One of the reasons for converting CD rips to DSD appears to be because DSD is inherently more linear for conversion back to analog. In other words, dacs that can operate in either PCM mode or DSD mode to play back CD audio may sound better when operated in DSD mode. What it suggests to me is that there is a problem is in the D/A conversion stage, not necessarily earlier in the reproduction chain.
Attachments
Well ESS DACs and probably most others convert DSD back to high sample rate PCM, then filter it, which turns it into multibit, and pipe it through the same multibit DAC as the usual PCM, so... uhmmmmm
AKM dacs such as AK4499 have a volume control bypass option for DSD mode that avoids remodulation. Even with ESS dacs, back a few years ago when I was experimenting with ES9038Q2M it sounded better to me and others in DSD mode during A/B listening comparisons. Anyone can try HQ Player for free in demo mode with an ESS dac to see what they think about PCM verses DSD. Roon also has a fairly good DSD modulator, although not quite as good as the best of HQ Player.
Regarding a need to convert DSD to PCM and then back to DSD, some work as been done to look at how some of that might be avoided.
I probably did, but didn't store any pictures of it, so I just generated two:
10 kHz, amplitude = 0.5 (maximum intended level), DFT bin size 1 kHz, Hann window:
The peak only goes to 0.25, presumably because of the way I implemented the Hann window.
Now with 3 kHz, amplitude = 0.5:
Instead of grasping for straws why not settle for the more obvious reason: you prefer the non-linearities and other colorations of your DAC in DSD mode.
Well, it's ones statements of ones personal experience that is commented on. If one don't want these comments, perhaps one (sick!) should not post the personal experiences...
//
Last edited:
Nonsense. By that reasoning if someone posts measurements and someone else doesn't happen to like what the measurements show then it would follow that its fine to attack the person who reported the measurements rather than to discuss the measurement results in a civil way. That can't be right, it would also be ad hominem.
Measurements can be in error. Measurement can be of the wrong thing. Measurements can be misinterpreted. No matter what, there is always a person involved somehow. Attacking the person amounts to ad hominem:
ad ho·mi·nem
/ˌad ˈhämənəm/
adjective
(of an argument or reaction) directed against a person rather than the position they are maintaining.
"vicious ad hominem attacks"
adverb
1.
in a way that is directed against a person rather than the position they are maintaining.
"these points come from some of our best information sources, who realize they'll be attacked ad hominem"
2.
in a way that relates to or is associated with a particular person.
"the office was created ad hominem for Fenton"
https://www.google.com/search?q=ad+...j0i10i512l8.4843j0j1&sourceid=chrome&ie=UTF-8
ad ho·mi·nem
/ˌad ˈhämənəm/
adjective
(of an argument or reaction) directed against a person rather than the position they are maintaining.
"vicious ad hominem attacks"
adverb
1.
in a way that is directed against a person rather than the position they are maintaining.
"these points come from some of our best information sources, who realize they'll be attacked ad hominem"
2.
in a way that relates to or is associated with a particular person.
"the office was created ad hominem for Fenton"
https://www.google.com/search?q=ad+...j0i10i512l8.4843j0j1&sourceid=chrome&ie=UTF-8
Maybe the phrasing could have been friendlier, but post #307 seems to be an alternative hypothesis rather than a personal attack.
It could be taken as someone's conjecture. Did occur to me to suggest rephrasing the general idea in a more impersonal way. However IMHO maintaining civility in online forums is an overriding issue in the big picture. Didn't want to dilute that point so kept the response narrowly focused.
Regarding such a conjecture as to preference, it is an idea that's been around for a long time in the audio world. Have written about what is likely wrong with it, but happy to discuss further if we can keep it civil.
Regarding such a conjecture as to preference, it is an idea that's been around for a long time in the audio world. Have written about what is likely wrong with it, but happy to discuss further if we can keep it civil.
Last edited:
So suggesting that Markw4's preferred DAC in DSD mode has non-linearities or colorations is now considered a personal attack 
All audio reproduction is imperfect. To suggest that someone prefers more distorted sound over less distorted sound is probably wrong for most people, including me. A lot of the problem may have to do with how distortion is typically measured and how those measurements may be misinterpreted. It may be recalled that typical FFT measurements have limited predictive value in terms of perceived SQ. Yet once a spectrum is looked at it may seem to explain everything about preference. Its WYSIATI. Its looking under the streetlamp for the car keys. Its a cognitive fallacy. Do we really need to rehash all of that here? And now that modulator imperfections are being measured, does it follow that anyone likes them? Or is it they are the lesser evil considering other tradeoffs?
Possibly consider that I could have responded by suggesting that Mr. Bohrok is the one truly grasping at straws, or perhaps more correctly, grasping at time-averaged steady-state FFT spurs. Where would that have gotten us to?
Possibly consider that I could have responded by suggesting that Mr. Bohrok is the one truly grasping at straws, or perhaps more correctly, grasping at time-averaged steady-state FFT spurs. Where would that have gotten us to?
Contrary to you I have never made any claims about sound. Measurements (including FFT) clearly and objectively pinpoint faults such as non-linearities or high noise which make these measurements valuable. Instead claiming that a device failing in basic measurements sounds good is an indication that you prefer non-linearities and colorations.