Article: My Lying Ears

[Jud]

I could hear a mouse fart during a Black Sabbath concert !

 

[miguelito]

I could hear a mouse fart during a Black Sabbath concert !

Jimmy sweetie... It wasn't a mouse!

[toddrhodes]

I wonder to what degree the lack of a "common language" holds back our ability to interpret and discuss what we like and dislike about audio products. In my own mind I try to break sound qualities into the following general buckets:

 

1. Frequency response (is it flat or at least linear if tilted down and to the right?) Although easily measured, room response effects can mess up an otherwise pristine starting point, so how sure are you about what you actually hear -- maybe a nonlinear piece of equipment well matches my non-linear room and sounds horrid in yours?)

 

2. Phase response and coherence Am I being negatively influenced by the arrival time of different frequencies at my ears? Most comments about good or bad imaging or sound stage probably fall into this category. Similar issue as #1 if incoherences are created by different pieces in the signal chain then the end result depends highly on equipment matching as some may cancel each other and others may reinforce the error. These are also measurable at the source, but can vary by listener position and ignores rear wave effects, particularly from dipole speakers.

 

3. Distortions (does one product seem to remove a haze that hides fine detail when compared to another?) This can include jitter, analog line noises, filtering artifacts, etc. so it's a big category but it seems clear that taking these away can "improve" the sound we hear). For example the "tighter bass" reported by most users here upon inserting an Uptone Regen in their system, I would interpret as having removed some form of clouding distortion. Similarly, a lot of the upsampling/filtering discussions appear to be about moving possible noise artifacts out of the audible range.

 

4. "Other" which can include stuff that is measurable and should really be its own category above (known others) and stuff that we think we hear but can't seem to either measure or define (unknown others). I would generally put comments like "more air," "smoother sound," "more analog sounding," "better prat," and "wider, deeper soundstage" into the unknown others category because we aren't really speaking to something that we can tie to a known or measurable equipment issue.

 

I would think that the more often we can assign a positive or negative experience to an identified bucket rather than the "other" category, the more likely that we can (a) agree about what we are hearing, and (b) address it. Stuff that stays in the other, and especially the unknown others category is likely to remain the subject of strong and often incoherent argument...

 

What other "defined" categories should be on my list above?

 

Going along these lines - those of us with just our own experience at our backs, no actual experience in mastering, recording, or significant "seat time" with systems of all makes, models, and budgets - I find this sort of thing to be analogous to cooking. How do you know a dish is too salty? Well, you've made it before with less salt. But if you've never experienced incorrect phasing, or a frequency response bump at 150 Hz, or at least didn't do so knowingly, recognizing those issues actually isn't that easy. In the same sense as the story in the OP, I've had a day or two where I was rocking out to my upstairs system, cooking some food (it all ties together in the end, really), and realized something kind of sounded odd. Yea - one speaker was unplugged. Was I listening critically? No. But I still enjoyed what I was doing and hearing, immensely.

 

My point is, finding flaws and issues, the very essence of trusting one's ears, means you've "been there, done that" on so many different levels that you can draw upon the experience of such an issue, recognize it, and fix it. This is where measurements are irreplaceable. But in the end, it's often said that ruler-flat response isn't all that engaging. And that's when we hit the subjective domain and posters should know to preface anything with "IMO." Because at that point, it's just opinion.

[speavler]

man i was hoping for something a little more interesting than the same old 'bits are bits' drivel. i never really enjoyed audio until i opened my mind and stopped listening to the 'everything sounds the same' crowd. it's a shame that they plague so much space on the internets.

[Teresa]

As to the original topic ... its always been my opinion that we have to be careful to divide observations (listening "proof") over what we enjoy, and observations that we present as evidence.

 

I personally do not accept other people's observations as having anything whatsoever to do with what I would hear, thus they are not evidence of any sort.

 

I have no problem with people saying "I enjoy listening to Music Player X more than Music Player Y" and that statement is there as their opinion and is not open to being questioned or requiring any justification. However when someone makes a statement along the lines of "I listened and Music Player X is clearly producing a better sound quality than Music Player Y" then I think that is being presented a "fact" and therefore is open to challenge over methodology used to come to that conclusion.

 

Why would you believe the second statement is presented as fact? The second statement only applies to the person doing the listening and if player Y produces better sound quality for them it does not follow that it would for you or I, thus is not a statement of fact but an observation.

 

And even more importantly why would you believe it without hearing whatever it is in your room, with in your system, with your ear/brain system? The best audio reviewers have aways said reviews are a guideline and if interested listen for yourself. For me, if enough people like something and it is something I'm looking for I will check it out and give it a long-term trial in my room as long as it has a money-back guarantee.

 

As an aside, I always feel such proof should include the fact that you can tell when things having changed. For example if you are "blind" testing two software players by having someone switch back and forward and asking you if can hear a differences, sometimes when they say they have changed, they should stick with the same player. Of course so many audiophiles will say any form of blind testing is too stressful and therefore the differences will always fail to be noticed...

 

In my opinion, any form of A/B testing either sighted or blind does not work. Long term listening over several weeks is the only way I have ever discovered to reveal important differences in the sound of music.

 

Five "human" things ensure why sighted or blind A/B testing fails to reveal statistical differences between nearly everything:

 

Cognitive bias

- your brain will fill in missing information thus making both samples sound the same on repeated listening.

 

Listener Fatigue

- switch back and forth too often and both music files will sound like crap.

 

Accumulative effects are hidden

- Accumulative effects on sound quality increase over time and remain hidden when switching back and forth between two music files, especially things such as strident/smooth, cold/warm sound, etc.

 

Soundstage and instrument placement

- it takes anywhere between 30 seconds to several minutes for my brain to map the soundstage and hear the instrument and vocal placement before I can judge anything. A/B'ing insures this never happens.

 

Confirmation Bias

- In addition sighted A/B testing has to fight confirmation bias, as some people think the major brand or more expensive item must sound better. This is not always true as sometimes the unknown brand or the least expensive item sounds the best.

 

Long-term listening will reveal those differences that your brain tries to hide from you. One has to use A/B in evaluating audio equipment prior to purchase. For me sometimes this can take many months, during which time all the stereo salespeople know me by name. I ask them to turn out the lights in the listening room, and if not possible to dim them to the lowest setting. And to leave while I listen to one complete song uninterrupted.

 

At home either sighted or blind I never A/B anything as I don't trust by ear/brain system's first impression. It takes a very long time and I hate comparing stuff and only do so when it is absolutely necessary.

 

I wonder if subjective reports put IMHO at the end would you accept them as observations instead of fact?

[spdif-usb]

I wonder to what degree the lack of a "common language" holds back our ability to interpret and discuss what we like and dislike about audio products. In my own mind I try to break sound qualities into the following general buckets:

 

1. Frequency response (is it flat or at least linear if tilted down and to the right?) Although easily measured, room response effects can mess up an otherwise pristine starting point, so how sure are you about what you actually hear -- maybe a nonlinear piece of equipment well matches my non-linear room and sounds horrid in yours?)

 

2. Phase response and coherence Am I being negatively influenced by the arrival time of different frequencies at my ears? Most comments about good or bad imaging or sound stage probably fall into this category. Similar issue as #1 if incoherences are created by different pieces in the signal chain then the end result depends highly on equipment matching as some may cancel each other and others may reinforce the error. These are also measurable at the source, but can vary by listener position and ignores rear wave effects, particularly from dipole speakers.

 

3. Distortions (does one product seem to remove a haze that hides fine detail when compared to another?) This can include jitter, analog line noises, filtering artifacts, etc. so it's a big category but it seems clear that taking these away can "improve" the sound we hear). For example the "tighter bass" reported by most users here upon inserting an Uptone Regen in their system, I would interpret as having removed some form of clouding distortion. Similarly, a lot of the upsampling/filtering discussions appear to be about moving possible noise artifacts out of the audible range.

 

4. "Other" which can include stuff that is measurable and should really be its own category above (known others) and stuff that we think we hear but can't seem to either measure or define (unknown others). I would generally put comments like "more air," "smoother sound," "more analog sounding," "better prat," and "wider, deeper soundstage" into the unknown others category because we aren't really speaking to something that we can tie to a known or measurable equipment issue.

 

I would think that the more often we can assign a positive or negative experience to an identified bucket rather than the "other" category, the more likely that we can (a) agree about what we are hearing, and (b) address it. Stuff that stays in the other, and especially the unknown others category is likely to remain the subject of strong and often incoherent argument...

 

What other "defined" categories should be on my list above?

Buckets (1) Frequency response and (2) Phase response and coherence are both part of bucket (3) Distortions.

 

That said, distortion type F is the most interesting type IMHO. Types of Audio Distortion

[mansr]

Yes, John and Gordon were just preparing for the day when they would get into the software business (which hasn't happened yet - perhaps in the next decade, eh?).

 

They sell their writing.

[Audio_ELF]

I personally do not accept other people's observations as having anything whatsoever to do with what I would hear, thus they are not evidence of any sort.

Teresa ... I know you approach things with a very different attitude to others. I don't disagree with anything you said because its about how you read other people's comments; however I also do not agree.

 

When I talked about how people write something, its not about single sentences but the whole impression of a posting or article. The subtext for want of a better word. Now perhaps how you read a particular post is right, and how I read it is wrong; but when (for example) its announced Sydney Audio Club to blind test audio software, formats | DAR Australia this is (to my mind) about something more than a casual "see what you prefer".

 

At the end of the day, many people want to put under analysis some of the claims of companies they are dealing with. You disagree with that and I'm not telling you you are wrong. But never should you be telling people who do want to do some analysis they are wrong.

 

Eloise

[Jud]

They sell their writing.

 

You're now being a little bizarre. In what sense were they "selling" comments on someone else's hosted Web forum, the same thing we're both doing now?

[Jimmypowder]

Jimmy sweetie... It wasn't a mouse!

 

Hmmm . Your right Miggy ! Lol .

[mansr]

You're now being a little bizarre. In what sense were they "selling" comments on someone else's hosted Web forum, the same thing we're both doing now?

 

They are professional writers, are they not?

[Jud]

They are professional writers, are they not?

 

So far as I'm aware that hasn't been either one's profession, no. They're engineers. Has either ever written an article for publication in return for compensation? It may have happened, though I've never seen it.

 

Regarding their primary professional pursuits, John designs the insides of ICs for a semiconductor manufacturer; and Gordon's Streamlength protocol is generally regarded as having begun the wide popularization of async USB input DACs, and by extension of computer audio itself.

 

So where you got "professional writers" from is a puzzle.

 

And then separately there is still the fact of Audirvana's source code having been open to the world for a long time and not a peep of criticism having been heard from anyone who looked at it.

[james45974]

This "lying ears" subject relates back in a way to the Lexicon fiasco of repackaging OPPO players and hiking the price. The nicer case or the price with the more zeroes has to sound better! Chris, those speakers you made can't have sounded any good, why they were...home made! No fancy CNC production or perfect lacquer finishes or unobtainium tweeters!

 

Thing is audiophiles like to use a live performance as a benchmark, a sort of absolute truth that is quite unobtainable without a live performance. No matter how your system is set up or how much it costs it is always coming up short, your ears are lying to you every time your turn your system on. The thing is to untangle your knickers and stop feeling that you and your system are inadequate and just enjoy the music!

 

It would be interesting to test one facet of this further. I have multiple test and setup recordings, on vinyl, CD and digital file. I think it would be interesting to do a test to see what differences there are in the sound of say an 80 hz test tone. With obvious differences in the format it would be interesting to see what the results might be. I bet you would find someone who would swear hearing a difference between the tone from the Nordost disc versus the Stereophile disc.

 

Maybe the jokers aren't far off when they wonder what Miles Davis had for breakfast the day he recorded Kind of Blue!

[james45974]

Interesting article on Oliver Sacks in todays LA Times. I have not read his book but is talks about “Musicophilia: Tales of Music and the Brain.”

 

But, he added, we’re all aurally overwhelmed whether wearing ear buds or not. “[F]or those of us not plugged in, there is nonstop music, unavoidable and often of deafening intensity, in restaurants, bars, shops and gyms. This barrage of music puts a certain strain on our exquisitely sensitive auditory systems, which cannot be overloaded without dire consequences.”

 

Is there a better argument for a moment of silence?

 

Oliver Sacks & music: On brainworms, hallucinations and sonic overload - LA Times

[mansr]

So far as I'm aware that hasn't been either one's profession, no. They're engineers. Has either ever written an article for publication in return for compensation? It may have happened, though I've never seen it.

 

Regarding their primary professional pursuits, John designs the insides of ICs for a semiconductor manufacturer; and Gordon's Streamlength protocol is generally regarded as having begun the wide popularization of async USB input DACs, and by extension of computer audio itself.

 

So they do sell stuff.

 

So where you got "professional writers" from is a puzzle.

 

There is a different John Swenson who writes various publications. When you insisted the one being referred to here didn't sell stuff I, incorrectly, thought that perhaps you were talking about him. Gordon spends enough time being interviewed by the audiophile press that he may as well be a writer.

 

And then separately there is still the fact of Audirvana's source code having been open to the world for a long time and not a peep of criticism having been heard from anyone who looked at it.

 

If it is bit-perfect there's obviously nothing wrong with the source code, so there's nothing to criticize.

[firedog]

So they do sell stuff.

 

 

 

There is a different John Swenson who writes various publications. When you insisted the one being referred to here didn't sell stuff I, incorrectly, thought that perhaps you were talking about him. Gordon spends enough time being interviewed by the audiophile press that he may as well be a writer..

 

You are really stretching to make your (not very strong) points. Gordon is interviewed a lot b/c he is considered an authority and b/c he expresses himself well. I'm sure he also turns down a lot of interviews in order to get work done.

Unlike you, many of us don't automatically dismiss someone's opinions just because he makes money in the industry. Some, like Gordon and John, have shown themselves to be helpful and useful to the rest of us ignoramuses - even when they have nothing to gain from it. Personally, I'd trust both of them over many semi informed amateurs with all sorts of axes to grind in online pissing contests.

[plissken]

In my opinion, any form of A/B testing either sighted or blind does not work. Long term listening over several weeks is the only way I have ever discovered to reveal important differences in the sound of music.

 

Industry luminaries Toole and Olive disagree

 

Five "human" things ensure why sighted or blind A/B testing fails to reveal statistical differences between nearly everything:

 

How come it works for various style of speakers?

 

At home either sighted or blind I never A/B anything as I don't trust by ear/brain system's first impression. It takes a very long time and I hate comparing stuff and only do so when it is absolutely necessary.

 

I wonder if subjective reports put IMHO at the end would you accept them as observations instead of fact?

 

During a cable burn in thread at another forum it was offered that two burned in and two non-burned in cables would be randomly labeled and sent out. 30 days the persons would have to not only group the cables into groups of difference but then move onto picking which group was burned in.

 

No test administrator, 100% sighted. So no SBT, DBT, or AB/X or other discriminate type testing and a long listening period measured in weeks and weeks. Guess how many believers stepped up?

 

Too many audiophiles put themselves squarely in the cross hairs of well thought out testing. Various forums are littered with this hilarity.

[Jud]

So they do sell stuff.

 

 

 

There is a different John Swenson who writes various publications. When you insisted the one being referred to here didn't sell stuff I, incorrectly, thought that perhaps you were talking about him. Gordon spends enough time being interviewed by the audiophile press that he may as well be a writer.

 

If it is bit-perfect there's obviously nothing wrong with the source code, so there's nothing to criticize.

 

But not player software. So no financial incentive. And "may as well be a writer" because he gives interviews - this is supposed to relate to financial incentives with regard to player software how, exactly?

 

Here you are throwing around accusations without having informed yourself about what John or Gordon or Damien actually wrote, nor do you have an idea what Damien's code does beyond transmitting the source bit perfectly if the user wants that. In another thread, you tried to disagree with Miska (who works for Intel) about the capabilities of DAC chips vs. CPU. In other words, while you have knowledge which can be helpful, you've often chosen to speak first and inform yourself later, if at all. How do you evaluate the credibility of people you know with that tendency?

 

May I therefore suggest simply seeking out what Damien, John and Gordon have said on the subject, informing yourself further about anything you feel you might need to or are curious about, and then coming back here and providing substantive comments from an engineering/software development point of view if you would like?

[Jud]

Too many audiophiles put themselves squarely in the cross hairs of well thought out testing. Various forums are littered with this hilarity.

 

I disagree that we can assume the tests are well thought out without knowing more about how the ear-brain and such things as auditory memory function. The way the brain processes sensory input is unbelievably specific (there are different visual area cells in the brain to process short vertical motions than process short horizontal motions, for example). So I'd caution against thinking we know enough about how this all works to design tests that are valid in multiple areas of the hearing experience.

[ednaz]

I personally do not accept other people's observations as having anything whatsoever to do with what I would hear, thus they are not evidence of any sort.

 

In my opinion, any form of A/B testing either sighted or blind does not work. Long term listening over several weeks is the only way I have ever discovered to reveal important differences in the sound of music.

 

Five "human" things ensure why sighted or blind A/B testing fails to reveal statistical differences between nearly everything:

Cognitive bias

- your brain will fill in missing information thus making both samples sound the same on repeated listening.

 

Listener Fatigue

- switch back and forth too often and both music files will sound like crap.

 

Accumulative effects are hidden

- Accumulative effects on sound quality increase over time and remain hidden when switching back and forth between two music files, especially things such as strident/smooth, cold/warm sound, etc.

 

Soundstage and instrument placement

- it takes anywhere between 30 seconds to several minutes for my brain to map the soundstage and hear the instrument and vocal placement before I can judge anything. A/B'ing insures this never happens.

 

Confirmation Bias

- In addition sighted A/B testing has to fight confirmation bias, as some people think the major brand or more expensive item must sound better. This is not always true as sometimes the unknown brand or the least expensive item sounds the best.

 

Long-term listening will reveal those differences that your brain tries to hide from you. One has to use A/B in evaluating audio equipment prior to purchase. For me sometimes this can take many months, during which time all the stereo salespeople know me by name. I ask them to turn out the lights in the listening room, and if not possible to dim them to the lowest setting. And to leave while I listen to one complete song uninterrupted.

 

At home either sighted or blind I never A/B anything as I don't trust by ear/brain system's first impression. It takes a very long time and I hate comparing stuff and only do so when it is absolutely necessary.

 

I wonder if subjective reports put IMHO at the end would you accept them as observations instead of fact?

 

As someone who's been a follower of Daniel Kahnemann's research (since the earliest days of Tversky and Kahnemann), and who's taught classes on cognitive biases, it's absolutely true that cognitive biases can raise hell with blind testing. If jamming a pencil in your mouth so that it pulls your lips back as if you were smiling can make you feel more positively about information... and if a person speaking to a huge conference hall full of people can influence how they feel about pieces of the presentation by ever so slightly nodding his head, or shaking his head (and by slightly I mean such that in a video of the event where the speaker and the presentation screen fill the frame, you can't tell he's doing it)... then any blind test where you're in the same room with the person cueing up the samples is garbage. I've done both the experiments mentioned here many times, along with many others described in Hahnemann's research. We are susceptible to the simplest of cues.

 

Second, I'd be VERY interested in what player each voter listened to regularly at home, and would like to compare that to their votes. There can be cues in the players that won't show up on a Tektronix screen.

 

I didn't look really closely at the setup, but I've been in many A/B testing situations (for digital cameras, printers, and audio systems) where small within range sample variations in some part of the system drove the outcome, confirmed by swapping the components in question. I also know, from far distant days of working in a stereo shop (now there's an ancient concept) that many speakers and components can sound WONDERFUL in an hour of listening, but when taken home and listened to for a few hours, are fatiguing as all get-out.

 

But I'd put my money on the cognitive biases. I've been able to have people watch a video where they saw something actually happen, and then made them believe the opposite in a few minutes of talking just with physical cues.

[plissken]

I disagree that we can assume the tests are well thought out without knowing more about how the ear-brain and such things as auditory memory function. The way the brain processes sensory input is unbelievably specific (there are different visual area cells in the brain to process short vertical motions than process short horizontal motions, for example). So I'd caution against thinking we know enough about how this all works to design tests that are valid in multiple areas of the hearing experience.

 

I know with the cable burn in that no one could find the chink in the armor so to speak. Fundamental and well thought out testing doesn't have to take into consideration anyone's ear-brain connect.

 

People were espousing the clear differences in a gross manner. That's what was being tested.

 

I really like that a subjectivist at JRiver forums finally admitted that the ear is terrible an figuring out what is going on in a complex system such as a speaker.

 

I've measured rooms where the owner thought the sound was great just to sort out the anomaly/s that were there and they were never going to get it by ear. The difference for them was improved fidelity.

[Jud]

I know with the cable burn in that no one could find the chink in the armor so to speak.

 

Did they try a "burned-in" cable on one channel and a new cable on the other channel simultaneously with well recorded mono to take the huge variable of auditory memory (which some research for at least one type of cue says is reliable for as little as two seconds) out of the equation? What about doing the same simultaneous testing with interconnects and listening through headphones to take the enormous variable of the room (which many people say is second in importance only to speakers, or perhaps even more important) out of consideration?

 

Auditory memory and room effects are completely unimportant and safe to ignore in your estimation? Or might those be considered "chinks in the armor" if you want to do really reliable testing?

[Jud]

I've measured rooms where the owner thought the sound was great just to sort out the anomaly/s that were there and they were never going to get it by ear. The difference for them was improved fidelity.

 

I think those sorts of measurements are extremely helpful if correctly done.

[dallasjustice]

I know with the cable burn in that no one could find the chink in the armor so to speak. Fundamental and well thought out testing doesn't have to take into consideration anyone's ear-brain connect.

 

People were espousing the clear differences in a gross manner. That's what was being tested.

 

I really like that a subjectivist at JRiver forums finally admitted that the ear is terrible an figuring out what is going on in a complex system such as a speaker.

 

I've measured rooms where the owner thought the sound was great just to sort out the anomaly/s that were there and they were never going to get it by ear. The difference for them was improved fidelity.

 

Thank you. I recommend anyone interested in the topics you raised to read this book.

 

[davide256]

Wow... interesting to see the same lie technique that sold me a pioneer receiver based on THD in the 70's is still being championed today... crock of s&%t. Maybe they think we all can be brainwashed to pay attention to marketing propaganda vs trust our ears.