• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required as is 20 years of participation in forums (not all true). Come here to have fun, be ready to be teased and not take online life too seriously. We now measure and review equipment for free! Click here for details.

SBCs as music servers / end points. Are they any good?

March Audio

Major Contributor
Manufacturer
Joined
Mar 1, 2016
Messages
6,382
Likes
8,929
Location
Albany Western Australia
#1
I was asked this question elsewhere. There is a view that the USB outputs from SBCs such as Raspberry Pi etc are inferior, noisy and inadequate for driving quality DACs. I dont share this view and decided to perform some tests and look at the output of my DAC1 and see if it is affected by different USB sources.

OK here are some measurements.

The SBCs are all running latest Volumio and Roon endpoint software. All playing via USB into my DAC1. I am playing a -60dB tone from Roon. If the PSU/SBC was causing USB noise problems you would expect to see it in the noise floor. Spot the difference.

Asus Tinker
Tinker -60dB.PNG



RPi
RPI.PNG


Khadas Vim2
Vim2.PNG


Intel Skull canyon NUC
skull Canyon.PNG



Sony Laptop
sony laptop.PNG




Essentially no significant difference. Well designed DACs should not have a major issue with using SBCs, they work very well. Of course I cannot comment whether any particular DAC you might have will suffer problems, but I very much see this as a DAC design issue and NOT a "noisy computer USB" issue. DAC should be expected to work without issue with a huge variety of USB sources.

Oh just for reference this is the sort of thing that can go wrong if you have gnd loops. I artificially created the problem by using the same computer that the measurement system is connected to to drive my DAC1.

8kHz USB packet noise and a bunch of other spuria.
Laptop with gnd loop.PNG
 

rebbiputzmaker

Addicted to Fun and Learning
Joined
Jan 28, 2018
Messages
915
Likes
288
#3
Maybe you are possibly misunderstanding the issue people have with the RPi sharing Ethernet and USB. Have not seen mention of the other systems you used as I do not think they have this issue of sharing the bus. Many do not seem to have a problem with this, but probably not best practice.
 
Last edited:

Krunok

Major Contributor
Joined
Mar 25, 2018
Messages
4,600
Likes
2,949
Location
Zg, Cro
#4
Maybe you are possibly misunderstanding the issue people have with the RPi sharing Ethernet and USB. Have not seen mention of the other systems you used as I do not think they have this issue of sharing the bus. Many do not seem to have a problem with this, but probably not best practice.
Michelangelo Guarise, the athor of Volumio, confirmed this is a known issue with RPI.
RPI versions prior to 3B+ have the same issues with HiRes playout while 3B+ can play without issuses but has issue when running BruteFIR.

Here is an example of folks complaining on the same issue: https://github.com/raspberrypi/linux/issues/2215
 

Soniclife

Major Contributor
Forum Donor
Joined
Apr 13, 2017
Messages
3,779
Likes
4,207
Location
UK
#5
Thanks, I've long suspected this to be true, nice to see measurements confirm it.
 
OP
March Audio

March Audio

Major Contributor
Manufacturer
Joined
Mar 1, 2016
Messages
6,382
Likes
8,929
Location
Albany Western Australia
Thread Starter #6
Maybe you are possibly misunderstanding the issue people have with the RPi sharing Ethernet and USB. Have not seen mention of the other systems you used as I do not think they have this issue of sharing the bus. Many do not seem to have a problem with this, but probably not best practice.
No I'm fully aware of it but that is a separate issue. What I am referring to here is the idea some have that a special Audiophile tweaked computer/SBC is required to eliminate noise or that it will magically improve sound quality.

It is absolutely correct that the Pi has limited ethernet bandwidth so can run into issues if you want to use very high sample rates or (pointless) DSD. However there are plenty of direct replacement alternative boards available that have gig ethernet not run from the USB bus. Some are shown above. FWIW I have been running a PI 3b with various dacs up to 192kHz with Roon without any issues.
 
Last edited:

rebbiputzmaker

Addicted to Fun and Learning
Joined
Jan 28, 2018
Messages
915
Likes
288
#7
No I'm fully aware of it but that is a separate issue. What I am referring to here is the idea some have that a special Audiophile tweaked computer/SBC is required to eliminate noise or that it will magically improve sound quality.

It is absolutely correct that the Pi has limited ethernet bandwidth so can run into issues if you want to use very high sample rates or (pointless) DSD. However there are plenty of direct replacement alternative boards available that have gig ethernet not run from the USB bus. Some are shown above. FWIW I have been running a PI 3b with various dacs up to 192kHz with Roon without any issues.
Oh OK.
 

Guermantes

Senior Member
Joined
Feb 19, 2018
Messages
424
Likes
451
Location
Brisbane, Australia
#8
Michelangelo Guarise, the athor of Volumio, confirmed this is a known issue with RPI.
RPI versions prior to 3B+ have the same issues with HiRes playout while 3B+ can play without issuses but has issue when running BruteFIR.

Here is an example of folks complaining on the same issue: https://github.com/raspberrypi/linux/issues/2215
I read through that thread with interest because, like @March Audio, I don't have this problem. I haven't attempted running BruteFIR, though.

My setup:
RPI 3B (Previously RPi 2B)
MoOde Audio Player
Topping D30 (I have also used a D10 via a powered USB hub)
Portable hard drive connected via powered USB hub containing music library

I often play back 96 kHz FLAC and DSD64 files. I've also tested 192 kHz FLAC and DSD128 files with no issues. However, I use Wi-Fi to communicate with MoOde, not ethernet.

The problem of interrupts causing audio glitching is not limited to the ARM-based SBCs. I've struggled with the same issue using pro-spec gear on Windows PCs in the past even using high bandwidth Firewire interfaces. Much depended on the hardware and drivers. I often found that video drivers were a culprit, so I'm not surprised if another process in Volumio is causing grief.
 

DownUnderGazza

Active Member
Forum Donor
Joined
Jul 28, 2018
Messages
103
Likes
261
Location
New Zealand
#9

Krunok

Major Contributor
Joined
Mar 25, 2018
Messages
4,600
Likes
2,949
Location
Zg, Cro
#10
I read through that thread with interest because, like @March Audio, I don't have this problem. I haven't attempted running BruteFIR, though.

My setup:
RPI 3B (Previously RPi 2B)
MoOde Audio Player
Topping D30 (I have also used a D10 via a powered USB hub)
Portable hard drive connected via powered USB hub containing music library

I often play back 96 kHz FLAC and DSD64 files. I've also tested 192 kHz FLAC and DSD128 files with no issues. However, I use Wi-Fi to communicate with MoOde, not ethernet.

The problem of interrupts causing audio glitching is not limited to the ARM-based SBCs. I've struggled with the same issue using pro-spec gear on Windows PCs in the past even using high bandwidth Firewire interfaces. Much depended on the hardware and drivers. I often found that video drivers were a culprit, so I'm not surprised if another process in Volumio is causing grief.
If you try to use Ethernet instead of WiFi you would run into dropout problems even without BruteFIR but with WiFi you can play 192kHz content from the network without issues (without BruteFIR).
 

Krunok

Major Contributor
Joined
Mar 25, 2018
Messages
4,600
Likes
2,949
Location
Zg, Cro
#11
Anyone interested in the whole SBC as a music streamer concept, I recommend our scientific advocate over at Archimago's Musings.
This is one of his more recent Myth-debunking / extreme DAC optimisation articles:
http://archimago.blogspot.com/2018/11/musings-raspberry-pi-3-b-touch.html
None of these helps with dropouts when you use USB DAC, pull music from network and use BruteFIR (which does upsampling).

Developer who ported BruteFIR to Volumio platform is running both on Odroid platform without any issues.
 

Ron Texas

Major Contributor
Joined
Jun 10, 2018
Messages
3,801
Likes
4,753
Location
A mysterious place with no name.
#13
I have a low end Intel NUC. It can play back music as well as videos. However, as cheap as this NUC was, a Raspberry Pi with Volumio is even less. In this hobby a lot of money gets thrown around without much benefit, so I am all for low cost solutions which really work. There are also some nice top had DAC's for the RP3. It's a world of choices.
 

Krunok

Major Contributor
Joined
Mar 25, 2018
Messages
4,600
Likes
2,949
Location
Zg, Cro
#14
I run my RP3 with Volumio directly into my DAC over USB.
I have my RP3 connected to my LAN over wired ethernet with wireless turned off.
I've never had a USB dropout... maybe I'm just lucky
Luck has nothing to do with it. You won't have droputs in that scenario unless you try to play HiRes or unless you install BruteFIR.
 

Krunok

Major Contributor
Joined
Mar 25, 2018
Messages
4,600
Likes
2,949
Location
Zg, Cro
#15
I have a low end Intel NUC. It can play back music as well as videos. However, as cheap as this NUC was, a Raspberry Pi with Volumio is even less. In this hobby a lot of money gets thrown around without much benefit, so I am all for low cost solutions which really work. There are also some nice top had DAC's for the RP3. It's a world of choices.
Sure. There's also Volumio for Intel NUC and that combo works flawlessly with HiRes and BruteFIR. I'm not aware of a cheaper platform that can run full blown FIR Digital Room Correction while still acting as a network player.
 

trl

Major Contributor
King of Mods
Joined
Feb 28, 2018
Messages
1,480
Likes
1,552
Location
Iasi, RO
#16
Nice - another myth busted thanks to your effort. :)
Not necessarily, some USB transports are sensitive to the USB's +5V power, like some Modi or even the older ODAC v1, especially to longer path/cables when connecting a 3m long USB cable to the front USB ports (inside the computer there are cheap USB cable pretty close to the RAM chips and SSD drives, also nearby the V-core CPU regulators that are basically powerful SMPS choppers).
 

trl

Major Contributor
King of Mods
Joined
Feb 28, 2018
Messages
1,480
Likes
1,552
Location
Iasi, RO
#17
Oh just for reference this is the sort of thing that can go wrong if you have gnd loops. I artificially created the problem by using the same computer that the measurement system is connected to to drive my DAC1.
How did you created this ground-loop, pretty please? It might be interesting for many to actually visualise the issue (it's like a warning: don't do this...or that). Thank you!
 
OP
March Audio

March Audio

Major Contributor
Manufacturer
Joined
Mar 1, 2016
Messages
6,382
Likes
8,929
Location
Albany Western Australia
Thread Starter #18
How did you created this ground-loop, pretty please? It might be interesting for many to actually visualise the issue (it's like a warning: don't do this...or that). Thank you!
Very simply by having the ADC and the DAC connected to the same USB bus on the laptop. So not necessarily something you will create but I have seen similar with certain usb "cleaners" no names mentioned making packet noise worse.

Always a risk with single ended systems that you get currents flowing in the ground signal conductor.
 
Last edited:

Music1969

Major Contributor
Joined
Feb 19, 2018
Messages
1,734
Likes
789
#19
Oh just for reference this is the sort of thing that can go wrong if you have gnd loops. I artificially created the problem by using the same computer that the measurement system is connected to to drive my DAC1.

8kHz USB packet noise and a bunch of other spuria.
Always a risk with single ended systems that you get currents flowing in the ground signal conductor.
Hi @March Audio ,so it's purely the creation of a ground loop (carrying ground currents) that brings up the 8kHz USB packet noise?

So to have no 8kHz USB packet noise issues, just make sure you have no ground loops?
 
Last edited:
Top Bottom