Re: Diffmaker
My comments on the Diffmaker approach (subtraction method), and those come from experience:
Completely useless with unsynced record/playback. That is, due to inevitable clock drift (and let alone speed drift of analog machinery) even an in other regards exactly identical playback will have a difference. For example, mostly I couldn't even overlay one single second of playback from the same CD-Player, two consecutive takes of the same piece of music. The residual (a very "phasey" one) from the clock drift would easily mask any subtle real difference.
And even with synced record/playback (that is, a dual soundcard-OUT-->DUT-->soundcard-IN chain, testing two different DUTs at a time) you will have the problem that even a miniscule magnitude/phase difference, say 0.1deg from the soundcard alone can mask any difference. And when in fact you would have a perfect channel match in the sound card (like what is assumed -- but has actually to be verified beforehand -- when you use sequential measurements with the same channel, only swapping DUTs), minscule sub-degree phase shifts (which can with all reason to be assumed as inaudible) will cause enough residual to mask what actually happens. The amplitude difference can be trimmed out to a certain degree *after* recording, but to trim phases is very hard. Therefore you need to trim out the phases *before* recording, which in the end requires an analog subtraction frontend allowing real-time trimming unless you're willing to spend hours for record/test/adjust cycles. Further, time domain averaging -- over many runs, preferably hundredths -- seems to be necessary, to i) improve signal-to-noise ratio in general and ii) to average out uncontrolled and uncorelated (w.r.t. stimulus) "spurious "events.
All in all the wave subtraction method has to be used with caution to not misinterpret false differences, especially from drifting clock rates (drifting playback and record "speeds") and from smallest but audibly completely irrelevant amplitude and especially phase differences. If all circumstances are well controlled and you have successfully verified that your difference really nulls if it is a true null and if you know your noise levels and resolution limits, then it is a very powerful tool. Otherwise it can be very much misleading, giving false positive results.
- Klaus
Salas,salas said:
My comments on the Diffmaker approach (subtraction method), and those come from experience:
Completely useless with unsynced record/playback. That is, due to inevitable clock drift (and let alone speed drift of analog machinery) even an in other regards exactly identical playback will have a difference. For example, mostly I couldn't even overlay one single second of playback from the same CD-Player, two consecutive takes of the same piece of music. The residual (a very "phasey" one) from the clock drift would easily mask any subtle real difference.
And even with synced record/playback (that is, a dual soundcard-OUT-->DUT-->soundcard-IN chain, testing two different DUTs at a time) you will have the problem that even a miniscule magnitude/phase difference, say 0.1deg from the soundcard alone can mask any difference. And when in fact you would have a perfect channel match in the sound card (like what is assumed -- but has actually to be verified beforehand -- when you use sequential measurements with the same channel, only swapping DUTs), minscule sub-degree phase shifts (which can with all reason to be assumed as inaudible) will cause enough residual to mask what actually happens. The amplitude difference can be trimmed out to a certain degree *after* recording, but to trim phases is very hard. Therefore you need to trim out the phases *before* recording, which in the end requires an analog subtraction frontend allowing real-time trimming unless you're willing to spend hours for record/test/adjust cycles. Further, time domain averaging -- over many runs, preferably hundredths -- seems to be necessary, to i) improve signal-to-noise ratio in general and ii) to average out uncontrolled and uncorelated (w.r.t. stimulus) "spurious "events.
All in all the wave subtraction method has to be used with caution to not misinterpret false differences, especially from drifting clock rates (drifting playback and record "speeds") and from smallest but audibly completely irrelevant amplitude and especially phase differences. If all circumstances are well controlled and you have successfully verified that your difference really nulls if it is a true null and if you know your noise levels and resolution limits, then it is a very powerful tool. Otherwise it can be very much misleading, giving false positive results.
- Klaus
Too bad that Liberty did not make it right. It could have really helped this discussion if we could have a common reference tool and exchange resulting files. Thanks.
fredex said:
Get you all subjective? Now why will I try to do that? 😀 😀
Some think that being subjective is like believing in magic and fairy tales, I also believe in measurements and explanations for everything, no magic, but after years of comparative tests I've learned that we can hear things that is difficult (impossible with my equipment) to measure. That's the reason for trying to understand how the brain analyse sounds, perhaps someone clever enough can come up with a good idea to do more worthwhile measurements. (I'm not saying FR and THD etc. is unimportant but I think it is not the only things that matter.)
André
SY said:
but there is also no reliable evidence to the contrary .....
which is where all the fun begins 😀
Andy, please repeat after me, three times:
It is logically impossible to prove a negative.
It is logically impossible to prove a negative.
It is logically impossible to prove a negative.
There. Do you feel more, I dunno, enlightened?
One cannot prove that there is no Santa Claus. One cannot prove that I don't have a dragon in my garage. One cannot prove that there is no tooth fairy. One cannot prove that freezing photographs of yourself will not improve sound.
It is logically impossible to prove a negative.
It is logically impossible to prove a negative.
It is logically impossible to prove a negative.
There. Do you feel more, I dunno, enlightened?
One cannot prove that there is no Santa Claus. One cannot prove that I don't have a dragon in my garage. One cannot prove that there is no tooth fairy. One cannot prove that freezing photographs of yourself will not improve sound.
Re: A Question Then....
That's easy! It was an uncontrolled audition. It has zero validity.
It is nothing to do with trusting the ears. The mind has the imagination.
thetubeguy1954 said:
That's easy! It was an uncontrolled audition. It has zero validity.
It is nothing to do with trusting the ears. The mind has the imagination.
Re: Re: Re: Re: Re: It's Obvious "IF" You're Open-Minded
AG, what is your proposed auditioning methodology, given that sighted listening is a joke?
Andy Graddon said:
AG, what is your proposed auditioning methodology, given that sighted listening is a joke?
Re: It's Obvious "IF" You're Open-Minded
very sensible indeed. I wish I had put it so succinctly
fredex said:
very sensible indeed. I wish I had put it so succinctly
SY said:
Hey, I'm on your side !! I agree with this completely.
SY. I think maybe you misinterpretted what I meant.
Re: It's Obvious "IF" You're Open-Minded
There are no non-listening objectivists. There are "objectivists" (who know what they are listening to) and "subjectivists" (who don't).
Andy Graddon said:
There are no non-listening objectivists. There are "objectivists" (who know what they are listening to) and "subjectivists" (who don't).
Re: Re: It's Obvious "IF" You're Open-Minded
You know, you are right.
Objectivists know what they are listening to and subjectivists know what they are listening for. 😀
tnargs said:
You know, you are right.
Objectivists know what they are listening to and subjectivists know what they are listening for. 😀
Andy Graddon said:
Andy, you believe SY when he says that it is logically impossible to prove a negative, yet you keep insisting everyone provides proof of negatives. There is plenty of evidence of the limitations of the ear, but you want people to provide proof that the evidence can't be wrong?
The whole point is to accept scientifically controlled experiments that have been validated by repetition, until scientifically controlled evidence is put forward that invalidates it.
But taking this approach to audio means "subjectivists" will have to abandon all the experiential "knowledge" they have built up over twenty, thirty, or forty years of uncontrolled listening. Only those with courage will do it.
tnargs said:
What you say is that science know everything? Sorry science know only as much as was discovered and much of that is still a mystery also.
What you suggest would lead to no further dicoveries or knowledge.
@ André
I just got the impression that subjectivists were a tricky bunch and I should be cautious. But your reply has convinced me otherwise.
I also believe in measurements, what's not to believe? Your listening experiences are probably similar to mine, Ive heard cables etc. In fact if you really listen for differences you hear them, it is quite magical. And of course we are all under pressure to hear things from the HiFi industry. Who wants to admit they can't hear a difference, it can only mean your ears are below par or your system is "not revealing enough". The second option can easily be fixed by spending more money. This means listening to more gear and now you are under real pressure to hear differences. It is your money on the line you don't want to choose bad sounding cables or amp.
Then you do a real blind test in your home and you discover that you actually can't tell by listening (which is the subjectivist's only truth) which gear is playing. Just try it.
There is nothing magical about any of our audio equipment. The magic that subjectivists don't believe in is between the ears.
Cheers all
I just got the impression that subjectivists were a tricky bunch and I should be cautious. But your reply has convinced me otherwise.
I also believe in measurements, what's not to believe? Your listening experiences are probably similar to mine, Ive heard cables etc. In fact if you really listen for differences you hear them, it is quite magical. And of course we are all under pressure to hear things from the HiFi industry. Who wants to admit they can't hear a difference, it can only mean your ears are below par or your system is "not revealing enough". The second option can easily be fixed by spending more money. This means listening to more gear and now you are under real pressure to hear differences. It is your money on the line you don't want to choose bad sounding cables or amp.
Then you do a real blind test in your home and you discover that you actually can't tell by listening (which is the subjectivist's only truth) which gear is playing. Just try it.
There is nothing magical about any of our audio equipment. The magic that subjectivists don't believe in is between the ears.
Cheers all
Andre Visser said:
More strawman. Science doesn't know everything, but it certainly has to know something. So tell us Andre, what science are you applying when you design and engineer "goodness" into your amplifiers and electronics? How are the "good cable" scientists designing and engineering "goodness" into cables? Are they far ahead of JNeutron with their testing methods in their deep underground Audio laboratories, where all these advances in "unknown" science is being made?
Andy, you seemed to have missed tnargs question. What is your proposed audio methodology? You are obviously using it. Please share it with us.
cheers,
AJ
Liberty *did* make it right. The problem is in the concept, in that it fails in many cases from systematic errors. It's up to the user to apply it under conditions that work, for that concept. It's only that most typical conditions don't work.salas said:
For the same reasons ideas like making microphone recordings of speakers, changing speaker cables, will not work well. Even if you take just the electrical signal at the speaker's binding posts one will have the described problems.
- Klaus
Andy doesn't know or doesn't want to expose his hearing data, he doesn't have any of his own suggestions except that all proposed test are not good enough. He's also trying to suggest tasks that are physically impossible or irrelevant to the subject.
Let me clarify that.
Measuring the response at the listening position is not a test simply because there's no standard or baseline that everyone can abide to. Every room or speakers system-room interaction is different and while this measurement may have a value for a particular user in particular environment it has no value as a general standard and just creates more ripples in the pond.
Near Field measurement of a driver with calibrated microphone can be done by anyone having software and hardware available to them. Being a Speaker DIY forum, plenty of people have that and can repeat their own measurements.
Let me clarify that.
Measuring the response at the listening position is not a test simply because there's no standard or baseline that everyone can abide to. Every room or speakers system-room interaction is different and while this measurement may have a value for a particular user in particular environment it has no value as a general standard and just creates more ripples in the pond.
Near Field measurement of a driver with calibrated microphone can be done by anyone having software and hardware available to them. Being a Speaker DIY forum, plenty of people have that and can repeat their own measurements.
Your personal feeling on this matter are absolutely correct and there is even hard evidence for it.
Maybe with the exception that sometimes it is not a subtle phenomen but a gross difference that remains undetected.
In another thread i´ve mentioned the the Locanthi/AES/Swedish radio - instance, Robert Harley wrote in the stereophile long time ago.
He repeats here:
forums.avguide.com/viewtopic.php?t=3914
I can´t follow his conclusions about blind testing in general, because these phenomenons are not restricted to blind tests. In fact it can happen in every test situation, be it sighted, double blind or whatever else.
And as we all know, it can happen in visual tests too.
This effect is named ´inanttentional blindness´ , some material on it can be found here:
http://viscog.beckman.uiuc.edu/djs_lab/demos.html
Afair modern neurobiology science has confirmed that test persons are leaving the socalled mental state of `awareness´ if concentrating to much on a specific task.
To detect unknown things this state of awareness is most efficient.
That´s why training is needed in reagard to listening tests.
On one side you have to train to stay in the state of awareness, on the other hand, if you want to confirm a known effect, you have to train to detect it (reliable) under unknown circumstances.
But you will only know, if you include controls in your tests. 🙂
First off shame on you for trying to debunk my statement when I clearly said it was NOT true, please go back an reread my post if you are curious why. And "easily" set up a test where people experience the outcome you think is true, maybe you didn't word this statement right but it sounds like you are talking about a rigged test.SY said:
Second, it seems like you believe that cables make a difference, in which case wouldn't it be rude to just ignore the claims that are being made? And why should we ignore subjectivist claims, because they can't substantiate them? I keep hearing people say open your mind and learn something, from both sides of the argument. Shouldn't that include you guys? You all have no evidence that there is a non measurable yet audible parameter, yet you use it to invalidate all science.
You think objectivists need to learn something, fine teach us something that you can support as factual.
- Status
- Not open for further replies.