acg Posted February 13, 2018 Share Posted February 13, 2018 Hey Mani, For my 2c, I think you just need to make the day low stress and interesting. Do your abx test, play with some USB cables, show your guests the difference in sound between a bloated operating system and one optimised by XXHE. Or don't. Then if you want to you could start an investigation into the 'why' of things or even attempt to record some things to share with the forum. Same goes for the guests as well. Don't feel unduly pressured, enjoy yourself, make new friends, keep an open mind. I helped run a multiple DAC blind test a few years ago and that proved to be a relatively stressful event for us who ran the show. It was, however, very interesting sitting on a perch and watching the participants, reading their assessments and then talking to them later. Cheers, Anthony. Superdad 1 Link to comment
acg Posted February 19, 2018 Share Posted February 19, 2018 17 hours ago, mansr said: No. Did you read nothing of what I wrote? There is no way to distinguish a near source from a more distant louder one. Any perception of depth or height in a stereo recording is our mind making things up based on what things usually sound like, possibly combined with room reflections. Have you ever noticed how the moon looks bigger when it's low in the sky? That's also our mind applying patterns that are usually right (objects overhead tend to be nearer than objects at the horizon) but fail in this extreme case. The moon looks bigger at the horizon because we are looking through more atmosphere. Look up refraction and you will get an idea why... it has nothing to do with 'applying patterns' as you suggest. Link to comment
acg Posted February 19, 2018 Share Posted February 19, 2018 54 minutes ago, Sonicularity said: I don't believe that is true from the information that I found about the topic. Where do you see evidence to suggest this? Light travels in a straight line...right? Only when the medium through which it travels is of homogenous density. Otherwise it bends from high density toward low density...slowly. In the case of the light travelling to the Earth it bends from high density atmosphere closer to the surface toward lower density atmosphere further up thus forming a curve rather than a straight line. There are all sorts of equations for it but basically the curvature is dependent on the density gradient and the distance which are in turn influenced by temperature and barometric pressure et cetera. It is basic physics and you can choose to disbelieve if you like. This is why telescopes point straight up into the sky. It is why any decent GPS will not use satellites close to the horizon. It is why a field levelling contractor can only be accurate about 400m metres from a laser. It's why surveyors have to make optical observations from both directions. And so on and so forth. Link to comment
acg Posted February 19, 2018 Share Posted February 19, 2018 7 hours ago, mansr said: If you measure the angle occupied by the moon when near the horizon and again when high in the sky, you get almost exactly the same value. Nevertheless, the rising or setting moon is often perceived as twice the size of the moon when (more or less) overhead. Atmospheric refraction can't explain this. Atmospheric refraction perfectly explains this. And I have measured it, more than once during my university degree, of both the sun and the moon. They both move quite quickly and are relatively difficult to manually track through an instrument. Link to comment
Popular Post acg Posted March 28, 2018 Popular Post Share Posted March 28, 2018 13 minutes ago, STC said: They were identical but how the DAC processed the two different SFS settings is something unknown and obviously it appears to make a difference. Mani already said that the files captured from the digital output sounded the same so we can conclude bit-identical files do sound the same. If there is anything to learn from this experiment, my take would be DAC processes files differently when the upstream process changes. I expect the analogue signal to show some difference. It also proved that the difference is not readily audible as Mani failed to recognize the difference with a different track. What matters here is @PeterStclaim that audible difference with SFS is real. Does this also mean others changes made to the bit-identical files before reaching the DAC could alter the sound? Maybe...but too insignificant to be reliably identifiable. It seems as though you are thinking about this one way when you could be thinking about it another way. In some ways the changed parameter, the SFS should be thought of as a red herring. The SFS did not change any bits and Peter has commented that SFS is reliant on many other factors before it becomes a factor itself for SQ. So instead of thinking about SFS, let's think about "changing the electrical processes in the AudioPC", which afterall is what XXHE software primarily does, and primarily by changing how the cpu uses current (there are other ways of course but that is the main way). So we have bit identical files...therefore the information (or bits) the dac gets is the same in each scenario. But how the AudioPC does its job is slightly altered by changing the SFS number. Quote If there is anything to learn from this experiment, my take would be DAC processes files differently when the upstream process changes. I expect the analogue signal to show some difference. My understanding is that the SFS dictates how large the blocks of RAM are that hold the information before it is sent to the dac, and not how the dac receives the "bits". I just bought into that red herring myself...sorry. Quote It also proved that the difference is not readily audible as Mani failed to recognize the difference with a different track. It did no such thing. It suggests that Mani managed to get confused when the original test protocol was used, but that all changed when ABX was used. This "confusion" was actually predicted by others earlier in the thread. manisandher, PeterSt and STC 2 1 Link to comment
Popular Post acg Posted March 28, 2018 Popular Post Share Posted March 28, 2018 16 minutes ago, rando said: Somewhere along the way it morphed into something very different. I dunno that it did morph into something very different. Mani and Mans needed to find something to test that would satisfy both of their objectives and sensibilities as well as the prerequisites for recording the digital and analogue feeds. It is not unsurprising that the exact things being tested have changed from those initially put forward, but at its core the invite was for a non-believer (I hate that word and all its connotations - sorry Mans) to come and either participate or manage controlled listening where the proponent (Mani) thought he could blindly and correctly identify differences in sound between two bit-identical musical tracks where the only difference is how that file is managed by the AudioPC. From the reports so far I think that objective has been achieved: using ABX Mani reliably identified a change in SQ between two bit-identical dac feeds when a computer software parameter changed. manisandher and sandyk 2 Link to comment
acg Posted March 28, 2018 Share Posted March 28, 2018 6 minutes ago, pkane2001 said: Why the insistence that only bits are significant? It is a well-known fact that poor timing of samples can cause jitter, and that large jitter can be audible. I insist on no such thing, quite the opposite in fact. Based on Mani's ABX results the bits are NOT the only significant things. 6 minutes ago, pkane2001 said: Before postulating noise or other less likely (and less measurable) causes of audible differences, can we please eliminate the well-known and obvious ones, first? Sure, go ahead. PeterSt 1 Link to comment
acg Posted April 10, 2018 Share Posted April 10, 2018 50 minutes ago, esldude said: I don't think 10 samples is enough. I like 30 minimum. I see too many things go astray not using enough samples. I know the math works even at 10, but some reputable statisticians also seem to like no fewer than 30. http://www2.psychology.uiowa.edu/faculty/mordkoff/GradStats/part 1/I.07 normal.pdf First thing I found handy that talks about the idea you need 30. And it is one of those rules of thumb without a definite mathematical explanation saying you need 30 (or according to some 32). I think with 30 ABX in the format that is necessary for this test (with the time delay et cetera) would be a true mind-f&%k for the person under test. 10 is statistically "good enough" at this stage such that I do not doubt that Mani can hear a change. Confused 1 Link to comment
acg Posted April 13, 2018 Share Posted April 13, 2018 13 minutes ago, pkane2001 said: But, that lack of jitter rejection or other issues within your DAC could have caused you to hear a difference, where a better DAC or a USB reclocking interface would eliminate them, I’m afraid you didn’t test for at all. Except the dac that Mani usually listens to is sota in the ways that you mention and SFS changes are audible with it as well... just not proven in this test of course. Link to comment
Popular Post acg Posted April 14, 2018 Popular Post Share Posted April 14, 2018 1 hour ago, jabbr said: Good. One would expect that as the isolation alone becomes better that the SQ differences would decrease. That said two different DACs may have different SQ responses to SFS that are not entirely due to isolation ie if one DAC is “more resolving” it might unmask subtle SQ effects that the other DAC dies not reveal — I’m granting this — nonetheless for any given DAC, as isolation improves, effects such as input jitter become less apparent — by definition ? I can see why you think that way. It's very 2d and assumes a straight line or linear relationship between the degree of isolation as you define it and SQ effects. But I strongly doubt that linear kinds of relationship actually exists. For example, before galvanic isolation Peters NOS1 was immune to usb cables but then better voltage regulators became available and some additional regeneration and galvanic isolation had been added to the dac and it no longer is immune to USB cables and Peter even makes and sells a couple of variants. Better isolation for sure, lower noise floor, a bit or so of extra resolution and better SQ but something that did not matter before now matters... and SFS changes if anything are more audible. Go figure! PeterSt and manisandher 1 1 Link to comment
acg Posted April 20, 2018 Share Posted April 20, 2018 It seems to be the sticking point at the moment Mani...maybe give it a go. This had been a most interesting thread for me. Thanks for your time and effort thus far. Link to comment
acg Posted April 30, 2018 Share Posted April 30, 2018 39 minutes ago, fas42 said: Okay, before anyone gets too excited - there's a tell at 0.9 sec in, which lasts for 0.025 secs. This could be a glitch in the recording process, for some reason - and is probably enough to 'learn' to recognise. It appears to be the only one, without looking more thoroughly at the difference file. STC did 50 ABX comparisons in 12 minutes...that's 14 seconds per ABX...it appears as though he could hear that "tell" Link to comment
acg Posted April 30, 2018 Share Posted April 30, 2018 20 minutes ago, PeterSt said: Once you know what to listen for, why ? Anthony, wouldn't you be able to do it "within the second" depending on the music and what XXHighEnd setting it is about ? I don't know about "within the second" but, yes, once I was familiar enough with the particular "difference" then I would be more confident of picking it blind...but that is not what I was so much concerned about. Frank suspects he has identified a tell at 0.9s into one of the captures and in this case STC is only listening to each A, B or X for about 4s on average. It certainly appears as though Franks "tell" could be having an influence on the results...at least it is within the realm of possibility. Link to comment
acg Posted April 30, 2018 Share Posted April 30, 2018 15 minutes ago, adamdea said: I think it's a shame that their appears to be no scope for reasoned discussion. At some point one reaches an impasse. I have run this back through and I'm afraid I can;t think of any way of moving forward if one can't see through an obvious flaw in reasoning when it's pointed out. I am struggling to remember what it was that you consider an "obvious flaw in reasoning". It was about the potential "tells" and whether they changed with the change from ABXXX to ABX...I think. There was talk of hearing clicks through two closed doors with a hallway in between...I am unclear...would you care to elaborate? Link to comment
acg Posted April 30, 2018 Share Posted April 30, 2018 33 minutes ago, adamdea said: <Groan>. Not "elaborate" it's all there already, more than once. But if you must.... It was suggested by Manis that his 9/10 Abx results could not be the result of a tell, because if they were, he would have got the ABxxxxx right. This is obviously not a sound piece of reasoning because the ABxxx and Abx tests call upon the listener to do two different things, The reason the ABXXXX test is a waste of time is to do with audio memory. It does not need any further explanation. As for a tell, it might or might not work the same way depending on the comparison which the listener were called to carry out, and because of learning effect. The key is to understand that all that is being said is that there may be something other than the direct effect of the software setting on dac's playback of the file. Unidentified protocol artefact is a better term. Either way it does not follow from the fact that the ABXXXX produced negative results that the Unidentified protocol artefact was not present. As Mansr pithily pointed out pages ago, if that reasoning were correct then one could just as easily point out that the alleged software effect was also supposed to be present. If you want to campare the Abx with the ABxxxx you have to ask what changed, and that is the test itself. As far as I can recall, that is the most effort you have made to voice your frustration, so thank-you. And I get it, for sure, makes sense, is logical. A "tell" is yet to be identified though... 39 minutes ago, adamdea said: Unfortunately than take the point. We have a had a tiresome disquisition with ineptly wielded irrelevances such as "logical" deductions and "occam's razor". It has been going on for pages and it is boring. It doesn't get any better with repetition. What you need is new data. If the effect can be recorded then it is much easier to repeat. You obviously have a different order of enquiry that you would like to see happen but this is an open forum and unfortunately the discussion goes wherever it wanders. The current discussion around STC's results is another small step in the right direction would you not agree? Someone else spending the time to have a go and let us know the results, identifying problems in the procedure and perhaps having another go. That is at least going towards "new data" is it not. Do not despair. One can only manipulate the direction the thread takes if one stays involved. 46 minutes ago, adamdea said: This is a dead parrot. This must be a coloquialism from your part of the world. I am not familiar with what it means. "Flogging a dead horse" perhaps? Link to comment
Recommended Posts
Create an account or sign in to comment
You need to be a member in order to leave a comment
Create an account
Sign up for a new account in our community. It's easy!
Register a new accountSign in
Already have an account? Sign in here.
Sign In Now