PS Audio DirectStream DAC

[wisnon]

Isn't it all about how you program the chip?

 

True, but the blurb makes it seem that all you need to do is use FPGA and you are home free...

[boatheelmusic]

So how should we think about this compared to the NAD M 51?

[k6davis]

The Sonore PureDSD DAC could be such a thing but nobody knows about launching...............

 

Thanks for pointing this out. It looks extremely interesting.

 

SONORE Rendu PureDSD DAC

 

The website is a bizarre combination of detailed and cryptic for now, but just knowing that there's another DAC coming that converts DSD without a DAC chip is encouraging.

 

But the Sonore also does PCM, and would need the hardware to do that.

 

I'd love to see something that would be the inverse of the forthcoming $14k Berkely PCM only DAC. Meaning something like an extremely simple, very inexpensive, DSD only, no-DAC chip, low-pass filter with quality component design, that would require you to convert your content to DSD in the computer.

 

It seems like very little circuitry would be necessary.

 

It would be the PS Audio without the FPGA.

[k6davis]

So how should we think about this compared to the NAD M 51?

 

Good question. The NAD is a kind of predecessor to the PS Audio. They both convert their inputs to a 1 bit datastream, but the PS Audio does so at a much higher resolution. PS Audio claims that their upconversion is the best available.

[matthias]

So how should we think about this compared to the NAD M 51?

 

IMO, the PS Audio Directstream DAC in the case it does DSD to PCM conversion has the most similarity with the Mola Mola DAC.

 

Matt

[Hiro]

True, but the blurb makes it seem that all you need to do is use FPGA and you are home free...

 

That PS Audio went with a discrete FPGA DSD DAC design is most commendable IMO.

 

The question is how good is the quality of their particular implementation, and if pure DSD D/A conversion is possible at all with their chosen signal path.

[matthias]

Thanks for pointing this out. It looks extremely interesting.

 

SONORE Rendu PureDSD DAC

 

But the Sonore also does PCM, and would need the hardware to do that.

 

From the Sonore Homepage:

 

Sonore Rendu PureDSD DAC - Ethernet input and analog output (for DSD playback) SPDIF BNC output (for PCM playback)

 

You need a second DAC for PCM or you have to convert PCM to DSD by software.

 

Matt

[wisnon]

but just knowing that there's another DAC coming that converts DSD without a DAC chip is encouraging.

 

But the Sonore also does PCM, and would need the hardware to do that.

 

I'd love to see something that would be the inverse of the forthcoming $14k Berkely PCM only DAC. Meaning something like an extremely simple, very inexpensive, DSD only, no-DAC chip, low-pass filter with quality component design, that would require you to convert your content to DSD in the computer.

 

It seems like very little circuitry would be necessary.

 

It would be the PS Audio without the FPGA.

 

The Lampizator had Chipless DSD from last September...as you no doubt are aware, but you did say vey cheap. The Schitt Loki fits the bill, but is not Chipless.

 

Chord Hugo and the Qutes use FPGAs for both PCM and DSD, but not sure they are either chipless or non-processing (LPF only), like the Lampi.

[ChrisG]

Good question. The NAD is a kind of predecessor to the PS Audio. They both convert their inputs to a 1 bit datastream, but the PS Audio does so at a much higher resolution. PS Audio claims that their upconversion is the best available.

 

Haha, what else would you expect Paul "Mr Hyperbole" McGowan to say?

[VandyMan]

Interesting tech article on implementing DSD on an FPGA:

Implementing a Direct Stream Digital Decoder on National Instruments FPGA Platforms - National Instruments

 

Makes me think that a lot of these DSD FPGA DACs are likely very similar in implementation.

[Superdad]

An FPGA is only that a Fully Programmable Gate Array. Since there are no specific DSD chips available these days, DAC designers need to roll their own chips by programming the FPGA. It's not strictly a DSP but can be programmed in if the need or want is there.

 

Actually, most FPGAs have many DSP slices set up in them just waiting to be programmed--typically with Matlab. At truly "empty" FPGA would be a modern programmer's nightmare. (Might be almost as bad as asking someone to create a web site using just assembly language code. Yes, I know some here are too young to even remember what assembler code is.)

 

As for Ted Smith's design, a lot of the SQ will depend upon the quality of both the SRC algorithms and on the DSM code. I don't recall reading anywhere that Ted has a background in writing SRC filters, and frankly it would be perfectly fine if they just tuned (by ear and measure) a good computer-based open-source one like SoX and then loaded those coefficients into the FPGA--adapted for the extra-high rate of course. (That's what Swenson did with the results from our filter tuning session at my house last month for the Bottlehead DAC; see post at: http://www.computeraudiophile.com/f8-general-forum/do-all-dacs-sound-more-or-less-same-18877/index12.html#post300613)

 

Also, really good delta-sigma modulator design, while not computationally intensive (at least compared to SRC), may be an art in itself. Ask Miska or the folks at ESS about the years they have spent optimizing their modulator code.

 

Smith's/PS Audio's running of the DSM to just single-bit--as opposed to a multi-bit/multi-level DSM--would seem to put more pressure on the performance of the analog output filter. Perhaps that's the real reason they went with transformers.

 

John Swenson and I have been sketching and hatching plans for a somewhat similar (very high rate) DAC-less DAC for quite a while now. But it may use a giant, $250 FPGA, and it will be multi-bit/multi-level, and with an output stage better than Smith's video-opamps (but those are cool too--John has already done one DSD design using those).

 

Anyway, I will talk to Swenson this weekend (I just sent him Ted Smith's white paper), and have him straighten me out about anything wrong in what I just wrote. I got a kick out of watching Ted's presentation at the announcement presentation as he reminds me of John in a bunch of ways (the good qualities, not the shoeless and bearded style!). Both have lots of experience, knowledge, and enthusiasm. McGowan was smart to tap him.

 

Cheers,

 

Alex C.

[One and a half]

Interesting tech article on implementing DSD on an FPGA:

Implementing a Direct Stream Digital Decoder on National Instruments FPGA Platforms - National Instruments

 

Makes me think that a lot of these DSD FPGA DACs are likely very similar in implementation.

 

Interesting article.

 

Even if you were to identically copy and download the NI code into your FPGA to save time in development, the software versus reality bites with the code having to deal with PC noise, PSU and analog output stages. The point is, the final implementation that matters of the DAC is a whole and not the sound of a chipset or even software code.

[One and a half]

No, it's Field Programmable Gate Array. Which is good because updates can be done by the end user.

 

You're right, sunshine. Field conjures up acres of wheat or barely though...

[One and a half]

To Superdad,

 

I just wonder whether the 10 x DSD rate is just a one upmanship after all. This level is some serious horsepower that's needed, better off in a computer, however, my retired music server i5 mini couldn't cope with the PCM2DSD128 conversion from Jriver, so how does this work in a dedicated FPGA...unless there are several of them working in parallel.

[Superdad]

Makes me think that a lot of these DSD FPGA DACs are likely very similar in implementation.

 

So far, not at all. Not only do present FPGA-using DACs have very different ancillary circuits (input schemes and isolation, clocking, power-supplies, output filters, and output stages), but the way they are programmed (SRC stages, and SDM) are different in each of the still fairly few models out there.

 

I hope that FPGA does not just become a buzzword. It is all about design and implementation, and there is still plenty of room for innovation. Not to mention that there can be many different architectures/topologies/DAC philosophies that just happen to use a CPLD or an FPGA. The power and economy of an FPGA is alluring to designers that don't mind writing some code. So while you are likely to see more products with them, they won't be equal.

[VandyMan]

So far, not at all.

 

I don't think we know enough to say, unless you have inside info. A lot of tech products that on the surface (and in the marketing) appear unique are based on OEM platforms. That does not mean that they are not highly customized. I doubt these small companies did their FPGA programming from scratch, but am just guessing. Anyway, I did not mean that they would all sound the same or that the FPGA code is identical. I'm just guessing that it is similar, but customized. Usually when we start to see a bunch of very similar products (dongle dacs, for example) people are either using an oem design or cribbing from each other.

[VandyMan]

I hope that FPGA does not just become a buzzword....So while you are likely to see more products with them, they won't be equal.

 

That is pretty much the situation with DAC chips right now. People say things like they prefer the SABRE DAC chip or they find it bright, when they are really talking about a specific implementation and shouldn't be generalizing from it.

[bmoura]

It will be interesting to listen to the final product. The fact that it switched PS Audio from DSD critics to believers suggests it will be one to evaluate.

 

I was also amused to hear that while Ted was inspired to design it after hearing a Sony Multichannel SACD demo, the unit is only 2 channel! I guess he was impressed, but not that impressed....

[john57]

It will be interesting to listen to the final product. The fact that it switched PS Audio from DSD critics to believers suggests it will be one to evaluate.

 

I was also amused to hear that while Ted was inspired to design it after hearing a Sony Multichannel SACD demo, the unit is only 2 channel! I guess he was impressed, but not that impressed....

 

I think the key thing that both Paul and Ted were trying to say is this. Listening to the master tape you will them realized how much more information there is to the sound. I was at a classical concert in Boulder, CO some years back and there was a professional recordists taping the concert using the Sony digital recorder. I was not sure it was PCM or DSD but when I heard it back on the AKG passive electrostatic headphones it just blew my mind on how clear the sound is. I just wonder why my hundreds of classical CD's that I have do not have the live quality that I just heard from a master tape right after the concert.

 

Before that the best commercial recordings I have came from a company called Barclay Crocker which made open reel Dolby B tapes from masters for my Teac open reel tape deck at the time. I was amazed with the sound and the LP records that I brought which were several generations from the master tape by the time the pressing plant gets the copy with compression and EQ much of the information is lost. LP's records do not even come close to the first generation master tape. With digital you supposedly do not get generation loss but I can not escape the fact when I that that CD in hand and play it back how much of it changed from the master tape.

If PS Audio can give me the same quality and magic that I got from the Barclay Crocker tapes then it would be something to look into.

[bmoura]

I think the key thing that both Paul and Ted were trying to say is this. Listening to the master tape you will them realized how much more information there is to the sound.

 

There are DACs out there today, like the EMM Labs DAC6e that Ted mentions in his video, that take PCM and upsample it to 2X DSD (DSD128) and then play it back for enhanced sonics.

 

The big question here will be does this new system better the results of something like a DAC6e. Ted says he has compared the two and the new unit wins the comparison, albeit in 2 channel mode. When the product comes out, that will be an interesting claim to evaluate.

[matthias]

Pretty much all questions are answered in this thread:

 

Discussions | PS Audio

 

From the technical point of view I must say that I am disappointed. I hoped for a pure and chipless DSD DAC but without any conversion of DSD to PCM.

Let us see what the future brings........

 

KR

Matt

[boatheelmusic]

That is pretty much the situation with DAC chips right now. People say things like they prefer the SABRE DAC chip or they find it bright, when they are really talking about a specific implementation and shouldn't be generalizing from it.

 

 

Do you have any examples of Sabre based Dacs that don't have the tizzy "sheen" ?

[Hiro]

Pretty much all questions are answered in this thread:

 

Discussions | PS Audio

 

From the technical point of view I must say that I am disappointed. I hoped for a pure and chipless DSD DAC but without any conversion of DSD to PCM.

Let us see what the future brings........

 

KR

Matt

 

It is a big letdown that their 'DirectStream DAC' can't really do direct stream....

[christopher3393]

It is a big letdown that their 'DirectStream DAC' can't really do direct stream....

 

Do you mean a big letdown from an abstract technical perspective? Surely you would want to hear it. What if it sounds great?

[Hiro]

Do you mean a big letdown from an abstract technical perspective? Surely you would want to hear it. What if it sounds great?

 

It is a big letdown because they were so close to getting it right. They chose an FPGA over an off-the-shelf chip, they had total freedom to do a proper direct stream implementation, they even named the dac 'DirectStream', and still failed to deliver direct stream playback.