Digital Vinyl: Superior Sound Compared To Just Digital?

[Al M.]

This is a continuation from the discussion on this page from the thread:

dCS Varese short review

Interesting point. And I have pondered it. I am in the fortunate position that I can compare similar SACD and vinyl issues back-to-back. A company by the name of TRPTK makes state-of-the-art digital recordings. Its recording techniques, and resulting recordings, are superior to anything I know...

www.whatsbestforum.com

 

[Al M.]

Everybody is correct.

During the mastering process of both digital and analogue, choices must be made. I call that "introducing distortion". By the way, mastering a digital recording is different from mastering an analogue recording, which says it all. The point is: analogue is a continuous signal, similar to what the human ear has been trained to listen to for millennia. Digital is not.

This is not how it works. The output from a DAC is no less continuous than analog: It is a signal with an analog waveform. There are no discontinuities such as "stairsteps" or the like: Stair steps are a myth.

If digital vinyl (vinyl sourced from digital recordings) sounds better to some than just digital, this is not because the waveform is more continuous. Transfer of digital to vinyl does not miraculously make it any more "continuous" than the signal at the output of the DAC itself (already an analog waveform) that was used to create the master from which the vinyl is cut.

Sony/Philips's choice of a "standard" 44.1 kHz digital sampling frequency proved to be a mistake, notwithstanding the theoretical background that justified the choice. I remember listening to the first cd players back in 1982(?) and running away in horror. At that time, I listened to vinyl on a Roksan Xerxes with SME V and Koestu Black. As a classical music buff, I used to go to the concert hall at least once a week.

Of course, early CD playback was horrible. Implementation of technology, not the principle of it.

The situation today is unchanged: analogue remains continuous, digital - no matter the sampling frequency - remains discontinuous.

Again, this is a misunderstanding of how digital works. Here is a good explanation which includes a demonstration (on an oscilloscope):

Meanwhile, a great deal has been learned about digital recording, on how to push digital artefacts produced by digitalization as far away from the audible range as possible etc... IMHO, these advances fully justify the complaints of some/many? audiophiles about digital recordings' insurmountable shortcomings. Perhaps, a time will come when these advances close the gap with analogue reproduction. But imho, it has not arrived yet. I found the best proof in excellent digital recordings that sound superior when they are cut to vinyl - which seems to minimize/ filter out? the digital artefacts generated in the D/A conversion. Here, I disagree with many vinylstas who seek the goodness of old Decca recordings. IMHO, today, vinyl from digital sources can be a joy, if done properly (I get frustrated with the hit and miss DG LPs of late: feed the fad, forget about sound quality; eg Joe Hisaishi's uplifting music massacred on LP).

An aside: a few years ago, a small German outfit - Stockfish - made a curious experiment. It cut digital masters through a D/A converter to an analogue direct metal master on a Neumann VMS-82 cutting lathe; it then read the copper master with an EMT 997 tonearm/TSD-15 cartridge and fed the analogue signal to a Meitner A/D converter. The obtained DSD signal at 2.8224 Mhz was used to produce SACDs. These SACDs (two in total) are amongst the best sounding SACDs in my collection.

Once more, this points to vinyl colorations, or the deliberate introduction of them as in that last experiment described, being the pleasant trick that for some listeners makes digital more listenable. Many other listeners enjoy digital in itself just fine.

 

[Ian B]

This is not how it works. The output from a DAC is no less continuous than analog: It is a signal with an analog waveform. There are no discontinuities such as "stairsteps" or the like: Stair steps are a myth.

If digital vinyl (vinyl sourced from digital recordings) sounds better to some than just digital, this is not because the waveform is more continuous. Transfer of digital to vinyl does not miraculously make it any more "continuous" than the signal at the output of the DAC itself (already an analog waveform) that was used to create the master from which the vinyl is cut.



Of course, early CD playback was horrible. Implementation of technology, not the principle of it.



Again, this is a misunderstanding of how digital works. Here is a good explanation which includes a demonstration (on an oscilloscope):

Once more, this points to vinyl colorations, or the deliberate introduction of them as in that last experiment described, being the pleasant trick that for some listeners makes digital more listenable. Many other listeners enjoy digital in itself just fine.

Monty is actually wrong, and makes a lot of assumptions about digital filters and linearity at very low levels being perfect. I did a lot of audio recording at various sample rates, 16 and 24 bit. They were all different, and none actually was capable of sounding identical to the source tape. DSD 128 was very close, though.

If you think in terms of the waveform, you are getting a continuous output that is smoothed by the filter. The irony is that many audiophiles prefer the stair steps and aliasing that comes with NOS filters like in Yggdrasil, Audio Note DACs, Aries Cerat, TotalDAC etc. This (and also DSD) actually points to the problem being the filters themselves required to reconstruct the audio. I've yet to hear a DAC that could sound as continuous as an analog source, and it does always seem somewhat like an atomized approximation, if a very good one. Who knows why.

Current crop of top DACs are probably pushing the limit of what a DAC can do with a digital recording, but they don't have control of the ADCs involved in the recording, and there seem to be some limits in terms of ultra-fine timing response with PCM filter design and bandwidth. At this point many things have been tried: higher sample rates, DSD, DSD upsampling, NOS filters, adaptive filters, apodizing filters, MQA. All help in some respects, and some are detrimental in others. But nobody has fully cracked the problem of getting the analog playback quality without some kind of performance compromise.

Still, we are at a place with satisfyingly musical and extremely high performance digital, so I don't lose sleep over it.

(BTW, my reference points are source tapes. Vinyl definitely adds a sound, and its own rhythmic quality. Maybe adding that sound is enough to satisfy many. But I'm just looking at whether and how digital mediums can perfectly reproduce tape or live microphone feeds, and there is one hard to describe dimension that seemingly nobody has mastered yet.)

 

[Rexp]

Monty is actually wrong, and makes a lot of assumptions about digital filters and linearity at very low levels being perfect. I did a lot of audio recording at various sample rates, 16 and 24 bit. They were all different, and none actually was capable of sounding identical to the source tape. DSD 128 was very close, though.

If you think in terms of the waveform, you are getting a continuous output that is smoothed by the filter. The irony is that many audiophiles prefer the stair steps and aliasing that comes with NOS filters like in Yggdrasil, Audio Note DACs, Aries Cerat, TotalDAC etc. This (and also DSD) actually points to the problem being the filters themselves required to reconstruct the audio. I've yet to hear a DAC that could sound as continuous as an analog source, and it does always seem somewhat like an atomized approximation, if a very good one. Who knows why.

Current crop of top DACs are probably pushing the limit of what a DAC can do with a digital recording, but they don't have control of the ADCs involved in the recording, and there seem to be some limits in terms of ultra-fine timing response with PCM filter design and bandwidth. At this point many things have been tried: higher sample rates, DSD, DSD upsampling, NOS filters, adaptive filters, apodizing filters, MQA. All help in some respects, and some are detrimental in others. But nobody has fully cracked the problem of getting the analog playback quality without some kind of performance compromise.

Still, we are at a place with satisfyingly musical and extremely high performance digital, so I don't lose sleep over it.

(BTW, my reference points are source tapes. Vinyl definitely adds a sound, and its own rhythmic quality. Maybe adding that sound is enough to satisfy many. But I'm just looking at whether and how digital mediums can perfectly reproduce tape or live microphone feeds, and there is one hard to describe dimension that seemingly nobody has mastered yet.)

Agreed, I actually like the sound of SAR ADC's in some phones, have you tried any of those? Apparently they can do 16/44.1 now.

 

[Al M.]

Monty is actually wrong, and makes a lot of assumptions about digital filters and linearity at very low levels being perfect.

Not exactly. Here is the 1 kHz sine wave at -90 dB of the Yggdrasil Analog 2 DAC (the precursor of my Yggdrasil LIM DAC) in green. The red line is the same measurement of the Yggdrasil Analog 1 DAC, the original Yggdrasil:



Try to get such a great low-level sine wave from vinyl or analog tape. If I am not mistaken, according to the dynamic range of those such a sine wave, with the coveted analog waveform, should be buried in noise. Yes, it seems that digital is more analog than -- analog.

The irony is that many audiophiles prefer the stair steps and aliasing that comes with NOS filters like in Yggdrasil, Audio Note DACs, Aries Cerat, TotalDAC etc.

The newer Yggdrasil DACs have a NOS button so that you can bypass the filter, but the default is the proprietary filter developed by Mike Moffat . My Yggdrasil doesn't have the filter bypass button.