There seem to be many questions about Picard at the moment and a lot of people who love the old tagger interface and are frustrated with Picard’s UI. So, let me address the feedback we’re getting from our users:
- “MBTagger was perfect. Picard sucks.” — Ok, this feedback doesn’t help much, but we understand that you are frustrated. We’re working to improve Picard as we speak and the latest release is a beta release, so please bear with us.
- “I don’t understand the Picard interface.” — We understand that there are some people who prefer to tag tracks, not albums. We also understand that the UI is not immediately intuitive. We have plans to fix this, and to allow Picard to be usable by people who prefer to tag tracks and those who prefer to tag albums. For our thoughts on this improved interface, check out this wiki page. For more documentation on how to actually use Picard, see HowToTagFilesWithPicard. And to calm down all the people who are threatening to leave the project if the old MBTagger goes away, we plan to improve the interface of Picard way before the old MBTagger stops working.
- “Picard keeps crashing on me.” — It works for us, which is sucky since we can’t fix your crashes if it doesn’t crash for us. We’re actively soliciting feedback from people who can reproduce crashes. If you have a case where Picard crashes for you consistently and repeatably, please file a bug report.
- “Picard screws up my files!” — First, check to make sure you’re writing the right version of id3v2 tags. Depending on what other programs you use, you may need to switch to version 2.3/2.4 or vice versa. We know about some incompatibilities between the MusicIP Mixer and Picard. We’re going to start working on these issues soon.
- “How do I get PUID’s into MB?” — This answer has two parts. First, to create a new PUID for a file that MB doesn’t have a PUID for currently, you need to download and run the file through the MusicIP Mixer. (Yes, we know not everyone wants to use the Mixer application — for you we’re going to create stand-alone analyzer applications soon). Once a file goes through the MusicIP Mixer, in theory it should become available to the MusicDNS.org service (that MusicBrainz uses) within 24 hours. In practice, this is not happening yet. The MusicIP folks are working on this! Once the MusicDNS.org service has a PUID, you can re run the file through Picard and it should pick up the PUID and prompt you to submit the PUID to MB. Again, there may be bugs in the process — we’re working to iron out the bugs as we speak.
What can you do if you’re affected by these issues? Here is a quick check list for you:
- Do you have an issue not addressed herein?
- Can you reproduce the issue?
- Yes: Please review the open bugs for picard and if this bug doesn’t already exist, please add a new one.
- No: Please stop using Picard 0.7.0 for the time being and go back to Picard 0.6.0 or the old MB Tagger. Then wait for a new release of Picard and see if your problems have improved.
Bottom line: If you’re frustrated by how things are working right now, go back to a non-beta release and give us more time to iron out these issues. We’re aware that you are frustrated and we’re working on it. But if people keep pestering us with the same issues over and over again, it only keeps us from fixing actual bugs.
Technorati Tags: musicbrainz, picard
MetaBrainz Blog 
Must say I rather use picards GUI anyday compared to old MBtagger. But then I am that mass-tagging type with a lot of songs from the same albums, and to me it seems like a waste to use a program that download information from up to three diffrent version of an album from the server when all your songs come from the exakt same version of the album.With that I do not say that there is not room for improvments becouse there are, but I still prefer Picards interface.
There is a very simple solution to this issue. Just add UTF16 support to the Classic MB Tagger, then just split the unidentifed tabs into two seperate tabs Albums and Single Tracks. Then for all the items that fit into one Album they’ll be under the Album tab and you can group tag them.
Reading up on good interface design would also be a suggestion
Just google good user interface design
Kat: Thats the exact thing robert does not want to hear, because it is not constructive, and not in line with the musicbrainz strategy.
If you have suggestions on how _you_ would improve the interface, you can do so, but else we’d like to ask you to be, and stay quiet.
I don’t understand everybody’s problems with the new interface. It works like a charm for me.
But I have a question about the PUID’s. You say in the blog post that there will be a standalone app for analyzing them, but can’t this just be integrated into the Picard tagger? I prefer to have an all-in-one program for my Musicbrainz needs.
–“but can’t this just be integrated into the Picard tagger? I prefer to have an all-in-one program for my Musicbrainz needs.”–
This is what I am hoping for too.
I would like to see it integrated in Picard too.
Can you go into more detail about the PUID? I understand that I have to use the MusicIP Mixer to get the PUID into the MusicDNS database. But then you say to rerun the file(s) in Picard and it will ask to submit to MB???? This isn’t clear to me. Isn’t the reason to have to wait 24 hours or so because DNS is transfering it to MB? Why would Picard ask to submit to MB?
I think I understand why the PUID generation has to be a seperate process — it took about five days to analyse 5000 tracks — generally the process is complex enough that full generation of a new PUID (it does frequency & other analysis of two minutes worth of a song) is perhaps best seperate.
I don’t quite get why Picard isn’t finding PUIDs for songs that do have PUIDs in the database (like Dixie Chick’s “Fly” that I’m just looking at now). Do I have to have run them through the MusicIP mixer?
It also isn’t clear to me why songs are less than 100% similar when matched in Picard — the vertical view of old/new data in the Tagger made this clearer.
Thanks for all your work, guys. This is a great system and database.
What does “run the file through the MusicIP Mixer” mean? Some more specific instructions would be helpful as I’m not sure what to do now that my files are listed in MusicIP Mixer.
Jason — Once the MusicIP mixer has all your music in its library, it will start to generate analysis and fingerprints for each song. I believe it sends that data to MusicDNS and uses it to generate a PUID for any tracks that are not already in MusicDNS.
So all you need to do is let the MusicIP mixer analyse your tracks. Under “Help->About MusicIP Mixer” it will list the # of songs left to analyse — if you see that it still has songs to analyse, check the Library menu. If the Library menu has a “Start Analysis” option, then Analysis is not currently running. It is running if “Stop Analysis” is in that menu.
Be warned, however, that even though the MusicIP Mixer runs analysis at a low priority, sometimes it can consume a lot of virtual memory, making your PC quite slow! Turn off the analysis if you’re doing anything else CPU-intensive, and turn it back on again when your computer is going to be otherwise idle.
It’s good that you created this new blog entry, but count me among those who still find the whole process very confusing. It’s not just that the process of creating new PUIDs is broken, but I’m unclear how it would work if it worked.
Running a track through MusicIPMixer creates a fingerprint and and analysis which can then be ‘archived” into the track as tags. Presumably, MusicIPMixer also sends that information to MusicDNS which creates a record for the fingerprint in their database along with a 128-bit record-locator (the PUID). Picard 0.7, I presume, can also generate fingerprints (or does it have to use fingerprints archived into the track by MusicIPMixer?) and then query MusicDNS to determine whether that fingerprint exists in their database and, if so, return the corresponding PUID. But apparently there is some lag (who knows how long since I’ve never seen it work) before the new fingerprint/PUID is propagated to the database which Picard can query. If and when that happens then the PUID can be used to look up the track in MusicBrainz if someone else has already submitted it; if not it can be submitted if the user can identify the correct track by some other means.
What I find frustrating is that (a) I’m not sure whether my understanding of the process is correct and (b) I don’t have any way of checking what’s working and what isn’t.
On several occasions I have taken tracks which I created with SoundJuicer and analyzed with MusicIPMixer and then given them to Picard; Picard quickly looks at them and associates them with the correct album, but the “Submit info to MusicBrainz” button is always grayed out. Does this mean the PUIDs aren’t available, or does Picard even generate fingerprints when MB tags already exist for the tracks? Given a fingerprint, is there any mechanism to check for the existence of a PUID.
trjh — Thanks for the clarification. It looks like I really didn’t have to do anything more than add the files to MusicIP Mixer. I’ll just let it run while I go to work so it can do all the stuff that it needs to do.
Call me crazy, but there doesn’t seem to be much point in this migration to a new method. With Picard, what sets it apart from FreeDB and Gracenote now? It’s easy to tag things I know the album for- I’m sure I’m not alone when I say my primary use for MB is to identify a song that has an incomplete tag or no tag and an inadequate filename to accurately identify it. From everything I’m reading, Picard not only fails to address this, but doesn’t even intend to. I hope I’m incorrect, but that’s how it seems.
There is a few things confusing me a lot. I am sure there is an easy way to do this – I just can’t figure out how.
1. Every time I run Picard – all my identified files is lost – and I have to reidentfy them all from scratch – even when I have used save and/or the force save option.
All Picard remembers is album names without songs.
Second – Placing songs into albums involves a lot of drag and drop when you have a huge acrvhive of songs and albums – could it not be a right click meny with a suggested move to album – or a seperate window with albums so it would be easier/faster to drag songs into albums?
It says that there is WMA support – but when I run Picard – it says no WMA support – and puts all those file in error – Too bad – half my music library is wma files.
Other than that I would say that Picard is an awsome application. Its a little like magic in fact – and I wish and hope for Picard a bright future!
Glory to Picard!
how do i submit PUIDs to MB, after all?
that’s what i did:
1. added a new artist to MB database (it soon got applied);
2. added a new release for that artist;
3. clicked on green ‘tagger’ icon in mozilla to open the album in running picard;
4. dropped the album tracks to ‘new files’, analyzed them and then dragged and dropped them one-by-one to the corresponding places in the album;
5. after that, the save icon is available, but ‘submit PUIDs’ icon near it is grayed out, i.e., disabled.
How to, actually?
i also have to say that i was able to add my PUIDs to other albums which already had PUIDs but did not find out why and what it depends on.
how do i submit PUIDs to MB, after all?
that’s what i did:
1. added a new artist to MB database (it soon got applied);
2. added a new release for that artist;
3. clicked on green ‘tagger’ icon in mozilla to open the album in running picard;
4. dropped the album tracks to ‘new files’, analyzed them and then dragged and dropped them one-by-one to the corresponding places in the album;
5. after that, the save icon is available, but ‘submit PUIDs’ icon near it is grayed out, i.e., disabled.
How to, actually?
i also have to say that i was able to add my PUIDs to other albums which already had PUIDs but did not find out why and what it depends on.
shapirus:
Please post technical questions to the mailing lists:
http://musicbrainz.org/doc/MailingList
This particular question is currently being discussed in the devel list.
I’ve looked at the wiki for the redesign.
– I would like to be able to improve the metadata of mp3s in my existing folders as automatically as possible.
– All the program needs to do is update the metadata tags.
– I would like a simple linear process with small steps that I can do every so often. Support for matching by albums means that there are fewer tasks, but each one takes longer and requires more thought. I’d prefer more smaller tasks than being interrupted in the middle of more complicated longer ones.
– I would like to see the folder and filenames of the mp3s in the program so I know exactly which file is being updated.
– When identifying mp3s that haven’t been 100% identified, there are often a number of similar possible matches. To make the right match, its useful to see as much info about the suspect track as possible: directory name, filename, existing meta tags…and at the same time being able to see the album and tracklistings of the possible matches.