PS Audio DirectStream DAC

[Booster MPS]

Yes, it includes a digital volume control implemented in the FPGA.

 

Then it would not have an analog line stage which some might desire.

[k6davis]

Yes, it includes a digital volume control implemented in the FPGA.

 

Thank you barrows.

 

Has a photo of the rear panel appeared anywhere yet? I'd like to know what kind of inputs and outputs it has.

 

Edit:

 

If it's predecessor, the PerfectWave is any indication, it'll have a full compliment of inputs.

 

The older unit has (2) I2S, and 1 each of USB, coax, optical & XLR inputs along with balanced XLR & single ended outputs.

 

The picture of the display of the forthcoming Directstream DAC clearly shows that it's set to "I2S Input 1". That's an encouraging sign that the inputs/outputs will cover all of the bases.

 

PS Audio to 'rescue' PCM with DirectStream PerfectWave DAC | Digital Audio Review by John Darko

 

Keith

[matthias]

A comment of Ted Smith from the PS Audio Forum:

 

"Volume control:

All inputs (whether PCM, DSD or DoP) are upsampled to 30 bits @ 10 x the normal DSD rate (28.224MHz). Then a 20 bit volume is applied (this is all the bits that are needed to represent the volume from 100 == full scale to 1 which is -49.5dB). All significant bits from the output of the volume multiply are used in the conversion to single bit double rate DSD. There isn’t any truncation, rounding or dither used or needed."

 

If i understand Ted Smith correctly DSD is converted to PCM 30bit / 28.224MHz......

 

KR

Matt

[Highfilter]

Part 2 video is now up,

.

 

Also found some solid information from Ted Smith over at the PS Audio forum, in the DirectStream thread.

 

Some general information about the DAC and its inputs etc.

 

There are multiple DACs (and many other products) which use FPGAs out there. There is little similarity between the DirectStream and, say DcS, certainly the architecture, algorithms, upsampling and conversion to DSD are entirely different. Outside of a few Xilinx library/macros the code is all original. Perhaps the closest might be the Playback Designs/EMM Labs, in that they also convert to double rate DSD. Tho Andreas and I have talked I doubt that there is much similarity otherwise.

 

The DirectStream digital board is indeed different than the PWD digital board. There is little in common except the connectors. Too a first approximation, all inputs go to the FPGA and the FPGA does its work and passes the DSD signal to the analog card. The most complicated thing outside of the FPGA on the digital card is the XMOS USB chip which gets its non USB inputs and clocks from the FPGA and hands its data back to the FPGA over I2S.

 

USB, I2S, S/PDIF, TOSLink, AES/EBU are all gapless and in fact there’s no gap, click or pop when you change from one input to another, change sample rates or go from PCM to DSD or back.

 

Ask Paul When we had the transport hooked up with TOSLink, S/PDIF, AES/EBU and I2S he thought the input select was broken since there weren't any clicks or pops when the selection was changed and further the character of the sound remained identical.

No matter how the bits get to the DirectStream DAC they will sound the same, whether USB, the bridge, I2S or S/PDIF, etc.

 

Indeed a transformer is use for audio output: its main function is to isolate common mode noise. Tho I too was initially skeptical about using a transformer, it’s benefits far outweigh it’s downsides (most of which are ameliorated by using a much bigger transformer than some might use.)

 

Regarding the use of a pre-amp.

 

Q: Paul, did you try the DAC without your tube preamp? Do you still prefer the preamp in the chain with this new DAC?:

 

Paul McGowan: Yes and it's now a real toss up. But for now the preamp stays in. Mostly because it's a lot easier to do A/B testing!

 

Q: Ted, I think Paul is pleading the 5th on the preamp question (or he has just tried - which I find unlikely). Can you weigh in on the question if adding a top notch tube preamp is still beneficial with the directstream (as it is with the PWD MKII in some systems).

 

Tho I appreciate the concept of simplicity and avoiding extra connections and cables if not needed I don't know why we should expect that every component has the equivalent of an equal cost preamp built in for free. In most systems I've found that things go better with.... oops, that a preamp gives a more lively and dynamic presentation.

 

Regarding the volume control:

 

All inputs (whether PCM, DSD or DoP) are upsampled to 30 bits @ 10 x the normal DSD rate (28.224MHz). Then a 20 bit volume is applied (this is all the bits that are needed to represent the volume from 100 == full scale to 1 which is -49.5dB). All significant bits from the output of the volume multiply are used in the conversion to single bit double rate DSD. There isn’t any truncation, rounding or dither used or needed.

 

The PS Audio PWD has a 100 step, 1/2 dB per step volume control. Using 20 bits at the FPGA represents the 100 steps with no more than about 0.03% error. Looking at it another way 20 bits is 120dB of range. Why would anyone need more? The actual UI range and the mapping from the UI range to the FPGA range is done in the UI code and if more range is needed it's a simple UI change. If a 20 bit volume range ever turns out to be too few bits it's only an FPGA "recompile" to get more. Perhaps you were thinking of the total resolution of the volume control which is the complete 50 bit product of the 20 bit volume control and the 30 bit high rate signal: we don't drop any bits on the floor as we convert to single bit double rate DSD.

 

In addition there is a selectable analog 20dB attenuator for higher sensitivity systems.

 

Comparison with the EMM DAC6e:

 

I too find the DAC6e among the best out there and have been happy with it for years.

 

The DirectStream and the EMM DAC6e are similar sounding (they both convert all inputs to double rate DSD.) Tho with a quick A/B I personally don't hear gobs of differences, for longer term listening I definitely prefer the DirectStream. It's more engaging and I find I listen to more music with it. When my daughter was pregnant she had hypersensitive hearing for a while. When I asked her to do a blind A/B between them every time I selected the DAC6e she said "Stop That!"

[wisnon]

Couple questions? Why do they need a chip at all for DSD when only LP filtartion is strictly needed. This talk of increased computing power for processing seems to be going the wrong way.

 

The other point is, CHORD already uses FPGA tech to do their DSD and while decent, it is by far NOT the best DSD sound in my house...just saying.

 

I suspend judgement till I hear the thing, but carry serious reservations. This seems like hype.

[Hiro]

The other point is, CHORD already uses FPGA tech to do their DSD and while decent, it is by far NOT the best DSD sound in my house...just saying.

 

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

[k6davis]

A comment of Ted Smith from the PS Audio Forum:

 

"Volume control:

All inputs (whether PCM, DSD or DoP) are upsampled to 30 bits @ 10 x the normal DSD rate (28.224MHz). Then a 20 bit volume is applied (this is all the bits that are needed to represent the volume from 100 == full scale to 1 which is -49.5dB). All significant bits from the output of the volume multiply are used in the conversion to single bit double rate DSD. There isn’t any truncation, rounding or dither used or needed."

 

If i understand Ted Smith correctly DSD is converted to PCM 30bit / 28.224MHz......

 

KR

Matt

 

The idea of conversion from DSD to PCM, then back to DSD surely sounds like a bad thing, especially from a company that so forcefully condemns PCM, but my understanding is that it isn't yet possible to perform DSP (like volume control) on a DSD datastream.

 

I have a large library with all different types of files and resolutions, mostly flac, HD-PCM & DSD. I use (and love) volume leveling and DSDx2 upsampling in JRiver. That means that my DSD files are converted to PCM (for the volume leveling) and then converted back to DSD. I have the option in JRiver to play them without any such manipulation, but I choose not to because I highly value the volume leveling - much more than I anticipated I would, but that's another story.

 

Of course, if I felt that these conversions adversely impacted the sound quality, I'd be conflicted, but the sound is subjectively fantastic. More importantly, there is an objective case that, when done properly, these conversions are not as harmful as they might seem.

 

This is from JRiver's Wiki:

The DSD to PCM conversion process converts from 1-bit DSD to 64-bit PCM at 1/8th of the sample rate. The total amount of data from this conversion grows by 8x, so the process is effectively lossless / perfect.

Once you have PCM, it will be 64bit @ 352.8 kHz for DSD, and 64bit @ 705.6 kHz for DSD 2x.

Therefore, JRiver's DSD to PCM conversions do not lose any resolution.

1-bit x 2.8224 MHz (DSD x 1) = 64-bit x 352.8 kHz PCM

1-bit x 5.6448 MHz (DSD x 2) = 64-bit x 706.6 kHz PCM

 

I haven't done the math on 30-bit x 28.224Mhz, but I expect that Ted Smith's process similarly wouldn't reduce the resolution of the input signal.

The main point is that the supposed benefit of DSD over PCM (besides the higher resolution) is that DSD much more closely resembles an analog waveform. Therefore the process of converting it to analog is simpler and should yield a superior result when compared to the more convoluted process of converting multi-bit PCM to analog. That specific benefit remains intact when the bits are converted to analog regardless of an "effectively lossless/perfect" conversion step in the middle.

 

I don't mean to start PCM vs DSD debates (although I'm firmly in the camp that converts everything to DSD128 in JRiver), nor do I have as much technical knowledge as many in this forum. I'm just looking to learn & further the conversation.

[Hiro]

The idea of conversion from DSD to PCM, then back to DSD surely sounds like a bad thing, especially from a company that so forcefully condemns PCM, but my understanding is that it isn't yet possible to perform DSP (like volume control) on a DSD datastream.

 

Isn't Miska's HQPlayer doing just that? That is native DSD DSP (including volume control) in SDM?

[k6davis]

Couple questions? Why do they need a chip at all for DSD when only LP filtartion is strictly needed. This talk of increased computing power for processing seems to be going the wrong way.

 

The other point is, CHORD already uses FPGA tech to do their DSD and while decent, it is by far NOT the best DSD sound in my house...just saying.

 

I suspend judgement till I hear the thing, but carry serious reservations. This seems like hype.

 

PS Audio's diagram of this new DAC indicates that it (basically) only uses a passive LP filter to output the DSD signal to analog. What they are claiming is new is their FPGA chip converts all input to DSD in way that is superior to what has come before. Therefore, all of your music (PCM & DSD) gets the benefit of DSD decoding without a DAC chip.

 

This is *exactly* what I'm already doing with JRiver DSD128 upsampling and the LampizatOr DSD DAC. How much better PS Audio's upsampling is than JRiver's upsampling and how their solid-state approach compares to the tube design of the LampizatOr remains to be seen.

 

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

 

Indeed.

 

I was really intrigued by the DSD x 10 upsampling PS Audio claims. But now I see that the DSD x 10 data is only internal to the DAC to minimize the impact of any DSP errors. The PS Audio DAC sends DSD x 2 to its output - the same resolution you can get from JRiver for "free". Now, I'm even more curious about how much better the conversion is than what I get from JRiver. Just how magical are those algorithms?!?

 

That means that, with all due respect to Ted Smith, the primary benefit of this new DAC could be that it decodes DSD without a DAC chip. I suspect that after this unit & the LampizatOr have done it well, that more units will follow, with lower price tags.

[k6davis]

Isn't Miska's HQPlayer doing just that? That is native DSD DSP (including volume control) in SDM?

 

I've read postings by here Miska & heard of HQPlayer, but I'm not familiar with it, so I don't know.

 

I've read repeatedly that DSD has to be converted to PCM in music production for editing and DSP such as EQ, mixing & volume control. My comments are based on that understanding.

[matthias]

Isn't Miska's HQPlayer doing just that? That is native DSD DSP (including volume control) in SDM?

 

+1

I have read a lot of Miskas posts with great interest and if I understand his comments correctly his HQPlayer is able to do volume control in SDM without conversion to PCM.

 

Paul McGowan and Ted Smith stated that the use of a preamp is beneficial. Why then is the volume control in the DAC necessary? In this case they could get rid of the conversion from DSD to PCM.........

 

Matt

[Hiro]

I've read postings by here Miska & heard of HQPlayer, but I'm not familiar with it, so I don't know.

 

I've read repeatedly that DSD has to be converted to PCM in music production for editing and DSP such as EQ, mixing & volume control. My comments are based on that understanding.

 

All DSP (EQ, mixing and volume control) can be done in SDM. The DSD->PCM conversion step in legacy digital audio workstations reflects only the limitation of the tools not of the SDM technology.

[k6davis]

+1

I have read a lot of Miskas posts with great interest and if I understand his comments correctly his HQPlayer is able to do volume control in SDM without conversion to PCM.

 

Paul McGowan and Ted Smith stated that the use of a preamp is beneficial. Why then is the volume control in the DAC necessary? In this case they could get rid of the conversion from DSD to PCM.........

 

Matt

 

Right, or at least make the volume control and extra conversion defeatable.

 

That way the user could decide if the purchase of a preamp on the level of this premium DAC is worthwhile.

[Hiro]

Paul McGowan and Ted Smith stated that the use of a preamp is beneficial. Why then is the volume control in the DAC necessary?

 

Good question, but even if someone needs the volume control on a DAC, why not implement it in the FPGA using native SDM/DSD processing?

[k6davis]

Good question, but even if someone needs the volume control on a DAC, why not implement it in the FPGA using native SDM/DSD processing?

Good point.

 

At least there's a possibility that they could provide that capability with a firmware update in the future.

 

Along these same lines, does this DAC run native DSD x 1 and DSD x 2 streams through their process that internally upsamples to DSD x 10, only to output at DSD x 2?

[Hiro]

What I find intriguing is the use of FPGA for DSD, cuz Chord's use of it has created incredblly real PCM music...

 

Yet, the Chord is not a PCM DAC.

[Hiro]

Along these same lines, does this DAC run native DSD x 1 and DSD x 2 streams through their process that internally upsamples to DSD x 10, only to output at DSD x 2?

 

I do hope that you can bypass the additional processing for DSD playback.

[Hiro]

A comment of Ted Smith from the PS Audio Forum:

 

"Volume control:

All inputs (whether PCM, DSD or DoP) are upsampled to 30 bits @ 10 x the normal DSD rate (28.224MHz). Then a 20 bit volume is applied (this is all the bits that are needed to represent the volume from 100 == full scale to 1 which is -49.5dB). All significant bits from the output of the volume multiply are used in the conversion to single bit double rate DSD. There isn’t any truncation, rounding or dither used or needed."

 

If i understand Ted Smith correctly DSD is converted to PCM 30bit / 28.224MHz......

 

KR

Matt

 

In fairness, they do not specify if the 30bit/28.224MHz processing is done in PCM, so there's still a chance it's a native 10x DSD (SDM) process...

[matthias]

That means that, with all due respect to Ted Smith, the primary benefit of this new DAC could be that it decodes DSD without a DAC chip. I suspect that after this unit & the LampizatOr have done it well, that more units will follow, with lower price tags.

 

Yes,

that is the point.

IMO the best way is to do all the conversion work to DSD in a good software player(HQPlayer) and after that to have a simple pure chipless DSD DAC. It would have been nice to have the PS Audio Directstream DAC as a pure DSD DAC for half the price.....

 

Matt

[k6davis]

In fairness, they do not specify if the 30bit/28.224MHz processing is done in PCM, so there's still a chance it's a native 10x DSD (SDM) process...

 

The PS Audio upsamples to a 30 bit data stream.

 

Isn't SDM only 1 bit?

[Hiro]

The PS Audio upsamples to a 30 bit data stream.

 

Isn't SDM only 1 bit?

 

Nope, SDM can be any number of bits.

[matthias]

In fairness, they do not specify if the 30bit/28.224MHz processing is done in PCM, so there's still a chance it's a native 10x DSD (SDM) process...

 

Okay,

bur they could be more clear about that in their white paper:

 

The DSD Engine

The heart of the DirectStream DAC is its DSD engine:

• 10X DSD rate. Regardless of input format, whether PCM or DSD, all data are upsampled to 30 bits running at 10 times the standard DSD rate and then back down again to double rate DSD for noise shaping.

[Hiro]

Okay,

bur they could be clearer about that in their white paper:

 

The DSD Engine

The heart of the DirectStream DAC is its DSD engine:

• 10X DSD rate. Regardless of input format, whether PCM or DSD, all data are upsampled to 30 bits running at 10 times the standard DSD rate and then back down again to double rate DSD for noise shaping.

 

I get your point. From this description it is not clear at all (at least for me) whether it's a DSD (SDM) engine or rather PCM engine.

[k6davis]

Yes,

that is the point.

IMO the best way is to do all the conversion work to DSD in a good software player(HQPlayer) and after that to have a simple pure chipless DSD DAC. It would have been nice to have the PS Audio Directstream DAC as a pure DSD DAC for half the price.....

 

Matt

 

Exactly!!

 

Their FPGA and Ted Smith magic has really got to be amazing to justify the price when DSD conversion is available in software.

 

If this unit takes off, I'd be willing to be that someone will come out with a simple DSD-only DAC that doesn't require a DAC chip for cheap. It would essentially be the right side of the PS Audio diagram and therefore should be easy and inexpensive to make.

[matthias]

If this unit takes off, I'd be willing to be that someone will come out with a simple DSD-only DAC that doesn't require a DAC chip for cheap. It would essentially be the right side of the PS Audio diagram and therefore should be easy and inexpensive to make.

 

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