DCS Vivaldi

[Phelonious Ponk]

It's a stack after all and you will have the whole cabling make a "huge" difference issue, by using different cables you can bring more "synergy" to the system , etc..

While I have the engineers in the room, I'll raise this one again, perhaps showing my ignorance: I've never understood why the basic principal you're implying here -- that more cabling, a stack, more stuff (again, the technical term) -- is more opportunity for signal degradation, doesn't apply (perhaps even more so) to analog as well. Why is a separate pre, phono stage, monoblocks, etc., be superior? Why don't all of those cables, power supplies, and redundant parts degrade the signal? Why wouldn't a well designed, highly integrated approach actually generate less noise and distortion?

Please remember to speak slowly and be gentle. And my apologies for veering off-topic.

Tim

 

[opus111]

I'll take all the simplification I can get, thanks. Even at that, my lips are moving as I read your posts, so allow me to leap to a grossly over-simplified conclusion for you to debunk: You seem to be of the belief (and there is admittedly much assumption going on here), that simpler is better -- no external re-clocking, no oversampling (unless it is integrated into the DAC chip itself to compensate for performance), no nonsense. A fair assessment? So would the theoretical ideal would be?.... an SOTA DAC chip + ? In other words...

Its more of a working hypothesis yeah, but belief is close enough. Its also my experience that Einstein's dictum 'Everything should be as simple as possible, but no simpler' holds true in audio development.

In your view, what constitutes "a better DAC?"

One that delivers more listening pleasure. I like to be 'wowed' or surprised and even sometimes shocked by what my DAC reveals of a disk. Of course a lot of disks don't surprise or shock but that's not a reflection of the DAC.

So...what does all of this mean? You think common jitter levels are low enough and don't really need further reduction?

Just speaking for my own audio journey here, the most annoying things about DACs (by which I mean audible artifacts) seem to have nothing to do with jitter. If I'm hearing the effect of jitter, perhaps its a slight defocussing of the placement of instruments. Soft focus, if you like and not at all unpleasant so I've not sought to minimize it so far. But once all the other artifacts of DACs are dealt with then I'll play with jitter reduction and see what happens.

You don't think jitter is a particularly audible distortion?

More that it doesn't seem to interfere with my listening pleasure. I get tons of fun out of listening to my DAC which has had precisely zero attention paid to jitter to date. It is SPDIF input, which allegedly is a 'flawed' interface with tons of nasty jitter, but you could have fooled me

Clearly it is not a priority for you, while it seems to be a huge issue, maybe the biggest issue, for high-end jitter designers. What's your position?

My position is (for now, until further data appears) that people confuse jitter with RF induced noise. 'Jitter' has become a kind of catch-all for anything nasty sounding in digital. The marketing guys seem to overplay it, many people in digital appear to suffer from some kind of jitter-phobia. This is one side of the coin, the mosquito being strained out of the market's collective beverage. The camel they're swallowing is the sound of the mainstream S-D DAC chips.

Oh and if not "high-end," what kind of DACs do you design? Under what brands?

So far not under any brand. I doubt any existing brand would like to have one of my DACs - I just design and build one-offs, don't manufacture. But in the longer term I'm looking for manufacturers and distributors for my DACs. They'll be aimed at DIYers initially - think along the lines of the ODAC but better sounding and potentially cheaper too. My aim is to get at least in the same ballpark as the high-end for sound, but at a much cheaper price. That then becomes disruptive.

 

[DonH50]

Handwaving alert.

Cables are a very small part of the preformance in most systems. Separate components are typically designed with higher-performance parts, beefier power supplies, offer better isolation from radiated and coupled noise and the like (e.g. the power amp does not modulate the preamp's power supply if they are separate boxes, and the digital circuits in a video processor do not add noise to the TT preamp if they are separate, etc.) Something like an AVR that includes DSP, low-level analog signal processing, data conversion (DACs and ADCs), and power amps provides lots of opportunities for noise coupling among various signal blocks. On the plus side, integrating more components in a single block allows shorter signal paths with fewer connectors to wear out and such. Everything's a trade. Audibility is always debated.

 

[opus111]

Call it by any name you please , naive will come to mind, this saddens when there is real progress to be made in there and at more than reasonable prices and this won't stop the manufacturers from offering more smoke and ever higher priced digital gear ..

It doesn't sadden me, I see this as an opportunity. 'Lament it or leverage it' is one of my mottos. A dysfunctional market such as this is ripe for being disrupted, which is what I'm intending to do.

It will help if the company has some real engineering and professional music background. It is no surprise then that Weiss and DcS are at the top of the digital high price heap... Oh Well !!

Weiss, when I last looked at a pic of their innards were using an off-the-shelf TI/BB DAC chip. I've heard at least one member of that family - the timbres of instruments are corrupted. How this chip family can be classed as 'high-end' is beyond me.

 

[AudioExplorations]

And one thing about volume control... Do you really think they mean you have 2v and 6v as the only options for "volume control"?

These are the coarse settings of the analogue output stage voltage when the digital attenuation is 0dB. Obviously the voltage can be further reduced below 2v or 6v with the digital attentuation. There are DAC's on the market (i.e. Weiss Medea) where the output stage voltage can be adjusted anywhere from 0.5V - 16V on top of the futher fine-tuning with digital volume control.

 

[AudioExplorations]

Weiss, when I last looked at a pic of their innards were using an off-the-shelf TI/BB DAC chip. I've heard at least one member of that family - the timbres of instruments are corrupted. How this chip family can be classed as 'high-end' is beyond me.

The latest Weiss products (DAC202, MAN301 and Weiss Medea Plus) use the ESS 9018 Sabre.

 

[opus111]

Why don't all of those cables, power supplies, and redundant parts degrade the signal?

My view is - they do. Fewer cables, fewer transformers, fewer mains leads is going to be better.

Why wouldn't a well designed, highly integrated approach actually generate less noise and distortion?

My view is - it would. But seems the market isn't interested in that, it likes its traditions of separate boxes. Perhaps bring to mind that cables have probably the highest margins of all for manufacturers and retailers? How would those retail guys survive if punters were no longer encouraged to 'devote 10-20% of your budget to cables' ?

Please remember to speak slowly and be gentle. And my apologies for veering off-topic.

S l o w l y e n o u g h ?

 

[opus111]

The latest Weiss products (DAC202, MAN301 and Weiss Medea Plus) use the ESS 9018 Sabre.

That's a step in the right direction for sure. ESS is about the only company openly acknowledging S-D pitfalls. I've not heard Sabre so I'll not comment on how it may sound.

 

[DonH50]

It takes about 100 ps of random jitter at 20 kHz (and much more at lower frequencies) to reach the lsb level of a 16-bit DAC. At 1 kHz it takes over 2 ns (see e.g. http://www.whatsbestforum.com/showthread.php?1322-Jitter-101). That's quite a bit, and that is to induce an error that is some 96 dB below full-scale. Add some music or a movie and it is unlikely you'd hear several ns of random jitter. My belief (shared by others) is that random jitter is essentially a non-issue for modern systems.

Signal-dependent (deterministic) jitter is another issue, with lsb-level jitter causing significant reductions in spurious-free dyanmic range. Same for adding a bit of clock or other signal to the input (or vice-versa). For example, lsb-level injection of a fixed tone can reduce the maximum spur from ~140 dB below full-scale to "only" 96 dB or so. See my series on Jitter (101, 102, plus a thread on cable bandwidth and jitter) in the technical area of WBF (you can start at http://www.whatsbestforum.com/showthread.php?2829-Don-s-Tech-Series and click on the links). Isolating the clock and signal is critical for high-resolution converters, and so a trade can be made between keeping the clock close to reduce added noise, and moving it away to keep it from coupling to the signal. However, the maximum clock coupling typically occurs inside the DAC, so keeping the clock source near the DAC is generally optimal.

Asynchronous DACs and other schemes serve to reduce clock jitter (random and deterministic) at the DAC, where the actual conversion takes place. On the output side, the analog buffer and image filters typically (almost always) dominate the performance.

The AES spec is for studios, where you want a number of components all sharing the same master clock to stay closely synchronized during recording, mix-down, and mastering sessions that can take hours and days. I don't know if you actually need 1 ppm/year stability; without running any numbers my guess it is another of those things chosen to ensure inaudibility of any artifacts induced by clock drift.

HTH - Don

 

[DonH50]

Weiss, when I last looked at a pic of their innards were using an off-the-shelf TI/BB DAC chip. I've heard at least one member of that family - the timbres of instruments are corrupted. How this chip family can be classed as 'high-end' is beyond me.

Delta-sigma or conventional architecture?
Do you think it was the DAC or something in the other circuit implementation?

 

[asiufy]

These are the coarse settings of the analogue output stage voltage when the digital attenuation is 0dB. Obviously the voltage can be further reduced below 2v or 6v with the digital attentuation. There are DAC's on the market (i.e. Weiss Medea) where the output stage voltage can be adjusted anywhere from 0.5V - 16V on top of the futher fine-tuning with digital volume control.

I know that My point was that opus111 put that as a limitation of the dCS gear, that it could "only" do 2 output voltages... That's why I asked him if he ever used a dCS product, because if he did, he would've realised there was volume control on top of the pre-set output voltages.

Again, it surprises me how people, specially "engineers", can disparage a product without ever even seeing it being used, let alone hear how it sounds...

 

[opus111]

Delta-sigma. I'm convinced it was the DAC - I had already played with the I/V opamp.

 

[asiufy]

My view is - they do. Fewer cables, fewer transformers, fewer mains leads is going to be better.



My view is - it would. But seems the market isn't interested in that, it likes its traditions of separate boxes. Perhaps bring to mind that cables have probably the highest margins of all for manufacturers and retailers? How would those retail guys survive if punters were no longer encouraged to 'devote 10-20% of your budget to cables' ?



S l o w l y e n o u g h ?

"Tradition of separate boxes"? How many separate "stacks" exist in the market, and how many "all in one" solutions are out there? I'd guess the "tradition" is for the integrated approach, not the other way around...

And I see that unfortunately this thread has been infected with the "cables don't matter" mantra. It's rather unfortunate because the dCS gear in particular benefits from quality cables.

 

[opus111]

I know that My point was that opus111 put that as a limitation of the dCS gear, that it could "only" do 2 output voltages...

If you take the time to read what I wrote I didn't mean that it could 'only do' 2 output voltages. My comments were about compatibility with pre and power amp inputs. And yes I do realize there's volume control beyond the 2 output sensitivity levels, even though I've not used one.

Again, it surprises me how people, specially "engineers", can disparage a product without ever even seeing it being used, let alone hear how it sounds...

It doesn't surprise me in the least that you would attempt to disparage an engineer by using scare quotes around the term.

 

[AudioExplorations]

I know that My point was that opus111 put that as a limitation of the dCS gear, that it could "only" do 2 output voltages... That's why I asked him if he ever used a dCS product, because if he did, he would've realised there was volume control on top of the pre-set output voltages.

To me this is a limitation of dCS gear! I run power amp direct and due to my amp and speaker sensitivity I need ideally 0.5V so that only minor digital attenuation is used. The Vivaldi (or any other dCS gear for that matter) would never work in my system as with the lowest 2V setting I would need to use large amounts of digital attentuation. I think this is what opus was saying - not everyone's amp/speaker sensitivities are the same and this latest so called "cost no object" dCS only provides 2 options.

 

[opus111]

"Tradition of separate boxes"? How many separate "stacks" exist in the market, and how many "all in one" solutions are out there? I'd guess the "tradition" is for the integrated approach, not the other way around...

It seems likely you are misreading the context. Tim is talking about analog systems here, not digital 'DAC stacks'.

And I see that unfortunately this thread has been infected with the "cables don't matter" mantra.

Yours is the first reference I've seen here to that mantra. "The best cable is no cable" happens to be mine.

 

[opus111]

To me this is a limitation of dCS gear! I run power amp direct and due to my amp and speaker sensitivity I need ideally 0.5V so that only minor digital attenuation is used. The Vivaldi (or any other dCS gear for that matter) would never work in my system as with the lowest 2V setting I would need to use large amounts of digital attentuation. I think this is what opus was saying - not everyone's amp/speaker sensitivities are the same and this latest so called "cost no object" dCS only provides 2 options.

Spot on - you got me in one.

 

[Phelonious Ponk]

"Tradition of separate boxes"? How many separate "stacks" exist in the market, and how many "all in one" solutions are out there? I'd guess the "tradition" is for the integrated approach, not the other way around...

Thread drift. My bad. What you're responding to here is Opus' response to my question about the possible benefits of integration in analog, where separate stacks are the high-end norm. And it's not so much a "cables don't matter" discussion as a "wouldn't no cables (and fewer noise-generating components) be better" one. And pretty theoretical at that.

Tim

 

[Phelonious Ponk]

It takes about 100 ps of random jitter at 20 kHz (and much more at lower frequencies) to reach the lsb level of a 16-bit DAC. At 1 kHz it takes over 2 ns (see e.g. http://www.whatsbestforum.com/showthread.php?1322-Jitter-101). That's quite a bit, and that is to induce an error that is some 96 dB below full-scale. Add some music or a movie and it is unlikely you'd hear several ns of random jitter. My belief (shared by others) is that random jitter is essentially a non-issue for modern systems.

Signal-dependent (deterministic) jitter is another issue, with lsb-level jitter causing significant reductions in spurious-free dyanmic range. Same for adding a bit of clock or other signal to the input (or vice-versa). For example, lsb-level injection of a fixed tone can reduce the maximum spur from ~140 dB below full-scale to "only" 96 dB or so. See my series on Jitter (101, 102, plus a thread on cable bandwidth and jitter) in the technical area of WBF (you can start at http://www.whatsbestforum.com/showthread.php?2829-Don-s-Tech-Series and click on the links). Isolating the clock and signal is critical for high-resolution converters, and so a trade can be made between keeping the clock close to reduce added noise, and moving it away to keep it from coupling to the signal. However, the maximum clock coupling typically occurs inside the DAC, so keeping the clock source near the DAC is generally optimal.

Asynchronous DACs and other schemes serve to reduce clock jitter (random and deterministic) at the DAC, where the actual conversion takes place. On the output side, the analog buffer and image filters typically (almost always) dominate the performance.

The AES spec is for studios, where you want a number of components all sharing the same master clock to stay closely synchronized during recording, mix-down, and mastering sessions that can take hours and days. I don't know if you actually need 1 ppm/year stability; without running any numbers my guess it is another of those things chosen to ensure inaudibility of any artifacts induced by clock drift.

HTH - Don

OK, Don. Opus is doing good, but you need to speak much.....more.....slowy.... And use smaller words. If you engineers want us to write the marketing copy for you, you have to give us language that we, and the reading audience, will understand. If you don't, being marketing guys, we'll just make something up, like, "the DonOpus Premier Ultra OneDac simply takes a deeper dive into the data on your disc or hard drive, extracting details that you never knew were there, that you've never heard before. But it accomplishes this Herculean feat without ever delivering that cold, etched hyper-detailed sound. It will transform your digital listening to one much more akin to SOTA analog. It will not sound like data playing through a system. It will sound like sweet, sweet music."

Tim VPBS, Marketing

 

[Phelonious Ponk]

Opus said:

I like to be 'wowed' or surprised and even sometimes shocked by what my DAC reveals of a disk.

Wait a minute, now you're confusing me. Earlier you said a DAC couldn't get any more information from a disk, it could only do it's job of converting the digits that are there....

Tim