PS Audio DirectStream (DS) DAC: Snowmass Upgrade


Over four years ago, in May 2014, I reviewed the PS Audio DirectStream (DS) DAC. I ended up purchasing it and I still use it as my reference. Many times over these last four years I have been approached by other audiophiles asking why I have a DAC as my reference that is not ‘new’ and ‘up-to-date’ as compared to other DACs that have the newest DAC chip. ‘Oh no’, I say, ‘you are wrong’, ‘my DAC is only at most months old, it is yours that is old.’

My response is in fact, a fact, and that is what is so special about the DS DAC: Because its electronics are software based not hardware based, you get a new DAC every time PSAudio updates the software—and updates are free of charge. The DS at $6,000 retail does not have a traditional DAC chip built within; instead it uses a Field Programmable Gate Array (FPGA) which is an integrated circuit board that is programmed to serve whatever function it is commanded, including non-audio applications such as medical instruments and defence, distributed monetary systems and security systems.

For the DS, the function of its FPGA is digital-to-analog conversion (DAC), and its remarkable programmer Ted Smith has spent (and continues to spend) lots of time re-writing the software (roughly, every 9–12 months) using his unique expertise. The newest update, ‘Snowmass’, was released very recently. As with its predecessors (such as Torreys, Yale, Pikes Peak, Huron and Redcloud), it is named after a mountain in Colorado (PS Audio is stationed in Boulder, CO). So, I now have a new DAC again—as of early November 2018. This Snowmass upgrade is very worthy of consideration; hence this review.

A new software update affects everything from sound stage and imaging to distortion levels, the airiness and textures of instruments, transients, bass control, jitter reduction and so on. This is totally different from a typical audio device’s ‘firmware updates’ which in general do not and can not affect overall sound quality—because it is hardwired. Instead, firmware updates affect, for example, dimming and other display features of a monitor, and (say) functionality of the device (such as its streaming support capabilities).


Once an update for the DS is developed and released by PS Audio it can quickly be installed as a replacement for the old one by downloading it onto a laptop from the PS Audio website, transferring it to an SD card which you then snap into the back of the DS and reboot. You are up and running with a new DAC within 5 minutes. And I reiterate: Each such update is free of charge from PS Audio.

The ‘DS’ in the DS DAC stands for ‘Direct Stream’ which is from ‘Direct Stream Digital’, known as DSD. That is because of one of the many ways that the PS Audio DS DAC uses its FPGA is to convert any digital input into a 1-bit signal upsampled at 20 x the nominal DSD rate. But the purpose is not to convert to a standard DSD file, such as DSD 64, DSD 128, DSD256—it is to allow for a final conversion to analogue by utilizing the 1-bit DSD idea together with sophisticated use of upsampling and filtering. Even an original DSD 256 file is upsampled in this way by the DS. The DS DAC is unique in the way it uses the FPGA with its DS method. There are other DACs that use FPGA such as Aqua Hifi’s Formula xHD at $17,000. It uses an FPGA in a far less comprehensive way—and costs about 3 times the price of the DS.

In my original review of the DS I asked Smith how unique was the DS’ method of using a FPGA in the DS DAC, and I asked him the same question again very recently; he confirmed the same answer, so I repeat it here:

I do not believe there are other DACs that use the FPGA for everything from ‘parsing’ the input bits to outputting single bit DSD. There are a few other DACs that use sample rate conversion of all inputs to a single output frequency, but they use asynchronous sample rate conversion to convert from the incoming clock to the local clock or use a phase-lock loop (PLL) to drive the local clock. We use no input PLLs or FLLs (frequency-locked loops) and we use one master clock (and derive all other needed clock rates synchronously from that). As far as I know we are the only ones that do that.


Appreciating the sound quality of my DS DAC over time has been coupled with other significant hardware upgrades in my audio system including a stand alone preamplifier (PS Audio BHK Signature) as opposed to ‘direct to amps’, a music server that is very revealing (it is ruthless: the Mojo Audio Deja Vu Music Server) and most recently a pair of new mono block amplifiers (Audio by Van Alstine DVA SET 600 Mono Block Amplifier) that offer a larger/deeper soundstage, remarkable 3-D imaging, controlled bass, and a dead quiet background. Thus my current system can discern changes in the DS software that might not have been possible before.

The increase in sound quality exposed from moving from the previous Redcloud to this new Snowmass was the best I have heard. Many digital files, even at CD resolution that I already deemed special to me now have become even more magical in sound quality. I think the most apparent changes are in soundstage, imaging, bass control—and reduction of noise. Snowmass throws a larger/deeper soundstage with a more stable imaging of instruments that synchs so well with my new amps; kindred spirits.

Here are some examples.

In my original review of the DS DAC, I explained my passion for and enjoyment of the sound of cymbals. Just the other night (December 8, 2018), I had the pleasure of attending a small live concert of Ben Allison & Think Free consisting of Ben Allison (bass), Steve Cardenas (guitar), Jeremy Pelt (trumpet), Frank Kimbrough (piano), Alan Mednard (drums).

Besides the overall excellence of the performance, Mednard’s drumming truly intrigued me with his exceptional use of cymbals displaying their gorgeous delicate lingering decay, his complex use of the metal rims on all his drums (not just the snare), and his use of percussion mallets. When I came home, I played their most recent album, Layers of the City (June 2017) at 16/44.1. The DS with Snowmass did a lovely job with those cymbals, allowing for their natural live sound and decay, and spacial imaging. Never sounded so good before.

For singing, I listened to the double album Dreams and Daggers at 24/96 PCM by Cecile McLorin Salvant (2017), containing many live cuts. Outstanding intimacy on display with eloquence and emotional candour, the airiness of her remarkably flexible voice and the peripheral in-venue sounds of the audience including clapping and laughing (yes, she is quite witty at times). Her performance of the classic You’re Getting to be a Habit With Me is a perfect example. The audience joins in by clapping to the beat and shouting; the bass, piano and drums backing her up with such finesse. And no distortion.

Piano? Check out the 2007 album Stone Rose by Ola Gjeilo with Gjeilo on piano, Johannes Martens (cello), Tom Barber (flugelhorn) and David Coucheron (violin). I have it as a 24/96 FLAC file (recorded in Sofienberg Church, Oslo, Norway), on 2L. (It is also available on Tidal HiFi as an MQA 24/96 unfold. I use the Bridge II with the DS utilizing ethernet instead of USB, in which case the DS DAC can unfold MQA.) Much of the theme in Stone Rose is about New York City, with tracks such as, Snow in New York, for Piano, and Manhattan, for Piano. So delicate and gentle with dead quiets; superb acoustics revealed.

When I chatted with Smith recently, he said he thought that Snowmass resulted in the largest delta in sound quality between successive updates; I could not agree more. I asked Smith what kind of music does he like to use to tweak the software upgrades. He mentioned orchestral music because it offers so many instruments at once helping him with imaging of separate instruments, and soundstage overall. He even mentioned to me using the track North Country II from the Stone Rose album; its use of piano and flugelhorn in particular.

I also asked Smith whether a new update after Snowmass could possibly result in such a delta as Snowmass, and whether a new hardware design of the DAC itself would be possible/desirable. He said he thought that the current internal FPGA was sufficient for him to create further improvements, but most interestingly he thought that the DAC itself could be made half the length of the current model with only a one half inch increase in height. Amazing. Finally, Smith gave me an interesting piece of advice knowing that I use the DAC with a preamp: Change the DAC’s volume output level to about 86 instead of 100 before it sends out to the preamp. Subtle but noticeable, that modification seems to optimize the use of the pair; reducing noise even further, if that is even possible? So, I have been keeping the DS volume at 86-90. When I play at low volume levels at night, I can hear clean details such as those cymbals I mentioned in the Allison recording.


Snowmass is yet further proof that using software for digital-to-analog conversion instead of a fixed DAC chip within a DAC is the future. The PS Audio Direct Stream DAC is still exceptional and unique.

Very highly recommended.

Further information: PS Audio