Oddly enough this is easier than it seems. The only change would be to base it on a 5 second difference as opposed to 4.

The Yate database viewer can search for duplicates based on fuzzy differences in the Title, Album, Artist and exact matches on Track, Disc and Year. It can also use a time difference of 5 seconds. All tests are optional.

So the process would be to create a database with sufficient information to do a 'search for duplicates'; search for duplicates and save the results and then run an action over the modified database to do the moving.

You can download an action to produce the database at:

https://2manyrobots.com/actions/Create-Database-Duplicate-Minimal-Database.zip

To produce the database do the following:

- start the Batch Processor Wizard from the Actions menu
- select the Create Database Duplicate Minimal Database action
- hit Next
- drag the root of your Master folder to the target
- drag the downloads folder to the target
- verify the folders via View Folders
- hit Next
- disable Auto save files
- enable Verbose log
- hit start
- go get a coffee

when the process finishes you should have the following files

~//Documents/Yate/DupsDB.csv
~/Documents/Yate/DupsDB.plist
~/Library/Application Support/Yate/BatchProcessorLog.txt

Send me all three files and I'll write the second action. If you'd prefer to not send me the database files just send the log. I'll tell you to how to generate the duplicates when I've done the second action.

At this time I can't automate running the dup search on the database from an action. If I add that in 7.1 we could automate the entire process.

For now you'd have to produce the database, run the dup process and then the action.

If you want to proceed now, zip up and send me the two database files to support@2manyrobots.com so I can see what's going on and write the action.

I'll look at automating the dup process in 7.1

I've had a quick look and your criteria involving only Title and Duration will not work. You have a lot of tracks which have no Title metadata.

In my testing, I used Title (80%), Artist (70%) and Duration (5 seconds).

1790 tracks were identified and there were 839 unique groupings. Some tracks had more than one duplicate.

What do you want done when:

- all matches of a track are entirely in Downloads? Copy 1 I assume.
- all matches of a track are entirely in 'the other folder'? Eliminate duplicates?
- matches are split between Downloads and 'the other folder'? Copy none?

What criteria should I use to determine which copy of a file to keep? Most metadata of Title, Artist, etc.? Best bitrate?

I would like an action which duplicate checks all tracks within the downloads folder against tracks within my master music folder.
If a duplicate is found in the downloads folder then it is moved to a duplicates folder.
If a track is found to be a new track and not a duplicate then it is moved to a new tracks folder.

I can take this literally in which case I will ignore duplicates only in 'the other folder'.

I suggest that you play with the duplicates functionality to determine your match criteria. As I said, I used Title (80%), Artist (70%) and Duration (5 seconds)

- Open the produced database and from the context menu do a 'Search for Duplicates'.
- Enter your criteria and Start
- The database will have a Results column where the same value represents 'part of a group'

Let me know and I'll get you something over the next few days.

No problem. The action will not be too difficult.