I believe the "challenge" I had in mind is open to any individual, regardless of industry association either professional or hobbyist. Of course, that is in the US... not sure what the Netherlands offer. 🙂keyser said:
MBK said:
Few people would argue that all amps measure the same. Those that do are simply not familiar with the subject. But that isn't the question... we all know that there will be variations in FR, distortion spectrum, noise, clipping behavior, etc. The question is at what magnitude those variations become audible.
If power output and noise were also similar, and both operated in the linear range, and that singular THD number was low enough (not atypically low but just low enough), of course they would sound the same.
As you point out, the best DBT will spread trials randomly over time, as that minimizes the impact of time-dependent variables such as environmental noise (traffic, etc.), mood and awareness of the test subject, and any possible changes in local power supplies (assuming such changes were audible). Don't be fooled however into the thinking that a long time spent getting "familiar" with the component in uncontrolled conditions will reveal things to you that a quick A/B blinded cannot. While it is true that the subject ideally should be very familiar with the characteristics (both having listened to at length and knowledgable about relevant objective metrics) of both A and B, it is fact that our hearing is most sensitive to rapid changes.
If you can't hear a difference in rapid A/B switching blinded, then you stand absolutely no chance of hearing differences with time delayed, casual usage, extended blinded listening sessions (notice the importance of remaining blinded).
Thank you RH for joining this thread! Thjngs are getting much more interesting, because you have been there where I may someday be.
Hope KYW has some good answers, 'cause (almost) all amplifiers sounding the same sure sounds dull, and boring to me!
This weekend I'll do a few Blind Tests, just for me. I'm goin to compare my NAD 216 (guess it's about $700(?)) to an el cheapo JB Systems AX-400. They almost give this amp away for free! (150 EURO
's buys you 135 Watt @ 8 ohm, 200 Watt @ 4 ohm) I'll post my findings on this thread, whether anyone wants to know (or might find it not proving anything) or not!
Hope KYW has some good answers, 'cause (almost) all amplifiers sounding the same sure sounds dull, and boring to me!
This weekend I'll do a few Blind Tests, just for me. I'm goin to compare my NAD 216 (guess it's about $700(?)) to an el cheapo JB Systems AX-400. They almost give this amp away for free! (150 EURO
's buys you 135 Watt @ 8 ohm, 200 Watt @ 4 ohm) I'll post my findings on this thread, whether anyone wants to know (or might find it not proving anything) or not!
Well... I do have my doubts as to how severely I could personnally hear the differences between standard amps. But I see no good reason why they should not sound different. I also see no reason why THD - beyond a certain threshold, likely in the 0.1-0.3% range, should make a good predictor of sound quality. To quote KYW:
But you say:
Sure. And all dishes with the same ingredients taste the same. And all wines from the same grapes taste the same, as SY would say. Not.
Have a look at the Geddes site , read the two presentations he has online there, and listen to the files. He compares audibility correlation of standard THD with audibility correlation of a different parameter he developed. And surely the Geddes metric also just describes a part of the whole.
A single number - THD or any other parameter - describing a complex system completely? That's soooo 19th Century, as a philosophy of science!
Have you never had the experience of that dripping faucet that you could barely notice when falling asleep, yet that later hammered you awake at night?
In my experience, yes, I know very quickly if something sounds "right" or "wrong", but in a second or two, I adapt and start to believe that I can't hear a change. And from then on it often takes me a long time before I come back to my initial gut feeling. I see that as a particular problem of testing - the necessary conscious attention, and pressure, to make a rational decision based upon subtle gut feeling.
But you say:
Sure. And all dishes with the same ingredients taste the same. And all wines from the same grapes taste the same, as SY would say. Not.
Have a look at the Geddes site , read the two presentations he has online there, and listen to the files. He compares audibility correlation of standard THD with audibility correlation of a different parameter he developed. And surely the Geddes metric also just describes a part of the whole.
A single number - THD or any other parameter - describing a complex system completely? That's soooo 19th Century, as a philosophy of science!
Have you never had the experience of that dripping faucet that you could barely notice when falling asleep, yet that later hammered you awake at night?
In my experience, yes, I know very quickly if something sounds "right" or "wrong", but in a second or two, I adapt and start to believe that I can't hear a change. And from then on it often takes me a long time before I come back to my initial gut feeling. I see that as a particular problem of testing - the necessary conscious attention, and pressure, to make a rational decision based upon subtle gut feeling.
One has to be careful about generalizing the Geddes stuff- there's a difference between what numbers you can get with artificially generated distortion and test signals versus what actual physical amplifiers can do. IIRC, Geddes was pretty clear about this in his paper.
I think it was Heyser who built a box that measured well by a restricted set of bandwidth, noise, and THD, but was totally unlistenable. It was a very clever trick.
I think it was Heyser who built a box that measured well by a restricted set of bandwidth, noise, and THD, but was totally unlistenable. It was a very clever trick.
SY said:
and another trick -- by Carver -- was to "black box" any amplifier by analyzing the transfer function and replicating it. must've been sometime in the 1970's back when I was using an Apple II+.
if you really want to focus on building a good hi-fi system, spend time listening to live music, particularly piano (not electric piano btw.)
Sure, how much does it matter in real life, here we have a different question.
But Geddes and others invalidated the proposition "same THD = same sound" a long time ago *on grounds of principle*. If same THD does clearly *not* imply same sound, even in a single experimental case, we must reject at least *this* hypothesis. 😉
But Geddes and others invalidated the proposition "same THD = same sound" a long time ago *on grounds of principle*. If same THD does clearly *not* imply same sound, even in a single experimental case, we must reject at least *this* hypothesis. 😉
MBK said:
Don't get overzelaous about what I claimed. I also icluded other qualifications about power, operating in the linear range, noise level, etc. I should also have included frequency response similarities.
The important point is that if the THD is low enough, differences become audible. I'm well aware that THD is a woefully inadequate spec on its own if you really want to know about the objective performance of an amplifier. I am also well aware that for modest THD levels it is possible that higher harmonics contributing significantly to that total could result in audible differences. However, if the THD is low enough that all relevant harmonic contributions are below the level of established audibility (as is the case when the THD is sufficiently below the level established as audible for the most harmonic humans are most sensitive to), then I can claim identicality between all such amps (and I note that such levels are easily accomplished by decent designs).
So don't read more into what I said than I have actually said. I would never claim that two amps measuring 1% THD would sound identical (though it is very possible that they would), because our sensitivity to some harmonics falls well below 1%.
a friend and I did the blind test yesterday. We compared my NAD 216 to my cheap JB Systems AX400 (pro amp) in a single blind abx style test.
I calibrated the levels by ear. Probably not within 0,1 dB, but it was close!~First we listened to both amps sighted. I actually thought the sounded more dynamic and had a more authoritive bass. Then one of us listened to an amp not knowing which played. We both did 10 tests. I was right 4 times out of 10, my friend 7 out of 10.
I don't think I could hear the difference, neither does my friend (even though he was right 70%). Even if my friend did hear it, and say I would have heard it if we did 100 tests, the difference can not have been much. Certainly no reason to upgrade to a better one, what means paying a few hundred dollars.
My next upgrade dollars will go to the speakers, I'll leave amps be!
I calibrated the levels by ear. Probably not within 0,1 dB, but it was close!~First we listened to both amps sighted. I actually thought the sounded more dynamic and had a more authoritive bass. Then one of us listened to an amp not knowing which played. We both did 10 tests. I was right 4 times out of 10, my friend 7 out of 10.
I don't think I could hear the difference, neither does my friend (even though he was right 70%). Even if my friend did hear it, and say I would have heard it if we did 100 tests, the difference can not have been much. Certainly no reason to upgrade to a better one, what means paying a few hundred dollars.
My next upgrade dollars will go to the speakers, I'll leave amps be!
Konnichiwa,
How close? Close enought to ensure that the level differences cause no issues?
Clearly your results show some interesting trends, but the dataset is by far to small to obtain any reasonable certainty of anything, ignoring even the fundamental flaws in your test setup.
Given that you have been given more than sufficient access to material on the methodology, limits and risks in DB Testing I find this at the very least surprising.
Why do you insist on making tests that are meaningless and I mean ENTIERLY MEANINGLESS? Do you enjoy wasting your time? Or are you merely interested in obtaining with better than fair reliability results you claim support your personal prejudices even if in fact they support NOTHING, as they are incapable of providing any useful data?
Sayonara
keyser said:
How close? Close enought to ensure that the level differences cause no issues?
keyser said:
Clearly your results show some interesting trends, but the dataset is by far to small to obtain any reasonable certainty of anything, ignoring even the fundamental flaws in your test setup.
Given that you have been given more than sufficient access to material on the methodology, limits and risks in DB Testing I find this at the very least surprising.
Why do you insist on making tests that are meaningless and I mean ENTIERLY MEANINGLESS? Do you enjoy wasting your time? Or are you merely interested in obtaining with better than fair reliability results you claim support your personal prejudices even if in fact they support NOTHING, as they are incapable of providing any useful data?
Sayonara
Kuei Yang Wang said:
Are you a talking about poorly executed "controlled" testing, or the uncontrolled casual sighted tests most audiophiles defend? While your questions fit either, the penetrating wording sounds eerily like you are talking about the latter.
For the level matching I would have used a multimeter and a test frequency in the low 100's where most meters give good results. Accurate level matching makes a huge difference for the credibility of results.
In any case, Keyser now found a basis for his own personal decision making, and if he's happy with it, why not? As long as no one extrapolates this as an absolute truth about the two amps I don't see a problem.
Upgrading the speakers certainly won't hurt, and who knows, after that, another test may yield different results 😉
In any case, Keyser now found a basis for his own personal decision making, and if he's happy with it, why not? As long as no one extrapolates this as an absolute truth about the two amps I don't see a problem.
Upgrading the speakers certainly won't hurt, and who knows, after that, another test may yield different results 😉
keyser, your result is no different than any other ever conducted. All anyone has to do that claims he can tell the difference between, say, tubes and transistors is to accept James Randi's paranormal challenge and take those million dollars. Will never happen of course. Just like my Volvo is faster than any Ferrari except when somebody else is looking.
I really don't have a Volvo. And I would never claim I paid $1,000 for a Versace trench coat and a Garrard 301 because they are better than some other trench coat or turntable. I bought them because I wanted them. And should I ever get an amplifier other than a Gainclone I wouldn't do it because I think it's better, but because I want one for whatever reason. I'm too old for fairytales. Nowadays I care only for how things really are. Everything else is just a waste of time.
I really don't have a Volvo. And I would never claim I paid $1,000 for a Versace trench coat and a Garrard 301 because they are better than some other trench coat or turntable. I bought them because I wanted them. And should I ever get an amplifier other than a Gainclone I wouldn't do it because I think it's better, but because I want one for whatever reason. I'm too old for fairytales. Nowadays I care only for how things really are. Everything else is just a waste of time.
Hi all,
I am new on this forum. I think this is a weird thread. I am sure there are audible differences between components. KYW is absolutely right. A sighted test is reliable enough in my experience. You can listen in a more relaxed way if you know the components under test.
I am new on this forum. I think this is a weird thread. I am sure there are audible differences between components. KYW is absolutely right. A sighted test is reliable enough in my experience. You can listen in a more relaxed way if you know the components under test.
- Status
- Not open for further replies.