Questions about the right steps to take when using Picard

I’ve never used this app, but I’m having problems finding a good starting point.

I did find this document: MusicBrainz Picard — MusicBrainz Picard v2.9alpha1 documentation, but it leaves much to be desired. I need to know where to look in this guide and what’s relevant to me as this is taking hours to go through and doesn’t explain everything I want to know. The “Lookup” and “Scan” pages just say how to move files over to manually the right side into albums.

I went through all the options, and it looks like it’s a good idea to change some. On top of that, there are three buttons that seem to be completely random:

  1. Cluster
  2. Lookup
  3. Scan
  4. Refresh (hidden in a context menu on the right side)

From what I understand, Lookup is for local file lookups and Scan uses online information. But I don’t understand how they actually work.

  1. Should I do all 40,000 songs at once or try to select different folders? I’m wondering if Picard actually looks through everything to make sure it’s assigning albums based on where you probably got the songs.
  2. Can I mess up tagging files by clicking the “Cluster”, “Lookup”, and “Scan” buttons in the wrong order?
  3. Do I need to click on the folder (“Unclustered files”, “Clusters”) or right-side album before clicking those buttons?
  4. If I do a Select-All on the right side and click “Refresh”, what does that do exactly? It’s taking hours to finish. That’s longer than it took originally.
  5. If I change options like the album art order, do I need to re-scan everything again?
  6. What’s the safest way to use the app? Only saving Gold albums and going through the rest by hand? Very few of mine are gold albums because many of them came in compilations like Amazon’s “99 greatest xyz” or are singles from places like OCRemix from way back in the day.

P.S. I have a directory with a bunch of singles with bunch of album art photos that came from Winamp way back. Should I remove those images? Does that sort of thing slow down Picard?

The Work Flow Recommendations section has some very useful information, but it doesn’t answer all my questions.

The general approach you should take first is the cluster → lookup workflow described at Lookup Files — MusicBrainz Picard v2.9alpha1 documentation

There is also a video introduction at What is MusicBrainz Picard? - YouTube that explain the basics.

With clustering Picard tries to group the loaded files into what seems to belong to the same release / album.

Lookup uses the existing metadata in the files to search for matches in the MusicBrainz database. If you have clustered before and use lookup on the clusters it tries to find matching releases. If you use it on single files it just tries to find matching recordings. Overall this is the first recommended approach to take if you have halfway usable tags already.

Scan for the most part ignores existing metadata but looks at the audio itself instead. It searches for matching acoustic fingerprints using the AcoustID audio fingerprint database. Use this if you have little to no metadata or lookup did not find anything.

Don’t do that. Especially when you are new to the software try to do 1 to 3 albums first. Have a backup of those files! Then try the different lookup approaches, see if you are satisfied how Picard saves and renames your files etc. Once your are familiar with how Picard works and your are satisfied with the settings you can do more albums at once. You’ll figure out how many you can handle. It also depends on the files and how well they are already tagged.

Depends. If you use Scan it has the tendency to scatter the files more over multiple releases. So use lookup first and scan as the fallback.

Scan and Lookup run on the selected files. Same for Cluster, but if nothing is selected cluster will run on all “unclustered files”.

It loads the data from MusicBrainz freshly and re-applies the plugins and scripts. You usually only need this if either you just changed something on MB and want to have the new data or you changed some scripting, cover art or plugin settings that affect the loaded data.

That’s what you can use refresh for.

Those gold albums are a good indication to safe first, definitely. It also depends how picky you are. If it is fine for you that Picard assigned it to any of the releases for this album with a matching track list you are fine. If it is important for you that it uses exactly that 1992 CD release printed in UK with barcode XYZ that you have on your shelf you should double check and choose the proper release (right click on the loaded release and choose “similar versions”).

In general Picard is meant to be used interactively with the user doing the final decision. So I’d recommend to definitely check those that are not gold.

The image files don’t in general don’t influence Picard much. But there is one thing to be aware of: If you have configured Picard to create a folder structure and move files there then you can also enable “Move additional files”. By default this includes image files. The feature is meant to move extra files in an album folder together with the rest, such images or PDF booklet etc. If you have a collection of random files of singles and each will be moved to a separate final folder this can result in the first file being moved to get all the images files moved along.


This is a really good explanation, and the discussion tells me that we need to perhaps add a new “Beginner User” section to the documentation. The new section should probably include most (if not all) of what’s shown here.

We really do need to clarify in the docs the difference between “lookup” and “scan” and emphasize that “scan” should really only be used as a last resort if all else fails.


Maybe also extend the FAQ area a bit. @Sawtaytoes had a couple of really good questions, many of them come up frequently in some form or another.


Awesome thanks!

I finally got into it, tried out a few albums, and spent hours working through my collection.

I actually read through more of the docs and found Exact Audio Copy. I spent all night re-ripping my CDs to FLAC after going through some other guide from ages ago about the “right way” to configure the app. My next step is going to Amazon, CDBaby, and Bandcamp and downloading FLAC or higher-quality MP3 where possible.

With Exact Audio Copy and Picard, it’d be good to have nicer defaults. Going through the options in Picard (especially in Exact Audio Copy), I was pretty confused what I should be changing and which options were better.

First impression suggestions (my opinions)

There are all my opinions. I’m not saying you’re doing a bad job or anything. Picard is a wonderful piece of software!

Most of the apps I use today are zero-config meaning the best configuration is the default, and you only change it for compatibility or if you have a strange scenario. That’s also my preference. When I was a kid, it was fun to go in an app and turn on all the good settings because the app was designed with compatibility in mind.

With Picard, I really needed it to have some great defaults or some kind of configuration wizard such as:

  1. “Choose your language and location.” English and United States for me. But that doesn’t seem to matter. Many of my Japanese albums get converted from English text to Japanese characters.
  2. “Do you need iPod or older media player compatibility?” (This would change the ID3v2.3 whatever. Stuff I had to lookup and spent a lot of time trying to understand).
  3. “If you can’t find what your album, would you like to use a less-reliable database as a fallback? (such as FreeDB; although, I’m not sure you can even do this)”.
  4. “We can rename and move files based on your preferred player: Plex, Kodi, Emby, Jellyfin, Custom”. Right now, it says “Default file naming script”, but it doesn’t say “Plex-compatible” for instance.

Then set some good defaults like:

  1. Saving the cover image as part of the audio files and as separate files.
  2. ID3v2.4 UTF-8 tags instead of 2.3.
  3. Preferred media format: CD.
  4. Add more files to copy when moving: *.jpg *.png *.pdf *.txt *.doc *.rtf.
  5. AcousticID fingerprinting on by default. I downloaded Picard to ensure I could find all the album data. I want fingerprinting.

Set these sliders to be better for most people. I want to put all my stuff in albums unless they’re singles, but you wouldn’t know that unless I added all 40K songs:

I dunno the best positions for those sliders, but I’m very sure Album is the one you typically want the most. For the rest, I randomly guessed where to put them. I’m not even sure exactly what this does.

In terms of a tutorial, it would’ve been nice to have a cheat-sheet somewhere online: “gold is good, silver is okay. Make sure you check every album’s songs before saving any changes as there could be discrepancies such as the wrong song names or missing files that are on-disc but didn’t get added by Picard”.


All-in-all, I just needed to play around with the app. After I found out I shouldn’t put all 40K songs in there, the app was a lot faster, and it was a lot easier to sort through albums.

It’s gonna take forever to fix these, but that’s also because I have a lot of singles, and even my CD albums seem to be too niche. Still haven’t figured out a good solution for those yet.


I feel like Picard could use some beginner help, maybe a first-time setup wizard, simple/advanced modes, asking if you want to open the beginners guide on first start up, something like that.

a couple things I’d add to the list above:

  1. “where do you want to retrieve cover art from”, because you can get it from
    •, possibly slightly modified, but it’s typically gonna be nice and clean for your shiny new library
    • release group cover, typically a representative cover in high quality
    • release cover, the cover for the particular version of the album selected
  2. there was another, but it slipped my mind… I’ll look through the Picard settings later today couldn’t find it… lol

I really like the idea of more default naming scripts, especially ones compatible with other media players and services. I’ll try and remember to make a ticket for this later today


Amazingly detailed feedback @Sawtaytoes! Really impressive.

If you’d like to follow along, or see that some of the suggestions don’t get lost, the ticket tracker is the place to put things for the devs to follow!

I’ve gone through your suggestions, there’s some existing tickets:

Default to ID3v24: Already implemented, in the most recent Picard beta release :tada:

Default sliders, put album up more (yes please):

Warn people who try load a lot of files at once:


Some suggestions I disagree with (defaulting to preferred media format CD, saving the cover images x2, AcoustID fingerprint submission on by default), but lots of your other suggestions would make tidy little tickets. If you don’t have time, and you’d like me to open some, then let me know! @UltimateRiff is already on the case as well.

p.s. If anyone else is here, wondering WTF all the buttons on Picard do, the ‘What is Picard’ YouTube walk-through might be for you. It’s long and unscripted, but hits the basics.


Thanks a lot for the detailed feedback. I agree, we could improve the initial setup. A lot of this could be handled by a first-run wizard, which has been talked about a lot but so far does not exist. There was an attempt to it many years ago.

Maybe CD, Digital Media nowadays.

Can you elaborate? The fingerprinting is enabled after default install and can be used directly without configuration. Only when submitting fingerprints you will need to configure a API token for AcoustID first (and need an AcoustID account for it), but Picard will prompt on first use.

Separate naming scripts IMHO only make sense if there are really specific requirements. As the mentioned tools base their information primarily on the metadata I think the default should work for all of them. Is there some specific issue with a specific tool that can be solved with a special naming script?

Apart from that I’m a bit wary adding too much third-party tool specific info directly into Picard. This is outdated pretty quickly. We have some minor exceptions like the iTunes specific work tags, but that’s mostly because Apple caused the mess behind it in the first place :slight_smile: