The message is a simple one: a consumer-grade PC or Mac’s circuits aren’t designed with electrical noise minimisation in mind. Inside your average Apple or Dell, numerous high-frequency fundamentals ring out before harmonics bring in the chorus. The result? A wide (frequency) spectrum of noise.
When deployed at the top of a digital audio playback chain, EMI and RFI can prove subtly troublesome to downstream connected devices. Getting more device specific: direct-connect said PC or Mac to your DAC via USB cable and that noise can work its way into the converter where it messes with the timing of sensitive (sample rate) clocks which in turn piles on the jitter.
The majority of USB receiver chips rely on the host computer’s 5V line for power. A direct electrical connection between the two is mostly unavoidable. And as we shall see, even a D/A converter that powers its USButler from within isn’t immune.
AudioQuest’s approach to tackling the noise that travels along USB’s power and data buses is to passively filter it. The JitterBug connects directly to the host computer’s USB port and attenuates internal electrical noise by sending (some of) it to ground. You can read this commentator’s findings on real-world JitterBug deployment here and here.
Also from California, Schiit Audio’s approach is a little more involved. Their US$99 Wyrd device – made entirely on US soil – is a single-port USB hub with low noise power supply. Sharp-eyed readers will already have noted that a second USB cable is required: one between computer and Wyrd and the other between Wyrd and DAC.
Company founders Jason Stoddard and Mike Moffat are (wisely) making zero noise (ha!) about what it can/will do for your DAC’s USB input and the resulting sound.
“Some listeners say Wyrd improves the sound of their system. We’re not going to make any such claims. We remain Swiss on the matter—we don’t do the hard sell by promising sonic nirvana,” they say.
But this is Schiit Audio where ‘no promises’ implies the opposite. More likely that Moffat and Stoddard (again wisely) don’t wish to be drawn into the keyboard warrior’s favourite trap: the substantiation tangle. The more you struggle the tighter it gets.
Moffat and Stoddard refer to the Wyrd as a ‘Decrapifier’ on their website. In person the term they use is far less PG.
Pronunciation? “Wired”? “Weird”? “Word”? Anyone?
The Wyrd’s job is two-fold:
1) It allows data to pass through unhindered but cauterises the incoming Vbus. From the supplied linear wall wart the Wyrd slipstreams its own 5V feed. According to Schiit’s promo blurb this AC adaptor features “precision low-noise (2.5uV) voltage regulators”. The goal here is to lower the noise reaching the DAC over the Vbus not by filtration but by substitution.
However…
2) The Wyrd also Reclocks* – or rather repeats – the USB data packets coming from the host computer, the aim of which is to maintain a more evenly timed flow of data between source and DAC and so improve overall signal quality.
Signal quality? Electrical noise riding along the data bus can cause the USB receiver chip’s physical layer (PHY) to work harder to determine a) the precise location of the signal’s edges and b) the arrival timing of those edges.
The USB signal repeating role of the Wyrd is for improving signal quality so that the PHY can make lighter work of reading the signal. A better quality USB data stream lowers the chance of the PHY activating its noise-creating pre-processing steps. Such thinking also holds true for Ethernet cables and the corresponding PHY in Ethernet receiver chips.
The theory being outlined here is both complicated and multi-layered. If you’re still struggling to understand all of it (as I am), don’t fret. I’ll be digging into it again when tackling Uptone’s Regen.
I’ve thus far avoided mention of dropped packets or data errors. Those of a ‘bits are bits’ persuasion might be surprised to learn that file transfer methods over USB – where time sensitivity isn’t an issue – differ from transfer methods used for digital audio where time sensitivity is an issue.
Transferring a file from computer to external hard drive deploys ‘bulk transfer mode’ which calls into effect data integrity checks, ensuring all packets sent are received (or else are resent). These data checks don’t take place within the isochronous USB data transfer method used for digital audio; it’s a real-time transmission method with no provision for packet resends. Transmission errors can and do happen.
For the sake of argument let’s assume all data transmitted by host computer (and repeated by Wyrd) is received by the DAC. When faced with a high-noise or high-jitter data stream the PHY has to work extra hard to play catch on the incoming data and in doing so can create more noise inside the DAC. That is, even when every last morsel of data arrives noise can still occur.
Enough theory, time to listen.
The Schiit Wyrd makes zero audible difference to DACs fronted by an Antipodes Audio DX, itself tuned for lower noise from the outset.You can’t attenuate/eradicate that which doesn’t present in the first place. If anything, the DX loses a little of the organic nature of its sound.
Improvements weren’t fully noted until the Wyrd played intermediary between a 2014 MacBook Air and a range of decoders – the Chord Hugo, the Hugo TT and the Resonessence Labs INVICTA – in a headphone setup headed up by HiFiMan’s HE-1000. Results were then confirmed on a late 2014 MacMini in the main system: a Vinnie Rossie LIO driving KEF LS50 standmounts where an Aqua La Voce subbed in for the Chord Hugo TT.
Splicing the Wyrd into the chain saw treble glare diminish, replaced by a better sense of clarity and avidity. To what degree proved to be DAC dependent. The original pocketable Hugo converter turned out to be the most easily seduced by the Schiit box’s charms.
The Wyrd matches the JitterBug’s improvements with music’s rhythmic suppleness and ease. With either in a playback chain committed to The Hold Steady’s Boys And Girls In America, Chad Keubler’s guitar work didn’t cut the air with quite as much edgy etch. Where the Schiit bests the JitterBug is on the amount of metaphorical liquidity a listener might ascribe to musical flow. Nowhere was this more apparent than on the extended electronic jams that make up the front half of Minilogue’s pulsing Blomma.
The mains powered unit also outstrips the less costly passive dongle on layer delineation, particularly front to back. Without the Wyrd, the KEF wall of sound is a little more two-dimensional and not quite as engaging in the long-term.
Do I think the Wyrd brings more of an improvement than the JitterBug? Yessireebob. Did it improve each and every scenario tested here? Yup, but the delta width varied.
In my listening tests, pre-pending the JitterBug to the Wyrd made not a jot of difference. If adding it post-Wyrd made the tiniest of improvements I didn’t always hear enough to put money where my mouth is.
Those hoping for a low-cost shortcut into Antipodes’ high-end performance territory will be disappointed: the Wyrd does not lift the MacMini to DX levels. Not even close. What it does do is strip away a good portion of the Apple computer’s treble glare (which I’d not properly noticed myself until reviewing the Antipodes DS Reference almost two years ago). You don’t know what you’ve got till it’s gone – isn’t that how the old saying goes? Here, gone is good and for which US$99 is money VERY well spent.
None of improvements exacted by the Wyrd are what you’d call huge. Feel free to call bullshit on anyone who trots out the night and day dichotomy. You get more than you might from swapping out a USB cable, less than changing the DAC itself. Even so, the Schiit earned its place in all manner of configurations here and there isn’t one in which I wouldn’t use it.
The Wyrd is staying put.
And whilst this ‘ere USB hub doesn’t render the Light Harmonic’s LightSpeed cable redundant it does erode some of Old Red’s edge. (Ironically, the Wyrd demands two). My advice? Get a Schiit Wyrd first, add exotic USB wire later.
Further information: Schiit Wyrd
*A USB data clock is not to be confused with a sample rate clock.