Blind Listening Tests & Amplifiers

Status
Not open for further replies.
SORRY ..... Long Post !!

Peter D writes: What I'd like to agree on, is that although two amps may seem to sound the same in a blind test to a person, they might not sound the same in subjective environment, the one we are accustomed to in our usual way of listening and this should be treated as "real" reality, more important than just aesthetics. This might not appeal to everybody and of course is not the same to everybody, but in some ways it's real.

The problem with a "lumped parameter model" is that it has been badly abused in the past, and this was much of the incentive NW had to create this thread in the first place.

The point being that if we separate out what is "truly audible difference" (and this relates to the individual in question - not someone else in another location), then we can accurately assign importance to all of the "other factors".

The reason I have been at pains not to say person x, y or z can or cannot hear A vs B is I don't want to muddy discussion of the method. I believe NW made a fatal mistake when he infered certain people would or would not be able to hear differences between specific products. This had the very predictable result of reducing much of the thread content to "4 year old sand-pit level" argument.

The placebo-effect is a very real thing. Not all people are subject to it to the same degree, but it IS real.

People's ability to discriminate varies, their inate hearing thresholds vary, etc, etc. Despite all of these differences the method remains valid because it simply allows better description of the components of our global perception. We know how much is "audibly discernable", the rest making-up the total "package".

I took a shot at Pan for this reason. It doesn't matter what he can or cannot hear. I don't care if he can hear me typing in Oz from Sweden. This is actually near irrelevant to a discussion of the merits of the process/method.

When assessing the merits of a "method of assessment" we should be concerned with discrimination (of the method, not an individual), reproducibility, inter-rater reliability, flexibility, usability, etc. Personal anecdotes do not make a scientific argument.

The biggest problem with blind testing (IMO) is in fact its "usability". By this I mean it is actually very difficult to set-up accurately and for most of us impractical.

The question was, is this a valid method ..... to which I say, YES. Of course, you may personally choose whatever you like at the end of the day.

cheers
mark
 
diyAudio Senior Member
Joined 2002
THE LISTENING ENVIRONMENT.

Hi,

What I'd like to agree on, is that although two amps may seem to sound the same in a blind test to a person, they might not sound the same in subjective environment, the one we are accustomed to in our usual way of listening and this should be treated as "real" reality, more important than just aesthetics.

Very good point Peter.

It actually supports the psychoacoustic theory to a great extent in that in the case of a home comparison of two amps (or whatever) you actually have adapted over time to that particular environment.

The outcome of your judgement may have been different in an unknown environment, maybe not.

The key is, IMHO, that when attending a listening test in a room with unfamiliar acoustics to our brain things can get rather tricky.

This may well explain the odd outcome of blind listening testing when participating in said tests where you actually "know" from experience in your own familiar environment that product X or Y under test sounds quite different at your own listening place yet hardly distinguishable under unknown acoustical environment testing circumstances.

I feel the brain needs time to adapt to the acoustics of the new environment and this factor alone may be a great contributor to the " falsification" of the test.

Maybe there's no truth at all in what is said here either but it makes sense to me.;)

Cheers,:cool:
 
diyAudio Senior Member
Joined 2002
AH...AT LONG LAST.

Hi,

I believe NW made a fatal mistake when he infered certain people would or would not be able to hear differences between specific products. This had the very predictable result of reducing much of the thread content to "4 year old sand-pit level" argument.

Absolutely...and if you follow the evolvements you actually get remarks such as: But?...it seems that actually you do agree here???

Someone shot himself in the foot and now you've got an endless thread as the endresult.

Not that I'm not enjoyng it....:cool:
 
Which sounds better ....

So in effect you are saying that it sounds different if I can see which amplifier is being listened to.
__________________
Which one sounds better?

There is actually nothing wrong with this, so long as you know why you appreciate A more than B.

I built an Aleph4 and Aleph5. I "enjoy" the later more because I spent more time and care in it's manufacture, it looks better (to me) and I can convince myself it actually sounds better.

Could I tell them apart reliably blinded. Don't know, haven't tried, but I strongly suspect not.

Is this a problem and do I need to do blind testing ...... clearly NO!

OK, lets take another example .........

I start a company, "Perfect Pitch Pyramids. We manufacture audiophile grade pyramid shields for your hifi. There is a clear problem with the Earth's magnetic field causing polarisation of the electron rotational spin in the components of your stereo and resulting in half-wave harmonic distortion which degrades the signal and is clearly audible to the most discerning ears ......
For an additional consultancy fee we will install the pyramid, aligning it with the Earth's true magnetic axis in your location so you may obtain full benefit from your pyramid......
Comes in standard form or the delux model with the blue LED light bathing your stereo and removing the noise of the background ambient light.
Price on application ..... how much do you have??"

Now, we all know I could probably sell quite a few of these. :devily:

Should we (the audio community) subject this to blind testing, or should we believe the hifi mag reviewer, who I took out to very expensive lunch yesterday?

regards
mark
 
Re: Which sounds better ....

mefinnis said:

Should we (the audio community) subject this to blind testing, or should we believe the hifi mag reviewer, who I took out to very expensive lunch yesterday?


If after 5 years, you would still be in business selling pyramids, I would suspect there must be something to them and maybe even buy one (but of course on a used market).:)
 
P-P-Pyramid ....

If after 5 years, you would still be in business selling pyramids, I would suspect there must be something to them and maybe even buy one (but of course on a used market).

Nope, you still couldn't afford one and I'm certainly not putting the plans on the web for you DIY hackers ..... you would only do it incorrectly and sully the good name of my company ..... ;)

mark
 
Re: Re: Which sounds better ....

Peter Daniel said:


If after 5 years, you would still be in business selling pyramids, I would suspect there must be something to them and maybe even buy one (but of course on a used market).:)

Guys like Dennis Lee and Joe Newman have been peddling perpetual motion machines for decades? Does that mean there's something to them?

I think everybody knows the feeling of psychological tension and the pressure of making the right choice when the prize is big

Really? What's the prize?

I note that blind tasters can reliably tell you the difference between one wine and another on the Master of Wine and Master Sommelier tests. The prize is a big one- credentialing for a career. Why aren't these folks inhibited from distinguishing a Napa Valley Cabernet Sauvignon from a Pauillac?

In blind listening tests, trained listeners can distinguish changes of level of 0.1 dB, and eq differences of the same magnitude. How do these supermen manage that feat?
 
Eric,

I apologise for being so lateto reply, but I have read Bruce Candy's claims and your statements, and would like to add something to these august pronouncements.

The difference between a trumpet played soft and hard has been measured - forget where, read it in an acoustics book - at around 0.05% more odd order, chiefly 5th, 7th, and 9th.

Yet the human ear hears this quite distinctly, even subliminally; a phenomenon which requires elaborate measuring instruments to detect.

I mention this because I believe it gives clear indication of the inordinate sensitivity of the ear to high order distortion. You might suspect I'm playing into Bruce Candy's hands, but in fact, I draw a different conclusion.

We keep hearing this word 'musical' when applied to hifi equipment, and I suspect it's highly relevant here. Psychoacoustic measurements (Dr. Isabelle Peretz, Uni of Montreal, 2001-2) have shown that consonant and discordant chords are recognized with identical fluency by people of all cultures and races; so there would appear to be something universal here and thus the word has wide relevance. Odd order is discordant; even order, consonant. There lies the rub.

There are thus two ways of producing good hifi amplification. One, the most obvious, is reducing distortion to infinitesimal levels, and the other is minimizing odd order harmonic distortion. I would opine that both are legitimate methodologies....

Bruce Candy opts for vanishingly low distortion, in the hope that odd order distortion is at sufficiently low levels as to be indetectable. This is valid, and the purist approach you take.

Others, myself included, opt to minimize odd order distortion, realizing that distortion is, to some extent, inevitable, and we should try to skew it toward even orders (consonant) rather than odd orders (discordant). So, taking on the real world, it should be possible to design our amps/preamps to favor even rather than odd. Such an amp should sound good.

Finally, there's one further argument. We all realize that in 100 CDs, perhaps 20 are well recorded and, for want of a better word, 'musical'. This is just a fact of life; but why is this?

Could it be that microphones are like loudspeakers, and tend to show erratic reponse to the higher frequencies off axis? If the recording engineer gets it right, all the harmonic spectrum of the sound is captured, but if not, some of the higher orders are left out, and give a 'thin' sound, so characteristic of poorly recorded material? Of course, this is simplistic, there are many contributing causes, not least the lashes of cheap ICs in the recording console, but it's one reasonable argument. Ergo, the higher harmonic content is attenuated somewhat, leading to a different harmonic composition compared to the original. Of course, there are room effects, and so on, and the usual DSP, which can be quite destructive, particularly of imaging.

So, might not a reasonable approach be to design an amplifier with very little odd order harmonic distortion, and a little of second (and maybe a little of H3, which not as unmusical as H5, H7 and H9, which are progressively more objectionable)?

This added even order, maybe H2/H4, would reconstitute some of the lost even orders during recording, and add some flesh to an otherwise skeletal performance.

For vindication of this wacky theory, look no further than the singled ended triode amplifier, or the legendary Pass Single Ended amps. They add quite a bit of H2, and people love 'em.

Cheers,

Hugh
 
diyAudio Senior Member
Joined 2002
CD ERA IS DEFINETELY HERE....

Hi,

Could it be that microphones are like loudspeakers,

Yes, some are but with limited bandwidth and working in "reverse polarity" if you like.

So, might not a reasonable approach be to design an amplifier with very little odd order harmonic distortion, and a little of second (and maybe a little of H3, which not as unmusical as H5, H7 and H9, which are progressively more objectionable)?

Nope, let them sort the problem at the source, blimey!!!

They add quite a bit of H2, and people love 'em.

I'd love them too if I didn't know any better.

You do know better don't you?;)
 
Aha, but Frank,

I do know better, but I work in the real world. As a Brookly friend of mine hilariously said, 'We can't have the angel in Versace, let's settle for the ***** in Dior'.

We are human beings. Our world is dominated by compromise. Let's go for the amp with virtually unmeasurable odd order, and a tiny, inevitable smidgeon of even order. Even then it will still be truckloads better than the SET, with its low power, lousy TID, and crippling inefficiency.

Give Benny my warmest. Lovely guy, he's been good to me.

Faustian (or is that Dorian??), I would say that it is almost inevitable we foist harmonics on a (un)suspecting listener. This is recorded music, and most amps don't give 25 parts per billion of distortion (oh, what harmonic was that, Bruce??),

Cheers,

Hugh
 
diyAudio Senior Member
Joined 2002
NO OFFENSE INTENDED..

Hi,

Give Benny my warmest. Lovely guy, he's been good to me.

Will do and Benny and yours truly have some common history in audioland, nice chap he is indeed.:)

I do know better, but I work in the real world.

That's the message you should advertise, Hugh (IMHO at least), in the meantime I'll kick the SE's -sound engineers, that is- ***** 'till they grow a pair of ears.:bawling:

FWIW, better a little even order harmonic distortion than odd order distortion only.
With that golden recipe in mind semicondustors can be made to sound a little more pleasing to the ear, don't they?

Cheers,;)
 
Re: Re: Re: Which sounds better ....

SY said:


Guys like Dennis Lee and Joe Newman have been peddling perpetual motion machines for decades? Does that mean there's something to them?

At least it's worth mentioning here, in this thread.


SY said:

Really? What's the prize?

In blind listening tests, trained listeners can distinguish changes of level of 0.1 dB, and eq differences of the same magnitude. How do these supermen manage that feat?


Really? I had not trained listeners in mind and you should know that I'm speculating a bit, since I never tried a blind test. Don't feel really tempted to try one either.;)

The prize I've heard about was $10,000. Not a big amount by some standards, but quite a lot for a simple listening session.

http://www.diyaudio.com/forums/showthread.php?s=&threadid=7200&highlight=thousand+dollars+challenge
 
THE LISTENING ENVIRONMENT IS PARAMOUNT.

fdegrove said:
The key is, IMHO, that when attending a listening test in a room with unfamiliar acoustics to our brain things can get rather tricky.

This may well explain the odd outcome of blind listening testing when participating in said tests where you actually "know" from experience in your own familiar environment that product X or Y under test sounds quite different at your own listening place yet hardly distinguishable under unknown acoustical environment testing circumstances.

I feel the brain needs time to adapt to the acoustics of the new environment and this factor alone may be a great contributor to the " falsification" of the test.

Maybe there's no truth at all in what is said here either but it makes sense to me.;)

Cheers,:cool:

Hi Frank,
YES this is what I said along time ago in this thread.
Yes, the brain does need time to learn a new listening environment, and of course this is one of the variables in a BLT, and in my view this is what invalidates a typical short period BLT.

As the closeness of the amplifiers increases, then too does the listening time required to diffrentiate fine differences.

I think we have all had the experience where upon listening to a new amplifier, subtle differences in musical passages are noted on by one, even to the point of noticing little nuances and arrangements that were previously un-noticed.

Because FR and THD+N figures do not fully describe any amplifier, it cannot be said that same measuring amplifiers sound the same.

In the case of an amplifier such as the Halcro where amplifier contributions become vanishingly small, then it MAY be true that a Halcro amplifier would be indistinguishable from an equally measuring amplifier, but as none of us have been able to make such a comparison, this is conjecture only.

I still expect that such amplifiers would be distingishable due to very fine differences in characteristics due to individual component differences, such as capacitors and resistors.

Fine listening is not about detecting THD - it is about detecting fine differences in patterns, attack and decay characteristics.
With a well trained ear in a familiar environment, these parameters can be reliably detected.

Eric.
 
Aussies In Agreement...

Hi Hugh,
Thankyou for your long and considered reply, and as usual I am in complete agreement with you.

Yes, the Halcro is one aqpproach, and seemingly a very good one, albeit well out of the financial range of the most of us.
For this reason alone, this is not my realistic first choice approach.

The Stereophile review makes strong comment of all the test recordings being revealled as exactly as recorded, and this I interpret as virtually complete LACK of masking and colouration due to LACK of clearly audible distortion mechanisms.

I completely agree with you that if you cannot completely eliminate distortions, then it is pertinent to steer the harmonic products to those that are relatively benign or innocuous.

Indeed, in my experience an essentially distortionless amplifier can sound dead and lifeless, and the addition of some low order harmonics can put the 'life' back in, and it seems to me that this is the basis of the success of your amplifiers.
Addition of higher order odd harmonics can produce a sound that drives the listener out of the room.
These higher order odd harmonics also contribute to tinnitus and hearing loss conditions in my opinion.

Amplifiers with THD ratings of 0.01% are standard equipment nowadays, but of course the standard blanket THD figure says nothing about the levels of individual harmonics, and due to the extraordinary sensitivity of the ear to higher order harmonics and especially higher order odd harmonics, in my opinion it is to be fully expected that 0.01% THD amplifiers will be perfectly distinguishable because of differing higher order harmonic behaviours, and this is my experience.

Finally, there's one further argument. We all realize that in 100 CDs, perhaps 20 are well recorded and, for want of a better word, 'musical'. This is just a fact of life; but why is this?
Many musicians are out of tune both vocally and instrumentally to my ear.
If the recording engineers are similarly Absoloute Pitch deaf too, then we get the top 40 recordings that we get !.
Could it be that microphones are like loudspeakers, and tend to show erratic reponse to the higher frequencies off axis? If the recording engineer gets it right, all the harmonic spectrum of the sound is captured, but if not, some of the higher orders are left out, and give a 'thin' sound, so characteristic of poorly recorded material?
Yes microphones are typically very much erratic in response both on axis, and worse off-axis.
Most certainly this will affect the captured sound whether in a sonically dead recording booth, or a live room.
So, might not a reasonable approach be to design an amplifier with very little odd order harmonic distortion, and a little of second (and maybe a little of H3, which not as unmusical as H5, H7 and H9, which are progressively more objectionable)?
Yes I agree completely - if you cannot eliminate distortions, at least make them pleasant.
This I feel is the problem with some high NFB designs, as these can give low figures for low order harmonics at the expense of higher order generation.
This added even order, maybe H2/H4, would reconstitute some of the lost even orders during recording, and add some flesh to an otherwise skeletal performance.
Yes this would be a better tone control - variable relationships of H2 and H4 would even better I expect.
For vindication of this wacky theory, look no further than the singled ended triode amplifier, or the legendary Pass Single Ended amps. They add quite a bit of H2, and people love 'em.
Not wacky, just sensible, and this correlates with theory.
Too much distortion is too much distortion in my books, and I find the Pass amps that I have heard (Aleph 0) to be coloured, but very pleasant, and both at the same time.

I am currently working on an interesting product that changes harmonic relationships very pleasantly, and could indeed be described as a harmonic tone control.

In the current absence of suitably resolving test equipment gear to confirm, to my ear odd order and higher order distortion harmonics are much reduced, and this gives a sonically much improved performance, with musical details and nuances more clearly audible, and silences more aparrent, and on the whole quite fantastically improved musical results from all pro-audio and home audio equipment trialled so far.
To all who have heard this technique, the response has been of universal acclaim, and all who have heard it desire it.
Also interestingly using this technique, it is possible to listen at insane SPL's, and NO tinnitus or aural irritations or fatigue conditions result.

Hugh, you are on the right track in my view.

Eric.
 
Status
Not open for further replies.