SOLVED: Is USB audio quality resolved with the Raspberry Pi 4? i.e. No need for SPDIF Output HAT's

Yes, as you say. It’s a shame we don’t have direct measurements of the disputed item.

When I look at the RPi, though, I see a very low-power device with very little going on. There are no motors, no fan or spinning disk, prime sources of electrical noise. There have been reports of HDMI devices leaking noise into the RPi, which is why in 2017, three years ago, @spockfish modified RoPieee to shut down the HDMI subsystem when operating. There’s no good reason why there should be noise on the USB lines.

So unless someone presents some evidence of this supposed noise, I think it’s most likely to be an urban myth, or even a myth advanced by manufacturers of overpriced streamers to keep their sales up. Like I say, I try to keep an open mind on these things. Show me some evidence. What evidence there is seems to argue that there’s no noise.

8 Likes

No, it is your “reality”. What is suspicious with Archimago’s results is that all ports on all devices look the same. That is not going to be reality…

Bits are not just bits. If only life were that simple. I suppose you think DACs are just DACs too…and that amps with the same measurements must sound the same and that cables make no difference.

There is no detectable noise from a RPi4 with USB out to a good DAC.

6 Likes

That is simply false.

So:
Tidal > Roon > RPi 4 > Digione Sig > Modi (Coax) > Saga + > Vidar > KEF LS50

I think sounded marginally better (blacker background, more instrument separation and clarity) to me than

Tidal > Roon > RPi 4 > Modi (USB) > Saga + > Vidar > KEF LS50

but not by much. Obviously I do not have high end equipment nor am I experienced at critical listening. I also have young kids which makes nothing easy.

Looking forward to getting the Modius with the Unison USB and comparing again. However, based on this test, I don’t think I’ll be buying HATs again, at least for my equipment.

3 Likes

It’s irrelevant if the Pi USB is noisey or not.

What’s relevant-if that affects the output of the DAC. Because that’s what we end up listening to.

DAC’s generally have noise rejection built in.
So far we seen multiple measurements showing that there’s no audible difference from these devices.
We haven’t seen measurements showing the opposite. Maybe someone will come up with some, but so far we haven’t seen them. Including from all the producers of those streaming devices.

Those are the facts.
Believe what you want. But you should take the facts into account.

What’s for sure - there isn’t some kind of magic going on here.

5 Likes

You cant determine audible differences with measurements. This might seem strange to some people, but you actually need to listen to determine audible differences.

I used a RPi4 to my RME ADI-2 DAC before, now I use an upgraded microRendu behind an opticalModule, and the difference is quite big. And the difference would be even bigger if I used an better transport, like Lumin, oR, sotm ultra, etc.

2 Likes

@Billy_McSkintos I reply here with No measurements Just Ears Only: I have tried Pi4 & Pi3 USB out direct to DAC with Very Little Difference in SQ if any to my ears. Older USB DAC’s I have found where better severed via COAX, AES than USB. Many modern DAC have much better USB implementation where USB/SPDIF are all but in distinguishable IMHO, so USB for higher than red-book or convenience.

I also have the Allo Digione Sig which really is a performer and to me out performs Pi USB and is a fine streaming solution. Also have the Allo USBridge Sig where this is all you need for good USB SQ. Putting the Digione Sig onto the USBridge Sig to me produced equal SQ.

Just to throw a spanner into all this I have found a Digital to Digital Converter (USB IN to SPIDF/AES/I2S OUT) and or a Streamer can better USB. It is not a giant leap yet to my Ears I use a Streamer in-between Ethernet and DAC for best SQ (to me keeping the signal clean post ethernet was key).

Now in a fast-changing better bang-for-buck digital domain, USB could well end up being the de facto SQ standard. Its hard to keep up and I am arriving at the point of “your main HiFi is good enough to just stop all this gear (all the gear and no idea) chasing and just enjoy the music”

6 Likes

It defies logic to say that the “difference is quite big” but it can’t be measured in any form. What’s the difference: less distortion, less jitter, or something else?
Those can all be measured. Please don’t say the difference is a “cleaner” signal.
What’s been cleaned up?
How do you explain no measureable difference of the output at the DAC? Surely if those devices are helping the DAC work better some difference should be visible: less distortion, less jitter, better SNR.

I’d challenge you to undergo a non sighted listening test and see if you can pick out those “quite big” differences consistently. I’d bet you can’t.

Of course it’s also possible your chain is simply introducing euphonic distortion into the playback that sounds good to you…

9 Likes

Agree with danny2. When you think what the original post was about, it’s a tad humorous to see how this has, yet again, exploded into a battle of dacs and NAA devices. Pi4 is good enough for the original poster’s question. He may well spend £1000s of pounds on the network transport and gain some marginal sound improvement.

2 Likes

Lots of technical. reasons why a RPi4 isn’t a perfect transport, but lets leave that for now. What is important is that I have actually tried and listened to both an RPi4 and a better transport, and in my system the difference is quite big. And yes, a blind test would be easy to complete 10/10 if I could select the tune to listen to.

From a price/performance, the RPi4 is pretty good, but don’t try to fool anyone that its as good as it gets because it isn’t if you have a reasonable transparent HiFi system.

Yeah, we always seem to be leaving that for now… Don’t explain, show evidence.

4 Likes

Actually I usually explain, but why bother. People who listen by looking at measurements is never going to understand that you actually have to use your ears to listen.

For people who might actually want to learn something: switched electronics (especially voltage regulators) and inaccurate clocks are two of the reason why an RPi4 is never going to be optimal for audio, unless using a HAT that fixes this of course. And no, I am not talking about clock for 44.1/48khz, but underlying 24Mhz USB protocol clock.

So:

  • The USB in the RPi 4 has been decoupled from the ethernet bus.
  • There is no measurable noise significantly impacting audible listening ranges out of the USB.
  • Most ‘newer’ DAC’s can filter/reject any such noise anyway
  • This ability varies from DAC to DAC

However, subjective listening may suggest:

  • The USB clock could cause problems with sound quality and so a HAT could improve it
  • Another NAA/Transport could have less of these issues and thus improve sound quality
  • These are subjective opinions based on one’s audio chain and environment

It seems that spending the money on a “better” DAC would likely improve both sides of the argument.

3 Likes

In what we laughingly call “The Age of Reason”, explanations are usually advanced after something has been observed in some way that lets others puzzle over the observations – that thing we call evidence. The interesting fact about high-end audio is that we keep getting these “explanations” about this and that, in response to no actual evidence. They seem, to me, to be more of rationalizations than explanations.

Or just plain myths, perhaps circulated by unscrupulous vendors who have to have some reason for selling simple electronics at rapacious prices.

Then there are the pitchmen, the folks who call themselves “reviewers”, who in olden days stood outside the booth of a carnival, barking at passersby to come in and see the wonders within the tent, “ya won’t believe it till ya see it with your own eyes!”, but nowadays have to lurk on the Web yelling about subjective “reviews” (ya won’t understand till ya hear it with your own ears!) to draw eyeballs to the ads of rapacious vendors who have to somehow sell their overpriced electronics.

But I think what you do have with your “24MHz USB protocol clock” is a hypothesis. Which might be re-stated as, “My hypothesis is that the RPi 4 will never be a good USB streamer, because of the underlying 24MHz protocol clock, which …” (There’s a missing piece there, which is about why the protocol clock should invalidate the use.) The next step, of course, would be to run some experiments and gather some evidence, to see if that hypothesis could be validated. I’d be interested in looking at that evidence, once it’s gathered. But until then, it’s just a hypothesis.

13 Likes

Until there is a proof that everything we hear can be measured, all audible conclusion based on measurements are also hypothesizes. In fact, as it is right now, no scientific conclusions can be drawn from measurements with regards to how it sounds except very basic ones like “this is silence”.

I said that on Amir forum (audiosciencereview forum) and was banned, which shows just how “scientific” they are. More like a religious sect that can’t accept people with different opinions :stuck_out_tongue_closed_eyes:

But what we all can do is use our ears and listen, and if I hear something sounding better and it keeps sounding better after a few days, its enough to convince me.

So you can’t prove that a piano or guitar are turned correctly using electronic devices? You can’t tell how loud or quiet music is? How much distortion is present? Or how much noise? You can’t subtract A from B and see if there is anything left?

Wow! Your fact is so revealing.

1 Like

Those are what I would call basic facts, but measurements are often used to make broad statements. A good example in this thread is measurements that show how RPi4 is “perfect” as a transport, and yet a better transport sounds more transparent with more air, bigger soundstage, added clarity etc.

Lots of other examples, I can take one that I have personal experience with: RME ADI-2 DAC and Chord Qutest (I have the RME DAC myself and has listened to Qutest). Both are good DACs, but they don’t sound the same, in fact they sound quite different despite having very similar measurements.

Measurements can be useful, no question about it, but what I object to is using measurement to determine how something sound.

All these buzzwords! Show me the recordings, please.

5 Likes

That don’t make any sense, you want to see the cover of the recordings? Or how can you show sound?

I see with my eyes and I listen with my ears, I suggest you try it :slight_smile:

1 Like