When Google Music shut down - I decided to use the opportunity to play with a great new open source music app - Navidrome and run it on Kubernetes (because why not).
Ampache’s been around a while and has been really solid for me. Wondering if there was a specific motivation for the Navidrome folks to start something new, other than new is always shinier of course.
I wrote MousikóFídi as an alternative to Ampache, the website has a small feature comparison table. I’ve only recently learned about Navidrome but plan to add that the table as well.
As for why would I write that: I too was a longtime Ampache user and was very happy with it, to be honest it just seemed like a fun project! After a week or so I had something reasonably usable so I thought: “let’s keep going!” I wrote in detail about it here.
looks neat, nice work! does it have subsonic api support? that would be a good feature to put in your table. Also, without database support, do you just build an in-memory index at startup? seems like that would take a long time on a large collection, but maybe it’s not as bad as I think.
Currently, there is no subsonic support. I don’t use it, nor do I have much of an interest in it, but many folks have asked about it. When I do get around to adding the plugin system, subsonic API support is for sure something I’d like to tackle.
Nothing is loaded at startup; instead, the application reads files and gets metadata (or tries to) on each request. This is running on a single-CPU virtual machine, to give you an idea of what performance might be like on a relatively low-powered host. When database support is added (it is planned as an optional plugin) much less work would be needed, of course.
ohh i see, it’s more file browser oriented. i basically only access ampache via the subsonic api, so that’s a requirement for me. i also use the “virtual” playlists features a lot (like Newest, Most Played, etc). I also need to be able to do things like “search by year”, “search by genre”. Those kinds of things seem like they would be very slow without a database, but if that’s on your roadmap that’s cool.
This is neat. I recently did something similar, but s/K8s/NixOS/ and I host at home as I already have hardware and sufficient network connectivity for the task. https://ww.telent.net/2020/12/1/advent_of_blogpost
Is there a comparison of Navidrome vs. Ampache?
Ampache’s been around a while and has been really solid for me. Wondering if there was a specific motivation for the Navidrome folks to start something new, other than new is always shinier of course.
Shameless self-promotion, but:
I wrote MousikóFídi as an alternative to Ampache, the website has a small feature comparison table. I’ve only recently learned about Navidrome but plan to add that the table as well.
As for why would I write that: I too was a longtime Ampache user and was very happy with it, to be honest it just seemed like a fun project! After a week or so I had something reasonably usable so I thought: “let’s keep going!” I wrote in detail about it here.
looks neat, nice work! does it have subsonic api support? that would be a good feature to put in your table. Also, without database support, do you just build an in-memory index at startup? seems like that would take a long time on a large collection, but maybe it’s not as bad as I think.
Thanks for the compliment :)
Currently, there is no subsonic support. I don’t use it, nor do I have much of an interest in it, but many folks have asked about it. When I do get around to adding the plugin system, subsonic API support is for sure something I’d like to tackle.
Nothing is loaded at startup; instead, the application reads files and gets metadata (or tries to) on each request. This is running on a single-CPU virtual machine, to give you an idea of what performance might be like on a relatively low-powered host. When database support is added (it is planned as an optional plugin) much less work would be needed, of course.
ohh i see, it’s more file browser oriented. i basically only access ampache via the subsonic api, so that’s a requirement for me. i also use the “virtual” playlists features a lot (like Newest, Most Played, etc). I also need to be able to do things like “search by year”, “search by genre”. Those kinds of things seem like they would be very slow without a database, but if that’s on your roadmap that’s cool.
Yes, they absolutely are - I experimented with DB-less search and it was a total disaster. Not that I expected otherwise, but I had to try it.
Thanks for the feedback!
This is very interesting, thanks. Am I right that your program works on a files/folders basis, not a metadata basis?
That’s correct! Just give it a list of directories and browse away.
This is neat. I recently did something similar, but s/K8s/NixOS/ and I host at home as I already have hardware and sufficient network connectivity for the task. https://ww.telent.net/2020/12/1/advent_of_blogpost