A simple script used to configure a docker-compose stack for Tidal Connect.
Please see the Goal
Please note that support goal is limited to cover running costs for subscriptions to music services.
If you use a Raspberry Pi 5, you might encounter an issue that shows with the following content in the container logs:
/app/ifi-tidal-release/bin/tidal_connect_application: error while loading shared libraries: libsystemd.so.0: ELF load command alignment not page-aligned
This issue can be solved by editing the file /boot/firmware/config.txt using the following command:
sudo nano /boot/firmware/config.txt
Add the following line in the all section:
kernel=kernel8.img
You will need to reboot, then restart the container.
You can now build a custom image, which will be based on debian:bookmark-slim.
In order to build the image, change the current directory to build, then execute:
docker build . -t my/tidal-connect
or execute the provided local-build.sh using
./local-build.sh
The resulting image will be tagged as my/tidal-connect. You will have to set the variable TIDAL_CONNECT_IMAGE to this image name if you want to use it instead of the default docker image.
Note that the image you build this way will not contain the tidal binaries. Refer to the following readme files in the bin, certificate, lib and lib-arm-linux-gnueabihf.
At the end of July 2024, Tidal has removed all the MQA content. This particular implementation of Tidal Connect could play hi-res files only up to 24/48 and for MQA content: in the latter case, an unfolding to 24/88 or 24/96 was implemented by the application itself, while the further processing was delegated to a MQA capable DAC.
As far as I know, content above 16/44 available as HI_RES at 24/44 or 24/48 is currently inexistent on Tidal, so I guess we are now stuck to standard redbook resolution (16/44).
I don't consider this limitation a show-stopper. Some people might miss MQA, and that is totally fine for me. On the bright side, at least from now on, we don't have to worry about this controversial format to be detrimental to the audio quality, as some people said.
All the hires content from Tidal is now available only via MPEG-DASH. Some of the alternatives mentioned below support this streaming schema and will be able to play hi-res content up to 24/192.
I am trying to document all the configurations I have encountered with my own dacs and through your tickets. Look here and here.
Look at the samples and userconfig directories. They might contain a good file for your dac. It something for you isn't already there, open an issue, and we will create a sample file for your dac, which will also help other users.
Also, consider the new mopidy-tidal alternative. Some things are not on par with the full-fledged web player or mobile app, but the developer is excellent, and I believe things are going to improve.
Visit this page for an example configuration using my docker image for mopidy. You will also find the references to the underlying projects.
This repository does not contain the Tidal Connect software. There is only a docker-compose.yaml file, a configurator script, and a modified entrypoint which is just used for selecting the desired audio output more easily and reliably.
All the hard work has been done by the owner of the repository mentioned in the References section and in the other repositories from which his one has been forked. I am just trying to provide a way to run their container more easily in certain environments (where the index of your audio device is not the same on every restart).
This solution is not and will probably never be supported by Tidal. It might stop working at any time.
If you need to have a supported solution, look at Tidal Connect enabled products. A cheap one is the ubiquitous Wiim Mini/Pro/Pro+/Amp/AmpPro/Ultra. Another option is represented by Google Chromecast/Chromecast Audio devices. Also, if you use Apple devices, you can already stream to AirPlay-enabled devices.
Even your current TV might already be used as an endpoint for Tidal via the embedded Chromecast functionality, or via AirPlay.
Alternatively, if you are not scared of some DIY, you might want to create an upnp/dlna renderer, maybe with upmpdcli (you might use my docker image for upmpdcli) and mpd (you might use my docker image for mpd), and then use some Android app like BubbleUPnP or mConnect Lite (this one is also on iOS/iPadOS), connect those apps to Tidal and then stream to your upnp/dlna renderer.
Starting late November 2023, BubbleUPnP supports hires flac, see this post on Reddit. This is something that, from what I know, is not supported by this Tidal Connect application, and will probably never be.
Another solution might be my Tidal Plugin for Upmpdcli but again, there will be no support from Tidal. This solution works well with Tidal HiFi or with HiFi+ with MQA DACs.
This solution supports native HiRes flac playback (HI_RES_LOSSLESS so up to 24/192) to and mpd/upmpdcli player. If you configure it to play only up to HI_RES (so up to 24/48 and unfolding on a DAC), it will work on pretty much any UPnP renderer.
See here for a configuration for the HI_RES_LOSSLESS enabled media server.
Another alternative, if you have a HiFi subscription and/or are ok with being limited to standard resolution or mqa-encoded tracks, is using Logitech Media Server and Squeezelite, for which you might use my docker image for squeezelite.
Even considering that after February 2024 the mysqueezebox online services will be shut down, there is a new, very promising, and I'd say already very usable plugin here by community heroes michaelherger and philippe44. This solution does not currently support Tidal HiRes while it supports MQA encoded streams: there will be no software-based unfolding, but a MQA enabled DAC should do that instead.
There is a new plugin for Mopidy called Mopidy-Tidal, which enabled Mopidy to stream for Tidal with full support for HiRes streams.
I have prepared a docker container image for this solution here with a strong focus on this plugin especially.
A simple configuration can be found here.
A commercial solution, Audirvana Studio, allows you to stream from Tidal to UPnP renderers.
I maintain docker images for Audirvana Studio and here you can find a simple configuration for using mpd and upmpdcli as a player for Audirvana.
Roon is another commercial solution which can allow you to stream to Tidal.
You can build an endpoint using Roon Bridge. A dockerized solution is here.
The contents of this repository entirely rely on this repository from GitHub user TonyTrump. A big thank you to the author for the great work.
It will also use his docker image.
I created this repository because it was very inconvenient for me to figure out the string to put for the PLAYBACK_DEVICE variable. I even failed for one of my DAC, I don't know honestly why, but as far as I understand, the ifi-pa-devs-get for some reason refuses to see the device. The most relevant issue I found about this issue is this, but it does not seems to have an easy solution.
Additionally, even if you can specify the correct string for your DAC, I found that the resulting configuration would be error-prone as the string reports both the device name and the device index. AFAIK the index can change across restarts, so outside of a known and controlled setup (which is probably represented by the I*i devices) this situation can an will lead to errors or unwanted configurations.
Keep in mind that the audio device index can also be changed because one time the (usb) device is powered on during boot, and another time it isn't.
This is my experience, unless I am missing something obvious. If so, I will be glad to be corrected.
The work in this repository consists in slightly altering the container startup phase (the entrypoint.sh file), in such a way that a custom /etc/asound.conf is created with (hopefully) the correct device index, regardless of the order of the audio devices, which can vary across restarts. The underlying application then always uses the default audio device.
You will need a single-board computer (or anyway, a computer) with an armhf architecture (arm64 should work as well), running docker and docker-compose.
A Raspberry Pi 3/4 will work. If you plan to use a usb dac and hi-res audio, consider at least using a Pi 3b+ or, even better, a Pi 4b.
I am also running this on as Asus Tinkerboard. With this hardware, my suggestion is to not allow it to scale down the cpu frequency too much, or you might experience every kind of crackling noises along with what will remain of your music. In my experience, I am having good result if I set the minimum frequency at least about 600MHz, but, of course, YMMV.
Docker is a prerequisite. On debian and derived distributions (this includes Raspberry Pi OS, DietPi, Moode Audio, Volumio), we can install the necessary packages using the following commands:
sudo apt-get update
sudo apt-get install docker.io docker-compose
sudo usermod -a -G docker $USER
newgrp docker
The third line command adds the current user to the docker group. This is not mandatory; if you choose to skip this step, you might need to execute docker-compose commands by prepending sudo.
The last line will login to the group named docker. The previous command (usermod) is of course required in order for the newgrp instruction to be successful.
You need to clone the repository. Make sure that git is installed using the following command on debian and derived distributions (again, this includes Raspberry Pi OS, DietPi, Moode Audio, Volumio):
sudo apt-get update
sudo apt-get install -y git
Move to the your home directory and clone the repository using the commands:
cd
git clone https://github.com/GioF71/tidal-connect.git
If you just downloaded the repository, you can skip this step.
If you previously cloned the repository, it might have been updated in the meantime. Move to the directory and pull the changes:
cd $HOME/tidal-connect
git config pull.rebase false
git pull
From the repository directory, just run the configure.sh bash script, specifying the following parameters:
| PARAM | DESCRIPTION | VARIABLE |
|---|---|---|
| -n | Sound card name (e.g. DAC), if not specified and also card index isn't, sysdefault is used |
CARD_NAME |
| -i | Sound card index, not recommended: if not specified and also card name isn't, sysdefault is used |
CARD_INDEX |
| -d | Sound card device, optional | CARD_DEVICE |
| -s | Card format, optional (S32_LE, S16_LE, etc) |
CARD_FORMAT |
| -l | Enables softvolume, defaults to yes |
ENABLE_SOFTVOLUME |
| -g | Enables generated tone, defaults to yes |
ENABLE_GENERATED_TONE |
| -f | Friendly name, defaults to TIDAL connect |
FRIENDLY_NAME |
| -m | Model name, defaults to Audio Streamer |
MODEL_NAME |
| -c | MQA Codec, defaults to false |
MQA_CODEC |
| -p | MQA Passthrough, defaults to false |
MQA_PASSTHROUGH |
| -r | Asound file prefix | ASOUND_FILE_PREFIX |
| -o | Force playback device | FORCE_PLAYBACK_DEVICE |
| -a | Name of the virtual sound card in the generated asound.conf file | CREATED_ASOUND_CARD_NAME |
| -t | Sleep time in seconds be, defaults to 3 |
SLEEP_TIME_SEC |
| -w | Restart wait time in seconds, defaults to 10 |
RESTART_WAIT_SEC |
| -e | Custom client id, defaults to empty string | CLIENT_ID |
| -b | log level, defaults to 3 |
LOG_LEVEL |
| -h | override default certificate path | CERTIFICATE_PATH |
| -j | Disable control app if set to 1, defaults to 0 |
DISABLE_CONTROL_APP |
| -v | DNS Server list, defaults to 8.8.8.8 8.8.4.4 (Google's DNS servers) |
DNS_SERVER_LIST |
| -k | Disable app security, defaults for false | DISABLE_APP_SECURITY |
| -q | Disable web security, defaults to true | DISABLE_WEB_SECURITY |
I recommend to use the -n parameter instead of -i, because the index of the devices might change across restarts.
If you already used the configure.sh command and you are experiencing issues (because of the card has changed its index), you can run the command again. In the latest version, the card index is calculated during the container startup phase and hopefully there will not be any need to use configure.sh again unless you change the audio device you want to use.
Configure for sound card named "DAC", using friendly name "Aune S6 USB DAC" and model name "Asus Tinkerboard":
cd $HOME/tidal-connect
bash configure.sh -n DAC -f "Aune S6 USB DAC" -m "Asus Tinkerboard"
If no error is reported, you will find a new (or updated) .env file.
If you find a spurious .asound.conf file there, it probably was generated by a previous version of the configure.sh script, and you can safely delete it.
So now you can run the docker-compose.yaml as usual:
cd $HOME/tidal-connect
docker-compose up -d
The container can be entirely configured using the environment variables listed on the following table:
| VARIABLE | DESCRIPTION |
|---|---|
| CARD_NAME | Alsa name of the audio card. Example for xmos dac might be DAC while e.g. it is D10 for a Topping D10 |
| CARD_INDEX | Alsa index of the audio card |
| CARD_DEVICE | Audio device, optional |
| CARD_FORMAT | Audio format, optional (S32_LE, S16_LE, etc) |
| FORCE_PLAYBACK_DEVICE | If set and if there is an asound.conf provided or selected via a prefix in userconfig, this will be the playback device |
| FRIENDLY_NAME | Friendly name of the device, will be shown on Tidal Apps. Defaults to TIDAL connect. |
| ASOUND_FILE_PREFIX | Search asound.conf with this prefix, a . is used as separator |
| CREATED_ASOUND_CARD_NAME | When creating asound.conf, use this as the declared device name |
| ENABLE_SOFTVOLUME | Generate a configuration with softvolume if set to yes, defaults to no |
| ENABLE_GENERATED_TONE | Generates a tone before starting the app, defaults to yes |
| MODEL_NAME | Model name of the device. Defaults to Audio Streamer. |
| MQA_CODEC | Can't comment a lot on this, defaults to false. |
| MQA_PASSTHROUGH | Can't comment a lot on this, defaults to false. |
| SLEEP_TIME_SEC | Sleep time before starting the real app, after starting tmux. Defaults to 3. |
| RESTART_ON_FAIL | Enables auto restart (see issue #16), defaults to 1 (which means restart is enabled). |
| RESTART_WAIT_SEC | Wait time in seconds before trying restart (see RESTART_ON_FAIL), defaults to 10. |
| CLIENT_ID | Set custom client id, defaults to an empty string |
| LOG_LEVEL | Application log level, defaults to 3 |
| CERTIFICATE_PATH | Override default certificate path |
| DISABLE_CONTROL_APP | Disable control app if set to 1, defaults to 0 |
| DISABLE_APP_SECURITY | Defaults to false |
| DISABLE_WEB_SECURITY | Defaults to true |
| DNS_SERVER_LIST | The DNS serves to be used, defaults to 8.8.8.8 8.8.4.4 (Google's DNS servers). |
| TIDAL_CONNECT_IMAGE | Alternative image |
Please note that if both CARD_NAME and CARD_INDEX are specified, only CARD_NAME will be considered.
Also, if both CARD_NAME and CARD_INDEX are not specified, sysdefault (the system default audio device) will be used.
Here is the list of volumes:
| VOLUME | DESCRIPTION |
|---|---|
| /userconfig | Might contain user-provided configurations. |
If you put an asound.conf file in the userconfig directory of the repository, this file will be copied to /etc/asound.conf, so you can implement your custom configuration. You might need to set the variable FORCE_PLAYBACK_DEVICE according to the contents of the provided file, unless the device you want to play to is named custom.
It is possible to use this solution for easy installation of Tidal Connect on Moode Audio and Volumio.
It is required to have a ssh connection to the Moode/Volumio audio box. In order to enable ssh on Volumio, refer to this page.
Those two platforms do not ship docker out of the box (unsurprisingly), so docker installation is required. See Docker Installation earlier in this page.
If you have just installed docker with the previous commands, it is probably a good idea to logoff your current ssh session, then log back in. Otherwise, just open a ssh connection to your box.
We need to configure the audio output you want to use for Tidal Connect.
If your device only has one output, or if that output is also configured as the default output, no configuration might be needed other than the Friendly and Model name.
On one of my boxes, I have a Hifiberry Dac+ Pro Hat, so when I use the command:
cat /proc/asound/cards
I get:
pi@moode-living:~/git/tidal-connect $ cat /proc/asound/cards
0 [sndrpihifiberry]: HifiberryDacp - snd_rpi_hifiberry_dacplus
snd_rpi_hifiberry_dacplus
Great, the operating system has just disabled the onboard audio and set the Hifiberry HAT as the default card.
So let's configure Tidal Connect:
cd $HOME/tidal-connect
./configure.sh -f "Living Aux1" -m "Raspberry Pi"
We are not specifying anything (not the card index and neither the name) because there is only one output available.
Replace the first and second strings to your liking. Once configured, start the service as usual:
cd $HOME/tidal-connect
docker-compose up -d
On another one of my boxes, I have an usb dac connected, so when I use the command:
cat /proc/asound/cards
I get:
moode@moode:~ $ cat /proc/asound/cards
0 [b1 ]: bcm2835_hdmi - bcm2835 HDMI 1
bcm2835 HDMI 1
1 [Headphones ]: bcm2835_headpho - bcm2835 Headphones
bcm2835 Headphones
2 [X20 ]: USB-Audio - XMOS USB Audio 2.0
XMOS XMOS USB Audio 2.0 at usb-0000:01:00.0-1.2, high speed
So in this setup, the operating system has not disabled the onboard audio. Even if you have configured Moode so that is will use the USB DAC, this might not be enough for Tidal Connect to automatically select that card.
The safest way (at least IMO) is to use the string that identifies the dac as card name:
So let's configure Tidal Connect:
cd $HOME/tidal-connect
./configure.sh -n "X20" -f "Desktop" -m "Raspberry Pi"
Replace the second and third strings to your liking. Once configured, start the service as usual:
cd $HOME/tidal-connect
docker-compose up -d
Tidal Connect will access exclusively your audio device if you select it in your desktop/mobile Tidal App.
If you want to play to the selected audio device from other sources, you will need to disconnect from the Tidal app you are using.
Similarly, you won't be able to play to a device if this device is already playing something from another source.
In order to check if a device is playing, follow the instructions in the next paragraphs.
Say you have selected your card by its index (not recommended), and that the index is 1, you can execute the following command:
watch cat /proc/asound/card1/pcm0p/sub0/hw_params
If this does not say "closed", it means that the audio device is being currently used.
Say you have selected your card by its name, and that the name is D10, you can execute the following command:
watch cat /proc/asound/D10/pcm0p/sub0/hw_params
If this does not say "closed", it means that the audio device is being currently used.
This solution will try to configure a software volume for your Tidal Connect device, so that you should be able to control the software volume from the Tidal App on any platform, without touching the hardware volume, which is something you might find useful if you have other players running on the same audio device, e.g. mpd, librespot, squeezelite, etc.
However, software volume will be enabled only if your audio device does not already have a control named Master. You can verify if your device already has such control using the following command, for the audio device at index 0:
amixer -c 0 | grep \'Master\'
The desired output is no output, as this would mean that there is no Master control already.
If, instead, you get a line like this one, you are (almost) out of luck:
Simple mixer control 'Master',0
Even under this condition, the solution will create the SoftVolume control naming it SoftVolume. This will allow the Tidal Apps (at least the current Android version) to control the hardware volume using its slider. This means that changing the volume from the Tidal App will affect all the other players running on the same audio card.
If you are in this situation (so if you already have a Master control defined in your audio device) and you don't want to use hardware volume, consider disabling software volume by setting ENABLE_SOFTVOLUME to no in your .env file, or acting on the correspondent switch of the configure.sh script.
See this issue for some context information.
Special "Thank You" to zamiere for providing useful information.
Remember that, should you change something to your Moode/Volumio setup, maybe replacing the audio-hat with an USB DAC, you will most likely need to reconfigure Tidal Connect accordingly.
Please be aware that this solution will not be completely equivalent to the built-in premium feature of Volumio. That solution (probably) allows the attached (touch) display to show the currently playing song, while this solution for sure does not allow that or any other related features.
Tidal connect won't work if your system does not support ipv6. See this issue.
Afaik, there is no solution or workaround available other than, somehow, enabling ipv6.
On DietPi (which I am running on my Asus Tinkerboard), you might need to enable avahi-daemon, if this is not enabled yet.
You might find the following on the logs:
[tisoc] [error] [avahiImpl.cpp:358] avahi_client_new() FAILED: Daemon not running
This can be fixed by installing the avahi-daemon. It is not installed by default on DietPi, so we can installing it with this command:
sudo apt-get install avahi-daemon
An already started tidal-connect container should start working immediately, at least that is what happened with my setup.
| Date | Comment |
|---|---|
| 2024-11-20 | Add support for own docker image |
| 2024-10-15 | Add newgrp instruction so a logoff/logon is not strictly required. |
| 2024-09-23 | Add sample for FiiO K11 (see issue #202) |
| 2024-09-23 | Corrected markdown table in userconfig/README.md |
| 2024-08-06 | Add sample for Chord Qutest (see issue #186) |
| 2024-08-02 | Add alternatives: Mopidy-Tidal, Audirvana (paid) and Roon (paid) |
| 2024-07-17 | Add documentation about software volume |
| 2024-07-17 | Create SoftMaster control when a Master control already exists |
| 2024-07-08 | Add sample env files for Fosi Audio DS1 headphone amp |
| 2024-07-08 | Fixed bug which would cause ENABLE_SOFTVOLUME to not be honored when set to e.g. "no" |
| 2024-07-08 | Add pi-headphones.asound.conf for RPI Headphone jack |
| 2024-07-01 | Add asound.conf for Apple USB dongle, see #187 |
| 2024-07-01 | Add support for 48kHz tone, see #187 |
| 2024-06-17 | Add presets for hdmi on raspberry pi |
| 2024-04-08 | Add sample config file for hifiberry dac plus |
| 2024-04-06 | Reverted removal of optional version property in docker-compose file (#160) |
| 2024-04-02 | Remove optional version property in docker-compose file |
| 2024-04-02 | Add sample config file for Yulong D200 USB DAC |
| 2024-03-27 | Add sample config file for hifiberry digi+ pro |
| 2024-03-13 | Support for disabling control app |
| 2024-03-13 | Support for overriding certificate path |
| 2024-03-13 | Support for log level |
| 2024-03-13 | Support for custom clientid |
| 2024-03-04 | Lowered default for RESTART_WAIT_SEC to 10 |
| 2024-03-04 | Add logs.sh, restart.sh and restart-watch.sh scripts |
| 2024-03-04 | Add -w for RESTART_WAIT_SEC to configure.sh |
| 2024-03-04 | Corrected configure.sh (sequence of opts) |
| 2024-03-04 | Add aune-s6 dac configuration with softvolume |
| 2024-03-03 | Fix software volume, avoid to give up for a device which only contains Master (see #136) |
| 2024-02-22 | Add support for software volume |
| 2024-02-22 | Add support for configuration self-test using a generated tone |
| 2024-02-17 | Add support for newer variables in configure.sh, see #108 |
| 2024-01-30 | Add support for ASOUND_FILE_PREFIX, see #101 |
| 2024-01-26 | Assume custom playback device when asound.conf is provided, see #90 |
| 2024-01-25 | Support custom asound.conf, support forced PLAYBACK device, see #80 |
| 2024-01-25 | Revert latest change, see (#78) |
| 2024-01-24 | Always create sysdefault in asound.conf and log device names, see #76 |
| 2024-01-23 | Add support for optional card device (CARD_DEVICE) and format (CARD_FORMAT), see #72 |
| 2023-09-12 | Clarify how to install on Volumio, see issue #29 |
| 2023-09-04 | Allow default audio card selection, see issue #22 |
| 2023-07-18 | Allow user-specified dns server(s), see issue #13 |
| 2023-07-07 | Fixed asound.conf generation from card index, see issue #2 |
| 2023-06-02 | First unfolding seems to be working |
| 2023-06-02 | Some effort to avoid resampling |
| 2023-06-02 | MQA passthrough defaults to false |
| 2023-06-01 | Using hardware mode |
| 2023-06-01 | Resolve device name at container startup |
| 2023-05-29 | First working version |