DSD comparison to PCM.

garylkoh said:

I'm not taking sides....... here's a new article by Lynn Olson on Positive Feedback
http://www.positive-feedback.com/Issue65/dac.htm

Click to expand...

From the article:

"Yes, [DACs with delta-sigma DAC chips] all sound modern and up-to-date; smooth, pleasant, free of the grit-and-grain that plagued early digital, and play 88.2, 96, and 192 kHz digital with the greatest of ease. But there's something missing. I don't have any fancy audiophile words for it, but it's an absence of life, of sparkle, of that intangible sense of being right in the room with the performer."

Pretty much sums up my experience with all the delta-sigma DACs I've heard too.

Back in 1990, I switched from a multibit Rotel RCD-955 CD player (which I loved) to the newer 1-bit RDC-965, based on raving magazine reviews of the newer machine. To my ears the new machine sounded dull, boring and lifeless compared to the older machine. I lived with it for a few years, but was quite frankly never happy. I then started work and treated myself to a Rotel RHCD-10 - a beautiful machine with PCM63K chips. Wow, the magic was back big time - stunning CD player. I kept this machine for well over ten years and in this time compared it to a bunch of other CD players. In all this time, the only one that I thought sounded better was the Naim CD555... with PCM1704 chips.

Maybe some of us are more sensitive to the side-effects of delta-sigma modulation. Dunno. But I've never been able to live with it.

Mani.

garylkoh said:

I'm not taking sides....... here's a new article by Lynn Olson on Positive Feedback
http://www.positive-feedback.com/Issue65/dac.htm

Click to expand...

Sigh. The ESS video mentioned in the article has been put up in this forum a few times by jkeny but each and every time, it was either ignored or dismissed. Most complained they don't have the time to wade through. When the exact time in the video was posted, the thread suddenly went dead. One member even went as far as to say why would he trust the data of a man/company who/which is pushing the superiority of its own product. I guess nobody likes to be told they might be wrong.

IMHO, not a lot of people are interested in looking for answers, just cherry picking whatever data/information to champion whatever format they are vested in. In another thread, I've suggested that if people really want to have a test of which format is best, hold a test that actually test each format at its best. To simplify the test, use a master tape for conversion and limit it to DSD, Hires PCM and RBCD via NOS 16bit. I've suggested using Playback MPD-5, Phasure NOS1 768 and Opus's dac. Mani indirectly have brought up a better idea, use a PM model 2 ADC to do the PCM files. But guess what, it died a natural death, as usual. There are vociferous defenders of DSD but every time, Lipschitz and Vanderkooy's calculation is brought up, the threads will go dead. See the pattern here.

I myself uses a NOS 1541 dac and am not blind to its drawbacks, frequency droop and imaging artifacts. It is in spite of this that I still prefer NOS because in my case (YMMV), its strengths outweighs the drawbacks. I am not under any illusion that it is the best (obviously it is not), it is just the best for me. But then I also believe what makes a dac is not just its digital output but the overall package, ie. input and output stage, power supplies and attention to emi/rfi artifacts. I don't know how many dacs I have seen that just slaps on a off the shelf SMPS power supply package and call it a day and yet every now and then, that dac is the champion of the day. All I can say is I will be very very skeptical.

I guess my point is most people argue for the sake of maintaining their vested interest but has anybody done any test rigorous enough to even come close to conclude which format is best? When you compare formats using a DSD dac, it goes without saying DSD is going to sound best. Bruce says the phasure is nothing special (i don't know because I've never heard it) but did he let it performs at its best? The phasure has been said to need the XXhighend and a very powerful PC to be at its best. Did he go to that length? Not likely. As far as I am concerned, no test has come near to conclude what got every one so excited about, ie DSD vs Hires PCM vs RBCD. And there will never be. If not, everyone will be listening to one dac, one amp and one speaker because it is the best.

Sigh. The ESS video mentioned in the article has been put up in this forum a few times by jkeny but each and every time, it was either ignored or dismissed. Most complained they don't have the time to wade through. When the exact time in the video was posted, the thread suddenly went dead. One member even went as far as to say why would he trust the data of a man/company who/which is pushing the superiority of its own product. I guess nobody likes to be told they might be wrong.

Click to expand...

It is all not simple of course, and only to some extend we are able to grasp real merits - but it may be better to say we don't understand anything of it anyway and next give up. Btw, I think that time in the video would be ~38 minutes into it.
So, thread dead again ?

IMHO, not a lot of people are interested in looking for answers, just cherry picking whatever data/information to champion whatever format they are vested in. [...]

Click to expand...

I think you nicely summarized what I said in that other post in a more complicated way. But remember my half-way conclusion in there which went something like "since we can't get hold of the technical merits we have to listen and compare - and that is impossible because ..." and then your story.

Bruce says the phasure is nothing special (i don't know because I've never heard it) but did he let it performs at its best?

Click to expand...

Allow me : Spot on. Also the very reason for that other post and I tried to squeeze in something like you say here, but found it too explicit. But to summarize this, two main topics :
1. NOS is still to be filtered;
2. When done the common-means way it can be regarded an OS DAC.
... which stupidly enough indeed leads to XXHighEnd only (I can't help that) because of its dedicated filtering for this (which btw is named Arc Prediction). Add to this that only through XXHighEnd the "old version" of the NOS1 can do 24/384 (otherwise it's 24/192) and just know how much this matters (which is why the NOS1-USB does 24/768 because again it vastly matters).
And now I must be careful that this is not about me or my products, but were it about the formats and how to judge them, it IS important.

That article is nice, although it seems to deviate from all subjects started while it began so nice about how SMDs sound all the same and different from PCM (R2R) converters. Anyway, including diving into some of the links in there and looking at that YouTube, this is what we should get from it :

- With SDM you listen to an algorithm;
- With SDM you listen to what any (software) engineer could make out of it and this is different for all engineers (manufacturers);
and my additions :
- Since all sound different none will be good (though subjectively we may like it);
- With PCM and any common-means filtering we again listen to an algorithm and again we listen to what one could think of. But :
- Today we can do this ourselves to a large degree, like e.g. iZotope provides quite some options.
and again but :
- Only with a DAC which doesn't do a thing herself this can be fully utilized. We name them NOS/Filterless.

Btw before it is misunderstood or not thought about : Any next upsampling step needs 1 additional bit. So if Redbook is the base, 2x needs 17 bits, 4x needs 18 bits, 8x needs 19 bits - and so on. And of course this sadly means that a TDA154x can not be used (oh it can, but the result will be sub-optimal and less).

Peter

It is in spite of this that I still prefer NOS because in my case (YMMV), its strengths outweighs the drawbacks.

Click to expand...

In my view it is most of the time overlooked that almost nothing in electronics can be compared for design decisions, and most certainly this can't be done when combined with digital (and while digital within itself again exhibits the same). Still referring to that important line above, here is a good example of it :

We want to (Async USB) noise isolate DACs from the PC. Well, first off we can find more points to do that. E.g. can be in between the USB cable, or can be in between i2s. Both make a difference, because the one is theoretically subject to jitter (i2s) and the other is not (Async USB). So let's take USB and a common solution : glass. We get ourselves an Adnaco solution (which works at the real higher USB speeds).
Too bad that the power supply needed for it at the DAC end exhibits more noise than the native USB connection ever contained (when done well) and now impedes jitter because of that.
So we take i2s. Analog isolation applications for this (MHz++) speed do not exist, but digital do. Fine. Too bad that any digital solution carries jitter (these chips come with jitter specs) and thus can not be used.

So read the quote above again and never ever anymore think that we will be able to easily see how DSD compares to PCM because it involves too many benefits but drawbacks at the same time, even jitter being amongst it (don't ask any engineer, but I claim that any 64 times higher sampling speed needs a 64 times better jitter spec).

Peter

Guys

let me get it straight, because the discussion is losing focus to me.

Given a piece or music e.g the famed Mercury Living Stereo. On SACD and on CD. I have two dacs about the same price: A Berkeley DAC Alpha and a Playback Design PD-3. Reproducing system is the same in both cases. Which one do you expect to sound better? Or is it that the only time PCM will sound superior to DSD is through a NOS DAC ? (Not my opinion but I would like the expert to chime) ...

Taking Lynn's experience, he doesn't have a NOS DAC as far as I'm aware, but it is multi-bit (PCM1704). I reckon its only necessary to have multi-bit to hear the difference, NOS would be more icing on the cake.

Incidentally I have boxes of 'Mercury Living Presence' and also 'Living Stereo' - the former sound great but the latter suck for dynamics. When I investigated as to why, I found the CDs in the 'Living Stereo' box are all DSD remastered.

bbb said:

(...) IMHO, not a lot of people are interested in looking for answers, just cherry picking whatever data/information to champion whatever format they are vested in. In another thread, I've suggested that if people really want to have a test of which format is best, hold a test that actually test each format at its best. To simplify the test, use a master tape for conversion and limit it to DSD, Hires PCM and RBCD via NOS 16bit. I've suggested using Playback MPD-5, Phasure NOS1 768 and Opus's dac. Mani indirectly have brought up a better idea, use a PM model 2 ADC to do the PCM files. But guess what, it died a natural death, as usual. There are vociferous defenders of DSD but every time, Lipschitz and Vanderkooy's calculation is brought up, the threads will go dead. See the pattern here. (...)

Click to expand...

I think you have a wrong idea of an audio debate, mainly at WBF. People are not looking for answers because they know that in these matters we can not have a definite answer just carrying a few experiences in a forum, and most of it there are no definitive answers. However, they can filter many posts with experiences and opinions from users and learn from them. And yes, people usually defend what they prefer ...

I have not invested in the DSD format and I do not care of the Lipschitz and Vanderkooy's calculations. But, if members with great systems, whose public opinions I respect since long as their experience agrees with mine in many other aspects of sound reproduction, tell me that DSD is a better format I am better prepared to accept it. It is what I learn from WBF. In my view, the big advantage of long threads and debates is that it allows members to know each other better and valuate others opinions accordingly. Incisive questions and challenges are most of the time just welcome triggers, that animate the debates.

FrantzM said:

let me get it straight, because the discussion is losing focus to me.

Given a piece or music e.g the famed Mercury Living Stereo. On SACD and on CD. I have two dacs about the same price: A Berkeley DAC Alpha and a Playback Design PD-3. Reproducing system is the same in both cases. Which one do you expect to sound better? Or is it that the only time PCM will sound superior to DSD is through a NOS DAC ? (Not my opinion but I would like the expert to chime) ...

Click to expand...

About losing focus ...
Suppose we take it as a fact that any conversion from DSD to PCM or the other way around is to be avoided.
Do you now have your own answer ?
And suppose you have, do we know whether that Playback Designs doesn't convert to PCM first ? (most DSD DACs do).
As far as I am concerned the Alpha sure does convert to DSD-like (because it's SDM based).

So ?
So I don't have the answer. It needs real digging and this starts with the origine of the recording. But assumed it is the same recording *and* same mix (chances ?), one of the two will have to lose. And with that no NOS is guaranteed to be the better one.
Some say that NOS (with proper filtering ) should win when it is known that the original recording was through a PMII because ...
... some say that this has a PCM ADC. It might have. Still, any Hires recording (which surely will have been applied when the PMII was in order) needs decimation of some sort (to 16/44.1 Redbook) which is very similar to that common-means "OS" filtering used through playback (like e.g. the Alpha uses that kind). And now the ringing and such is on the CD.

Do we now finally know the answer ?
I sure not !
Peter

microstrip

This business of appeal to authority is an argument fallacy. Someone can have the greatest system but still doesn't mean he or she holds the truth. I also do care about Scientific papers. Science is the basis of technology and so is empiricism ... I'll leave it at that.


@everyone, in particular Opus11, many, PeterSt, DonH50, AMirm and other I am forgetting .

Could you write/a little primer on the different path taken by designers on DACs. A way to summarize this DAC affair, a way for the layman to understand and thus make more informed decisions. brands and designers approach and your personal observations on thier sound.
Thnank you in advance

You'd like each of us to write a primer? I don't think I could write anything approaching a summary, but I'm happy to open a thread where I post up random musings about various topics in relation to DAC design. For example digital receivers, power supplies, DAC chip datasheets and how to read them, analog stages, filtering and noise control. How does that sound? One thing it won't be coming from me is succinct though, I tend to blabber on rather...

No Problem

There's need a need to catalog things in a way that would allow those that don;t have infinite access to gears to make informed decision. While implementation is key, all design choices aren't equally good. There are also gears of distinction that are totally unknown simply because the company or their designer cannot afford or don't know ho do the necessary marketing. being able to pinpoint these gens would help us all. Or so I think.

garylkoh said:

I'm not taking sides....... here's a new article by Lynn Olson on Positive Feedback
http://www.positive-feedback.com/Issue65/dac.htm

Click to expand...

That's great Gary! Required reading for sure.

Frantz,

To me this seems impossible. But for an impression, read this post but promise to stop after reading that one or otherwise you well get mad. There's over 400 posts in there like that one, although most from my hand are only half this length. It nicely touches a few of the subjects we talk about in here though, so you could try to read this one (with the length of a whole page in this forum). But not advised.
But to give the general idea : it still took 18 months before this (commercial) product was finished and the first DAC went out. After that I right away started on the USB interface, and along the project I worked hard on Firewire (which didn't make it).

"Design decisions" is a hard word. We are driven by the availability of parts, "constructions" from it possible, while at first all starts with a nice idea; at some stage you can't go back because too much time will have been lost otherwise, while *still* it is to wait about how the sound result will be. For me the objective was a bit constraining because I wanted to go for the very best only. I recall having spent 3 months in a row only to *afterwards* squeeze in a not-deteriorating volume control. It could not be done. Why ? wrong design decision. "Never use the PCM1704 for that". This, while it had been the base of it all (see the thread here about NOS and 24 R2R etc.) and most of the time was spent on the current to voltage conversion (I/V). My luck back at the time : there wasn't anything else anyway so I couldn't regard it as wrong design decision.

Regards,
Peter

bbb said:

Bruce says the phasure is nothing special (i don't know because I've never heard it) but did he let it performs at its best? The phasure has been said to need the XXhighend and a very powerful PC to be at its best. Did he go to that length? Not likely. As far as I am concerned, no test has come near to conclude what got every one so excited about, ie DSD vs Hires PCM vs RBCD. And there will never be. If not, everyone will be listening to one dac, one amp and one speaker because it is the best.

Click to expand...

I have some of the most powerful computer workstations available. I spent over a week listening to this DAC with some of the best native DSD and PCM material available (and unobtainium!). No one is going to sit here and tell me I didn't give it it's due. I spent my hard earned money and time seeking out 2 NOS DACs in the chase of the Holy Grail. It certainly wasn't in these DAC's. I sold one and the other is still sitting in the rack.
As far as I'm concerned, I already have the Holy Grail of DAC's in the MPS-5 and MSB.

Bruce, you used the 1st generation 24/384 NOS1 DAC - the current async-USB 24/768 is much better.

Did you use the current version of XXHighEnd to feed your 1st generation NOS1? Was it a fully activated version? Were you using it in 'Minimize OS' mode? How about 'Phase Alignment' - what settings did you use? I'll stop there. There's a lot more than I've described above. If you didn't even do this, you certainly didn't give it it's fair due.

In any event, if you're happy with your MPS-5 and MSB DACs, that's cool. You obviously like the delta-sigma sound.

Mani.

Why are some people jumping to conclusions about what Bruce did or didn't do without asking him first before assuming?

opus111 said:

Taking Lynn's experience, he doesn't have a NOS DAC as far as I'm aware, but it is multi-bit (PCM1704). I reckon its only necessary to have multi-bit to hear the difference, NOS would be more icing on the cake.

Incidentally I have boxes of 'Mercury Living Presence' and also 'Living Stereo' - the former sound great but the latter suck for dynamics. When I investigated as to why, I found the CDs in the 'Living Stereo' box are all DSD remastered.

Click to expand...

What do you think of Lynn's article? Do you concur with the analogy to non-feedback low powered Class A v A/B feedback...and the criticisms laid at the feet of S-D?

Any of you guys looked at the pro products from Grimm Audio?
Just mentioning because what makes them interesting is Bruno Putzeys (heavily involved in DSD back in 2004 at Philips Applied Technologies along with engineering in Class D) and Peter van Willenswaard.
Both have excellent background in digital and DSD.
http://www.grimmaudio.com/pro_products.htm

They both have written excellent articles on DSD and digital.
Cheers
Orb

Orb said:

Any of you guys looked at the pro products from Grimm Audio?
Just mentioning because what makes them interesting is Bruno Putzeys (heavily involved in DSD back in 2004 at Philips Applied Technologies along with engineering in Class D) and Peter van Willenswaard.
Both have excellent background in digital and DSD.

Click to expand...

I'm using the AD1 and OC1

I suppose that people opinions on these matters are strongly enforced by the systems they are using in their listening experiences. I have listened to a few top digital systems and I have found that my preference would be dominated by the remaining equipment. I would never feel able to speak about the generic sound characteristics of the chip that is being used based on listening a just a few implementations .

One of my more challenging experiences in digital was carried decades ago when I modded an old Sony X7 ESD - a marvelous machine - the drawer movement was something special. It used a BB PCM58 k grade - I think it was an 18 bits chip. I removed all the analog stage, IV converter included, and replaced it with a grounded 50 ohm high quality resistor followed by a home hand made coil and polystyrene capacitor implementing a 100 kHz LC filter. Perhaps it was DIY bias expectation, but it was by far the more sweet and detailed digital I had listened for a long time, happily my preamplifier was able to compensate for the low output. It sounded very different from the original. Just to tell that in this case the sound of the machine was dominated by the analog stage, not by the DAC . All IMHO.