Raspberry Pi Postscript: S/PDIF Vs. USB Sound Comparison
Pi versus Pi
By Steve Graham
Raspberry Pi 4B with S/PDIF Out versus Raspberry Pi 4B with USB Out
Confession time. In responding to those kind enough to comment on part 3 of my Raspberry Pi series, I mentioned that I didn’t have a decent USB DAC for comparison. WRONG. It completely slipped my mind that my PS Audio DAC has a USB input. A cream pie in the face for me.
To satisfy my own curiosity and maybe mend a few fences with readers, I’ve run a quick comparison between two Raspberry Pis. One is a Pi 4B with the HiFi Berry Digi+ Pro HAT (the Pi used in the 3-part series), which sends digital out over S/PDIF coaxial. The second is a Pi 4B (no HAT) sending digital out over a USB connection. In both cases the digital signal was routed to my PS Audio DirectStream DAC. The PS DAC did dual duty in the three-part series, receiving data from both the Pi with HAT and my NAD 50.2 player. The DAC on the RPi board was not used.
I had a heck of a time getting the PS DAC to lock onto the USB signal. The USB output of the Pi was selected in the Squeeze Lite setup menu. I tried several methods to get it to lock, restarting PiCorePlayer, rebooting PiCorePlayer, even powering down and a cold restart. It was just a bit of a lottery whether or not the DAC would lock. I could never figure out why it would or would not lock, it just seemed random. Thinking it might be the DAC, I tried a Meridian Explorer 2 and it wasn’t any easier. The Meridian locks on to the USB out from my mid level HP laptop without issues. The PS Audio DAC locked, without issues, onto the coaxial S/PDIF signal from the Pi with the HiFi Berry HAT board.
Once locked, I compared the three tracks mentioned below, first on the USB-out Pi, then on the S/PDIF-out Pi.
Through the USB connection, this track sounded kind of catchy. The bass was a bit on the phat side, guitars had a nice ring to them and high frequency percussion didn’t “bite”. But the soundstage seemed a bit diffuse and the whole thing sounded sort of “mid-fi” and imprecise.
Through the S/PDIF connection, everything just snapped into a clearer focus. Bass was tighter, guitars and vocals clearer and highs had more precise detail without sounding harsh. The soundstage was more precisely rendered too.
Again, the USB connection was a letdown. The piano sounded sort of clangy on louder passages, the sax less liquid than I’ve heard it sound on my system, and the muted trumpet had less bite, brassiness and clarity. Percussion was less clear and the bass less distinct too.
Via S/PDIF the instruments sounded much more real. The was piano clearer and less strained, the sax smoother and at the same time more detailed. The trumpet was more believable, bass and percussion improved in the same way too. The recording acoustic was more apparent as well. There isn’t much of a continuous soundstage in this recording, but the trumpet was more precisely placed in the centre, and was as if it inhabited its own distinct pocket of air.
Same sort of deal here too. The S/PDIF connection had clearer vocals and the soundstage, even though an artificial recording studio construct, was more dimensional, with instruments more precisely placed and layered. Overall clarity and detail were better too.
Before anyone sends me a nastygram saying I must have done something wrong or I didn’t use power supply X, cable Y or software Z, I can only report on what I hear. I don’t mean to down play any of these improvements people make or say they don’t matter. I can only try so many options or variations on a reviewed item before I have to move on. My experience with USB may be atypical. If, in your system, USB is more enjoyable than S/PDIF, that’s fine, I’m not trying to negate your preference. It’s all about what sounds best to you. As always, performance is very system dependent. To summarize my experience: the fifty-seven bucks I paid for the S/PDIF HAT board was the best fifty-seven bucks I’ve ever spent in the service of musical enjoyment.
After finishing the Pi vs Pi comparison, I switched my system back to the NAD 50.2 and played the three test tracks again. There were improvements on many fronts over the Pi with the S/PDIF HAT board. However, these improvements were more subtle than the upgrade going from Pi with USB out to the Pi with S/PDIF out via the $57 HAT board.
The Raspberry Pi is an OK performing music server/streamer on its own, but, and it’s a big but, it makes a very much better server/streamer with the addition of an audio-grade HAT board.
I enjoy reading the review comments people leave. Thank you for taking the time to post them. My policy is to respond once and if the commenter wishes to follow–up, I give them the last word. If they ask a question on a second communication, I will endeavour to answer if possible or at least say I don’t know. I don’t think the comments space is a good place for ongoing sound quality discussions. Web forums better serve that purpose.
Send a Donation
If you've enjoyed this content, consider supporting Wall Of Sound with a donation. It only takes a moment and will be greatly appreciated.
What case did you use with the pi4b and the HAT board? Are you happy with the case?
Sorry, I found this confusing. You did succeed in connecting the USB out of the Pi on a consistent basis to the DAC and that was your comparison to the HAT SPDIF to the DAC? Or the commparison was some other setup, not comparing the two on the same DAC?
Thank you both for your questions. I apologize for any confusion. I’ll try to clear it up.
As for the case, both Pi with HAT and Pi without hat used the black case shown in the pictures in Part 2. This is the case included with the Pi 4B Budget Kit from BuyaPi.ca and I think PiShop.US.
With regards to the sound comparison between the Pi with HAT sending S/PDIF to a DAC and the Pi running USB out to the DAC: both Pis were connected to my PS DirectStream DAC.
As the ability to lock to the DirectStream was a bit hit and miss I tried the USB-out Pi with the Meridian Explorer 2 DAC. It was hit and miss with the Explorer as well. I even tried a USB regenerator between the Pi and the Explorer DAC but that didn’t seem to help.
There was never a problem getting the DirectStream DAC or an old Adcom DAC to sync with the S/PDIF HAT-equipped Pi.
I hope this clears up any confusion.
There will be one final episode in the Pi adventure where I try a DAC hat on a Pi. I hope to have it out before the end of 2020.
Thanks for commenting.
I have a Pi with DAC HAT with RCA analog outs. I find it sounds okay, but not as good as a decent DAC. (Something that costs $200-$700). It’s definitely better than a typical computer sound card, or some other introductory units.
RE: Raspberry Pi (USB Out vs. spdif/coaxial)
Hi Steve: What an interesting -and unexpected- result !
In all of my years in hifi (40+), nothing has given me more heartache than
digital’s spdif transmission specification; it’s flawed -terribly so.
Embedding/sending the vital clock signal with the data (as it is with every spdif connection) requires the receiving component to recover it precisely, and that has proven problematic; clearly it’s not easy to accomplish with the precision demanded. Listening sessions has repeateadly shown it to be highly flawed. Nine out-of-ten times it simply sounds bad -or at least not very good ! lol
But this is and has been nown since spif’s inception: It’s a terrribly flawed interface. Remarkably, although much better digital signal transfer protocols exist (Wrd. Clk, i2S (HDMi, RJ-45,) very few DAC/CDP manufacturer’s utilize it. Understandable I suppose with entry-level hifi, yet there are mega-buck DAC’s that don’t offer these superior alternatives. Most bizzare.
With that said, there are some exceptions, including Auralic, T+A (and others) that use a i2S type communication line but proprietry in nature -only useable with the company’s ‘matching’ gear.
In any case, it has been my experience that spdif is best avoided if at all possible. But, you may get lucky, and find that ‘synergstic’ match between devices and cable.
Back to your experience/observations, I wonder if the PS audio has some type of fault in its USB input/circuit? It struck me odd that you had a hard time getting a locked (USB) signal.
Here at my place I have five (5) Pi 3B+’s feeding various (USB) DAC’s: Volumio is used for internet radio listening -five different radio stations.
In all cases, I’m fascinated (delighted) with the sound quality. Although most stations transmit at 128 kb/s (compressed), today’s MP3/AAC (compression) algorithms are astonishingly advanced -compared to early 320 kb/s MP-3. I’m continually surprised at how good it sounds.
In my experience (about 4-years/computer-USB audio), USB (asynchronous) has been a model of high performance; it sounds good. Really good.
I’d be interested to hear what other have to say of your observations.
Thanks for sharing.
Thank you for your comment Mr. Jasz, I’ve been expecting it. I briefly mentioned in my report that I tried a Meridien USB DAC that always locks to audio from my laptop. The Meridien had trouble locking to the Pi as well and if memory serves did so less often than my PS DAC. So I don’t think it’s a PS DAC problem. Hans Beekhuyzen, the YouTube audio reviewer, has said that the Pi 4B playing audio doesn’t sound as good as the 3B. His speculation was the 4B’s microprocessor runs at a faster rate than the one in the 3B and the 4B’s might be generating more noise electrical noise. At any rate, my patience is very short for things that work intermittently.
A glass of wine with some S/PDIF-delivered music always sounds better to me than wine and no music.
I came accidentally to this site. For people who are looking for reliable data driven comparison there is nothing in here. No measurements, no numbers no scientific data. Just anecdotal evidence. Additionally author is using PS Audio DirectStream DAC which has objectively bad measurements and vry high price. Btw Psaudio is known oil snake company in the industry, or to put it nicely caters to the vinyl crowd. Nevertheless I am grateful to the author of this article it is useful point of view in marketplace of ideas. Besides some people are interested in subjective points of view.
Btw I am microprocessor hardware engineer and only interested in hard numbers So pls understand where I come from take my point of view accordingly.
Hmmmm: “I came accidentally to this site.” (Who are you -have you a name?)
” ..Btw I am microprocessor hardware engineer and only interested in hard numbers So pls understand where I come from take my point of view accordingly.”
Then what is your point/purpose in responding/replying to this review ?
C’mon man. You can’t seriously get different audio output from USB vs. S/PDIF unless you do something terribly wrong. Both send binary data, so the DAC should get the same data in both cases and thus produce the same result. It’s 100% impossible to have tighter bass and clearer vocals when comparing two identical digital signals.
The only thing that should be different is electical noise, which you can compare by cranking up the volume knob without actually playing anything.
Thank you all for your interest in WOS.ca and for taking the time to comment.
Even folks far more learned than myself have come to the conclusion that measurements do not necessarily determine how a component will sound. In my opinion each audiophile’s ears must be the final arbiter of sound quality, however they define that for themselves. We don’t judge wine or coffee by how they measure, each person judges according to their own taste buds. I’m not saying measurements are irrelevant, just not the be all and end all. Making the measurements you suggest would be beyond my knowledge and certainly my ability to afford. If you would like to contribute any technical findings of your own Noam might be interested in publishing them especially if you can correlate them with listening experience.
I can only report on what I heard. As you say, “The only thing that should be different is electrical noise……” I am of the opinion that there is no perfect digital interface. In other words, they are all flawed to some degree or other. I’ve read that the USB buss on the Raspberry Pi is quite electrically noisy, the 4B more so than the 3B because of the 4’s faster clock rate. I don’t believe you will hear electrical noise with no music playing. As far as I know DACs will mute in the absence of a data stream (or not, I might have this wrong). In any event, it’s how electrical noise affects the clocking of the data stream that is the generally accepted explanation as to significant variations in digital sound. If you want to avoid both USB and S/PDIF interfaces have a read of my WOS report on the Pi with an Allo Boss I2S DAC HAT board. In my humble opinion that combination offers very good sound for an almost unbelievably low cost.
Best regards to all,