What do you do now?


You can go to the artists in your collection and social media accounts and ensure there is a link to wikidata (that links to wikipedia).

wikidata is a database behind wikipedia and is used to store facts and to wikipedia pages in different languages.
If I find a wikidata entry for the artist I add the musicbrainz artist id to link the entry back to us.
If you want to go further and edit wikipedia add the {{authority control}} template to the bottom to get this to link back to musicbrainz.

Another thing you can do is start to add works.
See the databases linked from https://musicbrainz.org/doc/Work where you can search for writing credits.


And if you do that, @loujin’s Wikidata Greasemonkey script is a great help.


I haven’t even started editing/adding my own stuff yet.
I thought it would be better to learn on other’s first.

Well, that, and the fact that I came here from Wikipedia - where you aren’t allowed to edit things that you have personal involvement with.

Right now, this morning, I am editing some foreign artists. Some of them have multiple names (english/hebrew/japanese) but they aren’t showing up in search because the aliases weren’t set up the right way.

A few weeks ago, I looked at two bands that have had a combined 100 members over the 75 combined years. It was too big of a project to add the members at that time. I am going to do that some day.


The classical player sounds great! How about one that you run on your laptop/desktop that uses your NAS as a data source for the music files and plays them on your stereo, not on the computer.

When I’ve tried that with Amarok or VLC, it is very slow to impossible. They seem to want to load the files onto my computer (from the NAS) and then send them to the network-connected amplifier. That is a huge waste of time and transmission, when all of the devices are on my local network and it should be easy to instruct one to send to another without sending the data aroud all over the place.


Those are some great ideas! I will be keeping up with the things I entered, of course, should I get error reports, and correct as needed.

I notice that numerous among you mention ‘works’. My interest is primarily classical music and this term has great significance there, since a composer writes a work and many different performers will perform it. I discovered ‘recordings’, which are the performers’ renditions of the tracks that make up a work, but I never really saw anything about the work as a whole.

How would one complete this aspect, for example?


You need them shared in some way the amplifier can understand, which is often pretty limited. Probably has UPnP/DLNA at least, but that (at least on my Onkyo has some issues—primarily, it lacks gapless). That will directly stream from the UPnP media server (which could be your NAS) to the amp, with you able to control it from another device. Best one I’ve seen is the Android app BuibbleUPnP, but it doesn’t really understand classical. And unfortunately UPnP has its own metadata standards which don’t, either.

The other one your amp might support (and it’s rare) is OpenHome, which at least means you probably get gapless and playlists stored on the amp.

Honestly, though, even if you’re streaming FLAC files around, that’s somewhere around 700kbps—so that from the NAS to a computer then to an amp shouldn’t stress a network at all… Even an old 802.11b wireless network should be fine with that. Curious why its causing an issue on yours.

Personally, I connected a Raspberry Pi to my receiver (via HDMI) and run mpd on the Pi. Which feeds digital (PCM) to the receiver.


Take a look at https://musicbrainz.org/release/f7df581c-8001-44cb-813c-ac6e5091c165 (which is a disc I grabbed from your edit history), see where it says “recording of: (something)” after each track? That’s the Work entry. If you follow it, you’ll get to (for track 1) https://musicbrainz.org/work/42d55796-b89d-3b28-87d7-d1379333ba57

You’ll also want to read https://musicbrainz.org/doc/Style/Classical/Works at least.


Yes, my amplifier has UPnP/DLNA. While I don’t know how good the implementation is, often I wish for a more sophisticated connection, but it does work. Gapless playback is one pain that has bothered my a lot, but I have gotten used to it (at the beginning, I was so bothered that I thought I might need to re-rip some of my albums into a single file per work, but the thought of that quickly calmed my bother :wink: I don’t know if the amplifier does gapless, though. Maybe it does, but the QNAP player might not be sendling that information? How would it know?

Actually, aside from the gapless playback issue, my system is working very well with the current UPnP/DLNA connection. I open the web interface to my QNAP NAS or I run the QMusic (or QFile) app on my phone and I tell the NAS to play files and send then to the amp. It works great, but sometimes things go awry and I wait for a long time before it starts playback or the first track gets skipped because of some connection problem, but this is relatively endurable (but shouldn’t happen!). I have wired connections from my router to both the NAS and to the Amp, so this isn’t WiFi. I recently also got a cable for the laptop, so only the phone is sending commands over WiFi.

I looked into BubbleUPnP, Twonky, Kodi, Plex, MinimServer, etc., but there are a number of issues. I prefer to use the QNAP apps, since they are supported by the company that built the hardware I have, and I am very satisfied with their performance. Secondly, these third-party apps are not supported by QNAP and there is no guarantee that they aren’t also doing nefarious things in the background, like collecting personal information or other things. And, most of these third party apps come in a basic flavour and a more sophisticated pay flavour. I don’t want to buy anything extra and have to keep paying and re-paying and re-paying for years on end. Gettitng out of this viciou$ payment cycle was one of the big reasons I switched to Fedora back in 1999 (and I have not used Window$ since). I purposely bought an Android phone for that reason, too, since all my hardware now used Linux and they all communicate perfectly. I am very, very pleased by the quality of the offerings.

Speaking of gapless: I notice that when I’m in my truck using my phone with the VLC (Video Lan Client) app, the tracks seem to be gapless! I wish I could load VLC onto my NAS! BUt I think NAP uses a modified version of VLC internally and they don’t support loading apps from anywhere other than their app store. WHile I have looked, I haven’t actually found any apps anywhere elese and I wouldn’t know how to load them anyway. Only last night did I manage, for the first time ever, to connect my computer to the NAS for file transfer. I had fiddled with WebDav and Samba (SMB) for the past year, to no avail, but last night I decided to just forget that stuff, since it wasn’t working (or I would get scrambled filenames!), so I tried NFS and it just worked!!! I used to use NFS to sync my desktop with my laptop or just to transfer files between them, and, lo and behold, the QNAP NAS understands this, too. Again, the QNAP operating system is QTS, which is a modified version of Linux, so it works perfectly with my other devices.

Since I tagged my albums (I only just finished Sunday night), I have been using the QMusic app to play my music and it is just super. MusicBrans made a big headache go away. I don’t think I will need to hunt down some other app like Bubble or Kodi or whatnot, since now things seem to be going great.

I wil be giving the setup some testing over the next eternity :slight_smile:


That sounds like a project! Before you start though, I suggest you check out Muso - http://klarita.net/muso.html. Not well advertised but by far the best classical music library manager with links to various players. I use it with squeezebox running on a Pi. It integrates with MB. For details of my setup including screenshots etc., see http://music.highmossergate.co.uk/index.php. I wrote the Classical Extras plugin for Picard specifically to take advantage of Muso’s multi-work level capability.


Definitely a project, will surely take 6–8 weeks to complete :stuck_out_tongue:

Interesting how we do things differently. Hand-drawn instead of using graphviz to fit in:

That skips a few git-annex steps, which I guess I shouldn’t have, considering they explain why for some of it (as I try to avoid extra file versions, as I keep them all). I rip the files making sure they already have MBIDs in them, so that I never accidentally get something mixed up. Then those get added to git-annex by a script I wrote (ingest-rip) which also checks the morituri log to see if AccurateRip verified the rip or not; if not they get put in a separate directory to re-rip on a different machine/drive to confirm (a process also scripted, verify-rerip). Verification also does some git-annex stuff (to add the log from the verify and annotate the original log, and of course to move them out of the to verify folder).

After the disc is ripped (and typically before its verified, if needed), I scan the cover art and upload it. Then, working with the scanned booklet/back cover/etc., I add the ARs to the release (2-monitor setup makes this easy—and using the scanned version helps with the tiny print typically used, especially if the room is fairly dark).

Now, potentially, comes a week wait while edits apply. This is to avoid having to re-picard something (and thus storing another version—unfortunately, versions aren’t delta-encoded, so each is another full copy).

Next, I add replaygain tags using the album replaygain tag as a per-work tag instead of per-album; I have a script that makes this fairly easy (flac-replaygain-group) though marking the work boundaries is manual (but so easy that I’ve never considered the time to automate it).

Almost done, I run files through Picard. Since there are already MBID tags in the file, it always grabs the right release. I embed all the cover art in the files (often doubling the size, unfortunately). After having Picard save them, I drag them back over to the to-identify pane and have it do AcoustIDs too, which I make sure to submit if needed. Then re-save the files to get the AcoustID tag.

Finally, I run a script rip-picarded on the rip. That script checks a bunch of things (required tags present, rip log present, album indeed was verified if needed), adds the mess to git-annex, removes silent track 0s (done now so that its stored in the history if ever needed), and finally moves it to a done directory.

The git-annex is sync’d (in full) to 3 machines, one off-site; partially to 2 more machines; and in full to a cloud backup. I do not ever want to have to do this again!


Wouldn’t it make more sense to get the cover art, the booklet, the track titles and timings, the relationships (who performed on each track, when it was recorded, etc) directly from the labels? They are the ones that have all of this information.


I would have thought so too.
But as far as I can tell metadata often hasn’t been “highly valued” in the recording industry.

There is that Dilbert strip where they’ve run out of the correct product and are all instead packaging the wrong product for distribution so the sales will get invoiced this month. They plan to clean up the customer relations mess next month.

This scenario captures my current view of commercial reality in the recording industry. I’m sure there are exceptions too.


If you can find a way to get correct info from the labels and transfer it into MB… please do! But I fear 6–8 weeks won’t be a enough for that project. And of course they’re not going to want to put much work into it, not like its going to help them sell more copies.


Another problem with that is the cover art often varies between releases in the same release group. Different art for different countries and larger font sizes on CD covers versus LP covers are pretty common. Also, going to a label’s website, you’re only going to see art for current releases. Those that are out of print will have to be sourced elsewhere.


For some labels (especially classical) this is actually completely doable and nice - see the “Label sites” section on my classical editor toolbox post. Sadly, for most music, labels don’t care that much.


Super! I believe it really does make their product more appealing to the consumer. It is worth spending time and money on. Many people like to buy the CD and rip their own lossless files, so they need to have this information on tap.


On another thread, adding disc ids was mentioned. Since I actually have the CDs for 100% of my collection, this is something I could easily—but VERY unhurriedly—do.

Anyone want to give me a detailed step-by-step rundown on the procedure?

Do I have to boot up picard to submit a disc id? Do I have to have the release displayed in my browser? Do I have to have my matching flac files in picard while I’m doing it?

I sure hope it’s easy. No messing with my files, please! :wink: Maybe just start picard and insert the CD?

I’m going to search docs on the MB site: maybe I’ll find something (but the search there is pitiful: I tried search docs many times and came up with nothing useful, so I learned by trial and error, mostly, and some editor suggestions).


Here you go: https://beta.musicbrainz.org/doc/How_to_Add_Disc_IDs


I managed to find that help guide through search :wink: When you’re searching for unnecessary work, it works lol

Ok, I just did one. I think it went ok. There’s a disc id now in the relase I had to create (since only a reissue had been entered—yes, a couple of times, I tagged with a reissue, since the performers, title, tracks and timings were identical and I had to locate a cover image manually, anyway… Now it bites me in the foot!).

I’ll have to try a few more to see how this goes before I can say with confidence that I’m doing it right.

While I was tagging and now that I’m submitting disc ids, I am wondering whether MB and the data being entered by users and stored in the MB database are being used to prevent and/or detect music piracy. I could imagine that it might be possible. These hash codes could be unique and allow the tracing of copies of files that get labelled by this software?


To the best of my knowledge, it isn’t. Also, DiscIDs are not really unique IDs, just hash of (parts of) the CD TOC, which basically just tell times apart (e.g., esp. singles with only a few (say 1–3) tracks on them will usually have DiscIDs that are identical to other releases). I guess if some labels put out pirated copies which have some off-by-1-frame changes, it might be potentially traceable, but I can’t recall having heard about any such cases.