Blue or red pill?

[adamdea]

22 hours ago, Audiophile Neuroscience said:

 

Adam, I have no idea what you are talking about but I suspect it makes no sense.

Edit: maybe Latin is more your thing?

 

22 hours ago, manisandher said:

 

You can't bring yourself to accept that I heard a difference in the sound of the music between A and B, despite the 9/10 ABX score, and zero evidence to the contrary.

 

Fine. Move on. Nothing for you here.

 

Mani.

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.

 

Equally, ever since the original 9/10 was reported, you have wanted to see this as proof of one proposition only, on the assumption that it must mean what you think it does, there apparently being (for you) only one explanation.

The 9/10 result is interesting and merits careful thought. It doesn't look as though it will get it. I agree with you on one thing though: time to move on.

[fas42]

I can't see or hear this.

 

 

Can you show us where you see it?

 

Mani.

 

In Audacity,

 

 

I can't hear it either - but since we were mentioning tells, it was a point that some might consider significant.

[acg]

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?

[mansr]

1 hour ago, manisandher said:

The differences between the timings of 23 and 24 really are minute (I'd say one or two 10ths of a second at most), and I'd be surprised if this were the 'tell'.

For an ABX between two files to be valid, they must be aligned to within a few samples. It's easy to do, so there's no reason not to.

[adamdea]

32 minutes ago, acg said:

 

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?

<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.

 

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.

 

This is a dead parrot.

 

 

 

 

 

[STC]

1 hour ago, manisandher said:

If anyone believes that A and B can be differentiated in a foobar ABX based on a 0.03s difference, then go ahead. I don't think so.

 

30 milliseconds is easily audible to most. My threshold is 10milisecond with headphones and strangely I can even hear 1milisecond difference with widely spaced speakers. 

 

That’s why it is important to capture exactly what’s coming out from the DAC’s output. Better still capture the signal reaching the speakers. How wide were your speakers placed and your distance when you did the ABX?

[mansr]

8 minutes ago, STC said:

30 milliseconds is easily audible to most. My threshold is 10milisecond with headphones and strangely I can even hear 1milisecond difference with widely spaced speakers.

Cuts in an audio track offset by much less than 1 ms can be readily audible depending on the content of the affected samples.

 

So here's the question, did you align and trim the test tracks to within 1 sample? If not, your 50/50 is meaningless.

[STC]

6 minutes ago, mansr said:

Cuts in an audio track offset by much less than 1 ms can be readily audible depending on the content of the affected samples.

 

So here's the question, did you align and trim the test tracks to within 1 sample? If not, your 50/50 is meaningless.

 

I have answered that. I only listened to the music content in between the start and within 10 seconds or much less. 

 

My question to you. Did you ensure the was no delay or  start of each track coming out from the speakers. The can be difference there. For an example the silence ( and noise including hypersonics) before beginning of the music could have formed some sort of ‘tell’ due to memory retention. Or whatever it is called. 

 

i also hope there wont be samples posted if they were not to be representative of what’s heard during the ABX. Now I guess the 70/70 is meaningless despite no one else tried. 

[acg]

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?

[mansr]

19 minutes ago, STC said:

I have answered that. I only listened to the music content in between the start and within 10 seconds or much less.

That doesn't answer the question.

[adamdea]

2 minutes ago, acg said:

 

This must be a coloquialism from your part of the world.  I am not familiar with what it means.  "Flogging a dead horse" perhaps?

 

[STC]

5 hours ago, manisandher said:

@STC, the tracks 23 & 24 aren't aligned and aren't of exactly the same length. If listened to all the way through, it should be trivial to identify either based purely on listening to when they end. B seems to end slightly sooner than A, just as a piano key is struck.

 

I never intended 23 & 24 to be used in an ABX, but if they are, they need to be aligned and cut to exactly the same length.

 

This was unnecessary during the 'real' ABX with Mans because each A and B were of randomly different lengths.

 

However, if you've already aligned and cut them, then I'm impressed.

 

Mani.

 

5 hours ago, STC said:

 

 I don't need the whole track to tell the difference. The first 10 seconds will do. Ok..I didn't align them because I thought that what was captured from the output and that what would have been reaching the amplifier. Is there a reason to align them? 

 

4 hours ago, manisandher said:

 

I had to start/stop the analogue captures manually - my recording software doesn't have auto start/stop (I tried to get Audacity to work but it refused to recognize my ADC). Considering it was done manually, I think I did a pretty good job to get them started at a similar time. But the stop time is clearly different.

 

If you're listening to just the first 10 seconds or so, I doubt there's enough of a difference between the start times to use this to differentiate them - but there is a difference.

 

The digital captures don't suffer from this, and should start at pretty much the same time. Oh and BTW, the digital captures sound identical to me too.

 

Mani.

 

9 minutes ago, mansr said:

That doesn't answer the question.

 

Please see above. Now about my question. I think the difference in the timing after the DAC is the tell. 

[mansr]

18 minutes ago, STC said:

Please see above.

So that's a no then. This invalidates your results.

 

18 minutes ago, STC said:

Now about my question. I think the difference in the timing after the DAC is the tell. 

What difference in what timing? You'll have to be a bit more elaborate.

[Audiophile Neuroscience]

1 hour ago, adamdea said:

Either way it does not follow from the fact that the ABXXXX produced negative results that the Unidentified protocol artefact was not present.

 

Lets just say we have a variance of opinion on a number of points.

 

Unless you have evidence to the contrary whatever mechanism that could have constituted a tell was held the same between the ABXXXXXX and ABX methodologies. IMO you assume that for some reason the very same tell that failed in the ABXXXX trials  mysteriously started working in the ABX trials. I am not saying it is *impossible* but I don't consider it likely and would need some evidence to that effect. I agree it could have been a learning effect but that is speculation, still possible.

 

1 hour ago, adamdea said:

 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.

 

The alleged software effect HAD to be present in both trials (not just supposedly present) because they continued alternating the settings by intention in exactly the same way between trials.So again, the only thing that changed was the test methodology. As you correctly pointed out one obvious explanation for the software effect to become apparent was due to variances in short term memory between the methodologies.Whatever the case it only served to demonstrate that the ABX test was a more sensitive test for whatever reason. None of this invalidates that the potential tell was the same between trials.

 

 

Quote

 

If you want to campare the Abx with the ABxxxx you have to ask what changed, and that is the test itself.

 

Yes correct the test methodology itself changed and once again, any potential tell did not, nor the way they manipulated the software.

 

Quote

Unfortunately than take the point. We have a had a tiresome disquisition with ineptly wielded irrelevances such as "logical" deductions and "occam's razor".

 

You may consider logical deductions and philosophical principles irrelevant but I do not. I am happy for you to disagree with me though.

 

 

Quote

 

  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.

 

This is a dead parrot.

 

For once we agree.

[STC]

@mansrBecause nothing there to show that you or Mani captured exactly what’s coming out from the analogue of DAC. I am not keen in accepting the the digital capture because it only capture the digital data of the original. What goes into the DAC and what’s come out is different whether you want to accept Mani results or not. Even with the so called difference in file 23 and 24 I don’t see many could identify the difference. 

 

IMO, analyzing the analogue output is not the correct way of telling what’s going on with the SFS if it did not capture the exact signal from start to finish as heard by the amplifier. 

 

If there were a small timing difference when playing the two different SFS settings, the difference in timing with the noise ( even inaudible ) would leave different signature imprint in your brain that can influence the music. 

 

Being familiar with carnatic music, I know the deliberate drone instrument  ( some sort of white noise like sound) influence the perception of the music. As a child I hated that but as I grew older I realized how important the influence is. 

 

So going back to this red/blue pill, start the investigation by capturing what’s exactly coming out from the DAC output. The different silence at each track even in milliseconds could be the results of DAC decoding the influence of the SFS which is part and parcel of the elements that made audible changes.  

 

If there is nothing for us to compare reliably and even passing the ABX with given files than there is nothing worth discussing except  some magic happened there in Mani’s house which you two witnessed and cannot be replicated here. 

 

I can always get the XXHE and hear the difference myself but there would be no way to tell why I heard the difference or validly prove to the objectivists. 

 

In on an unrelated matter, I had a weird reception problem with my satellite TV. The guys came over and used all the digital equipment and told me that the meter reading is showing reception and signal to the decoder is as per spec and what’s coming out too. It took me another week or two when I discovered that the problem happens when the two decoder function together it influence another channel. I managed to replicate the problem and then showed to the technician who then replaced the decoder. Somehow when I switch certain channel combination thenprovlem happens. Now you know why I doubt that what recorded by Tascam is not necessarily what’s coming out from the DAC. 

 

[STC]

5 minutes ago, Audiophile Neuroscience said:

Yes correct the test methodology itself changed and once again, any potential tell did not, nor the way they manipulated the software.

 

Some difference can be audible due to other factors. For an example, I perceived SQ difference between JRiver and Foobar. Initially, I thought it was due to gain difference but than after few day I thought something else was influencing the difference. It then I discovered that Foobar had higher noise level which was not audible at the normal volume level but it was there when I moved the volume much higher. 

 

So so in this ABX the silence different between the two setting can have an effect. It can only develope if enough exposure is given to leave a sonic signature. The different methodology used in can account for the different level of exposure where one could give you the tell sign. 

[mansr]

12 minutes ago, STC said:

@mansrBecause nothing there to show that you or Mani captured exactly what’s coming out from the analogue of DAC.

How would you do that other than by connecting the DAC to an ADC?

 

12 minutes ago, STC said:

I am not keen in accepting the the digital capture because it only capture the digital data of the original.

The purpose of the digital captures was to make sure the data bits entering the DAC really were identical, NOTHING ELSE DAMMIT.

 

12 minutes ago, STC said:

What goes into the DAC and what’s come out is different whether you want to accept Mani results or not.

Determining this was the purpose of the test.

 

12 minutes ago, STC said:

Even with the so called difference in file 23 and 24 I don’t see many could identify the difference. 

I have no idea what you intend to mean by that.

 

12 minutes ago, STC said:

IMO, analyzing the analogue output is not the correct way of telling what’s going on with the SFS if it did not capture the exact signal from start to finish as heard by the amplifier. 

So now recording the analogue output is wrong? Please tell us how you would go about capturing the effect, supposing one exists.

 

12 minutes ago, STC said:

If if there were a small timing difference when playing the two different SFS settings, the difference in timing with the noise ( even inaudible ) would leave different signature imprint in your brain that can influence the music. 

Timing difference in what relative to what?

 

12 minutes ago, STC said:

Being familiar with carnatic music, I know the deliberate drone instrument  ( some sort of white noise like sound) influence the perception of the music. As a child I hated that but as I grew older I realized how important the influence is. 

I fail to see the relevance of this paragraph.

 

12 minutes ago, STC said:

So so going back to this red/blue pill, start the investigation by capturing what exactly coming out from the DAC output.

So now you want to capture the output again? Make up your mind already.

 

12 minutes ago, STC said:

The different silence at each track even in milliseconds could be the results of DAC deciding the influence of the SFS which is part and parcel of the elements that made audible changes.  

Not making much sense there.

 

12 minutes ago, STC said:

If there is nothing for us to compare reliably and even passing the ABX with given files than there is nothing worth discussing except  some magic happen there in Mani’s house which you two witnessed and cannot be replicated here. 

 

I can can always get the XXHE and hear the difference myself but that would tell why I heard the difference or validly prove to the objectivists. 

 

In on an unrelated matter, I had a weird reception problem with my satellite TV. The guys came over and used all the digital equipment and told me that the meter reading is showing reception and signal to the decoder is as per spec and what’s coming out too. It took me another week or two when I discovered that the problem happens when the two decoder function together it influence another channel. I managed to replicate the problem and then showed to the technician who then replaced the decoder. Somehow when I switch certain channel combination thenprovlem happens.

TV tuners emit some EM noise that depends on the tuned frequency. If a broken tuner emits too high levels or is too sensitive to outside interference, two of them might well malfunction in the way you describe. Nothing mysterious.

 

12 minutes ago, STC said:

Now you know why I doubt that what recorded by Tascam is not necessarily what’s coming out from the DAC.

Nope, not at all. The ADC is obviously subject to its own limitations in noise levels and clock accuracy, but beyond that I have no idea what you're going on about.

[Audiophile Neuroscience]

4 minutes ago, STC said:

So so in this ABX the silence different between the two setting can have an effect. It can only develope if enough exposure is given to leave a sonic signature. The different methodology used in can account for the different level of exposure where one could give you the tell sign. 

 

I am not saying it is impossible whether a learning effect or something else as you describe but without evidence it remains speculation. lets see where the evidence leads.

 

 

[STC]

10 minutes ago, mansr said:

So now recording the analogue output is wrong? Please tell us how you would go about capturing the effect, supposing one exists.

 

It depends what you are trying prove? There cannot be difference?

 

@manisandher and you both telling the analogue captures were not time aligned because the start and stop of the recorder were done manually. So how you know the original sound that came out from the DAC is not misaligned itself like the SFS 200 comes out 30ms later than the SFS0.1. That’s what still puzzling me DAMMIT. 

 

 

[STC]

25 minutes ago, mansr said:

TV tuners emit some EM noise that depends on the tuned frequency. If a broken tuner emits too high levels or is too sensitive to outside interference, two of them might well malfunction in the way you describe. Nothing mysterious.

 

If that was as simple as that I would not even relate it here. The technicians were competent enough to determine that. It was some weird problem when the channel were changed in certain sequence then the problem happens. So channel x in 1 and channel U in 2, I got no problem but if I were to tune channel abc and x for 1 and channel U comes after certain combination than the problem starts. It was so unique that I have to list down the combination. It was a mystery. 

[mansr]

11 minutes ago, STC said:

It depends what you are trying prove? There cannot be difference?

Mani says he hears a difference through the speakers. I contend that if this is the case, some difference should be evident in a recording. I'm not really trying to prove anything.

 

11 minutes ago, STC said:

@manisandher and you both telling the analogue captures were not time aligned because the start and stop of the recorder were done manually.

Not what I said. I did say that before using those recordings to conduct an ABX test, e.g. using Foobar, you must align them to within one sample and trim the ends such that both contain exactly the same segment of music.

 

11 minutes ago, STC said:

So how you know the original sound that came out from the DAC is not misaligned itself like the SFS 200 comes out 30ms later than the SFS0.1.

Oh, the delay from pressing the play button to music coming out of the speakers differs by a lot, SFS 200 taking some 15 seconds longer. This is why I waited a "random" amount of time between plays. At first, I speculated that perhaps I hadn't randomised this delay sufficiently, but on closer examination, I could find no correlation between the delays and the identity of the X samples. Had I not added this random delay, but simply pressed play as quickly as possible, Mani would have easily been able to identify the setting purely from the time between plays.

[mansr]

8 minutes ago, STC said:

If that was as simple as that I would not even relate it here. The technicians were competent enough to determine that. It was some weird problem when the channel were changed in certain sequence then the problem happens. So channel x in 1 and channel U in 2, I got no problem but if I were to tune channel abc and x for 1 and channel U comes after certain combination than the problem starts. It was so unique that I have to list down the combination. It was a mystery. 

Having worked on satellite receiver software (DirecTV), I have no problem believing that. Those things are crawling with bugs.

[adamdea]

43 minutes ago, Audiophile Neuroscience said:

 

 IMO you assume that for some reason the very same tell that failed in the ABXXXX trials  mysteriously started working in the ABX trials. I am not saying it is *impossible* but I don't consider it likely and would need some evidence to that effect. I agree it could have been a learning effect but that is speculation, still possible.

 

 

The alleged software effect HAD to be present in both trials (not just supposedly present) because they continued alternating the settings by intention in exactly the same way between trials.So again, the only thing that changed was the test methodology. As you correctly pointed out one obvious explanation for the software effect to become apparent was due to variances in short term memory between the methodologies.

 

The only real point of contention is that for some reason you consider it less plausible that a tell should be affected by the change in test methodology than that the alleged software effect.

 

Unfortunately your above assumption had to be buried in pages of hand waving about logic and ontology whereas it is no more than an expression of opinion. For some reason you have insisted for pages in saying that someone was implying that the tell had appeared when it was perfectly apparent that all that could possibly be implied (aside from learning) was that the change in test methodology affected the ability to detect the tell.

 

The reason why the tell might work on the ABx  but not ABxxx (aside from learning) is precisely because in the ABx the listener was called upon to compare B with X which the listener can do, whereas in the ABxxxx, he couldn't really as A and B were too old. That still leaves the question of what exactly he is detecting in the ABx comparison which he could not detect in the ABXXXX.  I made this point right at the beginning. There is no reason I can see why a tell could not be affected by audio memory just as much as whatever the software is supposed to do.

 

In turn I find it implausible that any audible software  effect could be incapable of being recorded. And I happily stand by that. If the contrary can be demonstrated then new ground will have been broken. If however it can be recorded then it will be easy to find out whether it is audible.

[Sonicularity]

30 minutes ago, adamdea said:

The reason why the tell might work on the ABx  but not ABxxx (aside from learning) is precisely because in the ABx the listener was called upon to compare B with X which the listener can do, whereas in the ABxxxx, he couldn't really as A and B were too old. That still leaves the question of what exactly he is detecting in the ABx comparison which he could not detect in the ABXXXX.  I made this point right at the beginning. There is no reason I can see why a tell could not be affected by audio memory just as much as whatever the software is supposed to do.

 

The first X of the ABXXXX tests should have been identical to each of the 10 ABX tests conducted later, yet in both ABXXXX tests the first response was missed.   The listener would have been comparing B with X.

[manisandher]

3 hours ago, Sonicularity said:

The first X of the ABXXXX tests should have been identical to each of the 10 ABX tests conducted later, yet in both ABXXXX tests the first response was missed.   The listener would have been comparing B with X.

 

The listener has a name. And how the f**k do you know what I was listening for?

 

Mani.