[analogsurviver]

   
Is that based on actually analyzing the samples (converting both to e.g. 176.4/24 PCM format, and subtracting the Red Book version should result in no difference other than the higher quantization noise of the CD, and the ultrasonic content of the DSD), or do you just believe it is the case ? Do you even know if the playback levels are exactly the same ?

OK, guys, that does it. This post is CLEARLY meant to imply that anything above 44.1/16 is meaningless.
 
To all computer first/audio second guys (and gals, if any ) - converting, substracting, etc-ing is utter nonsense if you are at location and comparing various digital gizmos to the sound heard live - it IS audible that anything else than live feed from microphones is inferior to it and that magnitude of that deterioration is  sample rate related .
I do try to match the levels - it IS exactly the same whether I use any setting for resolution from 44.1/16 as the lowest and DSD128 as the highest available on Korg recorders - and subjective results are as described by now countless times. Although I do not practice it on regular basis, sometimes I do match the level of recorder to direct output from the microphone - because that means I have to set the recording level a couple dBs below the optimum for the recording, as output level of recorder is fixed and to match the level with the microphone feed an additional preamplifier to raise the level of the mike would be required - or at least padding down the output of the recorder one way or another would be required. Both solutions add additional elements in the chain and none of these elements are perfect.
After doing it a couple of times - and hearing the difference(s) - I simply do not bother with it anymore - my goal is to make recordings best I possibly can, not to fiddle with ABXing and statistics and such. With differences as they are, one would not worry about the exact composition of tyres on  cars being compared - if one is an average family car and another F1 car - would one ? It is true that getting from A to B in traffic congestion is no faster with F1 than family car, that F1 car is incomparably more costly, that it is a single seater, it can not carry any luggage, that it is utterly impractical for anything but racing, etc, etc - but if racing is the name of the game, it IS the answer. 
 
I am striving to present these differences best I can - but if there is no virtual ABX comparator that can satisfy basic blind testing requirements AND quality of ALL format file playback ( it does OK for PCM ) - then it becomes hard. ANY non virtual real hardware ABX "box" with matching within 0.1 db or less with quality switching IS going to be expensive - I reckon 500 + for reasonable quality, but there are switches and potentiometers considerably exceeding that price - EACH. It is an area where commercial audio products have to make do with whatever can be used that fits within the selling price - you can't use $ 300 potentiometer in a $ 500 amplifier, for example. Take a "simple" RCA for example - there are $ .50 connectors ( sometimes I do use them, because they do sound good ) - and there are $ 200+ NextGen RCA connectors, with matching females at same level of price. NextGen will create better contact from new to who knows how many connection/disconnectio cycles - reliably, each and every time. .50 connector might work well initially - but will deteriorate with repeated use much faster and will fail completely after far less cycles than NextGen will still be good as new. The difference in price is approx 400 times ... - and I can only dream about NextGens.
 
Once in PCM, one can mix/remix/upsample/donsamle/whatever almost at libitum - and will remain "the same". Even back in the days I have been recording with CD recorder directly to CD, I did not care for dithering - as real world noises of both equipment and enviroment where music is being played are generally above the level dithering and more than 16 bits can improve upon in meaningful way. Not so with sampling frequency; one does get an incremental improvement, the most obviously audible one by doubling the sampling rate of redbook to 88.2 or 96 kHz. Above that, the differences are ever harder to hear and/or justify in financial terms - but that does not mean they do not exist and are not audible at least on top level equipment. That top level equipment in headphone world can cost less than 500 $/euro - to sky is the limit with speakers. That is to say that equipment capable of taking advantage of higher than 44.1/16 redbook sampling frequency recordings is not out of reach of majority potentially interested listeners.
 
I have to stress that the format or resolution of the recording is not in itself any guarantee of the overall quality of the recording. A Sony WMD-6(C) pro cassette recorder will produce infinitely superiour recording to any of the latest/greatest digital machines - if it is fed from better (positioned ) microphones than new digital gear. NEVER forget what comes first, resolution of digital is meaningful only if and when there is something worth supporting in the first place.

 

[stv014]

Originally Posted by analogsurviver /img/forum/go_quote.gif
 
To all computer first/audio second guys (and gals, if any ) - converting, substracting, etc-ing is utter nonsense if you are at location and comparing various digital gizmos to the sound heard live

 
What is nonsense about wanting to know for sure if the samples in a high resolution vs. Red Book comparison are the same master, and that the differences are entirely due to the limitations of the latter format ? It is not hard to "prove" subjectively that apples and oranges are different, but what is the point ?
 

Originally Posted by analogsurviver /img/forum/go_quote.gif
 
I am striving to present these differences best I can - but if there is no virtual ABX comparator that can satisfy basic blind testing requirements AND quality of ALL format file playback ( it does OK for PCM ) - then it becomes hard. ANY non virtual real hardware ABX "box" with matching within 0.1 db or less with quality switching IS going to be expensive - I reckon 500 + for reasonable quality

 
You have just claimed that 96/24 PCM makes a "night and day" difference compared to 44.1/16. Do you have any equipment (external DAC with USB or S/PDIF input, or whatever) that allows for playing 96/24 in software and satisfies your high standards of quality ? If yes, then there is nothing else you need to perform the test, no hardware switching is required, because all the playback would be done on the same DAC in 96/24 format. If you do not have a DAC that can adequately play 96/24, then how do you know it is so much better ? Or are you claiming that any involvement of software (even if it leaves the high resolution stream bit perfect) unavoidably makes the test invalid, or, in other words, any form of high quality computer based music playback is physically impossible ?

 

[analogsurviver]

   
What is nonsense about wanting to know for sure if the samples in a high resolution vs. Red Book comparison are the same master, and that the differences are entirely due to the limitations of the latter format ? It is not hard to "prove" subjectively that apples and oranges are different, but what is the point ?
 
 
You have just claimed that 96/24 PCM makes a "night and day" difference compared to 44.1/16. Do you have any equipment (external DAC with USB or S/PDIF input, or whatever) that allows for playing 96/24 in software and satisfies your high standards of quality ? If yes, then there is nothing else you need to perform the test, no hardware switching is required, because all the playback would be done on the same DAC in 96/24 format. If you do not have a DAC that can adequately play 96/24, then how do you know it is so much better ? Or are you claiming that any involvement of software (even if it leaves the high resolution stream bit perfect) unavoidably makes the test invalid, or, in other words, any form of high quality computer based music playback is physically impossible ?

If you have musicians playing in front of you live in real time, with sound :
 
1.) heard live without any technical means (in the vicinity of the microphone )
2.) heard as live analog microphone feed (using IEMs or headphones with very high isolation )
3.) heard as output from recorder (set to different resolutions, or different recorder(s) level matched to 2.) )
      (using same IEMs or headphones as above )
 
Do you really think you would need to ask regarding master (which is being currently recorded in front of you) being the same for 1.) trough 3.)? This is obvious Nonsense #1. The Nonsense #2 would be hearing the difference between 1.), 2.) and 3.) - particularly if 3.) would be 44.1/16.
 
I use  http://ifi-audio.com/portfolio-view/nano-idsd/ - soon to be replaced by http://ifi-audio.com/portfolio-view/micro-idsd/  The micro can play any format that is yet likely to hit the market - let alone the lowly 96/24 .The nano is about half as capable - which is to say that it still matches the formats that any other DAC currently available is capable of.
 
What I say that it is invalid in current computer audio is ABX comparator software that has to convert DSD to PCM in order to be able to perform ABX. If one works in PCM alone, it is OK - trouble is, PCM is not up to DSD. Listening to DSD via PCM (for any reason) is not the same. DSD is bound to have distractors - if nothing else, because it can not be represented on computer screen and dissected in the same manner as PCM. This is also problem with mastering/editing - one has to leave 1 bit enviroment and go to more bits (usually 8 ) in order to edit - so it is also bound to find fierce resistance from those engineers that would like to see every second edited to death - for the sake of note to note perfection also demanded by the musicians and listeners alike. I have heard one of recent(ish) recordings of one of Bartok's concertos by a premier soloist and orchestra took 113 (or was it 131 ? ) takes so that in the end note by note perfection can be pasted together - killing the spirit of a live concert beyond non-existance. Thank you -   but - no, thank you.
 
DSD is something most comparable to direct to disk recording - enabling at least takes of songs one after another ad libitum, not having to perform for the duration of a length of a LP side (say 20 minutes ) in one go, with pauses between songs no more than say 10-15 seconds. It does put less strain on the musicians and engineers while still allowing for sonics comparable to and most likely superior to direct to disk ( analog disk limitations ).
So you are not likely to find many musicians or engineers capable and willing of doing it - because PCM and editing-to-death is sooooo much more convenient, easy - not to remain unmentioned, time in post production is MUCH less expensive than having to do it right in the first place, in a rented recording venue with all the musicians present; sound quality be damned. This is also why multimiking came into existance in the first place - because it allows clever producer to considerably cut the costs of the recording. With the assumption "everything" can be fixed in the postproduction...
 
People have grown accustomed to this "CD photoshoped" perfection that simply does not exist in real life - just the same models we get to see image of never have freckles, dimples etc - all edited out with photoshop. To the point that if someone is daring enough to present his/her true self for others to ask his/hers sanity ... 
 
In a way, native DSD can be regarded as truth-by-technology without the possibility to "photoshop" - IF it remains strictly in DSD 1 bit domain. Editing tools differ by how much of DSD has to be converted to PCM to execute that edit - entire track, just close vicinity around the edit mark, to how many bits should DSD be converted for editing ( the lower bit count lesser quality loss and tougher to perform the edit ), etc - but goal is always to have best SQ possible. 

 

[RRod]

So what exactly would a non-multi-miked, straight from ADC to DAC recording of a live performance captured at 24/192 not have that the DSD recording would have?

 

[stv014]

<wall of text>

 
Well, it does not look like this discussion is leading anywhere, and you do not seem to be willing to actually answer any of my questions, so I just leave it at that.

 

[analogsurviver]

  So what exactly would a non-multi-miked, straight from ADC to DAC recording of a live performance captured at 24/192 not have that the DSD recording would have?

It is still the ringing on the square wave that 192/24 has on BOTH start and end of the semi-cycle - DSD behaves very much like analog, any ringing remaining (mainly from filtering ) being limited strictly to the start of the semi-cycle. 
 
That ringing I hear subjectively as definitive limiting of the acoustics to less than recording venue has in real life. 44.1 is a flat plane, 88.2 has say depth of a few meters and height in vestiges, 192/24 has depth approaching the real size - but definitely NOT QUITE - similar with height. PCM does not have that relaxed feeling as heard live, there is "something" there that is not in the real thing. The difference between DSD64 and DSD128 is most instantly obvious in depth - with DSD128 being VERY close to the real thing. But without that precisely defined feeling of "limited cage/boundary" as with PCM.
 
DSD 128 no longer rings on square wave if left unfiltered - but above audio band noise is still too much if left unfiltered. From this point, it is understandable why DSD 256 ( almost there ) and DSD 512 (most probably there ) - WITHOUT ANY OUTPUT FILTERING. Any real world filtering does ring a bit - and with DSD 64 (SACD) it is absolutely indispensable, with DSD 128 one can get away with murder on certain type of music ( but not all) - by the time we reach DSD 512, both frequency response should be extended beyond what could possibly be regarded as limiting in any way while the noise would be low and far enough from music to be considered inconsequential.
 
DXD, which is a form of PCM, still has ringing both on start and end of a semi -cycle - but due to increased sampling frequency it is far lower in magnitude. The advantage is possibility to edit natively . It is a practical DSD or improved PCM - whatever suits one better.
 
Both of the above is >> 192/24. 
 
One can say the same through reproduction of pulse - with those insanely high sampling frequencies, be it DSD or PCM/DXD, the theorethical ideal does get approached to beyond what the most advanced electromechanical transducer ( loudspeaker or headphone ) is likely to be ever capable of reproducing - even plasma tweeters that work without any mass are limited by the speed with which air molecules can cool down - and that means that it is possible to reproduce the highest frequencies better by a speaker with an extremely light diaphragm driven by a very powerful motor. This allows for very well controlled response without any objectionable high Q peaking to around 200 kHz ( not a typo - in words: two hundred thousands cycles per second ) - at a price.
 
And one should NEVER hear a tweeter on its own - it is there to help everything else, not to steal the show. Removing a well integrated tweeter with response way past 20 kHz from the system after having spent at least say few days with it is not going to be pleasant experience - provided the source material in the first place has recorded information that can be conveyed by such tweeters ( and DACs, amps etc having the support required ).

 

[analogsurviver]

   
Well, it does not look like this discussion is leading anywhere, and you do not seem to be willing to actually answer any of my questions, so I just leave it at that.

I do not know of a better way to ascertain that master is the same than live sound being recorded to whatever format(s) and comparing them level equalized with the microphone feed - it can not get any better than that.
 
With a master from a third party, one can never be 100% sure it was really derived at as claimed. 
It is amusing but not funny - the first truly analog recording of Vienna Symphonic Orchestra in ages : http://www.project-audio.com/main.php?prod=gustavmahlersymphonien1&cat=vinyl&lang=en is available trough Project as LP - and until recently, when Linn discontinued selling downloads of other labels, as download up to 192/24. No DSD as Linn does not have (yet?) DSD player/server/whatever. And no LP - by the once premier analog oriented manufacturer. This goes to show the same recording does get marketed on various formats trough vendors that have commercial reasons to distribute it - including https://www.highresaudio.com/artist.php?abid=74904 For which of the above would you trust to be a "true" master - with the knowledge real raw master recording #00001 is on analog tapes ?
 
I did link the specs for the DAC I am using.
 
I did answer which part of computer audio is still lacking (ABX comparators of native DSD vs PCM ) - which does not deny anything mentioned in your question, of which it is all valid.
 
So it must be communication error, not my unwillingness to actually answer your question. 

 

[RRod]

  It is still the ringing on the square wave that 192/24 has on BOTH start and end of the semi-cycle - DSD behaves very much like analog, any ringing remaining (mainly from filtering ) being limited strictly to the start of the semi-cycle. 
 
That ringing I hear subjectively as definitive limiting of the acoustics to less than recording venue has in real life. 44.1 is a flat plane, 88.2 has say depth of a few meters and height in vestiges, 192/24 has depth approaching the real size - but definitely NOT QUITE - similar with height. PCM does not have that relaxed feeling as heard live, there is "something" there that is not in the real thing. The difference between DSD64 and DSD128 is most instantly obvious in depth - with DSD128 being VERY close to the real thing. But without that precisely defined feeling of "limited cage/boundary" as with PCM.
 
DSD 128 no longer rings on square wave if left unfiltered - but above audio band noise is still too much if left unfiltered. From this point, it is understandable why DSD 256 ( almost there ) and DSD 512 (most probably there ) - WITHOUT ANY OUTPUT FILTERING. Any real world filtering does ring a bit - and with DSD 64 (SACD) it is absolutely indispensable, with DSD 128 one can get away with murder on certain type of music ( but not all) - by the time we reach DSD 512, both frequency response should be extended beyond what could possibly be regarded as limiting in any way while the noise would be low and far enough from music to be considered inconsequential.
 
DXD, which is a form of PCM, still has ringing both on start and end of a semi -cycle - but due to increased sampling frequency it is far lower in magnitude. The advantage is possibility to edit natively . It is a practical DSD or improved PCM - whatever suits one better.
 
Both of the above is >> 192/24. 
 
One can say the same through reproduction of pulse - with those insanely high sampling frequencies, be it DSD or PCM/DXD, the theorethical ideal does get approached to beyond what the most advanced electromechanical transducer ( loudspeaker or headphone ) is likely to be ever capable of reproducing - even plasma tweeters that work without any mass are limited by the speed with which air molecules can cool down - and that means that it is possible to reproduce the highest frequencies better by a speaker with an extremely light diaphragm driven by a very powerful motor. This allows for very well controlled response without any objectionable high Q peaking to around 200 kHz ( not a typo - in words: two hundred thousands cycles per second ) - at a price.
 
And one should NEVER hear a tweeter on its own - it is there to help everything else, not to steal the show. Removing a well integrated tweeter with response way past 20 kHz from the system after having spent at least say few days with it is not going to be pleasant experience - provided the source material in the first place has recorded information that can be conveyed by such tweeters ( and DACs, amps etc having the support required ).

 
Ringing is a manifestation of the duality of frequency and time: to accept a sharp cutoff in one is to accept ringing in the other. But once again, the question of *objective* audibility comes into play. You have quite the catch 22 set up here: you claim you can here things like "less ringing" in DSD, then point out that DSD vs. PCM comparisons are a mess, then assume you must be right. You also simultaneously seem to both accept and reject hi-res PCM as a solution, so you're happy to point out better looking square waves in hi-res PCM but suggest they are still can't possibly be as *audibly* good as DSD, once again falling back on a lack of "good" DSD to PCM converters. What would you consider to be a "good" conversion? My gut is nothing, so what are we even talking about here? It's also odd that you talk about "standing in front of musicians" then rely on a square wave argument, when most live instruments I know of don't put out square waves. The fact is that many of us are pretty sure you couldn't tell a 16/44.1 square wave from even a 24/88.2 square wave in a legit ABX test, so it's hard for us to just accept that you can hear magic happiness in DSD. People who want DSD to not die (unlike Linn, evidently…) pretty much find every possible way to reject PCM, let alone poor old Redbook. And 200kHz is a joke; it reeks of hypertweeter marketing and a desire to kill bats.

 

[bigshot]

   
Well, it does not look like this discussion is leading anywhere, and you do not seem to be willing to actually answer any of my questions, so I just leave it at that.

I'm with you. 

 

[Don Hills]

  Please do refer me to the videos and papers proving that 16/44.1 can resolve time differences into the nanosecond range.
 
That regarding timeline of (non)availability of "digital" sharing a single DAC with consequent phase shift "since 90s" does not hold true. Some computers, notebooks, netbooks etc, as well as external DACs, have been available as recently as 2009 - if not longer. With these, reduction of phase shift is clearly audible AND visible on the oscilloscope by the use of higher than 44.1 kHz sampling frequency in any software, foobar2000 being the most commonly known and used. ...

 
 
Videos and papers:
stv014 has already referred you to the Shannon-Nyquist Sampling Theorem. It is "the law". It says sub-sample timing resolution works. No-one has ever proven it wrong. If you don't believe it, re-read it until you do.
 
You really need to watch Monty's show and tell video from beginning to end, but at about the 21 minute mark he demonstrates what really happens when you delay one channel relative to the other by less than one sample. You see a smooth change in the delay, not a jump from sample to sample. Note that (almost) everything in that video is correct and unarguable, with the arguable exception being his opinion that dither isn't always required.
 
From one of J Robert Stuart's white papers:
Coding High Quality Digital Audio
 

Even among audio engineers, there has been considerable misunderstanding about digital audio, about the sampling theory, and about how PCM works at the functional level. Some of these misunderstandings persist even today. Top of the list of erroneous assertions are:

i. PCM cannot resolve detail smaller than the LSB (least-significant bit).
ii. PCM cannot resolve time more accurately than the sampling period.
...
Regarding temporal accuracy, (ii), if the signal is processed incorrectly (i.e. truncated) it is true that the
time resolution is limited to the sampling period divided by the number of digital levels[2]. However, when
the correct dither is used the time resolution also becomes effectively infinite.
 
[2] e.g. in CD that is represented by the reciprocal of 44100 * 64K

 
Regarding phase shift, your claims are... extraordinary. Phase differences between channels in a system with non-shared DACs will be negligible if the system is competently designed. Any "DAC" with significant phase differences between channels is simply broken. Feel free to take any competent modern DAC, feed it time-aligned ("mono") digital inputs, and produce scope pictures showing the "clearly audible" phase shift you claim exists.
 
 
In a later post, you say:
 

 

   
Is that based on actually analyzing the samples (converting both to e.g. 176.4/24 PCM format, and subtracting the Red Book version should result in no difference other than the higher quantization noise of the CD, and the ultrasonic content of the DSD), or do you just believe it is the case ? Do you even know if the playback levels are exactly the same ?

OK, guys, that does it. This post is CLEARLY meant to imply that anything above 44.1/16 is meaningless.
 
To all computer first/audio second guys (and gals, if any ) - converting, substracting, etc-ing is utter nonsense if you are at location and comparing various digital gizmos to the sound heard live - it IS audible that anything else than live feed from microphones is inferior to it and that magnitude of that deterioration is  sample rate related .
 

 
His post explicitly states that, given competent equipment and operator, the only difference between a 16/44.1 KHz recording and a DSD recording will be that the Red Book version will have a slightly higher noise floor and be missing frequencies above 22 KHz. It only means something if you have (a) screwed up so that the noise flor becomes audible and/or (b) the source had meaningful information above 22 KHz.
 
As for the "night and day" on location, you can easily make most or all of it go away. Simply adjourn to another (sound and vision isolated) room with your monitoring speakers / headphones. Have the desk operator feed you (level matched) the original, or the digital, without telling you which, and switch them back and forth. (There's a bit more to the protocol than that, to avoid out-of-band cues, but the principle holds.) You might just sometimes be able to tell 16/44.1 or DSD-64 from original, but I think you'll fail on 24/96 or DSD128. (A/B testing on "live" music is less accurate than using recorded music, because you have no opportunity to pick out specific musical phrases and repeat until you can identify a difference.)

 

[Don Hills]

  It is still the ringing on the square wave that 192/24 has on BOTH start and end of the semi-cycle - DSD behaves very much like analog, any ringing remaining (mainly from filtering ) being limited strictly to the start of the semi-cycle. 
 
That ringing I hear subjectively as definitive limiting of the acoustics to less than recording venue has in real life. 44.1 is a flat plane, 88.2 has say depth of a few meters and height in vestiges, 192/24 has depth approaching the real size - but definitely NOT QUITE - similar with height. PCM does not have that relaxed feeling as heard live, there is "something" there that is not in the real thing. The difference between DSD64 and DSD128 is most instantly obvious in depth - with DSD128 being VERY close to the real thing. But without that precisely defined feeling of "limited cage/boundary" as with PCM.
 

You hear the "ringing"? Even at 44.1 KHz SR, the "ringing" is at 22.05 KHz. You must have the ears of a bat.
In any case, it is not "ringing". It is Gibbs Effect, explained very clearly in Monty's video that I linked to in a previous post. Strictly speaking, it is the absence of frequencies higher than the Nyquist limit. So you're saying that you notice when frequencies higher than 22 KHz aren't there?

 

[diamondears]

The old "the earth is round" canard. Also not appropriate for sound science.

Yes, I'm saying that legitimate controlled testing to date has demonstrated that humans cannot hear any differences between red book and any of the hi res formats.

Do you have any evidence to the contrary? Are you stating that you believe there are "night and day"/"substantial" audible differences in formats that can't be measured with modern instruments?

As I said, I don't know about the objective measurements, so I don't have any evidence except my own hearing.
 
I agree, based just on my ears again, that hi-res formats doesn't have difference vs red book. But not with DSD, and I hear the difference as very substantial.
 
So, forget other hi-res formats, but with regards to DSD128 or up only, you don't hear any difference vs red book/16/44? I seriously think you need to have your ears checked.

   
Subjective perception can and does lie, depending on the circumstances. Several examples have already been posted (probably more than once) in this same thread.
 
 
It is ironical that the "flat earth" belief is actually the result of naive subjectivism (after all, it obviously looks flat from where one is standing, so it must be true, right ?), yet keeps getting brought up in its defense.

 You didn't get the point at all. The point is---don't conclude that DSD128+ is not better than red book 16/44...keep an open mind, study it more...that's what scientists are for...time to exit for me...not a scientist...lol

 

[analogsurviver]

   
 
Videos and papers:
stv014 has already referred you to the Shannon-Nyquist Sampling Theorem. It is "the law". It says sub-sample timing resolution works. No-one has ever proven it wrong. If you don't believe it, re-read it until you do.
 
You really need to watch Monty's show and tell video from beginning to end, but at about the 21 minute mark he demonstrates what really happens when you delay one channel relative to the other by less than one sample. You see a smooth change in the delay, not a jump from sample to sample. Note that (almost) everything in that video is correct and unarguable, with the arguable exception being his opinion that dither isn't always required.
 
From one of J Robert Stuart's white papers:
Coding High Quality Digital Audio
 
 
Regarding phase shift, your claims are... extraordinary. Phase differences between channels in a system with non-shared DACs will be negligible if the system is competently designed. Any "DAC" with significant phase differences between channels is simply broken. Feel free to take any competent modern DAC, feed it time-aligned ("mono") digital inputs, and produce scope pictures showing the "clearly audible" phase shift you claim exists.
 
 
In a later post, you say:
 
 
His post explicitly states that, given competent equipment and operator, the only difference between a 16/44.1 KHz recording and a DSD recording will be that the Red Book version will have a slightly higher noise floor and be missing frequencies above 22 KHz. It only means something if you have (a) screwed up so that the noise flor becomes audible and/or (b) the source had meaningful information above 22 KHz.
 
As for the "night and day" on location, you can easily make most or all of it go away. Simply adjourn to another (sound and vision isolated) room with your monitoring speakers / headphones. Have the desk operator feed you (level matched) the original, or the digital, without telling you which, and switch them back and forth. (There's a bit more to the protocol than that, to avoid out-of-band cues, but the principle holds.) You might just sometimes be able to tell 16/44.1 or DSD-64 from original, but I think you'll fail on 24/96 or DSD128. (A/B testing on "live" music is less accurate than using recorded music, because you have no opportunity to pick out specific musical phrases and repeat until you can identify a difference.)

Thank you for the links. 
 
It is 2:48 AM here - will return later today. Time to sleep.

 

[jodgey4]

@diamondears Have you considered that perhaps the DAC you're testing with does not have the same performance between PCM and DSD? There are many different methods of implementing DSD.

 

[analogsurviver]

   
Ringing is a manifestation of the duality of frequency and time: to accept a sharp cutoff in one is to accept ringing in the other. But once again, the question of *objective* audibility comes into play. You have quite the catch 22 set up here: you claim you can here things like "less ringing" in DSD, then point out that DSD vs. PCM comparisons are a mess, then assume you must be right. You also simultaneously seem to both accept and reject hi-res PCM as a solution, so you're happy to point out better looking square waves in hi-res PCM but suggest they are still can't possibly be as *audibly* good as DSD, once again falling back on a lack of "good" DSD to PCM converters. What would you consider to be a "good" conversion? My gut is nothing, so what are we even talking about here? It's also odd that you talk about "standing in front of musicians" then rely on a square wave argument, when most live instruments I know of don't put out square waves. The fact is that many of us are pretty sure you couldn't tell a 16/44.1 square wave from even a 24/88.2 square wave in a legit ABX test, so it's hard for us to just accept that you can hear magic happiness in DSD. People who want DSD to not die (unlike Linn, evidently…) pretty much find every possible way to reject PCM, let alone poor old Redbook. And 200kHz is a joke; it reeks of hypertweeter marketing and a desire to kill bats.

To clarify a few things: I did never say that DSD vs PCM comparisons are a mess, I did specifically point only that present ABX comparators - like the one for Foobar2000 most commonly used - can not play DSD natively. One can still blind compare DSD vs PCM with the help of another person and switching of some sort - that could give fair results.
 
Foobar2000 does play DSD natively when not doing ABX - question is how well. My experience is approximately at the level of Korg Audiogate V3.0.x Light Load (free) version- and is bested by both jRiver and Audiogate High Quality (legal only with purchase of Korg DSD device - either recorder or DAC ). This difference is large enough to allow a decent PCM to "slide in between". So, all players are not created equal and can and do influence the outcome. This is another reason why I said computer audio is hard - there are too many setiings in too many versions of software to keep it uniform.  And a faster computer will always have the upper hand.
 
I do not assuring I am right - I wrote how I hear these things. Using other equipment may well change the outcome - I did prefer PCM with an amp that has dislike for high frequency noise of DSD. In that case, DSD did sound fuzzy and mushy - while PCM was "business as usual", in this case yielding better result. But generally I prefer DSD on equipment that can handle it..
 
Square wave difference among various sampling rates of PCM and DSD is quite significant. And although no instrument can reproduce square wave, it is the best signal to use in order to see how well complex waveforms are reproduced. I was forced to start listening ( at safe levels, well below 0 dB ) to  1 kHz square wave used in phono cartridge testing - and anything bad sticks out like a sore thumb. For most phono cartridges, these abberations ( ringing ) lies lower in frequency than it is represented in hirez PCM - but there are some that meet or exceed hirez. Although direct ringing at 20 kHz + is not directly audible, it can and does creep in trough noise and intermodulation - both within audible range. 
 
I will load tomorrow 1 kHz square wave recorded from analog signal generator with DSD128 recorder and compare PCM made from this in 44.1, 88.2 and 176.4 sampling rate, as these can be ABXed in foobar 2K. Or I can make direct recording in above three resolutions of PCM - will report how it will go. For comparison with DSD I would have to ask another person to operate the switch etc.
 
There is one record I have seen on oscilloscope to resemble producing a signal extremely close to a square wave :  http://www.discogs.com/Thelma-Houston-Pressure-Cooker-Ive-Got-The-Music-In-Me/release/1229176 It would be interesting to compare it with the later issue made from analog tape and not direct to disc : 
http://www.discogs.com/Thelma-Houston-Pressure-Cooker-Ive-Got-The-Music-In-Me/release/3072422 - both by ear and on an oscilloscope. I forgot exactly which song(s) are "close to square wave" - it has been 20 + years I have been curious enough to hook up an oscilloscope for musical signals after hearing things uncommon in normal records made from analog tape masters. I currently posses no version of this recording - except for CD.
 
Bats always creep up whenever talking of beyond 20 kHz response in audio. And I have no desire of killing any innocent beings.