How will I identify the fixed files in order to transfer them back to my master library?
Why did my first trial fingerprint only about 16% of my files? Can’t any/all files be fingerprinted?
Identifying fixed files in the input folder...
[quote=nickeaston]How will I identify the fixed files in order to transfer them back to my master library?
[/quote]
By default, files are not renamed unless you check the Rename files based on metadata when matched so in this case the file is not moved. If you do rename the filename should clearly identify the song, and/or the report generated shows the old and new names of the file, does this make sense ?
[quote=nickeaston]
Why did my first trial fingerprint only about 16% of my files? Can’t any/all files be fingerprinted?[/quote]
Yes they can (unless they are very short songs or drm protected audio), could you send email me the html report created so I can understand your scenerio better. Unless the issue is simply that you are using the trial and you expected the fingerprints to be stored in the files after running FixSongs. With the trial the report show you what will be changed but no modifications are actually made to your files unless you have a full license.
Paul’s promo says: “Fix songs based on the audio only using Acoustid”
As a several-year user of Jaikoz, I welcome SongKong as a much more user-friendly, lean-and-mean streamlined app with a high appeal factor to owners of huge but tag-challenged music libraries.
I have processed several thousand tunes since my purchase and SongKong appears to get more useful tagging hits than Jaikoz ever did. (MusicBrainz only)
Since I have about 30,000-50,000 tunes out of a 2.5TB library that are poorly- or un-tagged, I am curious about the future of fingerprinting technology…and maybe I’ll eventually have to delete all files with no apparent metadata.
JRiver’s YADB server claims to have an in-house developed fingerprinting capability in addition to access to one or more meta-databases. I have seen no evidence that YADB fingerprinting produces results on tunes with no apparent metadata.
I am going to try the AcoustID Fingerprinter today, but as a veteran tagger I haven’t seen significant results from either YADB (of which I am a daily user, since Media Center is my player and editor), nor Jaikoz.
Paul’s promo says “easily find duplicate songs in your collection with the Delete Duplicate Songs option” so my question is: On what technical basis are dupes determined? What does “actual audio” mean? Is there a user-adjustable comparison factor as with the Audio Comparer app?
Paul’s promo says “watch a folder and automatically fix songs added to that folder” so my question is: What would the practical maximum folder size be? (…and what definition of “duplicate” would be used?)
[quote=nickeaston]Paul’s promo says: “Fix songs based on the audio only using Acoustid”
[/quote]
Maybe that lines could be miscontrued, it doesn’t mean that SongKong will ONLY use Acoustids to match data if the songs have useful metadata as well, it means that if your songs hvae no metadata then SongKong can still match them using just acoustids.
Of course it can only match them if they exist in the Acoustid database, but the Acoustid database is an opensource rapidly growing database with contributions frrm many sources.
By actual-audio I mean acoustid. The idea behind acoustids is that if an original recording is recordied from CD, vinyl or cassette, and encoded as flac,mp3,mp4 ectera all these encoding will still have the same acoustid. But if the original recording is at all different they will have a different acoustid. There is no user-adjustable comparison factor, I dont see the point of such a thing.
You can delete duplicates based on their acoustids and/or metadata ids as follows.
Every option checked increases the number of things that have to be the same for a duplicate to be found, reducing the cages of a duplicate being found. A duplicate MB Recording Id means two songs match to the same recording in MusicBrainz and a duplicate MB Release Id means two songs match to the same release id, so if both these options are enabled it means a duplicate must have matched the same recording on the same release (unless a release contains the same recording twice). A duplicate AcoustId means that the audio fingerprint of two songs is the same , they are sonically the same (or at least extremely similar).
A Watch folder, watches for new files being added, the size of folder is not relevant. It watches for new songs and fixes them but does not de-duplicate them, however that is an interesting idea I hadn’t considered and I will consider further to see if it is practical.
A Watch folder, watches for new files being added, the size of folder is not relevant. It watches for new songs and fixes them but does not de-duplicate them, however that is an interesting idea I hadn’t considered and I will consider further to see if it is practical.[/quote]
I stumbled upon this early on and thought it was a bug. Once explained, it makes sense.
Like Paul said, it is only new files to that folder. If you select a watch-folder with 20k files already in it, nothing happens. If you drag an album folder in, it will get processed and the rest left alone. There’s no size limit cos it’s not loading the contents of the watch folder, only monitoring changes. There’s no limit on what can be dragged in either, except you will be waiting a while to process 500 megs.
Here’s the reasoning:
You point the watch folder to your main collection, nothing happens (we should assume it’s already been Fixed). You want to file some new songs, SK will process, name, and move the new stuff you just dragged in.
The size of your collection may warrant a different watch-folder, but not for directory-size reasons. My collection is large enough that I wouldn’t be able to see any changes file-wise. I prefer to use a “New Songs” directory as a holding dir and as my watch-folder. Here I can visually see that things are changing correctly and then I move them over to my main pile manually.