Hacky solution for more automatic/accurate track separation

[bleuzen]

Hacky solution for more automatic/accurate track separation

[cgarz]

Yes I saw that thanks. I actually enabled that option since I preferred it.

Sorry I'm still being unclear, what I mean in this case is that the temp file would no longer have the <trackName>.flac format and would instead have a <GenericName>.flac format. Then after recording is done you would rename the <GenericName>.flac to <trackName>.flac

The process would be as follows:

• SpotRec loaded with an option that enables this alternative silent track method.
• SpotRec would then switch from using the track name for ffmpeg to using a generic name.
  (Example: SpotRec.tmp.flac)
• SpotRec starts recording the track to SpotRec.tmp.flac
  (At this point the track is the silent track, no name is known.)
• SpotRec sends the next track command
• SpotRec can now get the track name, stores it for later renaming
• Once the track is finished. Silent track begins playing again
• SpotRec notices this. Finishes the recording
• SpotRec then renames SpotRec.tmp.flac to <trackName>.flac
  (or whatever the user wants the format to be)
• Then the process can repeat.

In this way entire playlists or saved track collections can be reliably recorded and will at worst contain a tiny bit of leading/trailing silence. Which is much easier to remove programatically in a batch process if needed.

[Bleuzen]

This would not work, because it is not possible for us to guess the next song.
To start a new recording, we need the song name. We can only get the current playing songs name. So SpotRec would have to wait until next song starts playing to detect it's name.
Since there is a little delay after song name detection, SpotRec goes to the beginning of the song again to not miss the first second in the recording.
Currently the only way to go to the beginning is to pause and previous. No other way we could use rn.
So this would not eliminate the need of this.
-> Would make no sense to implement.

[cgarz]

Ah yes that would be another issue. SpotRec would need to write to a temp file and rename once done.
Is that out of the question? If so, why?

[Bleuzen]

Not sure what it has to do with this, but to write to a temporary hidden file is already supported.
Just pass -t as argument when starting SpotRec.

[Bleuzen]

Yeah would be possible.
But not today or tomorrow, because:

• Is this useful? How many users would benefit from it? Is it worth the time to implement? Or would it just blow the script up to much?
• I have limited time and other projects and real life things going on, I won't do it in the near future. But it may be a good idea for some free days or a pull request.

[cgarz]

I understand, it is a pretty niche thing with alot of setup required. I might try to make it myself to get all my saved songs. I'm not a great programmer though so it would probably be ugly.

Any suggestions how best to do it? For example where should I put the track info query and the renaming part?

[Bleuzen]

Hi, sorry for not answering, I missed your last message somehow.
I will close this issue because nobody seems interrested in doing this. (You may still do it if you want.)

As a hint: If you want to listen to track changes, implement your code in the "playing_song_changed" method of the "Spotify" class. In this method you have acces to the new song (as shown in the log message there). You can use it to detect if it is the silent track.

[cgarz]

No worries I've been working on other things so didn't have a chance to try this yet anyway.

I would still like to try it though so thanks for the tip.