Impossible to differentiate for toc549 with a correct audio system.
With the old amp it's more difficult this time but I was able to get 11/12.
At the end of this test I observed an intermittent operation from this amp. When I hit the amp the sound disappears or returns. I will carefully dismount it to know the problem and I will post it.
With the old amp it's more difficult this time but I was able to get 11/12.
At the end of this test I observed an intermittent operation from this amp. When I hit the amp the sound disappears or returns. I will carefully dismount it to know the problem and I will post it.
Last edited:
Be sure not to fool yourself. I thought I could hear a difference in the violin files, one sounded duller than the other. But I could not pick them in the ABX, even tho I "heard" them.
Violins have extended harmonics. A lot of the warmth comes from way up in the harmonics. But those harmonics are so dense they are bound to mask crossover distortion harmonics. The interest is - do the distortion harmonics add anything or change anything in the tone of the violin?
tocr sounds a little more detailed and a little less compressed in Sennheiser HD600 headphones driven by the DAC-3. Don't know if its the one going through the amp or not.
Pano, for no harmonics example you have the sine 440Hz test here. So some instruments have more harmonics and some less. Makes no difference for the result. In this last test, you also have reverbs, sound reflections,, fading out and other audiophile candies. Still does not help. The amount of distortion added is inaudible and not discernable by statistically valid test result.
The amount of distortion difference between the two files is slightly audible and not discernible by a statistically valid test that we happen to have available, using 5 old men as test subjects, one of whom has hearing loss and tinnitus.
The reason is more prosaic, the distortion artifacts lie deep below the ambient noise of the recording, in case that ear masking curves are not taken seriously enough as an explanation.
BTW, Garrick Ohlsson is in Prague, going to visit his concert tomorrow 🙂.
BTW, Garrick Ohlsson is in Prague, going to visit his concert tomorrow 🙂.
The noise itself is distorted. As are the harmonics, volume dynamics, and pitch inflections of finger vibrato on instrument necks. Listen carefully, is all.
Maybe try counting the number of vibrato inflections on one particular solo instrument note, and pay attention to their depth or volume variation with each pitch inflection. Listen to the graininess of the "air" and noise between notes. There are audible differences all right, but they are small enough to be quickly forgotten if noticed at all, especially since there are so many of them. I looped a short section of a solo note and compared very subtle details before there was time to forget or to develop ear fatigue. Differences are there, and they are resolved by the DAC-3 and good headphones. What I don't know is whether the better sounding file to me is going through the amp or not. If it is, I would assume the amp is acting as buffer between the DAC and ADC which improves their performance (maybe the ADC is cleaner driven with a lower Z source, for example, or maybe not), and/or maybe the amp adds a little good sounding distortion. If more detailed file isn't the amp, then I would assume the amp is loosing detail.
In the foregoing I describe a process that works for me, even though I have hearing loss and tinnitus. On the other hand, I probably have the most accurate DAC and headphone amp combination. Somebody with better hearing an equally good equipment could possibly demonstrate with ABX, if ear fatigue doesn't prevent it. Although Foobar ABX doesn't allow looping, the first couple notes in the files are more than enough to reveal differences, so maybe aural memory limitations could also be overcome.
Maybe try counting the number of vibrato inflections on one particular solo instrument note, and pay attention to their depth or volume variation with each pitch inflection. Listen to the graininess of the "air" and noise between notes. There are audible differences all right, but they are small enough to be quickly forgotten if noticed at all, especially since there are so many of them. I looped a short section of a solo note and compared very subtle details before there was time to forget or to develop ear fatigue. Differences are there, and they are resolved by the DAC-3 and good headphones. What I don't know is whether the better sounding file to me is going through the amp or not. If it is, I would assume the amp is acting as buffer between the DAC and ADC which improves their performance (maybe the ADC is cleaner driven with a lower Z source, for example, or maybe not), and/or maybe the amp adds a little good sounding distortion. If more detailed file isn't the amp, then I would assume the amp is loosing detail.
In the foregoing I describe a process that works for me, even though I have hearing loss and tinnitus. On the other hand, I probably have the most accurate DAC and headphone amp combination. Somebody with better hearing an equally good equipment could possibly demonstrate with ABX, if ear fatigue doesn't prevent it. Although Foobar ABX doesn't allow looping, the first couple notes in the files are more than enough to reveal differences, so maybe aural memory limitations could also be overcome.
That's what I did - listened to just the first few notes. You don't have to loop it, just start and stop. I had thought the difference was more obvious there than anywhere else, but ABX said I was still guessing.
To be honest with myself I went back and forth, back and forth between A & B. Sometimes I thought there was a difference, sometimes I thought there was not. If I can't reliably hear a difference between A & B, how can I sort X from Y?
But whilst the fast switching does demonstrate that short term aural memory does exist, if there is no nasty artifact that can be recognized during actual listening to music we have a very interesting (albeit non scientific) result, which does not overturn any accepted psycho acoustic research.
fascinating and no woo and foo needs chanelling. What's not to like 🙂
fascinating and no woo and foo needs chanelling. What's not to like 🙂
I still do not understand why it is so difficult (for some people) to accept the fact that there are well known limits - threshold of hearing as a function of frequency and auditory masking as a function of frequency and SPL. This thread is, btw, a reaction to BT thread where we can read that "everything is audible" - resistor -140dB distortion, golden fuses, Bybee magic devices etc. etc. Yes everything is audible in case that the listener is self-convinced that it is audible. Even the biggest nonsense is audible then. Maybe also a sound wave in vacuum. But just take the opportunity to know what is tested out from the listener and all the magic automatically disappears and the only remaining are possible technical errors of his replay chain. And this is the situation where we can start to investigate the real limits of audibility, not influenced by beliefs and superstitions.
Last edited:
Excellent! Let us know how it was.
Yes, I will.
Pano, I had some issues with "recipient" when answering your PM. Please check your PM mailbox if you received my PM.
Pavel, You did invite me to listen to the 24-bit Toccata files, and you expressed confidence here in the forum that those requesting 24-bit files would join in. If you do that, and in response I try my best to find out if there is any audible difference in the files I can detect, you probably should allow that you could get an unexpected result. By the way, I don't claim to have heard any amplifier distortion in case I didn't make that clear. I only claim an audible difference in the files. What I heard could have been some artifact of buffering your data conversion chain, which I suggested as a possibility. Although I suggested it might have affected ADC performance, it could also be DAC performance or maybe both, or neither for that matter. Whatever it is, only you have the equipment at your disposal to investigate if you would like to do that.
Mark,
all true, I invited you and I created the new test files mostly because of your responses in this thread. I greatly appreciate that you tried the files and did your best. The same time I am saying that even a negative result is a valuable and valid result, at least such is my view and understanding.
Pavel
Pano, I believe you might be able to reliably sort A and B using a sorting test as I have suggested elsewhere. Of course, an experimenter would have to create multiple files for you to sort in some order.
Also, some people would probably like to see a secure and blind testing application written for any sorting test, although as I have mentioned elsewhere non-proctored testing cannot be completely cheat-proof even with Foobar ABX.
I anyone would like, I could describe a procedure to A/B test Pavel's files blindly, if one would like to confirm if one is fooling one's self or not without using Foobar ABX. I use a such a method myself, but not as proof to offer anyone else, just for my own use.
- Status
- Not open for further replies.