[spruce music]

  The idea that you can't hear below a noise floor is a poor one. The possible advantage of HiDef is more about timing than acoustic bandwidth or measured noise floor and probably why I prefer it. Remember thresholds of what is considered audible was done with listening tests of single tones before solid state. The same ones the stated 1db is the threshold of difference when we know 1/2 db used as EQ in something good is repeatably distinguishable. I've done it. What happens in music is much more demanding and complex and it's the perception of that complexity that makes our hearing sense quite special. 
 
Here's an interesting take on it: http://slidegur.com/doc/180896/hugo-dac-technical-master-class Consider that if in complex material as opposed to simple measurements, not everything times perfectly when your ears are actually this sensitive to time, which they are. It would give the sound a flatter, hashier perspective. Scratch your fingers together in front of the bridge of your nose. Now move it a fraction of an inch one way or the other. Very easy to hear. 

Oh NO! The garbage ideas that won't die.  Timing of undithered 16 bit redbook.  Accurate to 56 picoseconds.  Less than that with proper dither. So 4 microseconds is no sweat.  Other garbage myths that won't die.  Transients.  Transients perceptibility are determined by frequency response.  For the human ear steady state, complex signal any kind of signal it does not extend to the point you need extremely high sample rates.  The stair step output sample.  So sorry, a myth.  Looking at 20 khz from a CD will show as smooth sine wave, no stair steps. Increased sample rate 20 khz sine waves....same smoothness.  No increased smoothness or accuracy from the higher sample rates.  Sorry, you were taken in by advertising hype and believed it.

 

[watchnerd]

  The possible advantage of HiDef is more about timing

 
Huh?
 
Please explain this in terms of Nyquist theorem.
 
Because it makes no sense at first glance how increased bit depth or higher sampling rates affect timing, which is a clock function and where errors lead to jitter.

 

[StanD]

  But the latest fad is R2R DACs and they're only available in 16 bits. That means 24 bit DACs are out of fashion.

 
 

   
While I'm not a believer in the alleged inherent superiority of R2R, that's not actually factually correct.  For example, the AD5791 used in the Schiit Yggy is a 20 bit chip (+1 LSB).

I checked the spec sheet, only the 14 LSBs are R2R and the 6 MSBs are decoded using 63 matched resistors and 63 switches. Weird,

 

[watchnerd]

   
 

I checked the spec sheet, only the 14 LSBs are R2R and the 6 MSBs are decoded using 63 matched resistors and 63 switches. Weird,

 
Wow...that is weird...I've never heard of that, nor sure why one would build it that way...

 

[StanD]

   
Wow...that is weird...I've never heard of that, nor sure why one would build it that way...

Just imagine how many resistors and switches would be required if the entire DAC used the same method as the upper MSBs? The chip would be the size of a car.  I wonder what the R2R worshipers would think of this? We could start a new fad.

 

[watchnerd]

  Just imagine how many resistors and switches would be required if the entire DAC used the same method as the upper MSBs? The chip would be the size of a car.  I wonder what the R2R worshipers would think of this? We could start a new fad.

 
Of course, forgot about the physical size issues, and that they're not transistors.
 
But I think you're hit on something with the car-DAC.

 

[gregorio]

  [1] The idea that you can't hear below a noise floor is a poor one.
[2] The possible advantage of HiDef is more about timing than acoustic bandwidth or measured noise floor and probably why I prefer it.
[3] Remember thresholds of what is considered audible was done with listening tests of single tones before solid state.
[4] Here's an interesting take on it: http://slidegur.com/doc/180896/hugo-dac-technical-master-class

 
1. I never said one can't hear below a noise floor! In certain circumstances, say a pure tone in the critical hearing band and a digital noise floor which has been "shaped" away from the critical hearing band, then it is possible to hear many dB, even many tens of dB below the noise floor. However, you've unfortunately missed the point because I wasn't talking about tones or shaped noise, I was talking about white (or close to white) noise. The close to white noise floor of the recording, the close to white noise floor of the listening environment and the white digital noise floor of 16bit. It is impossible to distinguish/hear a white noise floor beneath another white noise floor which is many times higher in level. This is both common sense and something you can easily test yourself, so just saying it's "a poor idea" is not even close to being enough here in the science forum. If you're going to make such an extraordinary claim, you're going to have to provide some fairly extraordinary evidence, otherwise, you're just going to come across as someone who doesn't even have a grasp of basic common sense!
 
2. What makes you think HiDef has any more timing accuracy than 16bit or that human awareness/perception of timing accuracy exceeds the accuracy achieved by 16bit? Is it just that marketing material to which you linked and presumably some other marketing material which suggests roughly the same thing? If so, the obvious implication of your statement is that you prefer Hidef because of marketing material! Just to put what Spruce Music stated into context, in case you didn't understand it: The 4us threshold (used in the marketing material) = 4,000ns = 4,000,000ps. At 56ps, the humble CD is therefore capable of a level of timing accuracy which is roughly 70,000 times below the threshold of audibility quoted in the marketing material! This fact obviously contradicts their marketing aim, so they hide it by not mentioning the timing accuracy of CD and try to hoodwink gullible audiophiles by confusing sampling period with timing accuracy.
 
3. True but if you're going to mention this then you need to be honest and also "remember" that listening tests of audible thresholds did not just stop at the advent of solid state but continued up to the present day and that these tests did not only use single tones, they also tested with music and with signals even more complex than music, noise for example.
 
4. Interesting, yes, but only from the point of view of seeing just how far audiophile marketing material is willing to pervert science in order to con gullible audiophiles! The solution to this common problem is quite simple, don't be so gullible! The majority of the time being less gullible really isn't very hard to achieve. In this particular case, just a modicum of common sense and a few basic facts would quickly expose the linked document as just marketing BS rather than the "technical master-class" it's pretending to be. For example, the first few pages go on about science's lack of understanding of hearing perception, that psychoacoustic models are therefore effectively useless and that the solution to all this is "lots of carefully designed listening tests" (p.5). One simple fact proves this is BS; psychoacoustic models are all actually based on "lots of carefully designed listening tests"!! And, science's tests are almost guaranteed to be more "carefully designed" than Chord's tests, plus science has definitely done "lots" more of them. The people at Chord apparently understand more than the whole rest of the world of science and have created chips which can apparently do what "no computer yet designed can do". If this were true, how come those people at Chord are not splashed all over the news being showered with Nobel prizes for advancing scientific understanding AND for advancing computer design? There are a number of other examples which don't need more than a bit of common sense and a basic fact or two to debunk, look at the two last points on page 12 for example. And with a bit more knowledge, most of the rest of the document is also easily recognisable as nothing more than marketing BS. The document should be titled: Hugo-DAC-Marketing-Master-Class! Although to be honest, I've seen better marketing BS than this example.
 
While I can see how some of the more technical BS could have slipped past you, how did you manage to miss the far more obvious BS which only needed a bit of common sense to recognise?
 
G

 

[goodvibes]

Oh NO! The garbage ideas that won't die.  Timing of undithered 16 bit redbook.  Accurate to 56 picoseconds.  Less than that with proper dither. So 4 microseconds is no sweat.  Other garbage myths that won't die.  Transients.  Transients perceptibility are determined by frequency response.  For the human ear steady state, complex signal any kind of signal it does not extend to the point you need extremely high sample rates.  The stair step output sample.  So sorry, a myth.  Looking at 20 khz from a CD will show as smooth sinke wave, no stair steps. Increased sample rate 20 khz sine waves....same smoothness.  No increased smoothness or accuracy from the higher sample rates.  Sorry, you were taken in by advertising hype and believed it.

Could be but I absolutely hear it and passed blindtests convincingly. Perhaps I'm trying to find a reason but it's not a bad idea. Just adding to the discussion. Feel freeto be rude. I've nothing to prove. You guys can have your ball back and toss it around the circle. I also think EVERY USB dac is worthless for hearing significant differences. Fine for making the noise from a pic more palatable but thanks t's about it. You can now hate on me some more. Ciao.

 

[analogsurviver]

   
1. Of course, because the actual ("truthful") level of a standing ovation is much lower than the level you would have to replay the recording in order to hear the digital noise floor of 16bit. Did you actually read what you are responding to?
 
2a. OK, that's nonsensical. You think record labels should make recordings for just the say 0.01% of consumers, that tiny number of extreme audiophiles? You do realise that record labels are commercial businesses and not audiophile charities?
2b. We're not talking about squashing an orchestra, we're talking about reducing transient peaks which, as you say, last for tiny fractions of a second and it's inaudible if they are reduced somewhat. Why is reducing these transients a mastering practice? To make the music sound realistic!! There's nothing "realistic" about not being able to hear quiet sections of the music, so what you're saying is the absolute reverse of reality. And, even if we don't reduce those transients, 16bit still has more than 10 times the dynamic range required. So what you're talking about is the aims of mastering, NOT any supposed limitations of 16bit!
2c. Replay gain has nothing whatsoever to do with replaying at the level intended by the artist, so this statement is also nonsense!
 
3. No it's not, that's nonsense. Firstly, you now appear to be confusing transient peaks with musical dynamic peaks. Secondly, not even a student mastering engineer would "master out" a 10dB musical peak in classical music, let alone a practising professional!
 
4. Unfortunately though you've failed, as you are taking it out of context. Maximum musical dynamic at the conductor's position is roughly up to about 110dB or so. A concert hall has a noise floor of probably at least 50dB. The dynamic range is therefore about 60dB, that's about 36dB less than what 16bit is capable of!! If you're talking about a "truthful" recording then it HAS to include the noise floor of the performance venue, which is way, way higher than the digital noise floor.
 
5. That depends. Many cinema sound systems cost $100k or more and can only just manage a dynamic range of 60dB.
 
6a. Maybe not an arm and a leg, just a pair of ears!
6b. I realise that extreme audiophiles have either no concern for the realities and limitations of human hearing or simply no understanding of them but it would be incompetent and negligent for a record label to release a recording where to hear the quiet sections means replay the recording at dangerous levels.
6c. The instructions would be simple: If you want to continue to hear normal conversations, NEVER set your headphones to peak level (0dBFS) equal to 120dBSPL!!
 
7. For hundreds of years, since the very dawn of modern science, science has had absolutely zero affect on the limitations of human hearing. 16bit is both beyond the ability of technology to fully reproduce but more importantly, beyond the limitations of human hearing.
 
Please, no more nonsense and especially no more nonsense which advocates injury!!!
 
G

Oh dear... do I REALLY have to go to a CD store, grab - say - 10 of the recently released classical recordings at random , endure the chore of ripping them - and then post the result on audition or whichever PCM editor ? I started because of  rejecting the SQ of CDs recorded around the turn of the millenium onwards - the newer they were, the less realistic they sounded ! 
 
Now, TBH - how many of headphones , even those considered to be high end, CAN in fact play back an uncompressed recording of a piano - or even worse, live microphone feed - at realistic SPL ? Say the same SPL as measured in say 5th row of parter ? And then, how many loudspeaker systems can do the same ? 
 
Because record companies are releasing what they are, the percentage of equipment actually capable of doing it may well be in the 0.01%ish  range mentioned above.  And that needs to change - no more massive bass reductions ( which is the very first thing to sacrifice ) - or, releasing two versions of the same recording; one truthful, another stated as being compressed/EQed so that lesser equipment can play it with satisfactory result.
 
I agree regarding the noise floor of the venues ; the newer they are, the noisier they are. Lighting alone can be enough, in case where "climatic" devices add their drone in the LF, it can be unbearable for any serious recording. Actually, I found that recording with lower resolution in such troublesome environment actually produces better result - sometimes all the way to using MP3 as an original master medium ! But, I will go to any length to avoid to have to record in noisy environment - and then will use whatever maximum resolution system available.
 
16 bit may or may not be enough for playback - but it is nowhere good enough for recording. I did record a few CDs using CD-R, which, according to its specs, was recording with 14 bits. And, I found that pushing the recording level as far as it would go without clipping was extremely desirable - because a single dB reduction of recording level had drastic reduction of SQ.  Now, it did give me the skill to work much in the same way using better recording devices, including DSD - where it keeps the level of the recorded sound ratio to ultrasonic noise of the DSD as high as possible. In PCM, by going to 24 dB, one can have headroom of more than 10 dB while still having better resolution than 16bit recording driven close to 0dBFS. RBCD also pushes under the carpet any ultrasonic noise of ADCs and DACs - everything above 22.1K gets chopped off - end of "problems". 
 
However, I DID find out that converting DSD128 to PCM 192kHz/32bit floating point can bring an uncannyly low noise floor - which I have yet to hear on any other medium. Compared to that is RBCD - noisegenerator. I particularly remember one organ recording; during the pauses, the sheer sensation of the acoustics of the church was great on master DSD128, unbelievable on 192/32 - and, compared to the former two, poor on RBCD. 
 
Now, RBCD may well sound fine if one never gets the exposure to the former two ... - or enough live music. 
 
On purpose I did not mention response above 20 kHz as being mandatory for quality recording. Regardless whether we can or can not hear pure sine waves above certain frequency - we can "perceive" "somehow" these frequencies ( there were studies and more are needed to bring down the myth "20 kHz is enough" - for good ) - without them it just is not realistic enough. It may be timing, it may be I-do-not-know-what - but I know I regretted with all my heart the recently released harpsichord recording could not yet be mastered in DSD ( Pyramax DAW, it goes from 1 bit to 8 bit only very close to the cut, meaning >99% of the resulting master recording will still be native DSD ) but in 192/32 and then released on CD. It is still very good - but the true charm of the harpsichord that has output (on this recording ) up to approximately 45-50 kHz ( one has to convert DSD into 192/24 or 32, in order to be able to use spectrum analyzer up to half that frequency, 96 kHz ) ,  is diminished. Next time, I hope DSD DAW to be fully operational - so that true DSD downloads can be made available. 
 
If you have read the above correctly, there was nowhere any mention of any loud sounds. True, organ can get loud - but harpsichord is a quiet instrument, even close up, even in full cry. Going above RBCD 16 bit 44.1kHz limitations brings MUCH more believable reproduction - of both.
 
Yes, those > 20 kHz are WAY, WAY down in level - but they DO matter. Yes, I know it not only sounds, but IS expensive - but centuries of musical instrument making and development have produced what we have today and it is my firm belief that if the technology allows for it, it is not excusable not to use it to preserve as much of the original sound as possible.
 
Back in 50s, they used whatewer best they could ( including optical tape ) hold their hands on; that is why we are still discovering, using ever more sophisticated vinyl playback devices, what is truly lurking in those grooves. In 50 years time, no one could possibly claim the same for RBCD . While for sheer reproduction 16 bits ( after all the work in the studio with greater bit depth ) may be enough,  44.1kHz sampling most definitely is not.

 

[spruce music]

couldn be but I absolutely hear it and passed blindtests convincingly. Perhaps I'm trying to find a reason but it's not a bad idea. Just adding to the discussion. Feel freeto be rude. I've nothing to prove. You guys can have your ball back and toss it around the circle. I also think EVERY USB dac is worthless for hearing significant differences. You can now hate on me some more. Fine for making the noise from a pic more palatable but thanks t's about it. Ciao.

What kind of blind tests?  You hear it (what?).  Do you mean a difference in timing? Sorry for being rude.  I have seen this misguided idea about the length of sample periods being the limit of timing with digital thousands of times.  It was wrong the first time and still is.  If you hear something you think is due to timing that is one thing.  I can tell you however, whatever it is you are hearing is not due to timing limitations of even CD quality levels. 
 
Same goes for those stair step illustrations.  Amazing that one still gets used.  Amazing that people believe it.  If you have never watched it, watch this video.
https://xiph.org/video/vid2.shtml
Here using quality analog signal generators and analog oscopes and spectrum analyzers they show what an old cheap AD/DA converter does to high frequencies.  The result is they look just like the analog signal they put into it at the output. So watch it and learn.  Unless you think they faked the video, next time some marketing BS masquerading as a white paper or master class tells you about the uneveness of sine waves from low rez digital and shows you stairsteps, you will know they are lying to you.
 
And yes, your complaint about USB being worthless, let me guess, just by the marketing around such, it is about noise from the PC leaking into and corrupting the DAC right?  Yeah, right.  You really need to get better sources for technical subjects.

 

[gregorio]

[1] couldn be but I absolutely hear it and passed blindtests convincingly. [2] Perhaps I'm trying to find a reason but it's not a bad idea. Just adding to the discussion. [3] Feel freeto be rude. I've nothing to prove.

 
1. Are you really saying your hearing is 70,000 times more acute than the article you posted says is the threshold of human hearing? Does that mean you are claiming not to be human or some sort of super-human? 
2. What, are you saying it's not a bad idea to seriously tell the world that you're super-human?
3. Of course you've got something to prove, this is the science forum. If you're going to make outrageous claims then you need to back them up (and not with marketing BS!!). You come here and are rude to this forum's members by posting BS marketing and making BS claims and then you get all b*tthurt when the forum's members are rude in return? Do you have the worldwide monopoly on rudeness, only you are allowed to be rude?
 

  [1] Oh dear... do I REALLY have to go to a CD store, grab - say - 10 of the recently released classical recordings at random , endure the chore of ripping them - and then post the result on audition or whichever PCM editor ?
 
[2]16 bit may or may not be enough for playback - but it is nowhere good enough for recording. [2a] And, I found that pushing the recording level as far as it would go without clipping was extremely desirable - because a single dB reduction of recording level had drastic reduction of SQ.  
[3] Now, it did give me the skill to work much in the same way using better recording devices, including DSD - where it keeps the level of the recorded sound ratio to ultrasonic noise of the DSD as high as possible

 
1. You could try but how would the result in a "pcm editor" support all your claims? There is in fact a much better alternative, STOP spouting your nonsense in the first place, it's getting to the point of trolling!
 
2. Name any music recording mics which have a dynamic range greater than 96dB.
2a. This is an example of not only being wrong but of being the exact opposite of correct. Making a recording as hot as possible does not improve SQ, it reduces it. That's one of the first lessons any recording student is taught.
 
3. Another example of not just being wrong but being as far from correct as imaginable. DSD does the exact opposite of what you suggest, it has an absolutely terrible signal to ultrasonic noise ratio, due to the severe noise shaping which is essential to making DSD work. You would know this if you had even a newbie level of understanding of how DSD works!
 
I could carry on explaining why the rest of the points in your post are utter nonsense but why bother? All you'll do is just is just ignore the explanations and post another whole bunch of nonsense. You obviously have not even read the OP, let alone understood any of it. If you had, you would realise there is no resolution difference between 16bit and 24bit in terms of SQ, only in terms of an inaudible noise floor. If you disagree then fine, present your argument WITH some corroborating evidence but DO NOT just keep posting your off-topic suppositions, especially as they're all utter nonsense anyway! What you are doing is trolling, SO STOP!
 
G

 

[goodvibes]

 
What kind of blind tests?  You hear it (what?).  Do you mean a difference in timing? Sorry for being rude.  I have seen this misguided idea about the length of sample periods being the limit of timing with digital thousands of times.  It was wrong the first time and still is.  If you hear something you think is due to timing that is one thing.  I can tell you however, whatever it is you are hearing is not due to timing limitations of even CD quality levels. 
 
Same goes for those stair step illustrations.  Amazing that one still gets used.  Amazing that people believe it.  If you have never watched it, watch this video.
https://xiph.org/video/vid2.shtml
Here using quality analog signal generators and analog oscopes and spectrum analyzers they show what an old cheap AD/DA converter does to high frequencies.  The result is they look just like the analog signal they put into it at the output. So watch it and learn.  Unless you think they faked the video, next time some marketing BS masquerading as a white paper or master class tells you about the uneveness of sine waves from low rez digital and shows you stairsteps, you will know they are lying to you.
 
And yes, your complaint about USB being worthless, let me guess, just by the marketing around such, it is about noise from the PC leaking into and corrupting the DAC right?  Yeah, right.  You really need to get better sources for technical subjects.

 You can disagree but your assumptions of my experience are completely erroneous. 
 
Tried dozens of USB DACs including DAVE and DCS via asio or wasapi and Wavelab, Foobar, J River and some Hi-End memory buffer type players. Found Wavelab setup correctly sounds the best to me though not really friendly as a library player. None sounded as good via USB as they did with a Firewire interface of a Weiss INT 202 or Konnekt interface (either linear supplied)(Dave wouldn't interface with the Weiss for some reason but we didn't troubleshoot very long. Those don't sound as good as what I can get streaming from a selected dedicated server and steamer. I'm plenty experienced so chill. I work with an award winning recording engineer and constantly hear transfers of both analog and digital sources and have been at the venues during the process. Not our main thing of which includes some tech repair. Mostly simple mic'd 2 track acoustic recordings in real space. We are always trying to improve the PC interface to have a better presentation when editing. Quality or price of equipment isn't an issue. Difficult to prove audibility on the interwebs. Measurements are always lacking IMO. Never told me much about the sound of anything. That we can disagree on and perhaps it's the bridge you can never cross but in a science forum, I think it would be better to find out why something exists that to simple dismiss its existence because it cannot be explained to your satisfaction. That existence is based on experience as you stated and ours are obviously different. I can't change you mind without a demonstrations so perhaps we could agree to disagree but I suspect you can't accept that. I truly believe you are as misguided as you think I am.
 
I know how this goes, you don't accept anything by anyone that doesn't want to play by your specific rules and continue to attack to get someone defending himself when it shouldn't be part of a discussion. I'm out. Have at it and remember that when you stick your head in the sand, those ears need a good cleaning.

 

[analogsurviver]

   
1. Are you really saying your hearing is 70,000 times more acute than the article you posted says is the threshold of human hearing? Does that mean you are claiming not to be human or some sort of super-human? 
2. What, are you saying it's not a bad idea to seriously tell the world that you're super-human?
3. Of course you've got something to prove, this is the science forum. If you're going to make outrageous claims then you need to back them up (and not with marketing BS!!). You come here and are rude to this forum's members by posting BS marketing and making BS claims and then you get all b*tthurt when the forum's members are rude in return? Do you have the worldwide monopoly on rudeness, only you are allowed to be rude?
 
 
1. You could try but how would the result in a "pcm editor" support all your claims? There is in fact a much better alternative, STOP spouting your nonsense in the first place, it's getting to the point of trolling!
 
2. Name any music recording mics which have a dynamic range greater than 96dB.
2a. This is an example of not only being wrong but of being the exact opposite of correct. Making a recording as hot as possible does not improve SQ, it reduces it. That's one of the first lessons any recording student is taught.
 
3. Another example of not just being wrong but being as far from correct as imaginable. DSD does the exact opposite of what you suggest, it has an absolutely terrible signal to ultrasonic noise ratio, due to the severe noise shaping which is essential to making DSD work. You would know this if you had even a newbie level of understanding of how DSD works!
 
I could carry on explaining why the rest of the points in your post are utter nonsense but why bother? All you'll do is just is just ignore the explanations and post another whole bunch of nonsense. You obviously have not even read the OP, let alone understood any of it. If you had, you would realise there is no resolution difference between 16bit and 24bit in terms of SQ, only in terms of an inaudible noise floor. If you disagree then fine, present your argument WITH some corroborating evidence but DO NOT just keep posting your off-topic suppositions, especially as they're all utter nonsense anyway! What you are doing is trolling, SO STOP!
 
G

I am a regular human being with particular bias on sound quality. 
 
No, I do not have monopoly on anything - and would react as I did only if treated like I have been by the members on sound science thread(s). 
 
1. PCM editor would go to show how compressed are commercial recordings ( classical, from premiunm brands, most of the time including audiophile labels ) compared to something that was allowed to be left intact. It would have been visible from across the room !
 
2. I will have to look up for the exact model #, but there is a Neumann mike with noise low enough to be on the same order of magnitude as air molecules impigning on its membrane. It can not get better than that.  It is covered in the Neumann book ( Jubilee X0 years ? ) from approx 5 years or so ago. No, I do not have it or have seen it in flesh so far - but it does exist. This means it has a good chance of exceeding the 96 dB dynamic range. There might be others as well - but I am not familiar with them.
 
2a. Not true in all cases. Certainly not in the one cited. I am aware if and when pushing the level is detrimental to SQ - and act accordingly. It usually has to do with less than optimal analogue stages; ADC and DAC work just fine, only to "highlight" the cost cutting measures of the analogue parts of the recording chain.
 
3. I know VERY well how DSD works. By recording as hot as it goes the CONSTANT ULTRASONIC NOISE FLOOR is kept as low as possible - that much you should understand. Ultimately, it will take DSD256 - or even DSD512 - to allow for > 100 dB S/N up to at least 100 kHz ; with DSD64 and less so, DSD128, the ultrasonic noise can quickly become a problem if the recording level is low(er) than it could be. IIRC, you get 6 dB lower ultrasonic noise floor for each doubling of the sampling frequency, which also starts twice higher in frequency compared to half the sampling rate frequency. It is a tradeoff . Korg recorders I use (with TI ADCs ) are noise -wise decent to approx 50 kHz, then the noise starts going up - regardless if it is DSD or PCM mode(s). With faster DSD ( better processors ) , the need to push levels should get reduced. I am eyeing Mytek Brooklyn ADC ( a recorder ) at the moment; DSD256. But will only go for it if I get a spectral analysis of its actual performance up to at least 100 khz. Last resort is Mytek*s 30 days return policy - but I would prefer to know this important spec/fact in advance.
 
I think that terminology used as well as the fact that English is not my native language can at least partly be blamed for so heated a debate. I guesss if we could get together in a GOOD listening room with GOOD equipment, you could hear for yourself what I am "trolling" about.  The trouble is that really good equipment is really scarce - one is almost forced to beg/borrow/steal/modify/design/whatever by him/herself. No amount of real or virtual ink spent can replace a few seconds of audio that goes beyond the "officially accepted standard to be good". I do not get positive feedback from musicians and listeners alike for nothing. And I have been honest enough to post a recording I was clearly not satisfied with - and stating how and why it happened. You have latched on this one as if it were my best recording. You can disagree with me, you can call me crazy - that is OK - but I AM sincere and most definitely NOT trolling.
 
The same goes for the moderators - ANY, each and every one of them. 

 

[StanD]

So what's next a new audio high speed optical link for audio? Is anyone claiming to be able to hear RF yet? Not TOS, that as slow as pudd'in.

 

[castleofargh]

 

Oh NO! The garbage ideas that won't die.  Timing of undithered 16 bit redbook.  Accurate to 56 picoseconds.  Less than that with proper dither. So 4 microseconds is no sweat.  Other garbage myths that won't die.  Transients.  Transients perceptibility are determined by frequency response.  For the human ear steady state, complex signal any kind of signal it does not extend to the point you need extremely high sample rates.  The stair step output sample.  So sorry, a myth.  Looking at 20 khz from a CD will show as smooth sinke wave, no stair steps. Increased sample rate 20 khz sine waves....same smoothness.  No increased smoothness or accuracy from the higher sample rates.  Sorry, you were taken in by advertising hype and believed it.

Could be but I absolutely hear it and passed blindtests convincingly. Perhaps I'm trying to find a reason but it's not a bad idea. Just adding to the discussion. Feel freeto be rude. I've nothing to prove. You guys can have your ball back and toss it around the circle. I also think EVERY USB dac is worthless for hearing significant differences. Fine for making the noise from a pic more palatable but thanks t's about it. You can now hate on me some more. Ciao.

I'm not trying to blame you for getting a little mad, as you're answering a post that wasn't exactly gentle. so while I'd like it to stay civilized, because TOS, IMO it's 1-1 and all is good. but I just want to say that the "I've got nothing to prove" really doesn't belong to this subsection.
we're discussing technical stuff and audibility, not ego and charisma.
so instead of getting mad, if you really care about finding out why you're getting differences that others fail to notice, explain your setup,  and maybe your settings for the blind test.
about USB, I'm limited to my own reading on the matter, but I've seen numerous times mentioned that async USB was usually the best option for timing. so unless all the stuff I've read was wrong, you might need to reconsider your argument about timing being the strong point of highres and what makes the differences audible. or stop trash talking usb, but both at the same time doesn't seem logical.
 
edit:
ps: @goodvibes weren't you the one saying that panning wasn't a problem for you on IEMs in the er4 topic when I mentioned that I couldn't stand listening to anything mastered for speakers without at least some matter of crossfeed? does that make sense to you to notice and favor devices and streaming solutions for stupidly small timing reasons, and favor highres while finding that there is no need to get a proper ITD and proper panning in general? did I get confused somewhere about the idea of fidelity?