I saw something while playing an audio CD (which is rare lately) in Rhythmbox on Linux. This CD is a CD-R. Upon Rhythmbox opening, I see a alert type bar at the top as follows:
The reason I am reporting this is that I see this as potentially problematic, but I might be missing something too. First, I would not want to submit this album at all. The data would be invalid data and damaging to the MB database. If I click the button to submit, it brings me to the screen on MB where I would normally search for / enter and add a CD TOC.
The first issue I have with this is that it is a CD-R. The second issue is that there is nothing checking for, confirming, asking or anything to see if this is a actual album or just a compilation of recordings I put together. In my case, I fail both of those quality control checks … this is a CD-R and not a real album in content.
I guess I am not really asking anything, but reporting this. I am unsure what, if anything, MB will want to, or even can, do. But I felt MB should be at least aware of this behaviour.
The same can happen with Picard. You can see how some of the questions are framed on the forum. Some people just using Picard to guess what their files are. They don’t really care which specific release, just getting the tracks named is enough. This will also lead to duff data getting into the database. Dodgy AcousticIDs in that case where someone didn’t select which release - just hit the button to submit.
Bad data can appear from all kinds of angles. The nice side of the double checking going on by all levels of pedants on MB system is that really dubious stuff does tend to stand out. Drowned by the volume of quality data.
When this is being submitted from Rhythmbox what was the username used to submit? Did it submit as you or a Rhythmbox username?
Also with your specific case it wouldn’t really matter if your CD-R’s DiscID got associated to that release as it would just lead to one unique DiscID added to the release. This isn’t as bad as adding a Real CD DiscID to the wrong release as that would lead to bad matches.
Thanks for the reply. I understand there are many sources of bad data. All differences aside, I care greatly about data quality, thus my post on this. I still view MB as the best data source for CD releases, which is what this relates to.
What I have done is log out of my MB account so I can emulate a random user. When I click to submit the album, I am taken to this screen:
So it asks me to log in or create an account, or do it without creating account which has a quality check. So to answer your question, it is either submitted under the users account or with no account and will be pending approval of sorts.
I understand that bad data is overshadowed by good data, usually. But it is the same issue I have with AcousticIDs… it is a GREAT system but is toxic with bad data. I would still love to help rebuild it with original CDs and digital release albums but there are just too many issues… different topic.
As long as MB is aware we are in satisfaction of my post. Also, I did NOT submit RE: “Did it submit as you or a Rhythmbox username”… just to be clear, I did cancel out of that.
(Note - I’m just a user like you who loves to see quality data. I am nothing to do with MB. But I am here to also like you to work with quality data and be wary of the messy users who enter any old junk… like uploading wrong artwork just so they have a picture to tag with)
It is good to see what happens with Rhythmbox when logged out. If it is using the person’s normal account, then it will be down to the normal attitude of that user to data quality. So that should be a relatively safe route. If you have a MB account you normally care about MB.
If it is going in as Rhythmbox user this is even better. Then it stands out really well as an addition of unknown quality.
I also expect the total noobies will run away once they hit that Lookup CD page. It is so full of weird geeky terms that the average person who has never seen MB before will likely step away at that stage.
(I don’t want to go OT either, but the way Picard handles AcousticIDs is setup for bad data. So often a bad match happens with a release that offers very wrong release matches. Look at these people trying to tag thousands of files at the same time… the mess that must be caused if they hit the Upload AcousticIDs during those scans is horrific to me… At least with Rhythmbox the errors will be less trouble due to them being DiscIDs as one bad CD-R being added doesn’t negate the Good CDs already in the database. Uploading your CD-R as a DiscID will never cause anyone else a bad match)
There is not anything special about this. This is the normal disc ID submission implemented by many tools like CD players, CD rippers and tagging software, including e.g. Picard.
The software cannot determine if this CD you have placed into the drive is valid for submission. The software uses MusicBrainz to gather additional data about this CD. And exactly because the disc ID is not available on MB the software cannot get that data. Hence it asks the user for help.
If you have the impression we are getting too many submissions for home-burnt CD-Rs we should probably talk about how this can be made clearer on the MusicBrainz disc ID submission workflow. But this is entirely under the control of MB itself, nothing applications should have to worry about.
So far MusicBrainz has actively encouraged software integrating its services, including disc ID submission (e.g. the libdiscid library gives easy access to the submission URL to use). Without this we likely would have far less disc IDs. Disc IDs are special as they are data we can only reliable gather with the help of people who actually have the physical disc available. There is no better place to ask for this data then in the application making use of it.
