Delete Duplicates - What's the decision tree?

[wynlyndd]

Again, like always, I love this program. Things just seem to get better and better.

(Like today I loaded up 33K files at once and did some spot checks and weeded out duplicates)

Anyway, What is the decision tree that the “Delete Duplicates” takes because it doesn’t appear to be merely highest bitrate is the one that is kept. In this batch, there were quite a few 128kbps M4a files as well as 192kbps MP3s, and it appears in almost every case, the m4a was kept and the mp3 was chosen for deletion. (I say almost because I didn’t look at everything but I never found a case where this wasn’t true)

Other things I noticed almost every mp3 was joint stereo and all the m4a were 2 channel. Does this figure in the decision tree?

One thing that might be nice as a wishlist item is the ability to mark some folder as “Master” and so any file from another location as “Copy” unless it is “better”. However since your program can identify duplicates in a number of ways, this might not work too well.

[paultaylor]

Hi, bit rate is used but the first decision is the file format and Mp4 is always considered better than mp3 ( an mp4 with a lower bitrate can still sound better than an mp3 with a higher rate, although there are limits), likewise falc is conidered better than mp4.

[wynlyndd]

Would you be able to share the full decision tree? something akin to

ALAC > FLAC > m4p > m4a > mp3 > WMA > ogg ?

While I could create test files and attempt to figure this out via experimentation, I’d rather not.

(in the midst of possibly changing my default format)

[paultaylor]

FLAC > mp4 (mp4,m4a…) > m4p > Ogg -> mp3 > WMA >

i.e lossless first, then advanced lossy before older lossy

If two songs are same format then takes highest bit rate
If bitrate same takes larger filesize

[zeiss]

i think making this process more robust would be a great program enhancement (unless other people disagree). i’m reminded of another file dupe program i’ve used called noclone (for all files, not just music). in there, the user can “smart mark” for deletion based on a number of rules of precedence:
-file date
-file location
-file size, etc.

in this case, because music has additional meta data, one could add:
-file type (user priority)
-file bit rate (might actually want smaller)
-has art
-closer to true song length
-has genre
-has MB ID (if comparing some non MB collections)
or anything else one could want.

obviously if certain fields are required to be the same to be dupes, then those couldn’t be used as selection criteria. i have to admit, i’m hesitant to let a program delete my files if i don’t get to control how it decides what to delete. for example, i like mp3s for universality, and i like 128-192 bitrate to balance quality/filesize. and i want to keep files that are already organized into folders i’ve created, instead of files that might be in a large “dump” folder.

just thinking aloud. love the program.

[paultaylor]

There should be at least some of these items in the next release. I am working on allowing you to replace your deleted files with a link to the not deleted files so playlists ectera are preserved, or moving the deleted files to a special location instead of actually deleting them.