There are some periods where YouTube make changes frequently so that e.g. FreeTube stops working for some time, but for the most part it works reliably well. I would say it provides a much better convenience than watching on youtube.com logged out, as you have profiles, subscriptions, playlist and history. Including adblock, sponsorblock and dearrow, and granular control over what to show or not (comments, shorts, live etc.).

Freetube exists for Android also.

Syncthing is your friend. Freetube stores playlists, history, settings and subscriptions as .db-files which you can sync between devices. Android version also allows access to these files if enabled in settings.

This is my solution also. I listen to audio books on my way to work, and read on an ebook-reader in the evening. Can be tricky to sync when the chapter structure is non-traditional though (e.g. Discworld).

My laptop came preinstalled with Linux :)

Depending on how diverse your taste is, you could always try to branch out to something outside of “similar artists”. Just look up genre names and start checking them out. If you find something you like, you can use the same " similar artists" approach on an entirely new search space.

Hm, the Music Assistant at least does not quite accomplish what I hope and it seems to rely on services such as Spotify or YT Music to be integrated. I couldn’t quite evaluate the LinkPlay-solution, but his comment on SD card corruption with RPis made me a bit worried for the balenaSound approach. I guess there’s a lot of write operations in such a setup, that can easily corrupt the SD cards. I wonder how often they kept failing for him - maybe it’ll end up being some sort of a “subscription fee” 😅

My wife says no more toys at the moment, but if I were to implement this, I’d probably pick up one of those Up2Streams for each room and try out the LinkPlay integration.

Then you have something to put on your list for Christmas, if that is something you celebrate :)

As long as you are connected, I think network adb will stay active but if you leave your network f.e., you have to re-enable it in the developer options. But don’t take my word for it. Feels like google changed this behaviour every major release with android.

Oh, if that is the case, this will not work. I’ll test it out later this week to see.

On their blog they say: This project is made possible by the awesome work of various open source projects, including Shairport Sync for Airplay, Raspotify for Spotify Connect and Snapcast for multi-room audio sync. So they “just” glue existing stuff together which leaves you with roughly the same limitations as if you would do it yourself. It might allow spotify to be used with snapcast for multiroom but as I’m a yt music user I didn’t digg any deeper.

Hmm, maybe I misunderstood it. Here is a blog post that shows how it can also use Bluetooth. To me, it sounds like that makes an app agnostic solution as long as you are fine with using Bluetooth. My understanding is that you then just connect to one of your speakers with Bluetooth when you want to cast, and you can then control which speakers the audio should play from. I will research this more. I should order a 3.5mm jack extension to the Pi Zero W, which is the only part I miss to be able to set up a proof-of-concept at home.

Nice. I don’t have much experience with VMs yet. Is USB pass-through easily configured?

The Acrylic device also seems pretty nice - I’ll dig a little deeper on that one as well. The rabbit hole becomes ever deeper… :)

It would be nice to have other people being able to use it, but it is not a top priority for me. Also battery drain is not that much of an issue as it will only be used while at home. The last point is a bit more concerning though, so I will see if I can test this out and see how well it works. And yes, the connection seems to be the biggest issue here. But it seems that once configured, it only requires running ‘scrcpy’ on the recieving end. And KDE Connect can be setup to run commands remotely, for example I just set it up to open VS Codium in a specific folder from my phone. I can’t seem to add that command as a quick access tile (which would be my preferred option), but I could add it as a widget on my home screen for quick access so that conncetion is a button press away. How long the connection can stay for I don’t know, but I will see if I can’t test that out this week.

Did you check out balenaSound by the way? If so, what difficulties did you run into that made you discard it?

fcast looks nice, but if I understand it correctly it would require implementation in every specific application. I think if I were to jump onto the Grayjay-wagon, that could be nice, but I would love for my solution to be app agnostic.

Ok, thanks for clarifying that - as I feared. I elaborated on my progress into finding a solution to this below, in case you have any interest in it or any thoughts that could help me along.

Hehe, and for that I apologize! In case you are interested, here are the options I am so far considering. Still very much in the research phase, trying to figure out my specs before I buy any gear I don’t already have. I am not very experienced with this, so feel free to point out any baloney in the below text. Anything that could spare me time researching dead ends is gratefully received.

Option 1: Snapcast

Since I will be running a Raspberry Pi 4 with Home Assistant anyway (not yet set up properly), I would like to make use of the Snapcast integration in HA to run the Snapcast server and then set up a Raspberry Pi Zero W (with some 3.5mm extension) with all speakers I want to connect. Ideally these Zeros could be powered by the speaker themselves (through USB for example) to avoid two plugs, but I don’t know how realistic it is to achieve this, and I have not done much research into this yet.

The issue that I so far have is that I don’t know how I can stream audio from my Android device to the Snapcast server. From my understanding, and what I hoped to clarify with this thread, is that it requires a specific audio source that the Snapcast can recieve audio from. Here is a list. This seems to my limited experience much more doable from my laptop running Linux than from my Android device. But I don’t know…

Currently I will be investigating whether I can use audio streaming in scrcpy to stream audio to the Pi, and then route that via PulseAudio to Snapcast. I don’t know yet if this is a really cool idea or an incredibly stupid idea. I want to setup scrcpy for another purpose anyway, so why not try? :) It might introduce additional latency from my device to the speakers, but as this happens before the “distribution” from the server to the clients, I don’t think this would affect the synchronization. Also, I will never use this to speak on the phone with anyone, so that there is some latency doesn’t really matter to me. The biggest issue would be toggling this off and on - maybe via a remote command with KDE Connect or something like that. If I could set that up as a custom tile in the quick access menu in Android, that might work.

All in all it seems a bit too convoluted though, so I don’t have too much faith in this.

Option 2: balenaSound

So this is the solution I first learned about as an alternative to Sonos, but I was turned off by the need to connect my devices via the internet to the balenaCloud hosted by the developers. However, either I missed this in the first round of research, or they have released it since, but OpenBalena exists which has much of the functionality of balenaCloud (but not all) and can be self-hosted. If I could get the server to run on my Raspberry Pi 4, and then flash balenaOS onto each Pi Zero W, this could provide what I want as I understand that it allows to stream audio directly to the Zero Ws via Bluetooth (with subsequent sync to the OpenBalena server via WiFi). It would be a much simpler solution than the one above, especially in terms of toggling it on and off on my Android device.

A downside to this solution though, is that I believe I would not be able to install it on a Pi running Home Assistant OS (correct me if I am wrong), and that running HA through Docker makes installing new integrations a bit more cumbersome? Maybe that will pose no problem, as I don’t plan on using too many integrations anyway (Zigbee, Netatmo and a MQTT broker). I could perhaps also run a VM that runs Home Assistant OS?

I recently started organizing my music to use with Jellyfin and/or Navidrome. Since Jellyfin requires a particular folder structure, I used this, and I’ve also used MusicBrainz Picard to tag all my music so that it works better with Navidrome. I ended up just using Jellyfin as it suited my needs perfectly, and using it with a desktop client on my laptop (Feishin) and mobile client on my phone (Finamp).

The way Jellyfin requires it to be organised is the way I would’ve done it myself anyway:

Artist 1
|-- Album 1
||----Disc 1
||----Disc 2
|–Album 2
Artist 2
|-- Album 1
etc …

In my experience, if you try to organize based on genres, you need to have a very defined sense of what genres everything you have is. Either you stick with very broad genres (Rock, Jazz etc.) or you get tons of subgenres that you quickly lose control over if you don’t know exactly what is what. Since the clients I use have the possibility to sort by genre, I am planning on giving it an overhaul at some point, but then I will use very broad genres.