Raspberry Pi 4 sound quality

[HaroldWilkin]

I have Ropieee running Roon on a Raspberry Pi 4 (4GB RAM) with an Allo DigiOne Signature Edition hat into a Chord Mojo and Naim XS. The excellent sound has marginally more detail and improved dynamics than my Pi 3B with the same hat. 

I was hoping to achieve better sound quality using Moode Player. What is surprising and disappointing is the thin sound lacking in midrange and bass from the Pi 4/Allo combination when using the latest Moode Player 5.4 Beta2 instead of Roon. From memory this sounds noticeably worse than the Pi 3/DigiOne Sig. Ed. running Moode 4.

This doesn’t make sense to me unless the Pi 4 is capable of degrading the sound in some way, which is a major concern. Does anyone have comments on Pi 4 sound quality?

[Edward]

I have Ropieee running Roon on a Raspberry Pi 4 (4GB RAM) with an Allo DigiOne Signature Edition hat into a Chord Mojo and Naim XS. The excellent sound has marginally more detail and improved dynamics than my Pi 3B with the same hat. 

I was hoping to achieve better sound quality using Moode Player. What is surprising and disappointing is the thin sound lacking in midrange and bass from the Pi 4/Allo combination when using the latest Moode Player 5.4 Beta2 instead of Roon. From memory this sounds noticeably worse than the Pi 3/DigiOne Sig. Ed. running Moode 4.

This doesn’t make sense to me unless the Pi 4 is capable of degrading the sound in some way, which is a major concern. Does anyone have comments on Pi 4 sound quality?

I'll be commenting soon, I have a DigiOne Sig too as well as Roon. I've used Ropieee, I like it as an end-point. It seems to me that Arch-Linux all by itself sounds better than Debian when other variables are kept the same....for example Ropieee vs Raspbian w/Roon end-point installed. Even MPD via Arch sounds better than MPD via Raspbian. At least to me.

Of course, everything sounds better with my OS tweaks installed, even Ropieee.

[Biston]

I have Ropieee running Roon on a Raspberry Pi 4 (4GB RAM) with an Allo DigiOne Signature Edition hat into a Chord Mojo and Naim XS. The excellent sound has marginally more detail and improved dynamics than my Pi 3B with the same hat. 

I was hoping to achieve better sound quality using Moode Player. What is surprising and disappointing is the thin sound lacking in midrange and bass from the Pi 4/Allo combination when using the latest Moode Player 5.4 Beta2 instead of Roon. From memory this sounds noticeably worse than the Pi 3/DigiOne Sig. Ed. running Moode 4.

This doesn’t make sense to me unless the Pi 4 is capable of degrading the sound in some way, which is a major concern. Does anyone have comments on Pi 4 sound quality?

Surely it is all 1s and 0s? ;-) USB3 on the Pi4 does makes a difference to me. But only in getting the data to the DAC with less dropouts in some .dsf files. And that's likely because, to my understanding, the Pi3 runs ethernet and USB2 off the same interface whereas the Pi4 has two separate ones, and to my recent learning, there is some sort of issue with the ethernet connection on the Pi3. [For all I know it might still be there on the Pi4!] I do notice a substantial speed boost [as you might expect] when copying files to the USB3 drive attached to the USB3 interface, when running the Pi4 off ethernet. Versus the Pi3B+ running off ethernet and USB2.

When I say substantial, I very much doubt my router is pushing out more than 400Mbps to the Pi4 via ethernet, but a 2GB file that might have previously taken 10 minutes or so to be copied to the USB3 attached drive on USB2 on the Pi3B+ is now getting there in about 1 minute or so - barely much more time than having the USB drive connected directly to the Mac in typical day-to-day use [for me].

[Biston]

Disk speed test using G-Tech Slim 500MB external 5400. USB3. Formatted as ext4. Connected to USB3 on Pi-4.


Using Blackmagic disk speed test:

9MB/s read and write when Pi-4 is connected via WiFi only [In the U.K., both the Pi3B+ and Pi4 max out at 150Mbps on 5GHz AC WiFi.]

50MB/s read and write when Pi-4 is connected via ethernet. [About 400Mbps - which is what I understand the Airport Extreme puts out via ethernet]

85MB/s read and write when disk is connected directly to Mac.

When I shifted over to Pi I thought I could get away with keeping HFS+ on the external disk connected to the Pi. After numerous inabilities to write to the disk via the Mac, and read permissions issues, reformatting as ext4 has stabilised things.

[Miss Sissy Princess]

Disk speed test using G-Tech Slim 500MB external 5400. USB3. Formatted as ext4. Connected to USB3 on Pi-4.
...

So what you're really measuring is the performance of the Pi-4 as a NAS for your Mac, both on WiFi and hardwired, not its USB 3 disk interface speed, right?

If the drive with the music is directly connected to the Pi-4 by USB3, how the Pi-4 performs as a NAS is immaterial to the sound quality (and this is the Sound quality sub-forum).  What matters there is the local hard drive performance of the Pi-4.

If you do want to evaluate its performance potential in a NAS role, then adding in bandwidth-limited devices like the Airport Extreme just muddies the waters.  Directly connect the Pi-4 to your Mac, first via Ethernet, then via 5GHz WiFi, and finally by 2.4GHz WiFi.  Run the benchmarks in each configuration.  Don't throw the Airport Extreme (and all of your other network traffic) into the mix.

Why did you want the Pi-4 volume to be formatted as HFS+?

[Biston]

Disk speed test using G-Tech Slim 500MB external 5400. USB3. Formatted as ext4. Connected to USB3 on Pi-4.
...

So what you're really measuring is the performance of the Pi-4 as a NAS for your Mac, both on WiFi and hardwired, not its USB 3 disk interface speed, right?

If the drive with the music is directly connected to the Pi-4 by USB3, how the Pi-4 performs as a NAS is immaterial to the sound quality (and this is the Sound quality sub-forum).  What matters there is the local hard drive performance of the Pi-4.

If you do want to evaluate its performance potential in a NAS role, then adding in bandwidth-limited devices like the Airport Extreme just muddies the waters.  Directly connect the Pi-4 to your Mac, first via Ethernet, then via 5GHz WiFi, and finally by 2.4GHz WiFi.  Run the benchmarks in each configuration.  Don't throw the Airport Extreme (and all of your other network traffic) into the mix.

Why did you want the Pi-4 volume to be formatted as HFS+?

Hi,

It was a 'real world' test due to having three Pi4s streaming music. It was not a benchmark test.

Three Pi4s accessing the same music directory, plus a Mac uploading a file to the music directory, for example. The Pi3+B could not handle that due to limitations of USB2 [stuttering, sound quality etc.,] . The Pi4s [or one Pi4] with USB3 seem to be able to.

The Pi-4 volume [I presume you mean the external disk?] had been formatted as HFS+ for a couple of years and I did not want to reformat it. Call me lazy and all that... but I was just looking for a simple streaming solution to replace four Airport Expresses connected to four DACs.

I understand the rest of your [might I say slightly condescending...] post. Might you be able to advise me on: how I can get the Pi3 or Pi4 to output more than 150Mbps in either 2.4Ghz or 5Ghz mode? Because I can guarantee without running any test that is the most I am going to be getting from the USB3 drive via WiFi... circa 9MB/s.. :-)

[Miss Sissy Princess]

Disk speed test using G-Tech Slim 500MB external 5400. USB3. Formatted as ext4. Connected to USB3 on Pi-4.


Using Blackmagic disk speed test:

9MB/s read and write when Pi-4 is connected via WiFi only [In the U.K., both the Pi3B+ and Pi4 max out at 150Mbps on 5GHz AC WiFi.]

50MB/s read and write when Pi-4 is connected via ethernet. [About 400Mbps - which is what I understand the Airport Extreme puts out via ethernet]

85MB/s read and write when disk is connected directly to Mac.

When I shifted over to Pi I thought I could get away with keeping HFS+ on the external disk connected to the Pi. After numerous inabilities to write to the disk via the Mac, and read permissions issues, reformatting as ext4 has stabilised things.

Disk speed test using G-Tech Slim 500MB external 5400. USB3. Formatted as ext4. Connected to USB3 on Pi-4.
...

So what you're really measuring is the performance of the Pi-4 as a NAS for your Mac, both on WiFi and hardwired, not its USB 3 disk interface speed, right?

If the drive with the music is directly connected to the Pi-4 by USB3, how the Pi-4 performs as a NAS is immaterial to the sound quality (and this is the Sound quality sub-forum).  What matters there is the local hard drive performance of the Pi-4.

If you do want to evaluate its performance potential in a NAS role, then adding in bandwidth-limited devices like the Airport Extreme just muddies the waters.  Directly connect the Pi-4 to your Mac, first via Ethernet, then via 5GHz WiFi, and finally by 2.4GHz WiFi.  Run the benchmarks in each configuration.  Don't throw the Airport Extreme (and all of your other network traffic) into the mix.

Why did you want the Pi-4 volume to be formatted as HFS+?

Hi,

It was a 'real world' test due to having three Pi4s streaming music. It was not a benchmark test.

Three Pi4s accessing the same music directory, plus a Mac uploading a file to the music directory, for example. The Pi3+B could not handle that due to limitations of USB2 [stuttering, sound quality etc.,] . The Pi4s [or one Pi4] with USB3 seem to be able to.

The Pi-4 volume [I presume you mean the external disk?] had been formatted as HFS+ for a couple of years and I did not want to reformat it. Call me lazy and all that... but I was just looking for a simple streaming solution to replace four Airport Expresses connected to four DACs.

I understand the rest of your [might I say slightly condescending...] post. Might you be able to advise me on: how I can get the Pi3 or Pi4 to output more than 150Mbps in either 2.4Ghz or 5Ghz mode? Because I can guarantee without running any test that is the most I am going to be getting from the USB3 drive via WiFi... circa 9MB/s.. :-)

Hello.  I understand better what your goal was and what you were testing.  Thank you for taking the time to explain.  

I did not intend to be condescending; after about 35 years in engineering dealing primarily with other engineers, my social skills have become somewhat atrophied.  

The best WiFi benchmark I've seen is about 11MB/s, and I understand that the WiFi shares the bus with USB 2.0 so that's probably a serious bottleneck. The only options I can think of are external, perhaps a USB 3.0 WiFi dongle, which will probably cost as much or more than the Pi-4, or a router configured as a WiFi bridge.

[TheOldPresbyope]

Check out https://www.raspberrypi.org/magpi/raspbe...enchmarks/ which includes both Ethernet and Wifi numbers for the RPi4B and other recent models.

Note that a "real world" situation involving multiple WiFi clients communicating with each other and the external Internet via an access point will likely give much lower throughput because of the way WiFi works (and this before any consideration of your real-world environment---distances of separation, intervening materials, EMI from inside and outside your house, etc.). As well, the performance of your NAS will depend on its specific configuration (e.g., how it's tuned). Google is your friend if you want to pursue these subjects.

It's just one man's opinion, but I don't think setting up a NAS to service clients via WiFi is a good idea. In my case, I already installed 2 years ago an Odroid HC1-based NAS with an SSD hung on its SATA interface and connected to my access point/router via Ethernet. It's running Open Media Vault with lots of plugins installed. My players communicate with the access point via 5GHz WiFi for all those which support it and 2.4GHz to those which don't. There's lots of other WiFi traffic in the house too (Roku, laptops, smart phones, etc.). 

My only measure of performance is whether my several moOde players can play tracks simultaneously without interruptions. In this sense, my server works fine. I would not have saved any money with an RPi4B solution had it been available at the time. If I had to do it now, I'd likely start with the newer Odroid HC2 but there's no accounting for taste

And, seriously, how about moving this thread out of the Sound Quality section?

Regards,
Kent

[Tim Curtis]

Disk speed test using G-Tech Slim 500MB external 5400. USB3. Formatted as ext4. Connected to USB3 on Pi-4.


Using Blackmagic disk speed test:

9MB/s read and write when Pi-4 is connected via WiFi only [In the U.K., both the Pi3B+ and Pi4 max out at 150Mbps on 5GHz AC WiFi.]

50MB/s read and write when Pi-4 is connected via ethernet. [About 400Mbps - which is what I understand the Airport Extreme puts out via ethernet]

85MB/s read and write when disk is connected directly to Mac.

When I shifted over to Pi I thought I could get away with keeping HFS+ on the external disk connected to the Pi. After numerous inabilities to write to the disk via the Mac, and read permissions issues, reformatting as ext4 has stabilised things.

Disk speed test using G-Tech Slim 500MB external 5400. USB3. Formatted as ext4. Connected to USB3 on Pi-4.
...

So what you're really measuring is the performance of the Pi-4 as a NAS for your Mac, both on WiFi and hardwired, not its USB 3 disk interface speed, right?

If the drive with the music is directly connected to the Pi-4 by USB3, how the Pi-4 performs as a NAS is immaterial to the sound quality (and this is the Sound quality sub-forum).  What matters there is the local hard drive performance of the Pi-4.

If you do want to evaluate its performance potential in a NAS role, then adding in bandwidth-limited devices like the Airport Extreme just muddies the waters.  Directly connect the Pi-4 to your Mac, first via Ethernet, then via 5GHz WiFi, and finally by 2.4GHz WiFi.  Run the benchmarks in each configuration.  Don't throw the Airport Extreme (and all of your other network traffic) into the mix.

Why did you want the Pi-4 volume to be formatted as HFS+?

Hi,

It was a 'real world' test due to having three Pi4s streaming music. It was not a benchmark test.

Three Pi4s accessing the same music directory, plus a Mac uploading a file to the music directory, for example. The Pi3+B could not handle that due to limitations of USB2 [stuttering, sound quality etc.,] . The Pi4s [or one Pi4] with USB3 seem to be able to.

The Pi-4 volume [I presume you mean the external disk?] had been formatted as HFS+ for a couple of years and I did not want to reformat it. Call me lazy and all that... but I was just looking for a simple streaming solution to replace four Airport Expresses connected to four DACs.

I understand the rest of your [might I say slightly condescending...] post. Might you be able to advise me on: how I can get the Pi3 or Pi4 to output more than 150Mbps in either 2.4Ghz or 5Ghz mode? Because I can guarantee without running any test that is the most I am going to be getting from the USB3 drive via WiFi... circa 9MB/s.. :-)

Hello.  I understand better what your goal was and what you were testing.  Thank you for taking the time to explain.  

I did not intend to be condescending; after about 35 years in engineering dealing primarily with other engineers, my social skills have become somewhat atrophied.  

The best WiFi benchmark I've seen is about 11MB/s, and I understand that the WiFi shares the bus with USB 2.0 so that's probably a serious bottleneck. The only options I can think of are external, perhaps a USB 3.0 WiFi dongle, which will probably cost as much or more than the Pi-4, or a router configured as a WiFi bridge.

Integrated WiFi/BT chip connects to the SoC via the SDIO and UART interfaces and is thus separate from the USB bus.

[Miss Sissy Princess]

Integrated WiFi/BT chip connects to the SoC via the SDIO and UART interfaces and is thus separate from the USB bus.

That's what I get for trusting a site called notenoughtech.com, which wrote: "The WiFi shares the bus with USB 2.0 so I’m expecting some serious bottlenecks there."

https://notenoughtech.com/raspberry-pi/2...peed-test/

Thanks for setting the record straight.