USB Cable Poll: Redux

[ItemAudio]

We need a passing statistician to join in, but my gut feeling (having spent a lifetime working with statisticians on understanding and analysing huge data sets) is that chance could be playing a big part part here. Correctly identifying 3 of 6 in a non-random selection (ie the range of all available audio USB cables is very small) will have odds of being correct that would surprise us all by being much more likely than intuition would have us believe.

I'm trying to find a web site that has examples of this kind of thing that I used to use in staff training. The examples had real-life situations where your instinct is "that's impossible" but the statistics show the chances are quite good. The 'birthday problem' is a good example - you only need 23 randomly selected people in a room to have a 50:50 chance of them sharing a birthday. Stats, chance and guessing can be tricky... plenty of people make a living out of understanding more about that than most (Derren Brown is one of the few to admit it).

Humbug! At Hydrogen Audio they spend so much time obsessing over statistical analysis of ABX testing they ignore the elephant in the room. If you need a plumber, hire a plumber; if you're investigating perception, hire a perceptionist.

 

[Fox]

On balance you write and say what the **** you like and stick two fingers up at anyone who agrees or disagrees with you... Well it worked for Joyce anyway. HiFi journalists have developed a semi-invented language of their own over the years so why not go the whole hog?

 

[hc25036]

I'd like a job as a perceptionist! I wonder if the training is as hard as my friendly statistician's 10 years or more?

 

[ItemAudio]

We'd do the same with Ethernet cables, but that would be like painting a huge target on your chest, so I'm personally reluctant to investigate, in case the answer is a positive one.

That definitely would. Baby steps, Alan . . . you've risked much already.

Having said that, among your day-job acquaintances, is it a commonplace that USB cables are a legitimate part of the business? Barely any of the manufacturers we deal with find the idea objectionable - most are just keen to extract the best performance from their product and welcome audio USB cables as part of the armoury.

In my experience, this kind of ire tends to flash up where like-minded consumer warriors pool to defend the innocent from industrial predation. It's strangely compelling . . .

 

[Chefren]

None of that comments on audio quality - ie, whether Mac sounds better than PC, Pi or a Commodore 64.

Obviously a Commodore 64, essentially having an analogue synth as it's audio solution, sounds perfect in every way.

 

[Julf]

Obviously this is an extreme example, but all often when 'impossibility' is claimed, it's what A C Clarke called a failure of imagination.

Everything is possible... until proven otherwise.

All this stuff is on a spectrum, so it would be interesting to ask those who claim to find USB cables different exactly how different they sounded - different like speakers? Different like DACs? Different like amplifiers?

An even better question to ask is "can people hear differences between properly designed (as in "that satisfy the requirements of the standard") cables when other sources of information are eliminated?". A simple question, that once and for all would end all discussion. And one that shouldn't be too hard to verify empirically.

 

[Julf]

if you're investigating perception, hire a perceptionist.

And if you are selling an illusion, hire an illusionist.

 

[Chefren]

And if you're selling a piano, hire a pianist - wait that doesn't sound right..

 

[Julf]

I feel we are damned whatever the outcome:

If we hear a difference and suppress it, we're lying
If we hear a difference and report it, we are accused of lying
If we don't hear a difference and report it, we are accused of being deaf
If we don't hear a difference and suppress it, we are hiding our deafness

On balance, the best action is no action at all.

Or report the results, along with enough information to allow others to verify and replicate the results. Seems to work for scientists and engineers.

 

[Julf]

Obviously a Commodore 64, essentially having an analogue synth as it's audio solution, sounds perfect in every way.

Who needs 16 or 24 bits, when you can have 8...

 

[ItemAudio]

I'd like a job as a perceptionist! I wonder if the training is as hard as my friendly statistician's 10 years or more?

I will have worked as a preceptionist for a while, welcoming people who are going to have had their ears tested. I ran a book for appointments.

 

[Alan Sircom]

Or report the results, along with enough information to allow others to verify and replicate the results. Seems to work for scientists and engineers.

I don't see how that would eliminate any of the scenarios I mentioned.

People like maxflinn seem to have a passing interest in the science either way. He dislikes the concept on grounds that at best seem 'sciency' and would not accept the findings no matter how verifiable they were. Were we to report a difference, the beatings would continue.

On the other hand, if we reported there was no difference, we'd lose the trust of those audiophiles who want there to be a difference. They would consign the magazine to the 'nerd' pile and ignore it. As they did with Audio and Stereo Review.

 

[Chefren]

Who needs 16 or 24 bits, when you can have 8...

C64 doesn't do digital audio at all. There are no bits involved.

 

[ItemAudio]

On balance you write and say what the **** you like and stick two fingers up at anyone who agrees or disagrees with you... Well it worked for Joyce anyway. HiFi journalists have developed a semi-invented language of their own over the years so why not go the whole hog?

Unfortunately English is poorly equipped to report the auditory sense - many terms in use are borrowings from sight or touch. Invention in this area is overdue.

Everyone's entitled to an opinion: but I'm very interested in what others think about this subject - hence the poll.

I'm not against any blind testing per se - having started with Fruit Pastilles in the playground at age eight - but one of several difficulties it presents is exactly that of reporting: as you point out, it's profoundly troublesome to communicate audio sense impressions. Which makes Hydrogen-style obsession with statistical analysis a curious exercise in turd-polishing.

 

[Markus S]

People like maxflinn seem to have a passing interest in the science either way. He dislikes the concept on grounds that at best seem 'sciency' and would not accept the findings no matter how verifiable they were. Were we to report a difference, the beatings would continue.

maxflinn knows that Naim preamps don't change their sound if you substitute a different power supply. And he didn't even have to go to the trouble of an audition to know this.

 

[Julf]

C64 doesn't do digital audio at all. There are no bits involved.

I don't think that is correct. Part of the SID chip was analog, but the wavetable stuff was low-resolution digital afaik.

 

[Julf]

On the other hand, if we reported there was no difference, we'd lose the trust of those audiophiles who want there to be a difference. They would consign the magazine to the 'nerd' pile and ignore it. As they did with Audio and Stereo Review.

I see your problem :-/

 

[Julf]

it's profoundly troublesome to communicate audio sense impressions. Which makes Hydrogen-style obsession with statistical analysis a curious exercise in turd-polishing.

Have you actually tried to present your views on HA?

I think your statement reflects a profound misunderstanding of HA. HA is not a hifi/audiophile site. HA is a site mostly frequented by developers of audio software and codecs. The only relevant criteria for evaluating a perceptual codec is "is there an audible difference between the processed and unprocessed sound?". Thus the obsession with statistically significant, verifiable listening tests.

 

[F1eng]

I wonder if the explanation is exactly analogous to placebo tests.
Placebos work, in that they actually make some people better even though there is no medicine being administered, i.e. scientifically there is no reason why they work. Placebos work on some people even when they know they have been administered a placebo.
A very expensive placebo pill is more effective than a cheap one, some colours are more effective than others but that varies from location to location.

So maybe some people genuinely do hear a difference in something they expect will work, and is expensive, even if there is no scientific reason why it would.

Placebos do not work on all individuals.
Cables don't sound different to all people.

This is the most reasonable explanation of the observations yet IMHO.

 

[Joe]

I wonder if the explanation is exactly analogous to placebo tests.
Placebos work, in that they actually make some people better even though there is no medicine being administered, i.e. scientifically there is no reason why they work. Placebos work on some people even when they know they have been administered a placebo.
A very expensive placebo pill is more effective than a cheap one, some colours are more effective than others but that varies from location to location.

So maybe some people genuinely do hear a difference in something they expect will work, and is expensive, even if there is no scientific reason why it would.

Placebos do not work on all individuals.
Cables don't sound different to all people.

This is the most reasonable explanation of the observations yet IMHO.

For Gods sake ! ... don't start bringing common sense into the discussion/argument.... they were enjoying themselves before you butted in !