Probably just an illusion of the nonlinearity that produces the harmonic distortion. With music it will produce some HD and IMD, which at low levels can have the effect of sounding a little bright above the midrange. A clue would be that EQ doesn't really fix it. If due to such distortion, the midrange may be perceived as a little weak.
Last edited:
Imo it is a bit unfair to play the language card at this point, but who am i to spoil the party......
Of course it is an appealing idea; whenever the numbers don´t favour your beliefs or in case of limited understanding, just invoke the "spirit" of statistics, preferable (i guess) in a self-elaborated version?!
In fact, using the term "indistinguishability test results" for negative results violates exact that spirit of statistics (or the spirit of good experimental design), accordingly the next misguided conclusions follows as it is actually the other way round.
Doing _good_ listening experiments would lead to results that (be it positive or negative) truly add to the body of evidence, because it would be much more justified to draw further conclusions from it.
To believe that the results of differently (but seriously) flawed experiments can be wondrously transformed into "scientific gold" if only "enough" of different experiments are done, does not reflect the "spirit of statistics".....
@DPH,
Imo there is no offense taken in the dialog with billshurv.
Could be the effect size - although i really miss any (even implicit) mentioning of it, but i´m wondering that in case of "super obvious or night and day" i _must_ take that literally while being simultaneously attacked/critized for using too literal interpretations......
Stating something is super obvious is often an exaggeration, but stating the impossibility of a perceptual effect (because otherwise laws of nature would be violated) is apparently also often an exaggeration, so i´d also really appreciate somewhat more restraint using such terms (in both cases)
+1
Imo there is no offense taken in the dialog with billshurv.
Could be the effect size - although i really miss any (even implicit) mentioning of it, but i´m wondering that in case of "super obvious or night and day" i _must_ take that literally while being simultaneously attacked/critized for using too literal interpretations......
Stating something is super obvious is often an exaggeration, but stating the impossibility of a perceptual effect (because otherwise laws of nature would be violated) is apparently also often an exaggeration, so i´d also really appreciate somewhat more restraint using such terms (in both cases)
+1
One of the properties of single ended tube circuits (or single ended class A circuits in general) is that the distortion vanishes at low level, the Pass forum is full of these plots. From this description (by a recording engineer) in that respect they are similar.
So does "midband suck out" mean there is not enough distortion vs level or what? Is designing amplifiers as instruments with no deliberate input to output aberrations a waste of time?
I don't care about your sweeps. I have had many circuits measure ruler flat yet had all sorts of emphasis/de-emphasis spectral personalities, especially with tubes. I would need to know what caps, resistors and tube brand you used on this SRPP to have a reasonable shot at finding the culprit of what I heard. I asked about the output coupling cap before. No answer yet.
All good but at what amount of such distortion?
And who is raising his concerns here please? As Scott mentioned, vinyl source has much distortion to offer. Doesn’t this ‘suck the midrange out’?
PS1 How much of the vinyl distortion is already embedded in the groove (cut process) and how much is due to playback is a gray area of course.
PS 2 Attached FFTs from playing back on a certain TT set the “The Ultimate Analogue Test LP” from Analogue Productions 2006, Side 1, tracks 1, 4, 9
George
Attachments
Reconciliation in this case meaning not as much about conflict and offense but trying to tie together what you're saying and what Bill is saying as talking about two different things, even if outwardly they're in contention.
I.e. separated by common language and talking past one another. My understanding is that it's not outright belief that these "fun little tests" are rigorous, but simply that if some of the claims made by members would show up very clearly. Subtle effects, as you fairly object, would absolutely *not* show up confidently in these "fun little tests". Hence my comment about effect size.Agreed. I know that a lot of the objections by more incredulous folks (ahem), come from the exaggerated language, e.g. your "I changed my wire and I was screaming out of the room as it was so painful to me" (which, depending on the wire, might be from a new and fantastic oscillation...), but no one is free from that sin of language.
Can't say specifically. Its just that if Morinix said he heard something in the tube processed sound that wasn't in the straight wire, I would not necessarily jump to the conclusion that the listener's imagination is the only or most likely possible cause. It could possible there was some real audible perceptual effect that some listeners might notice. I have seen something like that before from a small-ish nonlinearity, is all. I don't see a reason to make assumptions about what it was in this case either way. If we really want to know which we might have to do some more work to find out. Do we really care all that much?
Last edited:
Are there designers/manufacturers who don't ever listen during development ?.
Ok, so what's your take ?.
Having a decent reference is useful, even mandatory as reality check.
Same here, of course we have all heard items/systems that measure good/sound bad, measuring 'bad' also always sounds 'bad', it is the degree and nature that varies and this can still be pleasing to an extent.
Dan.
It seems that the opposite should also be true - it only matters when people post statements about 'fun' listening tests & claim "no audible difference" rather than "I prefer to believe there is no difference" ?
Well, you know how ABX software and ABX boxes somehow reduce differences? (I think you may have described the effect in a fairly recent post.)
That sound effect you hear is exactly the loss of the sighted listening illusion. That's what it sounds like to lose it. Sorry to be the bearer of bad news.
What's left is what is the real audible sound difference, and it is much weaker and harder to tell apart than if sighted. That's okay though, as far as I am concerned. I would rather know the real difficulty than not know it.
Note: Although ABX happens to be mentioned above, it is not the most sensitive DBT protocol, so I happen to strongly prefer other DBT methods for discriminating small differences.
OK have a few mins whilst I rip some CDs. I know I am only opening myself up to being scored D- again, but hey ho
Let's be clear up front. I am not claiming any of these tests are setup with any rigour or review. They are done by DIYers (some who wear pro hats during the day) so that people can have a go and discuss things. Many have been done, with varying degrees of success, but never a large enough sample size for any hard conclusions. However egos have been bruised, cheats have been found (I mean cheating on a DIY listening test, seriously?) and the same argument comes around again, as if we are actually in hell pushing the boulder up the hill each day. But let me try and restate my somewhat flippant post in a different way and see if Jakob gets where I am coming from.
Whenever these sort of tests occur there are a number of types of people who take the test or comment on the test. They mostly break down into the following groupings
1. Those who believe they can hear the difference in almost anything.
2. Those who believe a difference should not be descernable
3. Those who take the test with an open mind
4. Those who feel they have too much to lose if they take the test
5 Those who really don't care as there is music to be listened to
6 Jakob and Ed (only kidding)
Note that groups 2 and 3 vary depending on the test. People of a more techical bent may change their preconceptions if they consider something to be clearly unlikely to be audible. human nature and all that. But group 3 is generally the largest and that is good.
But (back in hell with the boulder) each time this is done, one or more people step up with bold claims and chose the wrong answer (by whichever protocol they are using). Things then degenerate into the same old boring circular arguments with no progress. That was my (badly made) point. Group 1s will never go 'blimey that's unexpected'. Their beliefs in their abilities are so strong they have to find excuses and use in some cases many deflecting tactics to not face up to the (possibly) most obvious conclusion. Really I am just stating how the cycle goes, not consciously making a comment on the validity or statistical significance of the test itself.
To your specific questions
. But my personal view is that if someone doesn't believe that, with the information they have then can make an ears only judgement, then they shouldn't take the test. By that I mean that there are no glaring drivers of a bias, not that all bias is removed. I am not interested in if that reduces the value of the test as for me its entertainment only. I am not trying to get Dr Toole to sign off on the protocol. Neither am I expecting statistically valid data. I do wish for a day when the boulder does not need pushing up the hill though.
Let's be clear up front. I am not claiming any of these tests are setup with any rigour or review. They are done by DIYers (some who wear pro hats during the day) so that people can have a go and discuss things. Many have been done, with varying degrees of success, but never a large enough sample size for any hard conclusions. However egos have been bruised, cheats have been found (I mean cheating on a DIY listening test, seriously?) and the same argument comes around again, as if we are actually in hell pushing the boulder up the hill each day. But let me try and restate my somewhat flippant post in a different way and see if Jakob gets where I am coming from.
Whenever these sort of tests occur there are a number of types of people who take the test or comment on the test. They mostly break down into the following groupings
1. Those who believe they can hear the difference in almost anything.
2. Those who believe a difference should not be descernable
3. Those who take the test with an open mind
4. Those who feel they have too much to lose if they take the test
5 Those who really don't care as there is music to be listened to
6 Jakob and Ed (only kidding)
Note that groups 2 and 3 vary depending on the test. People of a more techical bent may change their preconceptions if they consider something to be clearly unlikely to be audible. human nature and all that. But group 3 is generally the largest and that is good.
But (back in hell with the boulder) each time this is done, one or more people step up with bold claims and chose the wrong answer (by whichever protocol they are using). Things then degenerate into the same old boring circular arguments with no progress. That was my (badly made) point. Group 1s will never go 'blimey that's unexpected'. Their beliefs in their abilities are so strong they have to find excuses and use in some cases many deflecting tactics to not face up to the (possibly) most obvious conclusion. Really I am just stating how the cycle goes, not consciously making a comment on the validity or statistical significance of the test itself.
To your specific questions
It's simply the 'person claims they can hear below the brownian motion in the room' and yet generally they will not do an ABX, preferring to name which file is which. And many times getting it wrong, which then cues a load of excuses. Note I fully accept that FoobarABX may not be the preferred protocol for many, but it is the only common thing we have that (in the latest version) precludes cheating. If the differences were as obvious as some say, then they would be able to get 100% on the test. One could argue that this makes the tool only useful for dividing between 'obvious to all' and 'not obvious' and I'm not going to enter into that swamp today. Those who cheat with Foobar have been weeded out with the newer version of ABX. We also have the subset who post after the reveal to say 'it was easy' with no evidence.
IMO doesn't matter but I wonder why anyone would do it other than for fun? People in group 3 approach this with the right frame of mind, enjoy taking the test and are often intrigued/amuzed/enlightened with the output. If someone isn't doing it for fun, why are they doing it? This is after all a hobbyist forum for people persuing a hobby. Things like the Sousa test, the potato test and the opamp buffer test have been fun and educational, but without adding anything significant to the science of pyschoacoustics.
Ah, now this is where I could accuse you of twisting words, but I will accept I was terse in my initial post and put it open to interpretation. These are not controlled listening tests. As mentioned the nearest we get to a control is the Foobar ABX plugin which group 1s don't like if they can't game it. The file peekers used to be the worst as they would post reveals and ruin everything. Luckily they behave these days. This forum can be like kindergarden at times
. But my personal view is that if someone doesn't believe that, with the information they have then can make an ears only judgement, then they shouldn't take the test. By that I mean that there are no glaring drivers of a bias, not that all bias is removed. I am not interested in if that reduces the value of the test as for me its entertainment only. I am not trying to get Dr Toole to sign off on the protocol. Neither am I expecting statistically valid data. I do wish for a day when the boulder does not need pushing up the hill though. This may be my ignorance, but to say 'here is a test, the purpose of which is to see if you can determine a straight wire from a tube buffer in ABX'. is actually a valid goal?
But I tend to listen to music. Masking is there. Is it not valid to ponder if, with real music a buffer with 0.5% low order THD is in fact transparent?
Aside (and honest question). For concert halls where reinforcement is added what is the additional THD that the sound system adds. I know that at covent garden it really messes up the imaging from the cheap sets, but cant say I can hear its amplified.
Typical LP setups at even the highest level, are not even remotely flat and have horrifying "spectral personalities" in any way you look at them. We're back to "obvious to anyone" differences do not show up with any measurement, we go full circle and circle the wagons yet again.
Last edited:
- Status
- Not open for further replies.