bug: Subsonic - Songs on Loop only count as one scrobble

[4rft5]

Please check existing knowledge before opening an issue

• I have checked the FAQ and documentation
• I have searched through existing issues and discussions.

Describe the Bug

When listening to a song on loop on a Subsonic client (in my case, Navidrome/Amperfy), playing a song on loop causes the percentage listened to rise over 100%. When the loop finally ends or another song is played, only one instance of the listen ends up being scrobbled. I assume this results from the limitation of the Subsonic API not reporting the playback location.

In these screenshots the looping song was eventually played three times but only shows as once in recently played.

Maybe it'd be possible if the percentage played goes over 100% and playback state changes (stopped/a new song is played) to calculate the amount of times a song was played from the percentage or track length and scrobble that many listens?

An alternative could be just sending the scrobble when it hits each increment of 100% so it keeps the scrobble times accurate. When playback ends that final percentage could then be applied to the scrobble rules to see if it applies or not. If it ends on say, 258%, it'd already be scrobbled twice, but would then be scrobbled a third time if the scrobble rules allow for over 50% since it ends at 58%

Just spitballing, there's probably a better solution.

Thank you

Platform

Docker

Versions

• multi-scrobbler 0.10.0
• navidrome 0.58.0

Logs

[2025-10-23 23:47:16.332 -0600] DEBUG  : [App] [Cache] [Scrobble] Saved cache to file
[2025-10-23 23:47:14.674 -0600] INFO   : [App] [Scrobblers] [Koito - unnamed-koito] Scrobbled (Backlog) => (Subsonic) Nedaj - CODE @ 2025-10-23T23:38:02-06:00 (S)
[2025-10-23 23:47:14.653 -0600] DEBUG  : [App] [Scrobblers] [Koito - unnamed-koito] Found 40 recent scrobbles
[2025-10-23 23:47:14.604 -0600] DEBUG  : [App] [Scrobblers] [Koito - unnamed-koito] Refreshing recent scrobbles
[2025-10-23 23:47:14.603 -0600] DEBUG  : [App] [Scrobblers] [Koito - unnamed-koito] [Upstream Refresh] Should refresh => newest queued scrobble playDate is newer than last refresh
[2025-10-23 23:47:14.451 -0600] DEBUG  : [App] [Scrobblers] [Lastfm - unnamed-lfm] Raw Payload: 
[2025-10-23 23:47:14.450 -0600] INFO   : [App] [Scrobblers] [Lastfm - unnamed-lfm] Scrobbled (Backlog) => (Subsonic) Nedaj - CODE @ 2025-10-23T23:38:02-06:00 (S)
[2025-10-23 23:47:14.127 -0600] DEBUG  : [App] [Scrobblers] [Lastfm - unnamed-lfm] Found 40 recent scrobbles
[2025-10-23 23:47:13.688 -0600] DEBUG  : [App] [Scrobblers] [Lastfm - unnamed-lfm] Refreshing recent scrobbles
[2025-10-23 23:47:13.687 -0600] DEBUG  : [App] [Scrobblers] [Lastfm - unnamed-lfm] [Upstream Refresh] Should refresh => newest queued scrobble playDate is newer than last refresh
[2025-10-23 23:47:12.803 -0600] DEBUG  : [App] [Sources] [Subsonic - unnamed] Last activity was at 23:47:12-06:00 | Next check in 10.00s
[2025-10-23 23:47:12.801 -0600] INFO   : [App] [Sources] [Subsonic - unnamed] Discovered => Nedaj - CODE @ 2025-10-23T23:38:02-06:00 (S)
[2025-10-23 23:47:12.800 -0600] DEBUG  : [App] [Sources] [Subsonic - unnamed] [Player 1-4rft5] Started new Player listen range.
[2025-10-23 23:47:12.800 -0600] VERBOSE: [App] [Sources] [Subsonic - unnamed] [Player 1-4rft5] (fBL6NFMnTFfw1FgtewMaYa) Nedaj - CODE added after met thresholds with tracked time of 540.00s (wanted 240s) and tracked percent of 298.00% (wanted 50%) and not matching any prior plays
[2025-10-23 23:47:12.799 -0600] DEBUG  : [App] [Sources] [Subsonic - unnamed] [Player 1-4rft5] Generating play object with playDateCompleted
[2025-10-23 23:47:12.799 -0600] VERBOSE: [App] [Sources] [Subsonic - unnamed] [Player 1-4rft5] New Play: (RDaBxisXbabORDlyZWiLPj) Foo Fighters - Everlong
[2025-10-23 23:47:12.798 -0600] DEBUG  : [App] [Sources] [Subsonic - unnamed] [Player 1-4rft5] Ended current Player listen range.
[2025-10-23 23:47:12.798 -0600] DEBUG  : [App] [Sources] [Subsonic - unnamed] [Player 1-4rft5] Incoming play state ((RDaBxisXbabORDlyZWiLPj) Foo Fighters - Everlong) does not match existing state, removing existing: (fBL6NFMnTFfw1FgtewMaYa) Nedaj - CODE

Additional Context

No response

[FoxxMD]

Thanks for the feedback. I think the second approach is pretty clever! And I already have some logic in place to handle a situation like that so this might be easy to implement. I need to think about edge cases as well to make sure this doesn't break existing expected behavior.

[FoxxMD]

Please try docker image foxxmd/multi-scrobbler:pr-386 and let me know how the new functionality works out for you. I don't think this first fix will be sufficient to prevent false positives but i'd like to get your feedback on if you experience any issues.

The main problem is that the only data subsonic exposes, and that MS has to use, for monitoring what is playing is this getNowPlaying endpoint that:

• does not report the player status IE stopped paused playing
• does not report the player position IE 0:45/2:40 into playing a track
• the now playing entry for any session is only updated once every 60 seconds

Since multi-scrobbler can't get data on those first two data points it has to assume that any existing "now playing" entry means that the player is still playing. Even after you stop the player, close the app, exit the navidrome web interface, etc... the entry exists for another 60 seconds.

So there will always be another 60 seconds of time MS thinks the player is going...if you are looping a short track then repeat detection + 60 seconds could produce an additional, unwanted scrobble for the track.

Without the repeat detection its not really a problem since it just scrobbles once with extra "listened to" time, if you are using a scrobbler that supports storing that data. But with the repeat you may get a false positive scrobble.

Let me know if you see this behavior or any other unwanted side effects.

[4rft5]

Seems to be working so far, thank you!

[FoxxMD]

@4rft5 everything been ok so far? If you haven't had any issues i'll go ahead and merge the PR

[4rft5]

@4rft5 everything been ok so far? If you haven't had any issues i'll go ahead and merge the PR

Yep, seems to be all good here from what I can tell.

[FoxxMD]

Released in 0.10.2