World’s First Valid Comparison of PCM versus DSD?

[Miska]

I had to look up picotech. Looks like they sell glorified external Soundcards to me.

 

Can I laugh now? 20 GHz bandwidth at 16-bit resolution is glorified external sound card. At 16k€ price. Yeah.

 

Do you have an actual oscilloscope (LeCroy, Tek, Agilent)?

 

Yes, but most of those are crappy due to low resolution display, poor CPU performance and 8-bit ADC.

 

But let's compare to Archimago's posts of UD-501.

 

Obviously 1k square at PCM was reproduced in NOS mode. Here's mine:

 

This is how 1k square at 44.1k PCM looks like with "sharp" filter setting in DAC:

 

Here's mine for DSD64:

 

And here's the same for DSD128:

 

Something to complain?

 

Now what was not shown, let's do same with 1 kHz tone with same settings and also look at spectrum:

 

This is how 1 kHz sine looks like in the same NOS PCM mode:

 

This is how it looks like with sharp filter setting:

 

And this is how it looks like with DSD128:

 

Complaining about some brand/model of measurement equipment is pointless, unless you can prove that some piece of equipment behaves incorrectly. Not seeing the kind of results you want to see doesn't prove anything. You are free to post your own measurements of TEAC UD-501 with equipment of your choice. I'm not going to spend 50k just to buy equivalent equipment of all brands.

 

Now take your pick, I've taken my already and use UD-501 exclusively in DSD128 mode.

 

BTW, what you show has significant distortion (e.g. visible) near the sinewave peaks.

 

Good eyes, because it's around 0.0048%.

[Miska]

Let's look at some rise times for TEAC UD-501 now, shall we?

 

PCM at 384/32:

[ATTACH=CONFIG]9643[/ATTACH]

 

DSD128:

[ATTACH=CONFIG]9644[/ATTACH]

 

Both are practically the same, and would be even more equal for PCM at 352.8, or DSD128 @6.1 MHz (48x128).

[Kevin M.]

This test is very useful though: it made me realize I shouldn't worry at all about what format I get. Either one. If one has a better master, that's the one I'll want.

Agree 100%!

[Miska]

And since I'm a nice person, I also took photos of my 100% analog scope for TEAC UD-501.

 

1k square at DSD128:

 

20k sine at DSD128:

[magnum innominandum]

Miska,

 

At ADC side you should always have anti-alias filter before the converter, regardless of sampling rate. Otherwise you get nasty effects due to aliasing of any RFI/EMI.

 

Even for DSD?

 

Sure, but less than filters needed for PCM at 192k or lower rates.

 

I don't know. If we use quad speed PCM and allow as much ultrasonic images as DSD outputs noise, the filter may be much less severe. It is simply something that has not been tried much.

 

I've seen it long ago, but not really interested. Something is wrong there because of symmetric overshoot of the square wave.

 

Then it is wrong with dCS converters.

 

My measurements have been made with 100% real scope.

 

So why do you keep posting stuff from a software 'scope?

 

Why are you comparing my results with those posted by others, different DAC and different test signals - different results.

 

I am comparing because you posted your results as answer to a request to post such. Surely they are comparable?

 

No it doesn't convert it to PCM, it converts it to multi-bit SDM at same sampling frequency, there is no sampling frequency down-conversion at all. And of course it is DSD because that's what the DAC receives.

 

Based on the patent that ESS claims to use for the ASRC in the ESS DAC chips, what you describe is pysically imposible if this ASRC is used. As DSD does not bypass the ASRC according ESS's own white papers, we must conclude that all ESS DAC's render DSD into something else, that can be processed by the ASRC used by ESS.

 

Otherwise, in your logic, if I set my J-River to convert DSD to 88.2KHz/24Bit orfor example (illegally) download the many torrents containing SACD (PS3) -> PCM88.2/24 and play those (which my USA made multibit DAC can play) it is DSD, because that is the data read from disk or at least the data ripped from the original SACD.

 

It matters not if the conversion to PCM happens in the software on my PC or inside the DAC chip. The result is still DSD -> PCM and then something else or not. Any other interpretation is quite living the dream.

[magnum innominandum]

Miska,

 

Now take your pick, I've taken my already and use UD-501 exclusively in DSD128 mode.

 

Sorry, I do not heave UD-501. Nor any intention to buy one. I have a fairly old US made Multibit DAC (long discontinued), that so far sounds better than anything else I have heard so far, at least up to the 96KHz limit it can handle.

 

Good eyes, because it's around 0.0048%.

 

I am not neccesarily referring to THD. If you look closely, you can see distortion on the peaks of the sine,which is not clipping, but seems to vary (semi?)-randomly.

 

As a rule, anything visible in a 'scope trace with traditional analogue amplifiers is >> 1% THD. Of course what we seeing in terms of waveform distortion is not classic HD (is it this magical noise shaping?), but to bee seen easily, it must be of a substantial magitude...

 

Your 'scope traces for DSD start to coverage more with what other post.

 

I wonder if you could care to do a 1KHz square and a 20KHz sine at -20dBFS in DSD64 and PCM176.4 (as these use comparable data rates)?

 

If you like, add DSD128 and PCM352.8...

 

It may be equally interesting to compare ESS9018 to the PCM1795 on these.

 

It may help folks to both understand DSD better and to choose the right replay gear.

 

Based on what you have shown, I may be tempted to try another ES9018 based DAC, because on it DSD looks actually very reasonable and to avoid BB based DAC's (though my own DAC uses BB DAC chips and sounds, as well as measures excellent with these).

[magnum innominandum]

Miska,

 

No, it is as pure delta-sigma DAC as one can be. With 24 equally weighted current sources.

 

I think you need an irony transplant.

 

I find it very ironic that a pure single bit system needs "24 equally weighted current sources".

 

That is like saying an amplifier using 24 Tubes in push pull Class A/B is an absolutely pure Class A single ended amp.

 

At least you are funny. I nearly sneezed out the eggnog I was drinking. The family & friends found my facial expressions most amusing though I did not find the experience nearly as funny at the time.

 

So your "as pure single bit system as can be" needs 24 or 32 equally weighted bits... So it is really a "as pure as possible 32 bit system"...

[Hiro]

Let's look at some rise times for TEAC UD-501 now, shall we?

 

PCM at 384/32:

[ATTACH=CONFIG]9643[/ATTACH]

 

DSD128:

[ATTACH=CONFIG]9644[/ATTACH]

 

Both are practically the same, and would be even more equal for PCM at 352.8, or DSD128 @6.1 MHz (48x128).

 

Miska,

 

The links don't work (at least for me)...

 

And BTW, since the bitrate of 32/384kHz CPM is the same as for DSD256 (12.288MHz) files, I think it would be more useful to see the comparison between DSD128 and 32/176.4 PCM....

[Miska]

The links don't work (at least for me)...

 

I don't know why the forum software didn't embed those images, but here are the ones through URLs.

 

PCM:

 

DSD:

 

If these still don't work, I'll re-upload those as jpeg or something.

 

And BTW, since the bitrate of 32/384kHz CPM is the same as for DSD256 (12.288MHz) files, I think it would be more useful to see the comparison between DSD128 and 32/176.4 PCM....

 

Yes, or I could make the same at 16/352.8 to match the bitrate, but doesn't matter much for this context as it's mostly limited by the analog filtering.

[Hiro]

And since I'm a nice person, I also took photos of my 100% analog scope for TEAC UD-501.

 

1k square at DSD128:

[ATTACH=CONFIG]9645[/ATTACH]

 

20k sine at DSD128:

[ATTACH=CONFIG]9646[/ATTACH]

 

Thanks for sharing these measurements.

[Hiro]

Yes, or I could make the same at 16/352.8 to match the bitrate, but doesn't matter much for this context as it's mostly limited by the analog filtering.

 

Either 16/352.8 or 32/176.4 would be O.K. for a fair comparison with DSD128 which happens to have the same bitrate as the two PCM variants.

[Miska]

I find it very ironic that a pure single bit system needs "24 equally weighted current sources".

 

Well, it's 4.644-bit system if you think in two's complement binary. But there's nothing ironic, those current sources can be used to form all kinds of linear-phase analog reconstruction filters. But if you make similar DAC for 24-bit PCM you'll need 16777216 equally weighted current sources. Or you can make it with 24 2^x weighted current sources, but then your MSB precision needs to be better than 5.96e-8 and MSB-1 needs to have precision better than 1.19e-7 and so on.

 

That is like saying an amplifier using 24 Tubes in push pull Class A/B is an absolutely pure Class A single ended amp.

 

Completely unrelated. But this is equivalent of running 24 class A/B amps in parallel such configuration that cross-over distortion of each cancels each other out.

 

At least you are funny.

 

Good, now I'll wait for some constructive input from you, like measurement results and implementations of your own, instead of something borrowed from other you don't even seem to understand.

 

Meanwhile, I just consider you as a troll.

 

So your "as pure single bit system as can be" needs 24 or 32 equally weighted bits... So it is really a "as pure as possible 32 bit system"...

 

32-bit two's complement binary system (PCM) can represent 4294967296 levels, while 32-bit unary system (SDM) can represent 33 levels.

 

The dCS system can represent 25 different voltage levels.

[Miska]

Sonic has re-registered with another nick or asked some fellow souls form diyhifi.org to participate!?

 

long discontinued

 

Let's stick to modern ones. But I'm already waiting for you to post scope plots and spectrum graphs (0-1.5 MHz) of 1 kHz square and 20 kHz sine. With most multi-bit PCM DACs, I can see higher level of distortion components in spectrum analysis due to stair-steps than I see noise in SDM DACs. Even worse, those PCM frequency components are strongly correlated with the signal (and thus create intermodulation components), unlike SDM noise components that are uncorrelated.

 

I am not neccesarily referring to THD. If you look closely, you can see distortion on the peaks of the sine,which is not clipping, but seems to vary (semi?)-randomly.

 

There's some amount of random HF-noise from noise shaping. All delta-sigma DACs have it, so if you'd hear it, you'd hear quiet background hiss. Since it's from 60 - 250 kHz in frequency your speaker or headphone transducers won't be able to follow it, but instead they'll average it and that will specifically give you the great low level linearity you cannot reach with multi-bit PCM DAC.

 

Class-D power amps have much more of that noise...

 

As a rule, anything visible in a 'scope trace with traditional analogue amplifiers is >> 1% THD. Of course what we seeing in terms of waveform distortion is not classic HD (is it this magical noise shaping?), but to bee seen easily, it must be of a substantial magitude...

 

Now you are thinking about some traditional low-res scope, not a high-res scope like in this case where details even down to -120 dB can be reproduced.

 

Scope plots are misleading if you don't know how to properly interpret it, so better to use wide-band spectrum analysis instead (I've provided it).

 

I wonder if you could care to do a 1KHz square and a 20KHz sine at -20dBFS in DSD64 and PCM176.4 (as these use comparable data rates)?

 

Now it's your turn. I may post something again when I have time and motivation.

 

If you like, add DSD128 and PCM352.8...

 

I did already, although I gave PCM 2x advantage by using 32-bit instead of 16-bit.

 

It may be equally interesting to compare ESS9018 to the PCM1795 on these.

 

No, you cannot compare DAC chips like that, you need to take the following analog stages into account too.

 

But I've posted measurements for Sabre with PCM and SDM and the performance is pretty much the same for both. Now it's your task to browse back the forum posts.

 

Based on what you have shown, I may be tempted to try another ES9018 based DAC, because on it DSD looks actually very reasonable and to avoid BB based DAC's (though my own DAC uses BB DAC chips and sounds, as well as measures excellent with these).

 

Take a look at CS4398 and you may be able to exceed performance of both if you use it correctly.

 

If you like sound of ESS' digital processing, good for you. If you want a bit-perfect DAC you may need to look elsewhere.

[magnum innominandum]

Miska,

 

Sonic has re-registered with another nick or asked some fellow souls form diyhifi.org to participate!?

 

Who is Sonic? Some Video game Character? I think my Son played that when he was young?

 

Let's stick to modern ones.

 

Why? Because they are worse than classic old PCM ones?

 

But I'm already waiting for you to post scope plots and spectrum graphs (0-1.5 MHz) of 1 kHz square and 20 kHz sine.

 

I will oblige the next time I take my Pass D1 to work. It's a bit of a chore to carry and the wife will complain (she plays music via it when I'm at work). I do think it was covered in Stereophile back then.

 

I did (for courisity) a full workup when I got it (2nd hand - cheap), maybe '99 or '00...

 

Very clean.

 

With most multi-bit PCM DACs, I can see higher level of distortion components in spectrum analysis due to stair-steps than I see noise in SDM DACs.

 

Hmmm. Maybe.

 

I just like the way it sounds of course. And it measures well enough to be okay, I think, but crucially, it measures the same no matter what "angle" you use.

 

FFT and 'scope agree. I did not notice this in your scrapes from the ESS.

 

But maybe you cannot see because massive averaging "hides" the noise, while it does not hide harmonic components?

 

But being blind to something, does it mean it does not exit or that we cannot hear it, does it?

 

Even worse, those PCM frequency components are strongly correlated with the signal (and thus create intermodulation components), unlike SDM noise components that are uncorrelated.

 

First, the microphones used in recording and the speakers used in playback have high levels of distortion strongly correlated with the signal. I cannot remember anyone complaining much.

 

There is an opposite angle.

 

SDM creates tons of ultrasonic noise (where classic analogue amplifiers have very poor linearity) that are steady state, meaning they do not change with signal. The result is IMD that will invariably have some content folding back into the audio range. Because one component is (semi)random the resultant IMD will be essentially equally (semi)random and a heavily averaged FFT measurement will simply "tune out" this noise. A singe shot 'scope trace will show it easily at -40dBFS or so...

 

So the crucial question is, do we hear the (AP2) 64Ksample average or do we hear the individual (heavily corrupted) sample. I suspect the answer is both "depends" and "inbetween".

 

There's some amount of random HF-noise from noise shaping. All delta-sigma DACs have it, so if you'd hear it, you'd hear quiet background hiss.

 

Or not so quiet, depending on the IMD performance of the system used.

 

Since it's from 60 - 250 kHz in frequency your speaker or headphone transducers won't be able to follow it, but instead they'll average it and that will specifically give you the great low level linearity you cannot reach with multi-bit PCM DAC.

 

Hmmmm, or it will give that very dirty noisy background you cannot reach with multi-bit PCM DAC.

 

Class-D power amps have much more of that noise...

 

Absolutely true. I have yet to hear a tolerable switching amp (never mind decent) including the much hyped ones and I will readily agree that Delta Sigma shares many sonic traits with those amplifiers, if a little less severe. Had they been around when Bill "I never had sexual relations with that woman" Clinton was Prez of Down South, he could have claimed "I never had any relationship with music using this Amplifier" and been totally unimpeachable for it....

 

Now you are thinking about some traditional low-res scope, not a high-res scope like in this case where details even down to -120 dB can be reproduced.

 

True, 'scope traces (not FFT) stop pretty much at-30dB, so if you cannot see a difference on a 'scope does not mean you cannot find in in an FFT.

 

Scope plots are misleading if you don't know how to properly interpret it, so better to use wide-band spectrum analysis instead (I've provided it).

 

What a load of rubbish, the opposite is true. A 'scope shows you the exact waveform you are looking at. FFT averages huge numbers of samples. Wide band FFT is what is misleading, because even a severely distorted waveform (single cycle) can produce a low reading using FFT.

 

One must indeed understand the possibilities and limitations of measurements.

 

A simple 'scope trace is "raw", simply what it is. It shows you what you get for the window selected.

 

Any FFT will have a "window" thousands of times wider than a single 'scope screen and anything in this "window" is averaged to produce the FFT. Mathematically speaking, any FFT that does not use an infinite number of samples is flawed anyway.

 

But I've posted measurements for Sabre with PCM and SDM and the performance is pretty much the same for both.

 

Precisely. Have you ever wondered why?

 

Take a look at CS4398 and you may be able to exceed performance of both if you use it correctly.

 

I did more than that. I had a listen to several devices using it. Baring that all of them did a very poor job, the CS Chip based ones are about the ones that sound the least like music of anything I heard.

 

If you like sound of ESS' digital processing, good for you.

 

Cannot say I do. Unless the alternative choice you give me is CS4398.Then I probably love ESS.

 

If you want a bit-perfect DAC you may need to look elsewhere.

 

Why? I don't need to look. I have one.

[Miska]

Why? Because they are worse than classic old PCM ones?

 

I am only interested on how to make current DAC technologies sound great, that means delta-sigma. I'm not really interested in classics other than for history sentimental reasons.

 

You can fight the windmills and try to convince rest of the world that delta-sigma DACs are bad. I don't agree either, because they are great.

 

I've had Marantz CD-60 that had SAA7220 + TDA1541A, it sounded horribly digital. I have a DAC based on DF1700 + PCM63P and a DAC based on DF1700 + PCM1700 and they sound digital too. I have a DAC based on CS4328 and it sounds as good as DF1700 + PCM63P with tenth of the total price and tenth of the total complexity and it has tenth of the IMD...

 

But maybe you cannot see because massive averaging "hides" the noise, while it does not hide harmonic components?

 

If you use averaging yes, if no, not. However noise sounds like noise, so if you hear it it's hiss. Summing up the noise FFT components and relating it to the fundamental gives you figure and it includes everything without any doubt.

 

SDM creates tons of ultrasonic noise (where classic analogue amplifiers have very poor linearity) that are steady state, meaning they do not change with signal. The result is IMD that will invariably have some content folding back into the audio range. Because one component is (semi)random the resultant IMD will be essentially equally (semi)random and a heavily averaged FFT measurement will simply "tune out" this noise. A singe shot 'scope trace will show it easily at -40dBFS or so...

 

If you have it at -40 dB then the reconstruction filter is not proper. But same goes for PCM stairsteps too. Limit your scope plot to 20 kHz bandwidth and it's very clean and roughly tells you what you can hear. If you have uncorrelated noise at 1 MHz the difference frequency to 20 kHz is still 980 kHz.

 

If there would be IMD noise, it would still lift up the overall noisefloor, however I don't see that happening. In fact I tend to get better low level reproduction of -120 dBFS signals with DSD than with PCM (IOW when using the DAC's built-in oversampling filter and delta-sigma modulator).

 

Leaky oversampling filters, or even worse (shrug) NOS DACs, tend to have much worse IMD because of intermodulation with the ultrasonic images and those are heavily correlated unlike delta-sigma noise.

 

A singe shot 'scope trace will show it easily at -40dBFS or so...

 

So now give me a PCM DAC that has both clean time domain response and clean frequency domain response at the same time. What I usually see is this kind of time domain response:

...or this kind of frequency domain response:

 

DSD is the only format I've seen so far widely commercially available that gives reasonable compromise between the two at the same time.

 

While waiting for something better to be introduced as digital download consumer format, I'll stick with DSD.

 

PCM has been already spoiled at the ADC side with brickwall filters.

 

What a load of rubbish, the opposite is true. A 'scope shows you the exact waveform you are looking at. FFT averages huge numbers of samples. Wide band FFT is what is misleading, because even a severely distorted waveform (single cycle) can produce a low reading using FFT.

 

Any distorted components will however fire up corresponding peak in the FFT. Even if it's just one bit of single sample.

 

Precisely. Have you ever wondered why?

 

Because it all goes through same last stages, DSD just bypasses the digital oversampling filter stages. So DSD just gives wider bandwidth and thus better time/frequency domain response.

 

I did more than that. I had a listen to several devices using it. Baring that all of them did a very poor job, the CS Chip based ones are about the ones that sound the least like music of anything I heard.

 

I think then you didn't listen CS4398 in Direct DSD mode with external digital filters and modulator, bypassing everything else on-chip and only using the actual D/A conversion stages.

 

Cannot say I do. Unless the alternative choice you give me is CS4398.Then I probably love ESS.

 

Somehow I have a feeling that I have seen this before somewhere else. I think there are bunch of DAC manufacturers and people who disagree. But it's your opinion anyway.

 

Why? I don't need to look. I have one.

 

OK, so you have a NOS R2R DAC and you claim it measures great? Quite contradictory unless it's something like Phasure NOS1 that is intended to be run with off-board oversampling. I'm even more eagerly waiting to see some measurement results and hear why such great device would be discontinued.

[Miska]

SACD has been struggling from the first and it currently has pretty close to no support. Blue Ray, on the other hand, has become ubiquitous. If SACD/DSD were actually better, like Blue Ray, it would have become ubiquitous. The very existence of this thread tells me that if there's any difference, it's a "why bother" difference.

 

How many music-only Blu-rays there are out there (no picture). And how many Blu-ray players there are out there that you can conveniently use without having display attached. I have two Blu-ray players, PS3 which is pretty OK as such and Samsung which is something that I cannot find proper curse words for.

 

And yes, you can get pretty decent performance out of DSD, and also PCM at 352.8/384k rates.

 

The main difference is that DSD usually gets much less heavy lifting inside the DAC (and ADC) than PCM.

[sandyk]

So now, here we are days later and 270 comments on comparing DSD with standard CDs - Diogenes

 

PCM has been already spoiled at the ADC side with brickwall filters.- Miska

 

 

Miska

Why do you keep throwing that in ? That's comparing apples with oranges.

It certainly isn't a problem with material like the 24/192 PCM recordings from Barry.

SACDs ultrasonic artifacts are deliberately rolled off in commercial players, and perhaps that smoothing is preferred by many SACD devotees? i.e. it sounds less "digital".

Alex

[Hiro]

Miska

Why do you keep throwing that in ? That's comparing apples with oranges.

It certainly isn't a problem with material like the 24/192 PCM recordings from Barry.

 

Aren't most of Barry's remasterings available on CD only i.e. 44.1kHz PCM?

 

SACDs ultrasonic artifacts are deliberately rolled off in commercial players, and perhaps that smoothing is preferred by many SACD devotees? i.e. it sounds less "digital".

 

I don't think that the inaudible frequencies are responsible for the audible differences in SQ at all. The streamlined processing of DSD plays a much bigger role here IMO.

[Teresa]

Aren't most of Barry's remasterings available on CD only i.e. 44.1kHz PCM?

 

No, Barry Diament owns Soundkeeper Recordings and makes new original audiophile recordings at 24 bit 192kHz PCM, not remasters which are available in the following formats:

• CD
  
• CD-R
  
• 24/96 audio only DVD-R
  
• 24/96 AIFF or WAV files-on-disc DVD-R
  
• 24/192 AIFF or WAV files-on-disc DVD-R
  

Perhaps you were thinking of someone else?

[Hiro]

How many music-only Blu-rays there are out there (no picture).

 

I don't think the number of such discs is anywhere near 9000, and that's the number of SACDs released so far...

[Hiro]

No, Barry Diament owns Soundkeeper Recordings and makes new original audiophile recordings at 24 bit 192kHz PCM, not remasters which are available in the following formats:

• CD
  
• CD-R
  
• 24/96 audio only DVD-R
  
• 24/96 AIFF or WAV files-on-disc DVD-R
  
• 24/192 AIFF or WAV files-on-disc DVD-R
  

Perhaps you were thinking of someone else?

 

I was referring to his classic rock/pop remasters (not the four 24/192 recordings of his), to my knowledge they are all available on 16bit CDs only.

[sandyk]

Aren't most of Barry's remasterings available on CD only i.e. 44.1kHz PCM?

 

 

I don't think that the inaudible frequencies are responsible for the audible differences in SQ at all. The streamlined processing of DSD plays a much bigger role here IMO.

 

Of course you would say that. Many members have previously suggested that you are on Sony's payroll. Even if that isn't correct, it sure comes across like that to many. Sony had to introduce that HF rolloff due to happenings like amplifiers being destroyed originally, back in the days when Sony even sold tweeters with HF extension to 100kHz, especially for their new

(at the time) SACD format. It's a matter of record about such problems, which should still be able to be found using Google.

 

Barry's recordings are supplied to order at 24/192 on a DVD.

I have 2 such DVDs from Barry. They are of course meant for PC use, although they can also be used for creating your own 24/192 DVD-As.

[Hiro]

Of course you would say that. Many members have previously suggested that you are on Sony's payroll. Even if that isn't correct, it sure comes across like that to many. Sony had to introduce that HF rolloff due to happenings like amplifiers being destroyed originally, back in the days when Sony even sold tweeters with HF extension to 100kHz, especially for their new

(at the time) SACD format. It's a matter of record about such problems, which should still be able to be found using Google.

 

No argument = you work for Sony, and did I mention blowing amplifiers - how many of these were reported in the last 10 years?

 

My point is simply that the audible differences in SQ are the result of the removal of downsampling, decimation, brickwall filtering, upsampling, and oversampling from DSD recording/playback chain, which were time and again proven to impact sound quality. The frequency response of SACD (with 50kHz low pass filter) and 24/96 is basically the same.

[Hiro]

Yes, or I could make the same at 16/352.8 to match the bitrate, but doesn't matter much for this context as it's mostly limited by the analog filtering.

 

Really curious about the results of this comparison (limitations of the analog filtering notwithstanding), and even more so about this one, 32/176.4 PCM vs DSD128, as the 16/352.8 pcm thing doesn't exist in nature

[sandyk]

Super Audio CD - Wikipedia, the free encyclopedia