Need help with a FLAC sort script

Development Code
,
1

I realized I am going to need a good amount of assistance learning and creating this script. So here is the idea. I have this music library of both high and low quality FLACs right? Its all FLAC and all of the meta data is accurately placed in it by MusicBrainz Picard.

Picture example

image
image924×833 130 KB

As we can see above its sorted by artist>release-album>files.flac

The idea here now becomes what duplicates do I keep. I wish it were as easy as just choosing the higher sample rate FLAC but it isnt. We all know that when it comes to true masters of music 24 bit depth is chasing fidelity where 16 bit depth is driving volume so to speak. The sample rate is a different story. I want to make sure im not deleting my ONLY copy of a release.

So here is the concept I want to employ is that we need to pull this information for every single file.

#➤ [Bon Jovi - 2020 (2020) [24B-96Khz]] metaflac --show-sample-rate --show-bps 01.\ Bon\ Jovi\ -\ Limitless.flac 
96000
24
#➤ [Bon Jovi - 2020 (2020) [24B-96Khz]]

So the first thing I need to do is create an array or data set with ALL the files full directory paths. I am assuming find -type f > musicPaths.txt In the directory would work but thats not storing the directories in an array to step through? This is an array I can pull from for each directory and run the above command on it and store whatever is 24 bit and 16 bit into respective smaller arrays

From there the script would have to go in and sort each of the 24 bit and 16 bit arrays into smaller arrays of which are which sample rate. Sorting them into named arrays 16B44.1K, 16B48k, 16B96k, 24B44.1k and so on and so forth

From there the script would then need to sort for duplicate MAIN song title names excluding stuff like live etc…

From there it would need to choose the least quality duplicates and place those in an array. Then rm would have to just remove each file from the directory as listed in the deleteThis array. I want to make sure im not deleting my ONLY copy of a release.

How would I start. Where would I start. I have a sort of idea of what to do. If I need to clarify this please expound on where you need clarification in order to assist me in brainstorming.

2

First I would have the script make a tsv of releases, bit depth and sample rate.

Then I would sort and identify similar releases. Then you can use that to identify candidates for removal.

1 Like
3

so CSV
<dir>,bit depth, sample rate per row for each song

Each song?

I dont know commands well enough to do this admittedly lol … off to google :smiley:

This sort spots gonna be hard I think. the format is generally but not universally $Author - $Song name [$otherCrap]

4

Learn the commands. Their man pages are very verbose. You’ve got this

https://tldp.org/LDP/abs/html/testconstructs.html

What you’re asking for is not particularly difficult, but it is an excellent learning opportunity. I could probably write a suitable script in 30 minutes, but that’s no fun for you.

1 Like
5

thanks

I just :thonk: 'd

Add values to the CSV for author and song name… put dir as last value. So I can ease the sorting through the directories and stuff? Man I feel like I am creating a database

1 Like
6

I’d go author then album if you can grok it out.

1 Like
7

Im going to have to figure out how to metaflac those out real quick

P.S music library size 4.6TB OOPS

8

You could do it entirely on the command line storing your flac inventory in flat text files.

Step 1. Make inventory

find . -type f -name '*.flac' | parallel -j 10 --bar --eta metaflac --show-foo --show-bar --show-file {} | sort -k2,3,4 > $HOME/inventory.txt

Step 2. Make unique inventory

uniq --skip-fields=1 $HOME/inventory.txt > $HOME/inventory.unique.txt

Step 3. Find duplicates

comm -23 $HOME/inventory.txt $HOME/inventory.unique.txt

Try on a smaller set of few thousand files first. (this kind of command line stuff is something i just made up because i work in Linux all the time, not something i tried before, test it… it may not work, but none of the above is destructive)

2 Likes