Non Oversampling DACs versus High Rez Hype

[opus111]

I can see some parallels yep. Both measure badly by traditional measurements (in differing respects) but are nevertheless praised for their sound.

NOS DACs have something of the simplicity of SETs but don't suffer from their impracticability.

The question that fascinates me is whether they sound attractive because of their measurements or despite them. My hypothesis is its the latter which opens the way towards getting the same sound by different means and hence somewhat improved measurements.

 

[LL21]

Its funny...as soon as i heard MEP's post, i thought...yes. I had not considered that before, but it seems NOS DACs and SETs probably do have some parallels as more specifically outlined by Opus111. NOS DAC designers seem to hold that there are certain technical specs (phase aberrations, less noise shaping, digital filters), etc where NOS DACs are actually superior...and thus it seems they are questioning which are the right measurements to take to truly assess digital equipment.

As you say, Opus111, "they sound attractive...despite [their poor measurements]."

Thinking about more measurement, linear response-oriented listeners (say recording studio people), i am surprised that one of the more well known remastering guys (Steve Hoffman) is well known to use Concert Fidelity NOS Tube DAC, (along with Wavac amps) in his systems.

In an extreme example, are SET/NOS DACs like caricatures...and more linear, better measuring equipment more like photography? Clearly, the photograph is 9.9x out of 10 going to look more like the person...but when you get a great caricature, despite its 'poor measurement'...it looks EXACTLY like the person capturing their expression, personality. And sometimes a photograph captures a person in such a way...you look at it and you literally say "it doesn't look like you".

Is this a fair characterization...or do well designed NOS DACs actually measure very well...if you know what to measure?

 

[DonH50]

The best solution might be a conventional (R-2R/segmented/etc.) DAC that is in fact oversampled to reduce both the in-band roll-off (-3.54 dB at Nyquist for 50% duty cycle IIRC) and allow simpler filters with fewer in-band artifacts. ADI did that with their "NOS" DACs for some time (and I assume still allow that option).

This assumes by "NOS" you mean conventional DAC architectures vs. a delta-sigma approach.

 

[Phelonious Ponk]

Thinking about more measurement, linear response-oriented listeners (say recording studio people), i am surprised that one of the more well known remastering guys (Steve Hoffman) is well known to use Concert Fidelity NOS Tube DAC, (along with Wavac amps) in his systems.

Steve is one of a rare breed -- captial A Audiophile engineers. It's not at all surprising that he prefers NOS Dacs. He prefers tubes. He prefers analog. He thinks vinyl out-performs digital, in spite of all the evidence to the contrary that his engineering education can't possibly be missing. And he creates digital masters with a goal of making them more analog-sounding. All of this according to Steve. I only own one of his re-masters, Jackson Browne's "Late For The Sky." A CD remastered to sound more like vinyl is exactly how I would describe it.

In an extreme example, are SET/NOS DACs like caricatures...and more linear, better measuring equipment more like photography?

Caricatures has negative connotations that I'd personally avoid, but painting vs photography? Maybe. To put precision on the analogy, though, they're both photographs. The same photograph. In one case the photo is filtered.

Tim

 

[LL21]

The best solution might be a conventional (R-2R/segmented/etc.) DAC that is in fact oversampled to reduce both the in-band roll-off (-3.54 dB at Nyquist for 50% duty cycle IIRC) and allow simpler filters with fewer in-band artifacts. ADI did that with their "NOS" DACs for some time (and I assume still allow that option).

This assumes by "NOS" you mean conventional DAC architectures vs. a delta-sigma approach.

Not sure i followed all that being a non-techie, but i am starting to learn. Thanks for that.

 

[LL21]

Steve is one of a rare breed -- captial A Audiophile engineers. It's not at all surprising that he prefers NOS Dacs. He prefers tubes. He prefers analog. He thinks vinyl out-performs digital, in spite of all the evidence to the contrary that his engineering education can't possibly be missing. And he creates digital masters with a goal of making them more analog-sounding. All of this according to Steve. I only own one of his re-masters, Jackson Browne's "Late For The Sky." A CD remastered to sound more like vinyl is exactly how I would describe it.



Caricatures has negative connotations that I'd personally avoid, but painting vs photography? Maybe. To put precision on the analogy, though, they're both photographs. The same photograph. In one case the photo is filtered.

Tim

Thanks, Tim. Here's a controversial question...okay, painting or photo...one is more accurate...which one captures the person better?

 

[LL21]

I can see some parallels yep. Both measure badly by traditional measurements (in differing respects) but are nevertheless praised for their sound.

NOS DACs have something of the simplicity of SETs but don't suffer from their impracticability.

The question that fascinates me is whether they sound attractive because of their measurements or despite them. My hypothesis is its the latter which opens the way towards getting the same sound by different means and hence somewhat improved measurements.

Opus111 -I have question for you on noise floor of TDA1541A.

I am happy with the voice of the Zanden DAC. the only area where i really find newer units (DCS Scarlatti, Stahl-Tek Vekia Opus) are superior is detail, decay, lower noise floor. i prefer the Zanden voice. Thus, i am really focused on improving the Zanden's detail, decay, ability to separate out complex passages...in other words simply hearing MORE of the voice of the Zanden.

So here is MY QUESTION:

If the designer really focuses on shielding, separating power supply, lowering noise floor...will a TDA1541A-based DAC be able to have a comparably low noise floor to these latest SOTA DACs Or does the TDA1541A chip itself have too high a noise floor?

sorry if a dumb question...i am trying (as as non-techie) to figure out where i may be going with the Zanden. Thanks!!!!!

 

[jkeny]

Steve is one of a rare breed -- captial A Audiophile engineers. It's not at all surprising that he prefers NOS Dacs. He prefers tubes. He prefers analog. He thinks vinyl out-performs digital, in spite of all the evidence to the contrary that his engineering education can't possibly be missing. And he creates digital masters with a goal of making them more analog-sounding. All of this according to Steve. I only own one of his re-masters, Jackson Browne's "Late For The Sky." A CD remastered to sound more like vinyl is exactly how I would describe it.

Tim, this sounds at best faint praise & at worst a disdainful view of Steve's choices. Am I correct in my assessment?

I believe what Opus is asking/saying is that despite current measurements SET & NOS seem to sound good!

 

[mep]

Steve is one of a rare breed -- captial A Audiophile engineers. He thinks vinyl out-performs digital, in spite of all the evidence to the contrary that his engineering education can't possibly be missing.
Tim

Tim-You act like Steve Hoffman is the only guy on the planet that thinks vinyl outperforms digital. There are lots and lots of people who own both formats (vice people who are wedded strictly to digital) and firmly believe that vinyl contains more information and sounds better than digital. There are lots of those people right here on this forum and I count myself among them.

 

[Phelonious Ponk]

Thanks, Tim. Here's a controversial question...okay, painting or photo...one is more accurate...which one captures the person better?

I think the answer to your question was right in front of you Lloyd, in the clarification of my statement:

Caricatures has negative connotations that I'd personally avoid, but painting vs photography? Maybe. To put precision on the analogy, though, they're both photographs. The same photograph. In one case the photo is filtered.

And if it puts the same filter on every picture, or the same mask on every painting, no, it doesn't capture the person better.

Tim

 

[Phelonious Ponk]

Tim, this sounds at best faint praise & at worst a disdainful view of Steve's choices. Am I correct in my assessment?

I believe what Opus is asking/saying is that despite current measurements SET & NOS seem to sound good!

I can neither praise nor damn Steve's work. I just don't know it well enough. To my knowledge, I only have one of his masters. I disagree with is views on media. Nothing disdainful about it. I spent quite a bit of time on his music forums for awhile. I still go back and search through threads for clues to the best masters. I like Steve.

Tim

 

[Phelonious Ponk]

Tim-You act like Steve Hoffman is the only guy on the planet that thinks vinyl outperforms digital. There are lots and lots of people who own both formats (vice people who are wedded strictly to digital) and firmly believe that vinyl contains more information and sounds better than digital. There are lots of those people right here on this forum and I count myself among them.

Not at all Mark, I said he was of a rare breed -- capital A Audiophile engineers. That's quite different from the only guy on the planet that prefers vinyl. I've known more audio engineers than most people, i suspect. I've found that most of them think audiophiles are a bit nuts. Steve is the exception. And of course, in absolute terms there are lots of people who prefer vinyl. Thousands, I'd guess. Maybe even tens of thousands. Lots. And lots.

Tim

 

[jkeny]

I can neither praise nor damn Steve's work. I just don't know it well enough. To my knowledge, I only have one of his masters. I disagree with is views on media. Nothing disdainful about it. I spent quite a bit of time on his music forums for awhile. I still go back and search through treads for clues to the best masters. I like Steve.

Tim

Good that you are keeping an open mind on these matters, then.

 

[opus111]

I am happy with the voice of the Zanden DAC. the only area where i really find newer units (DCS Scarlatti, Stahl-Tek Vekia Opus) are superior is detail, decay, lower noise floor. i prefer the Zanden voice. Thus, i am really focused on improving the Zanden's detail, decay, ability to separate out complex passages...in other words simply hearing MORE of the voice of the Zanden.

From what I've read about the design, the Zanden is somewhat detracting from the voice of the TDA1541A. So I take it you'd like to hear more of the DAC, allowing less to get in the way of that.

If the designer really focuses on shielding, separating power supply, lowering noise floor...will a TDA1541A-based DAC be able to have a comparably low noise floor to these latest SOTA DACs Or does the TDA1541A chip itself have too high a noise floor?

The answer to that first question is 'Yes, for redbook CD source material'. The TDA is too noisy only for hi-res material (though some of that still has a noise above the DAC, being transcribed from RBCD or analog tape). RBCD's inherent noise (due to being 16bits) is comfortably above the DAC's own noise.

sorry if a dumb question...i am trying (as as non-techie) to figure out where i may be going with the Zanden.

Not a dumb question - the only dumb questions are the ones asked not for curiosity's sake

 

[opus111]

And if it puts the same filter on every picture, or the same mask on every painting, no, it doesn't capture the person better.

That's a big IF in the current context. Does it in fact put the same filter on every picture (speaking here of a NOS DAC) ?

 

[opus111]

In an extreme example, are SET/NOS DACs like caricatures...and more linear, better measuring equipment more like photography?

Its a debatable subject whether the more modern S-D DACs are indeed 'more linear'. They rely on requantization of the digital signal, typically down to 4-6bits, in the case of DSD, one bit. Then use lots of processing (in a feedback loop) to undo the damage of what's inherently a non-linear process (quantization).

Its interesting to note that such a signal chain goes against current wisdom in the design of power amplifiers. There, the guys who do this for a living will tell you 'make the circuitry as linear as possible before adding feedback'. In the case of S-D, there's the deliberate introduction of a gross non-linearity which its assumed can be sufficiently corrected for using feedback.

Is this a fair characterization...or do well designed NOS DACs actually measure very well...if you know what to measure?

Yes - and the vast majority of us don't know what to measure yet. But its going to be measurements which differentiate different time-domain behaviour, rather than frequency domain.

 

[jkeny]

Now your'e making too much sense, Opus111 & someone will jump in a "prove" mathematically that S-D DACs measure better than NOS DACs, ignoring, of course, what you just said about CURRENT measurements not being a good indicator of what's important, sonically! I just thought I would say this before the inevitable posts start - it just might help improve the signal to noise ratio?

 

[opus111]

Here's some reading material loosely related to my last comment about time-domain measurements, rather than frequency-domain. Those who'd like to learn more, feast your eyes... http://www.temporalcoherence.nl/index.php?pagina=artikelen&lang=en

 

[LL21]

I think the answer to your question was right in front of you Lloyd, in the clarification of my statement:



And if it puts the same filter on every picture, or the same mask on every painting, no, it doesn't capture the person better.

Tim

ACtually, a caricature does not put the same filter on every picture...it simply focuses on capturing a very different part of the person...an Al Hirschfeld caricature is an extreme example...but Al Hirschfeld captures far more of a person than a police mug shot of the same individual.

Now i am NOT condoning we go ahead and use caricatures as an appropriate direct analogy for audio. But i AM using it to illustrate the point that there may be elements of SET/NOS or caricatures...or paintings that capture something you cannot measure with a ruler, but that yOU CAN measure somehow else...that still relates directly back to the original recording/tape.

For example, some of us having probably seen those huge pixelated photos (examples above)...somehow the human brain knows it is Albert Einstein. I leave the other one anonymous but anyone can easily google it if need be or PM me. It is remarkable...the brain clearly is picking something up that can recognize these individuals with a blob...one could not possibly measure that photo and say accurate...but the brain KNOWS who it is.

By measurements, a computer would clearly tell us there is NO way we should know anything about who these people are. The measurements would be ridiculously bad. Off the charts bad. But our brains know. Thus, our brains are measuring something or picking something up we have not yet measured that can assimilate certain video or audio cues.

Coming back to your analogy of painting...then the question still applies, and one could easily take a photo of someone...and none of us would recognize that person standing right next to us. And you might have the right painting and be able to pick that person out of a moving crowd of 25.

 

[LL21]

Its a debatable subject whether the more modern S-D DACs are indeed 'more linear'. They rely on requantization of the digital signal, typically down to 4-6bits, in the case of DSD, one bit. Then use lots of processing (in a feedback loop) to undo the damage of what's inherently a non-linear process (quantization).

Its interesting to note that such a signal chain goes against current wisdom in the design of power amplifiers. There, the guys who do this for a living will tell you 'make the circuitry as linear as possible before adding feedback'. In the case of S-D, there's the deliberate introduction of a gross non-linearity which its assumed can be sufficiently corrected for using feedback.



Yes - and the vast majority of us don't know what to measure yet. But its going to be measurements which differentiate different time-domain behaviour, rather than frequency domain.

Nice one Opus111...thank you. That is very helpful!