PS Audio DirectStream DAC

[Ted Smith]

what is the rationale for the upsampling to PCM 30 bit / 28 MHz at all ? Volume control?

 

Why not direct to DSD128?Matt

 

Howdy

 

I think I addressed this above, but the 28MHz comes from wanting to do integer synchronous sample rate conversion of both 176400 and 192000 input sample rates: it's their least common multiple. I zero stuff 146 zeros / sample @ 192000 and 159 for 176400 then I lowpass filter with an IIR filter which uses wide samples internally. I chose 30 bits between it and the sigma delta modulator sort of arbitrarily: only keeping, say 24, seemed like a waste. I use an IIR because all of my experience says they sound more natural than FIR filters. I know what I'm doing so I can avoid the potential instabilities of IIR filters.

 

-Ted

[bmoura]

I know you are kidding but I have a MC system at my house and really enjoy it compared to two channel. I have literally thousands of MC SACDs and enjoy most of them.

 

My MC transport has given up the ghost and my grandkit poked his finger thru one of my rear beryllium tweeters so till I get some breathing room I'm stuck with two channel

 

Once the Stereo product is out the door, any plans to put all of this into a Multichannel version? That would be preferable to buying say three Stereo DirectStream DACs. Not to mention the issues in interconnecting and syncing multiple units for 5 and 5.1 channel DSD playback.

[Ted Smith]

I doubt these small companies did their FPGA programming from scratch, but am just guessing.

 

Howdy

 

Except for some Xilinx library routines (the FIR filter core, for example) I've done all of the FPGA programming from scratch in Verilog. I have the background to do everything, but why reinvent the wheel and write/debug the filter core from scratch? The real skill in FIR filter design is choosing the filter coefficients. I initially used my HP-48 calculator to get them I don't want to trust library routines from, say MATLIB, to have done the numerical analysis to prove their accuracy, let alone sound good.

 

Similarly tho there are libraries of sigma delta modulators out there (and some very complete ones in MATLAB) I have written all of mine from scratch. It's not trivial to get a clean sounding sigma delta modulator. I recently saw a very good talk by the ESS guys talking about some of the issues: check their web site. (FWIW my experience directly matches theirs.)

 

-Ted

[wisnon]

Exactly. I was only referring to the converter's "DirectStream" playback, which in actuality the DAC can not deliver, as it subjects DSD to two conversions internally: DSD>PCM , PCM >DSD.

 

Hiro, Lampi DSD is as direct as you can get. LoL

 

Emm Labs Dac2x and PDB MPD5 do not sound as good to me...

[ted_b]

Not to mention the issues in interconnecting and syncing multiple units for 5 and 5.1 channel DSD playback.

 

Aw c'mon Brian, that's easy. (Wink).

 

I do love that Ted, like you, me, DigiPete and a few other souls, is a multichannel guy. And a preamp guy! Ted, your preamp response on the PS Audio thread was music to my ears....

[bmoura]

Welcome Ted! (btw, does the name "Ted" mean you like DSD! Seems to. )

 

I think you'll find us fair and thoughtful. We can't ensure that there are no trolls, but we do a decent job of starving them.

 

Agreed. It's tamer here than the Audio Asylum and its High Rez Highway forum covering SACD and DSD. Although they continue to have some pretty hilarious moments on AA - witness the recent discussion of upcoming Channel Classics releases in SACD and DSD!

 

Jared has done it again! New babe violist in hi-rez - Chris from Lafayette - Hi-Rez Highway

[Ted Smith]

Once the Stereo product is out the door, any plans to put all of this into a Multichannel version? That would be preferable to buying say three Stereo DirectStream DACs. Not to mention the issues in interconnecting and syncing multiple units for 5 and 5.1 channel DSD playback.

Howdy Brian

 

I have to beg off and not speak for PS Audio and their future products. That said they are a two channel company and I don't see them going whole sale into the AV world.

 

Tho I've put a little thought into a MC version, I didn't want to compromise jitter performance by adding hooks for MC to the DirectStream.

 

Tho personally I want a MC version at some point and may eventually talk PS Audio into it, I'd rather be working on other projects which might replenish my retirements funds a little faster

 

-Ted

[ted_b]

Agreed. It's tamer here than the Audio Asylum and its High Rez Highway forum covering SACD and DSD. Although they continue to have some pretty hilarious moments on AA - witness the recent discussion of upcoming Channel Classics releases in SACD and DSD!

 

Jared has done it again! New babe violist in hi-rez - Chris from Lafayette - Hi-Rez Highway

 

I asked Jonas...he DID get her number by the way.

[bmoura]

Aw c'mon Brian, that's easy. (Wink).

 

I do love that Ted, like you, me, DigiPete and a few other souls, is a multichannel guy. And a preamp guy! Ted, your preamp response on the PS Audio thread was music to my ears....

 

Well, Ted, you are the DAC interconnect guru with those 3 linked Mytek Stereo DSD DACs! I'm more a fan of the all-in-one box approach to Multichannel. I'm sure you have plans to review a few of the DirectStream DACs soon for Multichannel DSD testing, for Computer Audiophile.... right?!

 

And I do agree on the preamp question. I'm using the EMM Labs Switchman II here - works great.

[wisnon]

Its very nice of Ted Smith to come here and answer question directly. Kudos.

[bmoura]

Howdy Brian

 

I have to beg off and not speak for PS Audio and their future products. That said they are a two channel company and I don't see them going whole sale into the AV world.

 

Tho I've put a little thought into a MC version, I didn't want to compromise jitter performance by adding hooks for MC to the DirectStream.

 

Tho personally I want a MC version at some point and may eventually talk PS Audio into it, I'd rather be working on other projects which might replenish my retirements funds a little faster

 

-Ted

 

Well, you did get PS Audio to move from criticizing DSD to championing it. That alone has me very interested in auditioning the product! It also suggests that there is hope in getting PS Audio to at least considering a Multichannel version in the future.

 

I'll have to keep an eye on the PS Audio site to see where the unit is going to be demo'd in the future.

[Ted Smith]

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.

 

Howdy

 

I'll let your ears be the judge of the quality of implementation, but "pure DSD D/A conversion" is just a low pass filter and I think it'll be hard to beat a good passive low pass filter...

 

If you are referring to the fact that I sometimes use extra precision in my DSP instead of just one bit, well perhaps we see things a little differently. Extra precision is nothing to be ashamed of. The real goal of running my own sigma delta modulator is that they aren't trivial, I don't like the sound of many of them and I take the opportunity to use more accurate math and better tuning to my ear and to the output hardware.

 

More details are in some of my posts above.

 

-Ted

[Ted Smith]

Well, you did get PS Audio to move from criticizing DSD to championing it.

 

Howdy Brian

 

To be fair they were converted before I showed up. You can watch their video of Gus Skinas and separately Paul talking about it. I never asked what changed their minds

 

I do admire Paul in that when Gus dragged him over to listen to the prototype Paul and the rest of his company were instantly on board. No NIH, no balking support, just welcoming arms and a can-do attitude.

 

-Ted

[ted_b]

Well, Ted, you are the DAC interconnect guru with those 3 linked Mytek Stereo DSD DACs! I'm more a fan of the all-in-one box approach to Multichannel. I'm sure you have plans to review a few of the DirectStream DACs soon for Multichannel DSD testing, for Computer Audiophile.... right?!

 

And I do agree on the preamp question. I'm using the EMM Labs Switchman II here - works great.

 

Just kidding...my exaSound all-in-one is the multichannel DAC of choice, for real world playback. I still do the stack, but more for testing and such.

 

As afar as DSD upsampled/converted to 30bit x 10xDSD, isn't that what much-heralded DSD experts like folks like Bruno are doing with the plan for the Mola Mola DAC? And Bruno knows a thing or two about DSD and A/D and D/A's (aka Grimm)? Isn't it about ending up in DSD, anyway (assuming the tough math is done in the deep stratosphere as opposed to our miniscule elementary 24/176k conversions to date). I mean, great beautiful DSD music that all of is have come to love have almost always been edited (or certainly cut and spliced) in a much smaller sample rate of 24/352.8k (hence the name DXD) before going back to DSD, and no one who is complaining here seems to complain about it....it would be great to do that at MUCH higher samples

[Mike Mcsweeney]

nvm

[Ted Smith]

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.

...

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

 

 

Howdy

 

The digital volume control is in my DAC there because it was "free" and non-destructive to add it to the volume control-less design. I explain above why, even before my DAC could do PCM, I did my own sigma delta modulator.

 

The JRiver argument is flawed. Or at least not everyone buys it. Even tho, for example, the Merging Technologies Pyramix converts DSD to 352.8 at 32 bits to do their editing (and storage) of DSD some of us claim to hear a slight degradation when doing so. To be fair I also think the Sonoma editing card isn't quite as good as going out to an analog mixer and then rerecording in DSD.

 

My experience indicates that lowering the sample rate is almost uniformly bad. Conversely I have no reason to believe that widening the precision of each sample is ever bad. Now how you narrow the sample width with dither in PCM or requantize in DSD is an issue and you can loose info there.

 

-Ted

[Ted Smith]

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.

 

Indeed the DirectStream is the exact same form factor and has the exact same inputs and outputs. I use a PW chassis, power supply, display and processor board for development.

[Ted Smith]

Howdy all

 

I'll stop digging earlier in the thread. If I missed your question, misspoke, or you want further excuses (oops, I mean clarification) please ask. I'll try to poke my head back in periodically for a while.

 

-Ted

[matthias]

As afar as DSD upsampled/converted to 30bit x 10xDSD, isn't that what much-heralded DSD experts like folks like Bruno are doing with the plan for the Mola Mola DAC? And Bruno knows a thing or two about DSD and A/D and D/A's (aka Grimm)? Isn't it about ending up in DSD, anyway (assuming the tough math is done in the deep stratosphere as opposed to our miniscule elementary 24/176k conversions to date). I mean, great beautiful DSD music that all of is have come to love have almost always been edited (or certainly cut and spliced) in a much smaller sample rate of 24/352.8k (hence the name DXD) before going back to DSD, and no one seems to notice or care..

 

Yes,

the Mola Mola samples PCM and DSD to 32 bit / 3,125 MHz, then converting to 1 bit / 100 MHz, then to analogue.

 

Matt

[bmoura]

Just kidding...my exaSound all-in-one is the multichannel DAC of choice, for real world playback. I still do the stack, but more for testing and such.

 

The exaSound e28 is part of my system as well. They may get competition someday. But for now, it's the way to go for Multichannel DSD playback.

[ss4]

I am personally pretty excited about this. I looks like in Canada it will sell for $6600, not exactly chump change. But, to me it is worth it if it will made my library sound better...

 

PS Audio to release the all new DirectStream DAC

[MikeJazz]

I would like to know if the conversion PCM to DSD is a lossless process....

[Ted Smith]

I would like to know if the conversion PCM to DSD is a lossless process....

 

Howdy

 

No, but the losses are almost all outside of the audio band. The digital DSD conversion process in the DirectStream has better S/N than the hardware. The benefits clearly outweigh the problems so I feel good about converting PCM to DSD for conversion to analog. The S/N rises to only about 100dB @ 60k, but that's far from a real problem.

 

-Ted

[opus101]

The benefits clearly outweigh the problems so I feel good about converting PCM to DSD for conversion to analog. The S/N rises to only about 100dB @ 60k, but that's far from a real problem.

 

Have you characterised the dynamic shifts in the noise floor with signal level (aka noise modulation)? Lipshitz and Vanderkooy pointed out this issue in their AES convention paper in 2001 which they said is due to the impossibility of correctly dithering a single bit modulator.

[Ted Smith]

Have you characterised the dynamic shifts in the noise floor with signal level (aka noise modulation)? Lipshitz and Vanderkooy pointed out this issue in their AES convention paper in 2001 which they said is due to the impossibility of correctly dithering a single bit modulator.

In fact there's a link on ESS's website to a pretty nice presentation on this problem (and their related experience which mirrors mine.)

Noise modulation is a tough problem and I did have to take quite a bit of time and try a lot of various "patches" before I got rid of any effects audible to me. I've found a few different techniques that help. In my quest I found more than a few SACDs which let some thru. One SACD in particular caused me a lot of trouble till I found that the noise was in the recording not my algos.

 

P.S. I just read your other posts here and I'll not answer any more questions from you.