeper concern however that the 'truth' is not being served, that advancements in audio will only come with 'real changes' in the research and development of technologies and equipment. Things that can be verified. There are parallels in medical research where people go to great pains to eliminate the effect of placebo influencing outcomes. On the other side of the equation the placebo effect, likely to apply to each and every one of us, is harnessed for its therapeutic power. A boost of around 30% efficacy is not to be sneezed at... and that goes double when treating hay fever ! It can however be potentially dangerous when it is the sole foundation for results if it distracts or diverts from other treatments that may be needed to address serious illness.

 

Is Audio truth, then, a matter of greater 'accuracy'. If we get more clever in our measurements and our gear improves to faithfully reproduce what we measure then it stands to reason that more accurate sounds better ? Does the more accurate CD sound better than Vinyl ? Does halving Total Harmonic Distortion have the inverse affect on sound quality ? Does eliminating all jitter sound better ? Is there a 'sweet spot' in increasing digital resolutions ? Or is there something in the flawed nature of analogue sound that appeals to us ? Stay tuned for the answers…..but not from me.

 

Is audio truth, then, a function of double blind listening tests ? Well as scientific testing goes I would suggest it is not exactly a model of the ideal test rig but it seems it is entrenched in our thinking. It certainly has face validity. It may be of genuine value depending on the methodology and limitations of the interpretations and conclusions. When examining the place of testing methods is it that truth comes in the form of pitting reliability, specificity, and prevalence together in a 2 x 2 table of false positives/negatives and true positives/negatives to determine Positive and Negative Predictive Values ? The whole DBT debate is a topic all on its own, just mentioning here with a view to deferring to another time.

 

Whether in medicine or music my view is that sometimes the evidence is right and sometimes it is wrong. Sometimes the effect is largely psychological, sometimes it isn't. Sometimes you’re the bug and sometimes you’re the windshield..er, got carried away. It does however trouble me when people start talking in absolutes on either side of the fence. A little red flag goes up in my brain saying, they don’t know what they're talking about. Then again , I could be wrong ! ;-)

 

 

The neurobiology is interesting on this subject. Psychological influences can not only 'colour' the way we see things but it seems also to be able to change the very hard-wiring in the central nervous system that processes sensory information. A kind of self reinforcing and self propagating loop that physically changes our perception of the world. At the centre of this is the neuroplasticity of the nervous system that adapts to changes not just in relational to the physical milieu but also to our psychological state, or if you like, our reaction to the world. This can have positive benefits in many domains including shaping our abilities to learn new things, discern or refine our perceptual skills. It can also be destructive, sensitising our responses to pain in a negative way. It is a very powerful construct, analogous to fire, it can warm you when your cold of burn you if applied incorrectly.

 

So just perhaps, when it comes to listening to music, it appears there is something more to "trust your ears" than 'meets the eye'.

 

Enjoy the music.

[wgscott]

The placebo effect is demonstrably real, but I'm not sure that really gets us any further.

 

I'm probably one of the most right-wing extreme "objectivists" on this forum (at least according to my detractors), but was chased off a different one rather quickly for pointing out that the self-annointed objectivists seemed to be slavishly aping what they wrongly assumed to be the aims and methodologies of the physical sciences.

 

I think there is a real lack of understanding about what "objective knowledge" actually entails. This is probably our own fault, as scientists. Very briefly, I think the problem can be traced back to LaPlace, specifically his subjective interpretation of probabilities as a quantification of one's lack of knowledge or understanding. This leads to a bizarre idea in information theory that entropy is a quantification of the state of our ignorance of a physical system, rather than a thermodynamic state function that describes the degeneracy of the system itself. After that, the whole thing just went to hell.

[Snowmonkey]

In the context of a physical system:

 

degeneracy = disorder

disorder = randomness

randomness = unknowability

unknowability = ignorance

 

The map is not the territory.

[chriss71]

Some things I can measure without before I have to hear it.

When I compare two versions of a CD, then I make a frequency analysis and when I get the a zero line over the complete frequency one is for sure: The two CD's shares the same mastering. The same thing, when I look at the WAV against FLAC Debate. Pure simple science. If the MD5 checksum on both files is the same, IT IS identical.

 

So, and now one cames up and say it sounds different. This is simply not possible (if the WAV have the same MD5 checksum as the FLAC - they are bit for bit the same data in it).

 

We get fooled by our perception so many times, that I believe more the scientific than my own ears...

So, most of the time I double blindfold test different masterings. Then I can say, what FOR ME sounds better....

[Snowmonkey]

So, most of the time I double blindfold test different masterings. Then I can say, what FOR ME sounds better....

Different masterings is easy. How about different resolutions?

[Zakus]

Is audio truth, then, a function of double blind listening tests ? Well as scientific testing goes I would suggest it is not exactly a model of the ideal test rig but it seems it is entrenched in our thinking. It certainly has face validity. It may be of genuine value depending on the methodology and limitations of the interpretations and conclusions. When examining the place of testing methods is it that truth comes in the form of pitting reliability, specificity, and prevalence together in a 2 x 2 table of false positives/negatives and true positives/negatives to determine Positive and Negative Predictive Values ? The whole DBT debate is a topic all on its own, just mentioning here with a view to deferring to another time.

 

 

Enjoy the music.

 

Your thoughtful contribution from the point of view of a practicing scientist is appreciated. I look forward to your next contribution re DBT's. I followed wgscott treatment on that other forum where DBX testing is the temple of all truth. I have often thought that their use in assessing music reproduction is majorly flawed. Can the results be accurately reproduced across so many variables? do they stand up to statistical analysis? I'm in the medical field but not as a scientist and look forward to your views.

[Audiophile Neuroscience]

When I compare two versions of a CD, then I make a frequency analysis and when I get the a zero line over the complete frequency one is for sure: The two CD's shares the same mastering. The same thing, when I look at the WAV against FLAC Debate. Pure simple science. If the MD5 checksum on both files is the same, IT IS identical. So, and now one cames up and say it sounds different. This is simply not possible (if the WAV have the same MD5 checksum as the FLAC - they are bit for bit the same data in it).......I believe more the scientific....

 

It shows you that a file measures the same as another file using your measuring tools. It shows you that an audio signal measures the same as another audio signal using your measuring tools. What it doesn't show you is how it sounds....numbers or signals are not sounds. Only a brain can hear sound, see colour, feel warmth.

 

You can make reasonable extrapolations at predicting the sound based on analysis of the data and determining if their is a correlation which is statistically significant to what is heard. Even then, correlation does not mean necessarily a causal relationship, that A causes B.

 

If you "trust science" then it means you follow through with scientific methodology no matter how much you want to believe that same measurements means hearing the same sound. The latter is an interpretation of the data. The only way you can test hearing is by hearing not by measurements used as surrogate values of hearing.

[chriss71]

What it doesn't show you is how it sounds....numbers or signals are not sounds. Only a brain can hear sound, see colour, feel warmth.

 

The problem is, we are speaking about two different things. I don't have said anything about HOW it sounds!

Can you agree with me on the following sentence: FLAC and WAV sounds the same as long as the FLAC is MD5 equal as the WAV (and I don't have to hear both of them).

[Audiophile Neuroscience]

Very briefly, I think the problem can be traced back to LaPlace, specifically his subjective interpretation of probabilities as a quantification of one's lack of knowledge or understanding. This leads to a bizarre idea in information theory that entropy is a quantification of the state of our ignorance of a physical system, rather than a thermodynamic state function that describes the degeneracy of the system itself. After that, the whole thing just went to hell.

 

I am sure not an expert on entropy or thermodynamics but I guess entropy can be seen as a number of possibilities for any given state and with that brings the probability of any one state over another. If probabilities defines your chances of being right then I suppose it conversely defines your chances of being wrong (quantifying ignorance).I don't know ?

 

It appears that in the physical sciences truth lends itself more to a binary concept, right or wrong, with chemistry and physics being governed by mathematical formulae of thermodynamics.

 

When the atoms and molecules get arranged into giant conglomerates called human beings things start becoming far less predictable than pure thermodynamics even if our internal systems are still governed by these rules. Probability rather than absolutes tend to apply and sometimes one plus one equals three, the whole being more than the sum of the parts.

 

I look at the practice of medicine and music as being arts based on science. You can study pc audio technology in isolation to human hearing as you can study physiology in isolation. The trouble with humans is they don’t read the text books and do as they're supposed !

 

Hey my vote for whacky Frenchman to blame our scientific misadventures would be Descartes and his theory of mind-body dualism. It is exactly the sort of reductionist thinking that underlies the subjectivist-objectivist debate.

 

[Audiophile Neuroscience]

The problem is, we are speaking about two different things. I don't have said anything about HOW it sounds!

Can you agree with me on the following sentence: FLAC and WAV sounds the same as long as the FLAC is MD5 equal as the WAV (and I don't have to hear both of them).

 

That is what you are saying, "FLAC and WAV sounds the same". Numbers and signals are not sound, they are surrogate values of sound.

[chriss71]

Yep, because the DAC get's the same binary stream on zero's and one's. I don't have to hear it...

[4est]

That is the antithesis of scientific, and a faith based opinion. Faith that the only thing that matters is the md5 check sum,

 

Yep, because the DAC get's the same binary stream on zero's and one's. I don't have to hear it...

[Audiophile Neuroscience]

Yep, because the DAC get's the same binary stream on zero's and one's. I don't have to hear it...

 

 

Its tempting and you want to subjectively believe it but it is not logical or scientific, therein lies the irony of the whole subjective-objective debate. Your hypothesis requires testing. You cannot test it against itself....the numbers are the same, therefore it must sound the same....because the numbers are the same

 

[chriss71]

That is the antithesis of scientific, and a faith based opinion. Faith that the only thing that matters is the md5 check sum,

 

Yeah, last post here. I am network technician and you can believe me, because it is so simple to prove. Get a Disassembler and a Debugger (I have used IDA Pro) and cut the bitstream before the DAC. FAITH BASED OPINION... LOL

[Audiophile Neuroscience]

I am network technician and you can believe me,

 

that would require my faith ,lol ;-)

 

anyways won't challenge your belief system further, we'll agree to disagree.

[chriss71]

we'll agree to disagree.

 

Yep, exactly... I don't understand what should SOUND different, when the DAC get's the same bitstream? It is not so, that when I put up some Dire Straits CD I hear some Mozart...

I have proved with scientific that before the DAC get's the binary stream, WAV and FLAC are the same... (and that means SOUND the same)

 

Now my conclusion: If all of the other things in the chain are the same, it sounds the same. This I have tested with my own ears on double blindfold conditions.

 

So, this said, you think numbers don't make the sound. Well, why we hear something when we put in a CD. What dou you think is on the CD? Answer: Numbers... (Zero's and One's). I am not sure if we talk about the same here....

[Audiophile Neuroscience]

I followed wgscott treatment on that other forum where DBX testing is the temple of all truth. I have often thought that their use in assessing music reproduction is majorly flawed. Can the results be accurately reproduced across so many variables? do they stand up to statistical analysis? I'm in the medical field but not as a scientist and look forward to your views.

 

Yeah things aren't as easy to prove as one would think. If they were, lipid experts around the world would not be still arguing about the best treatment for high cholesterol or which is fragment is the really "bad" one if the one we thought was bad is lowered and people still have heart attacks.Thats a debate going back to the Framingham studies decades into last century.Just the same Im not stopping my statin just yet :-)

 

[Audiophile Neuroscience]

... I don't understand what should SOUND different, ....

 

That doesn't mean it can't ;-). Understanding why is not a pre-requisite for reality.

[4est]

You are talking to a brick wall. Hey, he is a network technician. I am sure (s)he understands everything there is to know about PCM audio.

That doesn't mean it can't ;-). Understanding why is not a pre-requisite for reality.

[chriss71]

Until the bitstream enters the DAC, we are here in a complete digital domain. The signal is always a bitstream. So, if that bitstream is equal, it can't sound different. The DAC get's the same information...

 

We are here in pure network field. It is the same (and it doesn't matter if we are talking about network data, or music data, or anything else - all are a bitstream). And now to my question: You think it sound different - on the same bitstream?

If this is true we'll get a nobel prize...

[chriss71]

You are talking to a brick wall. Hey, he is a network technician. I am sure (s)he understands everything there is to know about PCM audio.

 

In a digital domain, there is no PCM Audio... (only a bitstream)

[4est]

Excuse me, but I meant that as a differentiation from analog. Interestingly however, those digital signals are all analog signals "representing" digital states. In the past I have gotten into trouble for saying this, but the more involved I am with digital audio, the more it seems like analog. It would seem that Gordon Rankin and Charles Hansen are of a similar mind, but I am sure that you understand more than they being the network technician that you are...

 

In a digital domain, there is no PCM Audio... (only a bitstream)

[chriss71]

I am not sure what you mean here, @4est. Until the bitstream hits the DAC, there is no analogy to analog. Digital works completely different than analog. Many times I see that people are taking analogys that was making in the analog aera to the digital aera. But that doesn't work.

 

I hope, you can get a clue of that, written above...

 

In the past I have gotten into trouble for saying this, but the more involved I am with digital audio, the more it seems like analog.

[The Computer Audiophile]

I am not sure what you mean here, @4est. Until the bitstream hits the DAC, there is no analogy to analog. Digital works completely different than analog. Many times I see that people are taking analogys that was making in the analog aera to the digital aera. But that doesn't work.

 

I hope, you can get a clue of that, written above...

 

 

Chris71 - You're not going to like this one -> There's no such thing as digital: A conversation with Charles Hansen, Gordon Rankin, and Steve Silberman | AudioStream

[chriss71]

@Chris: Thanks for the link, will read it with interest...