Install headless Audirvana Studio for Linux on Raspberry Pi 4B with DietPi Linux and a DAC connected over USB-A

[phas02]

Sharing step by step instructions for installing Audirvana Studio on a Raspberry Pi 4B board as I could not find them in this form Hope this is helpful.

It is fairly simple and will most likely work the same for other versions of the Raspberry Pi board. The DAC in this case is a RME ADI-2 that is connected to the Pi with a USB cable to ensure maximum potential for transferring the highest quality of sounds. Audirvana will take care of connecting with Qobuz and then forwarding the bytes to the DAC, which will take care of the conversion to analogue sound. I decided to not use a HAT or anything similar. The Raspberry Pi board should have a minimum of 4GB RAM but for streaming only it might also work with less, just needs to be tried out and monitored. In this case the board has 8GB of RAM which is more than enough.

This might also work with the Raspberry Pi Lite OS as it is also Debian based but I did not try it.

== PREPARE DietPi minimal Linux ==

1. Download and install DietPi on a microSD, as described on their website: How to install DietPi - DietPi.com Docs. Use the xz file directly to flash with balenaEtcher, it unpacks it while doing so.

2. After the validation, the OS (in this case macOS) unmounts the microSD card. Add it back in to change some files before starting the OS on the Raspberry Pi. It is easier to change the files on a desktop than over SSH.

3. For Wifi configuration open dietpi-wifi.txt with a Text Editor, eg. TextEdit on macOS. Then add your Wifi SSID to aWIFI_SSID[0]=‘’ between the ‘’. Add the cleartext Wifi Password to aWIFI_KEY[0]=‘’ between ‘’. If the Wifi uses WPA-PSK, save and close the file. Otherwise change other configurations or add more Wifis according to your Wifi(s) configuration.

4. Open dietpi.txt in the same Text Editor. Change locale and timezone if you want, it is not mission critical. If you want Wifi only set AUTO_SETUP_NET_ETHERNET_ENABLED=0 and AUTO_SETUP_NET_WIFI_ENABLED=1, or another combination based on your preferences. Set AUTO_SETUP_NET_WIFI_COUNTRY_CODE to your country code, also not mission critical. Set AUTO_SETUP_NET_HOSTNAME to something a bit more audiophile, or leave it. AUTO_SETUP_HEADLESS=1 as there won’t be any video streaming or HDMI usage in this case. This is the basics.

5. For login into the system over SSH and then later install Audirvana we will need a root password or a public key. Public key is preferred and much easier to use. Good examples for configuration of SSH on the laptop/ desktop with a public key can be found online. Add the generated pub key part to AUTO_SETUP_SSH_PUBKEY.

6. Set AUTO_SETUP_AUTOMATED=1 to automate installation of necessary software during first start-up this can also be done manually. Comment out AUTO_SETUP_GLOBAL_PASSWORD if SSH with a public key is used. Set SOFTWARE_DISABLE_SSH_PASSWORD_LOGINS=1 also in that case.

7. Change other things according to your own preferences and taste. I am sure there are more things that could be optimised.

8. Safe and close dietpi.txt

9. unmount the microSD and add it to your Raspberry Pi and start it up

== Finalise installation of DietPi OS and install Audirvana Studio ==

1. Download Audirvana audirvana-studio_X.X.X.X_arm64.deb from the Audirama website. It is the Debian one for ARM.

2. Make some tea and wait until dietpi installs und updates itself.

3. Log into the system with ‘ssh root@CHOOSENNAME’ to check on it

4. Upload audirvana-studio_X.X.X.X_arm64.deb with “cat audirvana-studio_X.X.X.X_arm64.deb | ssh root@CHOOSENNAME 'cat - > audirvana-studio_X.X.X.X_arm64.deb’ ”

5. On the Raspberry Pi you should be in /root, the file should be there, check with ‘pwd’ for location, and ‘ls’ if the file is there

6. Install audirvana-studio_X.X.X.X_arm64.deb with ‘apt-get install -f /root/audirvana-studio_X.X.X.X_arm64.deb’, ignore the warnings

7. Install Avahi-Daemon either over dietpi-software or with apt-get. This is necessary to run audirvana studio

8. Configure audirvana studio to start on start-up: /opt/audirvana/studio/setAsService.sh enable

9. Start it: /opt/audirvana/studio/setAsService.sh start

Now Audirvana Studio should be recognised with Audirvana Remote and be ready to be configured. The DAC will need some configs, eg. 0db PCM boost but that is personal taste and depends on the DAC and its configuration.

Sidenote: Yes, there is no ALSA installed or a sound card configured. In this case, this was not necessary to make it work as the sound card was connected over USB.

[jmtennapel]

Hi, thanks for the guide. Some points to consider:

• The WiFi transmitter in the Pi causes a lot of electrical interference on the Pi board, it is not isolated at all. That’s fine for general computational use, but not so fine for audio use. If you can, use Ethernet and leave the WiFi chip off.
• The USB out of the Pi4 is not very adequate for audio use, electrically very noisy which has a detrimental effect on the DAC connected to it. You utilise a RME DAC, which is exceptionally well isolated, so you might get away here.
• To counteract the ‘noisy’ USB, even for a RME DAC, you can either use a USB audio HAT like Allo uses, utilise a better and quieter 5V power supply like iFi creates in different varieties or put a DDC in between the Pi and the DAC, like Mutec or Singxer offer. For the RME you don’t need a reclocker version, but it might be the best quality option and most expensive option for most DACs if you like to use an USB connection.
• There are USB and LAN filters for audio use. One might try those as well, but my experience so far doesn’t favour these options, because it often leads to a bit of a sterile result to my ears. But the reader’s experience might be different.

The charm of DietPi is to have as least amount of services running as needed.

[Sailor]

Archimago measured the output of a simple Raspberry Pi 3 feeding a USB DAC and foubd it to be clean! See

MEASUREMENTS: Raspberry Pi 3 as USB Audio Streamer (with recommended CRAAP...

[jmtennapel]

Not very relevant test method, but the PI3 has always been valued as a better player for audio than the PI4. It has and still is be used by many manufacturers who take a PI board as a base for their equipment.

[Sailor]

Why not relevant?

To me the measurements look fine. But perhaps you see more into the data than me.

[jmtennapel]

From the article:

“As usual, what is important is to check the analogue output of the DAC and make sure there’s no evidence of excess noise or distortion. Also, we’ll have a look at the jitter test measurements.”

Measuring on the analog output port of the DAC says absolutely nothing about the quality of the USB output of any connected device. It tells you how good the DAC is at eliminating noise.

Also, the measurements done (in the table) tell you nothing about the digital output, it only tells you something about the TEAC UD-501 analogue output. That’s measured.

The jitter test is not specified how it is performed.

I bet with 100% confidence that when you swap out the used 5V SMPS that powers the Pi and the Odroid for, let’s say, a LHY 25VA of 5V and you measure again, the measurements won’t show any difference, but you will hear a difference.

The method used is simply not correct, and thus the article has no value, unless you are interested in the TEAC UD-501.

[Agoldnear]

@jmtennapel

The jitter test is not specified how it is performed.

The Dunn J-Test

TN-118_Measuring_Jitter_with_J-Test.pdf

Apparently the measurements do tell you about the digital output… It is plainly evident in the J-Test comparisons, although it is questionable that these differences are audible in the final audition.

An audio-centric SMPS generally will perform better than a generic LPS in any given digital-audio implementation… This has been shown to be true by Benchmark Media and other DACs and power amplifiers.

Obviously, intrinsic noise in a platform design architecture must be taken into consideration… Yes, ultimately it is the analog output that we audition and assess regarding the aesthetic sound-quality. The only true digital-audio “bits” are those encoded on or in the storage media before being read and lifted from the storage media for transmission through the amalgamation of analog electronics topologies, to the D/A circuitry and subsequently the amplifiers and transducers that propagate that energy into the psychoacoustic domain.

The referenced article sheds light on the subject of ‘noise’ on the digital-audio signal (jitter) as it may influence the audible perception of sound-quality in the real-world, using a common reference…