Thanks for the technical explanation and correction. This does not take away the fact that you don't need a NOS DAC for great soundstage depth and layering. Anyone who does not believe me: come listen to my system. The audible evidence will speak for itself.
Just looked at your system & see you are using a Berkeley Alpha 2. Indeed soundstage & timbre seem to be reported highlights of this DAC. Looking into this it uses a AD1955 with external filter (own code on FPGA) rather than the digital filter on AD1955.
So it may be that these digital filters are not detrimental to soundstage, per se but having them on the same silicon as the DAC may introduce some aspect that masks low level cues - noise fluctuation would be my best guess.
So maybe a NOS DAC avoids these aspects although I think Opus has done experiments with upsampling prior to NOS DACs & found it detrimental to the sound, AFAIR
Just looked at your system & see you are using a Berkeley Alpha 2. Indeed soundstage & timbre seem to be reported highlights of this DAC. Looking into this it uses a AD1955 with external filter (own code on FPGA) rather than the digital filter on AD1955.
So it may be that these digital filters are not detrimental to soundstage, per se but having them on the same silicon as the DAC may introduce some aspect that masks low level cues - noise fluctuation would be my best guess.
On all the systems that I heard that portray soundstage depth, it seems to happen with all recordings (e.g. even multitrack pop/rock recordings with overdubs), not just ones with honest spatial cues.
No, tough for digital to sound real except unless the analog is badly set up or the vinyl quality is really poor. And the best digital recordings will be far behind decent vinyl recordings, even if you take a 100k digital and throw it at a 10-15k set up.
That is correct in my view. An image far behind the speakers but without any front-to-back information is simply a recessed image. This has nothing to do with depth.
I have to partially correct myself here. I have, for example, a choral CD that sounded upfront before I had my external power supplies for the amps. After insertion of those in my system, the image greatly receded into a large space far behind the speakers, because now the low-level information about the recorded space came through. While there is depth of image, there is no depth within the image -- all voices sound on the same recessed plane.
Yet my system does not systematically move all images to the back, which would be just an artifact and not true depth portrayal. It can also portray performers far upfront and direct, if so recorded. And at the same time it can portray other performers on the same recording far in the back, if required by the spatial information on the recording (e.g. layering of intruments in an orchestra from front to back).
When I first got full blown soundstage presentation I was amazed, but now see it as the automatic endpoint of a system, any system, which is "competent". I've played with a variety of gear over the years, delving deep into the cheapest of the cheap, just to see what could be extracted; and this has clearly shown that it's all about the quality of the sound emerging from the drivers, which in turn is totally dependent on the integrity of the complete playback chain. Everything else is secondary - will have some impact, but is a minor player.
The width, depth, layering, detail of the recording will emerge to a level that the system playback is capable of at that stage of tune - and can always be better than what one gets at any point of refinement !! 'Artificial' recordings are the most amazing of all - can have a staggering complexity to them; and completely thrash "audiophile" recordings, IME ...
On all the systems that I heard that portray soundstage depth, it seems to happen with all recordings (e.g. even multitrack pop/rock recordings with overdubs), not just ones with honest spatial cues.
Yes, and it can be aggravating. The ability to portray honest spatial clues alas also shows the spatial artifacts in pop/rock recordings, such as lead voices with some reverb around them which then sound uncomfortably recessed.
Sometimes I wish I could flip a switch to just direct sound 'in your face'.
Thanks for the technical explanation and correction. This does not take away the fact that you don't need a NOS DAC for great soundstage depth and layering.
Agree absolutely. In the post John linked to he'd written about our experience together at his friend's set up, we listened to both the Soekris and his own DAC and neither of those is NOS. Its like Lloyd said - when immense effort towards purging the whole system of noise (EMC-wise) is taken, the depth is there.
So maybe a NOS DAC avoids these aspects although I think Opus has done experiments with upsampling prior to NOS DACs & found it detrimental to the sound, AFAIR
Yeah I tried running a simple DAC at both 1X and 2X oversampling where other aspects were controlled and found that oversampling led to a greyer sound. I put this down to the DAC being at the 'wrong' output voltage more of the time when its updating more frequently. However this experiment was done in the days before I'd fully realized the RF sensitivity of my I/V stage so part of the greyness might not have arisen from the DAC itself. I think I'm overdue to repeat the experiment with my discrete I/V stage.
Yeah I tried running a simple DAC at both 1X and 2X oversampling where other aspects were controlled and found that oversampling lead to a greyer sound. I put this down to the DAC being at the 'wrong' output voltage more of the time when its updating more frequently. However this experiment was done in the days before I'd fully realized the RF sensitivity of my I/V stage so part of the greyness might not have arisen from the DAC itself. I think I'm overdue to repeat the experiment with my discrete I/V stage.
Ah, very good - always good to realise what the possible shortcomings of previous experiments/results might be & update one's knowledge. Something that is missing on many audio forums we know
No, tough for digital to sound real except unless the analog is badly set up or the vinyl quality is really poor. And the best digital recordings will be far behind decent vinyl recordings, even if you take a 100k digital and throw it at a 10-15k set up.
It was my opinion three years ago. Changing somewhat meanwhile. In adequate systems sometimes the difference is now becoming a question of preference, as digital now has some positive aspects that can make a difference in realism over vinyl.
One thing I would suggest is very important is not to be trapped into thinking that just one aspect of a system has to be sorted for everything to fall into place: a bottleneck, on the potential quality, may exist somewhere which is not all obvious - the difficulty will often be to tracking down the underlying culprit, as opus is suggesting.
As an example of my own, 30 years ago I used a battleship Perreaux, and it turned out that the monster smoothing caps with screw down terminals at the top, were badly terminated internally to the core of the capacitor. This finally became clear when the whole screw assembly turned when tightening the cables! - a weakness in the design of a key part ...
One thing I would suggest is very important is not to be trapped into thinking that just one aspect of a system has to be sorted for everything to fall into place: a bottleneck, on the potential quality, may exist somewhere which is not all obvious - the difficulty will often be to tracking down the underlying culprit, as opus is suggesting.
Yes absolutely, Frank. I purposefully chose the wording "this is essential to get it right", as opposed to "this is the single one essential thing to get it right". Low-level noise control in electronics is a necessary but not sufficient condition. If the DAC does not operate with linearity from loud down to extremely soft signals it will all be of no help*). Or if the room is 'noisy' in the sense that unwanted reflections blur and mask the low-level spatial info in the recording. Oh, I forgot, you don't believe in the importance of the room .
_______
*) if I am not mistaken, non-linearity of DACs was a main culprit for the early 'dry' CD sound that suppressed ambience information. It's interesting historically that back then critics ascribed the problem to an allegedly too low sampling rate...
Or if the room is 'noisy' in the sense that unwanted reflections blur and mask the low-level spatial info in the recording. Oh, I forgot, you don't believe in the importance of the room .
_______
*) if I am not mistaken, non-linearity of DACs was a main culprit for the early 'dry' CD sound that suppressed ambience information. It's interesting historically that back then critics confused the problem with an allegedly too low sampling rate...
Tsk, tsk, ... yes, it all adds, in an absolute sense - obviously the room will alter the information to some degree, but so far I've relied on a clean reproduction chain to get me there. What happens, for me, is that the "room" of the recording takes over, completely - my listening space completely ceases to exist, subjectively; so then the low level info of the recording has full rein to express itself. I tend to listen at elevated levels, which helps me hear 'into' the recording - perhaps other people are more disturbed by room effects, this may be a very individual reaction to the overall sound picture.
The non-linearity of DACs figures in part, because many implementations have terribly long warmup periods. Cheap components have taken hours, even days, to lift their game - opus has pointed to S-D chips being notorious for flat sound, and a "solution" is to drive them hard for extended periods ...
The vintage CD players used DACs with fairly impressive linearity (TDA1541(A)) which when in NOS form do impressive ambience retrieval. I think its far more likely that the crummy opamp I/V and filter stages with amazingly lack-lustre power supplies did the damage in those early players.
Considering digital masters have never been outright bad at any point, as a whole, there's no doubt it's a matter of implementation since they've successfully transferred to vinyl. The basis of the technology clearly has not been the problem.
However it sure did seem like in order to avoid costs there's been numerous attempts at high sampling, different bit uses on multiple chips/dies, etc, basically attempting to circumvent spending the dough for good power supplies etc by inventing new chips constantly. There's a reason why only a handful have ever been notable, and never without mods or spending enough money initially to make a quality device.
However it sure did seem like in order to avoid costs there's been numerous attempts at high sampling, different bit uses on multiple chips/dies, etc, basically attempting to circumvent spending the dough for good power supplies etc by inventing new chips constantly.
'Artificial' recordings are the most amazing of all - can have a staggering complexity to them; and completely thrash "audiophile" recordings, IME ... 
