PS Audio NuWave DSD


Researchers at MIT recently discovered a “music channel” in the human brain. These neural pathways respond to all kinds of music—and only to music. “A listener may relish the sampled genre or revile it,” Natalie Angier wrote in the New York Times. “No matter. When a musical passage is played, a distinct set of neurons tucked inside a furrow of a listener’s auditory cortex will fire in response.”

“Why do we have music?” Nancy Kanwisher, one of the scientists behind the research, asked in an interview with Angier. Is our “sensitivity to music . . . tunable with experience? These are the really cool first-order questions we can begin to address.”

To me, the research provokes other questions: What is it about an auditory experience that tells our brain it’s music? Would a serious audiophile’s functional magnetic resonance imaging—the technique used in the study to scan the auditory cortex—look the same as other people’s fMRIs? Might it someday be possible to observe differences in the brain as it responds to different presentations of the same music?

For the past several weeks, I’ve been auditioning PS Audio’s NuWave DSD digital-to-analog converter ($1299), and the experience has me waxing philosophical on one of my favorite subjects: the connection between reproduced sound and our perception of music. The NuWave DSD, I’ve found, presents music in a striking and characterful way. What aspects of its sound—and the underlying electronic design—cause the music to come across the way it does? I’ll return to that question in a moment—but first . . .

The NuWave DSD presents music in a striking and characterful way.

Drums and wires

PS Audio’s flagship D/A converter, the DirectStream DAC ($5999), created a stir in the months following its 2014 debut— not only for the quality of its sound, which was widely deemed excellent by reviewers, but also for the novelty of its design.2 Instead of relying on PCM-based input processors, as most DACs do, the DirectStream converts all incoming data streams to DSDx10, which, conveniently, is the least common multiple of two common sampling frequencies, 176.4 and 192kHz. Instead of a dedicated DAC chip, the DirectStream uses a field-programmable gate array (FPGA), which allowed designer Ted Smith to use, as he told me via e-mail, “accurate math instead of the approximations that most DAC chips and upsamplers use.” The FPGA approach also allows the DirectStream to be upgraded in the field by updating its firmware, as has already happened three times in the months since the model’s launch, each upgrade reportedly dramatically improving the sound. The DirectStream’s output stage includes expensive transformers as part of a low-pass filter that smooths the reclocked, downsampled DSD data to make music.

1 Natalie Angier, “New Ways Into the Brain’s ‘Music Room’,”New York Times, February 8, 2016:


Description Solid-state digital-to-analog converter. Digital Audio Inputs: Coax, TosLink, USB, I2S. Sample Rates: Coax input: PCM, 44.1–192kHz. Optical input: PCM, 44.1–96kHz. I2S: PCM, 44.1–192kHz; DSD, single rate (2.8MHz) and double rate (5.6MHz), raw. USB: USB2.0 Asynchronous Audio; PCM, 44.1–192kHz; DSD single rate (2.8MHz) and double rate (5.6MHz), DoP. Analog Audio Output: 2 unbalanced (RCA), 2 balanced (XLR). Output level, maximum (RMS): 2.8V unbalanced, 5.6V balanced. Output impedance: 100 ohms (unbalanced), 200 ohms (balanced). Frequency response: 10Hz–20kHz, +0/–0.3dB. THD: 1kHz, 0dBFS, <0.01%; 20Hz–20kHz, 0dBFS, <0.2%; 1kHz, –10dBFS, <0.01%. Noise: <–80dBV. Format: PCM, DSD. Power Requirements: Input power (model specific): 100, 120, or 230VAC; 50 or 60Hz. Power consumption: 15W.

Dimensions 8.2" (210mm) W by 2.4" (61mm) H by 14" (360mm) D. Weight: 11.9 lbs (5.4kg). Finishes Silver, black.

Serial number of unit reviewed ANW-A-5J0621.

Price $1299. Approximate number of dealers: 65. Warranty: 3 years.

Manufacturer PS Audio, 4826 Sterling Drive, Boulder, CO 80301. Tel: (720)406-8946. Fax: (720)406-8967. Web:


Marketing copy on PS Audio’s website says that in designing the NuWave DSD they’ve “taken the lessons learned from one of the world’s best DACs, the DirectStream, and applied them to a smaller-scale, lower-cost unit.”3 I don’t doubt that claim, but the NuWave DSD and the Direct-Stream are very different DACs. The NuWave DSD does not convert incoming signals to DSD. It does use a dedicated DAC chip (a Sabre32 Hyperstream, from ESS Technology). And while the NuWave does employ a programmable integrated circuit— a complex programmable logic device (CPLD)— that component is smaller and less capable than an FPGA. The CPLD, PS Audio says, “discovers sample rate and format, reclocks all incoming data, reduces jitter, waveshapes data output to the DAC chip, and utilizes high speed/low gate count logic to reduce propagation delay for faster throughput.” Both DACs use passive filtering at the output, but the NuWave DSD does it more cheaply.

The NuWave plays DSD files, whether they arrive in native form or via DSD-over-PCM.

In contrast to the DirectStream, “The NuWave DSD cannot be updated in the field,” according to Bill Leebens, PS Audio’s director of marketing. “It can be reprogrammed at the factory, but we don’t anticipate any firmware updates on the basis of sound quality; we’re happy with that.” He added, “you can only go so far for a grand and change.”

2 See

3 A few days before I turned in this review, PS Audio announced the DirectStream’s true trickled-down heir: the DirectStream Junior ($3999).


The NuWave DSD plays DSD files (hence its name), whether they arrive in native form (at single or double rate) via an HDMI cable serving as an I2S interface (you’ll need a similarly equipped transport, of which few exist; PSA has announced one for October release), or via DSD-over-PCM (DoP), in which the DSD stream is broken into chunks and buried in PCM code transmitted via USB. Some say DoP is inferior to native DSD, but if the reassembled DSD stream is identical to the original stream (which it is), and both streams are reclocked in the same way (which they are), there should be no difference in the sound.

Of course, the NuWave DSD accepts PCM data, too, up to 24 bits and 192kHz via S/PDIF, asynchronous USB, and I2S. (The Tos-Link input accepts 24-bit data up to 96kHz.) Some converters go higher in bit depth and frequency, but 24/192 should be more than enough for transparency.

For years, PS Audio has focused on AC power problems, because dirty power lines can compromise the performance of even the best audio component. The company’s most sophisticated standalone solutions to those problems, their Power Plant AC regenerators, store the energy from the rough, 60Hz sinewave that comes out of your wall and use it to make a much cleaner sinewave. It’s not surprising that a company that so emphasizes power is attentive to the power driving its audio components. The NuWave DSD employs seven voltage regulators and “massive power supplies”—the phrase is from an e-mail from Paul McGowan, CEO of PS Audio—to ensure a consistent and precise rail voltage.

“There’s nothing we’ve ever made that eliminates jitter,” McGowan said in another e-mail. “And so cables still matter, though less so.” I was determined to give the NuWave DSD a completely fair audition, so I asked AudioQuest to loan me a high-quality USB cable. They sent three: their Cinnamon, Carbon, and Coffee, each 10' long. (Coffee is the company’s secondbest USB cable, second only to Diamond. I like their priorities.)

We got the beat

I listened to ripped CDs and other PCM files up to 24/192. I tested DoP with Kim André Arnesen’s Magnificat, with Anita Brevik conducting the Nidaros Cathedral Girls Choir and the Trondheim Soloists (DSD128 download, 2L 2L-106). It worked fine and sounded great. I wasn’t able to lay hands on a source that could deliver DSD over HDMI, so I wasn’t able to try DSD in native mode.


Operationally, the NuWave DSD performed without flaw: It swallowed whatever I fed it in whatever format via whatever input, and it output music, without one glitch or pop. I tested all the inputs (except I2S over HDMI), but did most of my listening via USB, as I suspect most listeners will in the Tidal era.

Most audio reviewers praise long-term listening as the gold standard for auditioning an audio component. I agree, but there’s no better time to notice a difference in the sound than just after you’ve made a change in a familiar system. When I plugged in the NuWave DSD, hooked it up to my early-2015 MacBook Pro, and began playing music via the playback application Roon, the most striking and immediate changes were rhythmic. Drums thwacked. A woodblock sounded like a woodblock, not a recording of one. When I put on one of my favorite classical CDs, Clifford Curzon’s 1962 recording of Brahms’s Piano Concerto 1, with George Szell and the London Symphony Orchestra (CD, Decca 289 466 376-2), the clicking of fingers on keys grabbed my attention as never before, and I wasn’t even paying close attention.

The DirectStream added enough bass weight and authority to be meaningful.

With Lucinda Williams’s excellent West (Lost Highway 9858348, 16/44.1 via Tidal), the grunt and, especially, the gravel in her voice really came across. “Wrap My Head Around That,” a repetitive Delta-blues rap that Robert Forster called “a song of unremitting spite and pain,” evoked a grim, countrified “Mars Eats Guitars” or a twangy, down-tempo “Burning Down the House.” (Weird comparisons, I know.) Not that this was an especially challenging test, but the NuWave DSD convincingly sold Williams’s spiteful laugh at the end.

So far I’ve emphasized the NuWave’s rhythmic qualities, but there was more to this DAC than percussion. In the Gryphon Trio’s recording of Beethoven’s Piano Trio 1 in E-flat, Op.1 No.1 (9 CDs, Analekta 295108), I heard a spacious acoustic and strings with a pleasant, meaty buzz, nicely complementing an appropriately percussive piano. I heard something similar with Marc-André Hamelin’s Live at Wigmore Hall (CD, Hyperion CDA66765), one of my 2012 Records to Die For: a great sense of the acoustic space, and a solid, impactful piano with nice transients on the leading edges. (I played both of these recordings from actual CDs using my Ayre Acoustics CX-7eMP CD player as transport, connected, via the Ayre’s AES/EBU digital output to the NuWave’s S/PDIF input, by a Canare transformer and a Stereovox HDVX coaxial cable. It worked flawlessly.) Generally, with good recordings, the density and placement of images—from side to side and from front to back—were very good.

My best friend’s girl

Toward the end of the listening period, JA loaned me his DirectStream DAC—he’d bought the review sample—to compare with the NuWave DSD. First impressions revealed a smaller change than I’d expected, but over the course of an evening the DirectStream grew on me—and grew. Listening to the first movement of Per Nørgård’s Symphony 1, with Sakari Oramo conducting the Vienna Philharmonic (Dacapo 6.220574, 16/44.1 via Tidal), the DirectStream added enough bass weight and authority to be meaningful. I heard more separation of instruments in space, on a wider, deeper soundstage—and was there also a bit more air?

The most important improvement with the DirectStream was that the music just seemed more relaxed. With this sense of relaxation—better spatial presentation may also have played a role—the music became more believable. Listening to “Light Blue,” the first track on Thelonious Monk’s At the Five Spot (probably Milestone 0888072470439, 16/44.1 via Tidal), I was struck by the feeling that I was attending a live performance. I don’t mean that I listened closely, analyzed the sound, and decided that it sounded “live.” Rather, I was transported. The tonal balance wasn’t quite natural—this 1958 recording is good, but not that good—but there I was at the Five Spot, near a wall or around a corner from the stage. That’s what $5999 can buy.

Downsides? The NuWave DSD is intended for use in a home system; there’s no headphone jack, no volume control. Be careful with the top plate—it looks as if it would scratch.


Spirits in a material world

I’ve been out of the digital game for a while, listening to terrestrial radio (WKCR, the excellent Columbia University station) and my collection of LPs. I haven’t missed digital sound. But recent developments—Tidal, MQA, the paucity and rising prices of good used vinyl, and maybe my lifelong tendency to reject whatever’s currently cool— have me listening more to digital. With a DAC like PS Audio’s NuWave DSD, whose sound possesses many of the virtues of my old-school analog rig—texture, vividness, corporeality, rhythmic assurance, surface noise (kidding!)—listening to more digital wouldn’t be much of a sacrifice. In fact, in the last few weeks I’ve had a blast with computer audio. Maybe it’s time for a reboot.

The clicking of fingers on keys grabbed my attention as never before.

Nearly 10 years have passed since I wrote my first review for Stereophile. Every review I write still teaches me something new about audio, and about reviewing. When I started out, I thought that, owing to my scientific training, I was better prepared than your average poet to write reliable, objective reviews. Yes, I was arrogant, and wrong: much more important, I’ve found, is experience.

Will I—will we—ever be able to discover exactly what sonic characteristics are responsible for certain subjective impressions: why the NuWave DSD makes music seem more rhythmic, or what aspect of the sound of the DirectStream yields the sense of relaxation that seemed to transport me to the Five Spot ca 1958? We’ll see. I’ll not give up easily on the notion that sound science, in both senses of the phrase, can be used to deliver peak emotional experiences—what a cool idea that is!

But, I’ve learned, there’s a basic problem: When you focus analytically on the sound—when you try to make listening more objective—you close yourself off to the music’s subjective character, its emotional message. That’s more or less the difference between a formal wine tasting—in which you take a small sip, swish it around your mouth, then spit it out—and sharing a good bottle over dinner with someone you love. (Does the brain have a wine channel?) The former can be done in a way that’s more or less objective, but what you learn from it is of limited importance. The latter is irredeemably subjective, and it’s the only thing that matters.