Like the concept of second and third order intercept in RF the harmonic structure grows at different rates for different harmonics so the harmonic structure is different at each output level.
Thank you for giving it a try - this is a very usual situation
I think that DBT and ABX tests are, more or less, useless. Why? Because the musical signal is too complex for the brain to analyze in this way. That reminds me of driving school teacher who told me that people who drive well in training sometime don't pass the test and those who drive bad in training pass the test because responsible persons are scared of test and irresponsible people are careless and drive relaxed at the test. ABX is like some test for the PUT (person under test) and PUT is not relaxed during testing. Furthermore, brain is easily confused if you play complex pieces of information, which is musical signal, more than once. Memory struggles to process that complex info comparing too complex signals and usually fails to compare them properly.
It's much more relevant when you play samples just once while relaxed and the impressions you get than is right.
But to be honest, sample that I found distorted can sometimes be perceived as better sounding. For instance, my friend has Harbeth loudspeakers which are voiced so that they have some HF roll of. My friend always likes amps with more distortion better than linear amps because distortion adds presence to sound, it acts as some kind of HF tone control. Even high levels of crossover distortion serves it's purpose well in cases like this. Even moderate oscillation and instability can sometimes be perceived as better sound.
Last edited:
I know that the engineers would like to have some exact and objective test that would make design of audio gear simple and straightforward but human perception simply does not work that way. Most of the processing power of brain is reserved for vision, not hearing. That explains why blind people have excellent hearing. All the processing power is available for hearing. Also, hearing is influenced by many factors and is changing rapidly. And last but not the least all human senses are easily deceived.
The more I think about ABX test the more it looks to me that it's deliberately invented to confuse the PUT's brain so that no human brain can solve it successfully. PUTs are tested but nobody told them that the human brain does not have the capacity for the task. The PUTs are meant to fail at the test. The test is invented in order to prove that we can't hear the differences. But it's not that we can't hear the differences, the test is meant to confuse us.
One can arrange ABX that it *is* exactly like that. Just a matter of prodedure.
Hi Klaus,
sorry for my late reply to your PM, just sent it a minute ago.
The poll will expire in about 2 days, I will make a disclosure then. No manipulation except for level equalization was done with the files.
And, I have to admit, that I have used something I would call a "proper output level" from the tube amp not to give a chance for guesses based on subjective noise level.
We have had not a single positive ABX protocol, yet. To me, when the link stage is properly used, there is not a chance to tell the difference, omitting sighted impressions. All the non-hearing based sighted biased impressions have to be exlcluded from the listening tests, that is my answer to ivanlukic as well. There was no intention to head towards a null result, the null result is only a logical impact of excluded bias.
sorry for my late reply to your PM, just sent it a minute ago.
The poll will expire in about 2 days, I will make a disclosure then. No manipulation except for level equalization was done with the files.
And, I have to admit, that I have used something I would call a "proper output level" from the tube amp not to give a chance for guesses based on subjective noise level.
We have had not a single positive ABX protocol, yet. To me, when the link stage is properly used, there is not a chance to tell the difference, omitting sighted impressions. All the non-hearing based sighted biased impressions have to be exlcluded from the listening tests, that is my answer to ivanlukic as well. There was no intention to head towards a null result, the null result is only a logical impact of excluded bias.
This only happens if the person has previously announced to the world that they can hear obvious differences between A and B, or at least firmly believes this to be true. Those who make no assumptions are under no pressure at all. Anyway, this issue has been done to death in other threads so no point in reviving it here.ivanlukic said:
PMA has not forced you to do the test, and he has not even forced you to use ABX, so there can be no pressure at all. You do the test willingly and totally relaxed, or not.
If you see a poll where in every option "have an ABX test" or "have not an ABX test" is included (so exclusive mentioning of one specific protocol) to conclude that there is no "force to use ABX" is slightly underestimating how things work, isn´t it ?
Not at all. A choice between ABX and no ABX seems to me to be a free choice. Someone could use their own favourite test protocol and enter the result as 'no ABX'. There is also the choice 'do not do the test'. Hence any pressure or 'test stress' must be self-imposed.
It is always amusing to see the wriggling and squirming which takes place after the reveal when someone makes the 'wrong' choice and then feels obliged to 'explain' what is wrong with the test. Those who make the 'right' choice will usually give the credit for this to their superior ears. 'Right' and 'wrong' are of course according to their own assumptions, which may be different for other people.
It is always amusing to see the wriggling and squirming which takes place after the reveal when someone makes the 'wrong' choice and then feels obliged to 'explain' what is wrong with the test. Those who make the 'right' choice will usually give the credit for this to their superior ears. 'Right' and 'wrong' are of course according to their own assumptions, which may be different for other people.
That's perfect and me too
All those HE exclusivity supporting stories are you know what about
foo_abx 2.0.5 report
foobar2000 v1.4.1
2018-11-14 19:29:28
File A: ella1.wav
SHA1: 6e2606e184d13abb9f6e8cb15d932eb7cddf4618
File B: ella2.wav
SHA1: 11f2942400a533c5a2909ad54932d348a2494bcf
Output:
DS : Primary Sound Driver
Crossfading: NO
19:29:28 : Test started.
19:30:39 : 01/01
19:31:00 : 02/02
19:31:11 : 03/03
19:31:28 : 04/04
19:31:52 : 05/05
19:32:03 : 05/06
19:32:24 : 06/07
19:32:36 : 06/08
19:32:36 : Test finished.
----------
Total: 6/8
Probability that you were guessing: 14.5%
-- signature --
d3b9b7d7eebe8a0fb0c0b7b3ab11821c43ae8626
Not going to say I prefer one over the other or which is which, because they really are so very close. But on open headphones with the wife blasting the TV in the background I scored 8/10. I can pick up the difference pretty easily once I found the right 1 second spot in the vocals. I can hear it both in the intensity of the sibilance on T's and the overtone's of Ella's voice. But honestly in the real world I'd probably never be able to tell A to B. Not to mention toeing in my speakers three degrees, or the condition of my sinuses that day is likely a bigger difference than the distortion here. It would be interesting to see what % becomes really super obvious without concentrating or studying it.
foobar2000 v1.4.1
2018-11-14 19:29:28
File A: ella1.wav
SHA1: 6e2606e184d13abb9f6e8cb15d932eb7cddf4618
File B: ella2.wav
SHA1: 11f2942400a533c5a2909ad54932d348a2494bcf
Output:
DS : Primary Sound Driver
Crossfading: NO
19:29:28 : Test started.
19:30:39 : 01/01
19:31:00 : 02/02
19:31:11 : 03/03
19:31:28 : 04/04
19:31:52 : 05/05
19:32:03 : 05/06
19:32:24 : 06/07
19:32:36 : 06/08
19:32:36 : Test finished.
----------
Total: 6/8
Probability that you were guessing: 14.5%
-- signature --
d3b9b7d7eebe8a0fb0c0b7b3ab11821c43ae8626
Not going to say I prefer one over the other or which is which, because they really are so very close. But on open headphones with the wife blasting the TV in the background I scored 8/10. I can pick up the difference pretty easily once I found the right 1 second spot in the vocals. I can hear it both in the intensity of the sibilance on T's and the overtone's of Ella's voice. But honestly in the real world I'd probably never be able to tell A to B. Not to mention toeing in my speakers three degrees, or the condition of my sinuses that day is likely a bigger difference than the distortion here. It would be interesting to see what % becomes really super obvious without concentrating or studying it.
I remember for a given single ended topology and phono preamp, the power supply dominating (?) the circuit's signature and would change going from choke input to cap input to cap multiplier, etc. (and adding RC-decoupler to 1st stage from cap mutiplier) High level stages reasonably matched in level from sounded different too feeding run of the mill solid state receiver - a stepped attenuator sounded different than an Alps pot and both different from an IC stage with NFB.
Nothing can be further from the truth. Everybody knows that when your doctor measures your blood pressure it is enough to see the person in white coat and blood pressure jumps for several points immediately. The patient is not relaxed when doctor measures blood pressure. Seeing the doctor makes blood pressure higher. But my main point was that the brain is confused by the test and that is the only reason why it's impossible for most people to discriminate between samples.
That's exactly my point. You can't get positive result because nobody can! Your brain is bombarded with complex impressions and confused. Just like everybody else. Brain activity is influenced by many factors. It's not unusual that at first hearing a person finds "valve sound" more "involving" and "lifelike" but after few days distortions start to annoy and makes listening tiresome. To try to conclude anything relevant with ABX test is sheer lunacy. To reduce such complex brain activity, as hearing, to some mechanical test procedure is nonsense.
I think that the difference is easier to perceive if samples are played once than when samples are repeated several times as in the ABX test. But you obviously find the way to overcome confusion by finding exact moment in the sample when distortions are very high.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.