Valin's new MSB Reference dac & transport review, AS Product Year Award

[Al M.]

I represent MSB and I wish to clarify exactly what is upgradable in all of our current DACs so there is no ambiguity or confusion (Discrete DAC, Premier DAC, Reference DAC and Select DAC).

Software is all independently upgradeable via user installed firmware upgrades. This includes the user interface processor, Digital Core Engine, FPGA based digital processor and supplemental processors located in input modules such as the computer in the Renderer Input or the USB module processor. Input modules are all upgradable in all products (tool less lever) except the built in S/PDIF inputs on the discrete DAC (there are still two upgradable slots on the discrete DAC however). The clock modules in all products except the Discrete DAC are upgradable (requires tools, Premier DAC is tool less). The DAC modules in all products are upgradable, replaceable modules (requires tools). The output modules which in all products are user upgradable (tool less lever). The power supplies are external for all our DACs so those are upgradable as well.

Our 25 years of DAC design and manufacturing experience we have found that it is primarily input and format compatibility that makes DACs “age” so we made the software and input modules the most easily upgradable parts of our current DACs. We spent a lot of effort to make the input module replacement as painless as possible. A user can simply power the DAC off, pull out locking lever, pull out the module, insert a new module, push in the locking lever and power up the DAC. We found that some types of input modules also can also suffer a higher than average failure rate due to “tweaks” like custom USB cables that can damage USB receiver chips. It is much easier to ship a new module to a customer than have them send their whole DAC back to us for repair. We also found that developing multiple versions of each new input was wasting a lot of design effort (especially the design of complex inputs like the Renderer) so we standardized on one extremely capable input module format across all of our product levels. It was one of the major reasons we recently completely revamped our product lineup. The modular input design also allows up to optimize each input for the best noise isolation, best data recovery and lowest jitter possible for each input format.

We write all of our software in house and we are constantly improving and refining the software for all of our DACs. Improvements to the user interface and input module support are the most common but improvements to the core digital processing and software configurable DAC modules are released at least once a year as well.

There are some limitations to the basic hardware on all of these DACs without motherboard and DAC replacement (which is possible). They have been designed to anticipate future trends as much as possible to ensure the longest life span out of these DACs as possible. The DACs, Core Engine and input module slots support multibit PCM of up to 32 bits and up to 6Mhz sample rate. The DACs, Core Engine and input module slots support Native DSD to DSD 16x (DSD 1024). Multibit Delta Sigma (2, 4 and 6 bit) is also natively supported to 12.288Mhz sample rates even though this is not a currently an available format. Input modules each have a varying capability that is dependent upon reliable support of each format. For example S/PDIF is limited to 24 bits and 192Khz PCM sample rates (good for 1x DSD also). Our current renderer supports PCM at 32bit resolution and 768Khz sample rates (good for 4x DSD also). Our Pro ISL input is our most capable current input and supports 32bit PCM resolution at up to 3Mhz sample rates (good for 16x DSD also).

Thank you for this extensive and informative reply.

I have another question:
One thing that currently keeps me from seriously considering an MSB purchase is that all MSB DACs are both PCM and DSD capable. However, I have absolutely zero interest in DSD. All my playback is CD. If I would ever go the server route -- not planned at this point --, in addition to Redbook 16/44.1 I would only be interested in PCM of higher bit/sampling rate, not in DSD. And even that is questionable since, let's face it, most music is just on CD.

What I like about the Yggdrasil DAC is that it puts its resources into PCM only, and I did not have to pay for extra capabilities that I simply do not want. If I understand correctly, MSB DACs have separate PCM and DSD converters. Would it possible to obtain a 'PCM only' MSB DAC?

Thanks
Al

 

[KeithR]

Thank you for this extensive and informative reply.

I have another question:
One thing that currently keeps me from seriously considering an MSB purchase is that all MSB DACs are both PCM and DSD capable. However, I have absolutely zero interest in DSD. All my playback is CD. If I would ever go the server route -- not planned at this point --, in addition to Redbook 16/44.1 I would only be interested in PCM of higher bit/sampling rate, not in DSD. And even that is questionable since, let's face it, most music is just on CD.

What I like about the Yggdrasil DAC is that it puts its resources into PCM only, and I did not have to pay for extra capabilities that I simply do not want. If I understand correctly, MSB DACs have separate PCM and DSD converters. Would it possible to obtain a 'PCM only' MSB DAC?

Thanks
Al

This is absurd. If you want MSB to build a specific dac for you, maybe you should look at other dacs.

 

[Al M.]

This is absurd. If you want MSB to build a specific dac for you, maybe you should look at other dacs.

Keith, I asked DMSB a question. I did not ask you for your opinion.

 

[microstrip]

(...) If I understand correctly, MSB DACs have separate PCM and DSD converters. (...)

Can you tell me where you read it? I did not get this idea, maybe I am wrong. In fact I have got very little details about the MSB DACs - I could not understand why they are called "hybrid".

 

[DaveyF]

Al, why don’t you have an interest in DSD/SACD? To me, the standard Redbook is a little limiting, and many of the best sounding CD’s are in fact dual layer, with the SACD layer included. Why not have the ability to decode both?
I know that’s where I will go next in the digital system.

 

[Al M.]

Al, why don’t you have an interest in DSD/SACD? To me, the standard Redbook is a little limiting, and many of the best sounding CD’s are in fact dual layer, with the SACD layer included. Why not have the ability to decode both?
I know that’s where I will go next in the digital system.

Until a few years ago I used to think too that the standard Redbook is a little limiting, but in the meantime I have discovered just how much information there is on a Redbook CD. Even with my Yggdrasil DAC I am absolutely stunned about the resolution of micro-detail in the timbre of, for example, solo strings (the rest of the system must be up for that as well, of course, and the acoustics of the listening room must not mask detail either). Soundstage depth we have discussed already. And at least with a top level MSB DAC it would probably be a safe bet to assume that its decoding of Redbook CD is of a greater quality than SACD on the best players at the time the new format was introduced, and back then hailed as the second coming of digital that finally allowed for fulfillment of the promises of the technology. It would be interesting to actually perform such a comparison.

The problem with CD is mostly just practical from an implementation point of view: the format is so 'on the edge' in terms of theoretical minimum for full transparency that lesser decoding is bound to fall short. So-called 'hi-rez' formats have much more of a safety buffer built in in terms of bit and sample rate, so that they can be more easily implemented to satisfaction on lesser equipment. Of course, an SACD may still sound better than the equivalent CD on the same top-level DAC, but again this will be more a matter of practical implementation than of theoretical limitations in the resolution of the CD format.

When SACD was introduced I knew it would be a commercial flop, and the amount of music released on it a negligible fraction of all the music available on CD. So I never bothered. I'm extremely glad I didn't, also for the reasons just discussed.

Furthermore, there is very little native DSD, on SACD or otherwise. In order to be able to mix and master you need to convert to PCM (usually done for DSD mastering at the so-called DXD level). So if in the vast majority of cases your DSD format has extensively 'seen' PCM during various stages, why even bother with the format and not simply listen straight to PCM?

 

[Al M.]

Here is an interesting white paper by Grimm Audio on native vs. non-native DSD:

DSD Myth

 

[microstrip]

(...)
Furthermore, there is very little native DSD, on SACD or otherwise. In order to be able to mix and master you need to convert to PCM (usually done for DSD mastering at the so-called DXD level). So if in the vast majority of cases your DSD format has extensively 'seen' PCM during various stages, why even bother with the format and not simply listen straight to PCM?

On a significant number of recordings with music I appreciate a lot the DSD layer sounds better in my system than the redbook layer. Why shouldn't I like DSD? Please note that systems with dedicated SACD transports send the encoded DSD data to the transport in the native format.

 

[microstrip]

(...) When SACD was introduced I knew it would be a commercial flop, and the amount of music released on it a negligible fraction of all the music available on CD. So I never bothered. I'm extremely glad I didn't, also for the reasons just discussed.
(...)

Well, I deeply regret not having bought the many SACDs with music I enjoy - nowadays they are terrible expensive and some become collector items, as people with appropriate players found they can sound great.

Even as a financial investment it would have been a wise decision - some SACDs sell now for a few hundred of euros!

 

[DaveyF]

Al, every time I have heard the SACD layer, I have noticed a significant increase in SQ over the standard layer. 100% of the time.

While I think that some CD's can sound very good..particularly the ones IME, that are HDCD layered, I still think the SACD is a step up over even those. YMMV.
BTW, I still don't think that the extreme layering that analog can bring to the overall SQ is attainable yet in the digital realm, at least I have never heard it---regardless of the DAC being utilized. There is in fact mention of this by RH in the summary section of the new Absolute Sound where the MSB DAC gets its award.

 

[Al M.]

Please note that systems with dedicated SACD transports send the encoded DSD data to the transport in the native format.

Which was not my point. You cannot speak of native DSD once the recorded signal has been through PCM processes -- *before* it has been re-encoded as DSD on the SACD and is read as such by the transport and transmitted as such to the DAC.

I would encourage everyone to read the DSD white paper that I linked to.

 

[Al M.]

BTW, I still don't think that the extreme layering that analog can bring to the overall SQ is attainable yet in the digital realm, at least I have never heard it---regardless of the DAC being utilized. There is in fact mention of this by RH in the summary section of the new Absolute Sound where the MSB DAC gets its award.

Do you have a link please?

 

[Mike Lavigne]

Here is an interesting white paper by Grimm Audio on native vs. non-native DSD:

DSD Myth

i wonder what Andreas Koch of Playback Designs (who knows a little about dsd and sacd's) would say about that white paper?

i'm not inferring that Mr. Grimm is even a little wrong, only that that white paper is but one man's opinion. i'd enjoy knowing what Andreas thinks.

i'm just a lowly listener enjoying both PCM and dsd. and so far with lot's of each format i find the source code for any recording is usually the format where the best sound is found. whether it's an sacd rip or a pure native dsd file. if it starts out as dsd then that is going to sound best.....better than any PCM version. just my viewpoint.....based on 18 years of SACD and dsd file listening.

there certainly are cobbled SACD's that don't sound that good for whatever reason. but they are the exception.

of course; if one avoids sacd or dsd how would you know that?

 

[DaveyF]

Do you have a link please?

Sorry Al , so far only available in the hard copy.
The quote from RH regarding the MSB Reference DAC and JV's review in the 'overall product of the year' section of the magazine is as follows: " It is weird that this DAC has turned his head, since it doesn't do three-dimensional imaging and bloom the way record players do ( no digital source does). "

 

[Al M.]

i wonder what Andreas Koch of Playback Designs (who knows a little about dsd and sacd's) would say about that white paper?

i'm not inferring that Mr. Grimm is even a little wrong, only that that white paper is but one man's opinion.

These are not opinions, but technical facts. And the folks at Grimm know a little about DSD too, and are in no way against it: they produce the AD1 converter, specifically designed to make great SACD masters from analog tapes.

I'd suggest you read the paper.

 

[Al M.]

of course; if one avoids sacd or dsd how would you know that?

I have heard DSD elsewhere, and while some of it sounded excellent, it did not strike me as exceptional compared to PCM.

 

[Mike Lavigne]

These are not opinions, but technical facts. And the folks at Grimm know a little about DSD too, and are in no way against it: they produce the AD1 converter, specifically designed to make great SACD masters from analog tapes.

I'd suggest you read the paper.

i scanned it, but don't have the technical chops to judge it's credibility. which is why i mention Andreas. or maybe Ted Smith would be another person to comment. Ted is a local friend (worked with Andreas on the Sonoma back in the day, and designed the PS Audio dac). he really knows dsd.

but unless you are a guy like one of those guys, you end up accepting someone else's interpretation. i'd rather hear it from someone who knows it.

and Grimm could be completely spot on, or......maybe it's not that cut and dried.

 

[Al M.]

Sorry Al , so far only available in the hard copy.
The quote from RH regarding the MSB Reference DAC and JV's review in the 'overall product of the year' section of the magazine is as follows: " It is weird that this DAC has turned his head, since it doesn't do three-dimensional imaging and bloom the way record players do ( no digital source does). "

Thanks, Davey!

In any case, if you are ever in the Boston area I'll gladly invite you to listen to my system and experience the spatial layering and 3D imaging. You may be very surprised.

 

[DMSB]

Thank you for this extensive and informative reply.

I have another question:
One thing that currently keeps me from seriously considering an MSB purchase is that all MSB DACs are both PCM and DSD capable. However, I have absolutely zero interest in DSD. All my playback is CD. If I would ever go the server route -- not planned at this point --, in addition to Redbook 16/44.1 I would only be interested in PCM of higher bit/sampling rate, not in DSD. And even that is questionable since, let's face it, most music is just on CD.

What I like about the Yggdrasil DAC is that it puts its resources into PCM only, and I did not have to pay for extra capabilities that I simply do not want. If I understand correctly, MSB DACs have separate PCM and DSD converters. Would it possible to obtain a 'PCM only' MSB DAC?

Thanks
Al

Maybe this will be confusing but I will try to explain how the hybrid DAC works. All data converters no matter what the format are collections of the same basic components. Precision references and switches. A normal DAC wires these basic components in only one way allowing only one type of data conversion. The Hybrid DAC is different, it is a software defined analog system that can be rewired on the fly to dedicate all of its precision analog components in almost any way. So when a Hybrid DAC module is converting PCM all of its recources are dedicated to native PCM conversion. When native DSD is converted all of those same recources are rewired for DSD conversion. So when PCM is being played back the hybrid DAC is a 100% PCM converter and when DSD is being played the hybrid DAC is a 100% DSD converter. Same components, different wiring. The prime DAC module used in our lower priced DACs operates in an identical way, it just has fewer precision analog recources than a hybrid DAC module. So to answer your question any of our current DACs are pure PCM DACs if you are playing PCM and any of our current DACs are pure DSD DACs when playing DSD. The software is responsable for the exact form the DAC actually takes so it can convert most any form of data natively. The main limitation is the ability to actually stream the data into the module and the limitation of the clock subsystem to generate the correct sample rate. That is limited to about a 500Mb per second data rate per module and sample rates to those divided from the two independant oscillators avalable in the femto clock modules.

 

[Al M.]

Maybe this will be confusing but I will try to explain how the hybrid DAC works. All data converters no matter what the format are collections of the same basic components. Precision references and switches. A normal DAC wires these basic components in only one way allowing only one type of data conversion. The Hybrid DAC is different, it is a software defined analog system that can be rewired on the fly to dedicate all of its precision analog components in almost any way. So when a Hybrid DAC module is converting PCM all of its recources are dedicated to native PCM conversion. When native DSD is converted all of those same recources are rewired for DSD conversion. So when PCM is being played back the hybrid DAC is a 100% PCM converter and when DSD is being played the hybrid DAC is a 100% DSD converter. Same components, different wiring. The prime DAC module used in our lower priced DACs operates in an identical way, it just has fewer precision analog recources than a hybrid DAC module. So to answer your question any of our current DACs are pure PCM DACs if you are playing PCM and any of our current DACs are pure DSD DACs when playing DSD. The software is responsable for the exact form the DAC actually takes so it can convert most any form of data natively. The main limitation is the ability to actually stream the data into the module and the limitation of the clock subsystem to generate the correct sample rate. That is limited to about a 500Mb per second data rate per module and sample rates to those divided from the two independant oscillators avalable in the femto clock modules.

Excellent, thanks!

This is a very satisfying concept.