Proposal:New Feature Suggestions: Difference between revisions

From MusicBrainz Wiki
Jump to navigationJump to search
No edit summary
 
(3 intermediate revisions by 3 users not shown)
Line 1: Line 1:
{{proposal_failed
[[Image:Attention.png]] This page is a [[Big Mess|BigMess]] of mixed very old suggestions and recent ones. It's badly outdated. Sure [[Needs Updating|NeedsUpdating]] and [[Needs Editing|NeedsEditing]]. Maybe it could be merged with [[Technology Ideas|TechnologyIdeas]] (at least partly). -- [[User:dmppanda|dmppanda]] 15:20, 19 February 2008 (UTC)
|status=Officially closed as Abandoned, March 24, 2010
|champion=None
|trac=
}}


=Feature Wishes and New Feature Suggestions=
=Feature Wishes and New Feature Suggestions=
Line 14: Line 18:



{{proposal_failed
|status=Officially closed as Abandoned, March 24, 2010
|champion=None
|trac=
}}


==Suggestions for New Catalog Links/Ids==
==Suggestions for New Catalog Links/Ids==
Line 26: Line 24:


----
----



==Support for Subdirectories in the MusicBrainzTagger==
==Support for Subdirectories in the MusicBrainzTagger==
Line 316: Line 316:
----
----


Have more [[Server Access Paths|ServerAccessPaths]] for cgi-bin/mq_2_1.pl: In particular, allow fuzzy searching rather than requiring an ID. Why? Say I want to link to something in [[RDF]] but I don't want to spend exhorbitant amounts of time looking for a TRM/MBID - or I have an automated service which *cannot* do it. This would mean that simple [[RDF]] clients like Foaf Explorer could explore [[MusicBrainz]] without having to know SOAP to discover fuzzy IDs. If (when) this is implemented, please contact me -- [[User:DanielOconnor|DanielOconnor]] ([mailto:daniel.oconnor@gmail.com mailto:daniel.oconnor@gmail.com])
Have more Server Access Paths for cgi-bin/mq_2_1.pl: In particular, allow fuzzy searching rather than requiring an ID. Why? Say I want to link to something in [[RDF]] but I don't want to spend exhorbitant amounts of time looking for a TRM/MBID - or I have an automated service which *cannot* do it. This would mean that simple [[RDF]] clients like Foaf Explorer could explore [[MusicBrainz]] without having to know SOAP to discover fuzzy IDs. If (when) this is implemented, please contact me -- [[User:DanielOconnor|DanielOconnor]] ([mailto:daniel.oconnor@gmail.com mailto:daniel.oconnor@gmail.com])


----
----
Line 346: Line 346:
|champion=None
|champion=None
|trac=
|trac=
}}
}}


==Replace AccurateRip==
==Replace AccurateRip==
Line 359: Line 359:


* In Picard, go to Options -> Options -> Cover Art -> Check "Save cover images as separate files", change the name for those files as you please, and you're done. --[[User:HumHumXX|HumHumXX]] 17:18, 25 February 2010 (UTC)
* In Picard, go to Options -> Options -> Cover Art -> Check "Save cover images as separate files", change the name for those files as you please, and you're done. --[[User:HumHumXX|HumHumXX]] 17:18, 25 February 2010 (UTC)

[[Category:To Be Reviewed]] [[Category:Proposal]] [[Category:Picard]]

Latest revision as of 00:26, 28 March 2010

Status: This page describes a failed proposal. It is not official, and should only be used, if at all, as the basis for a new proposal.



Proposal number: RFC-Unassigned
Champion: None
Status: Failed, due to Officially closed as Abandoned, March 24, 2010
This proposal was not tracked in Trac.


Feature Wishes and New Feature Suggestions

This is a brainstorming page. You can add a feature here and hope that somebody picks it up. But this is not likely to happen at all. This page contains one long messy list and it is not used for active referencing very much.

Therefore if you want to help with Turning Your Ideas Into Reality then it is better to:

  1. Contact us in the Mailing List and discuss your ideas with us.
  2. Write a Feature Page (see How To Write Feature Suggestions).
  3. If you are willing to work on it (or have found someone to do it for you) then list (and describe!) your Feature Page on Future Work or Current Work.
  4. If what you want is an instrument added to the AR attribute dropdown list, please go to Talk:Advanced Instrument Tree and discuss it there, please.


Suggestions for New Catalog Links/Ids

  • Descarga (http://descarga.com) is probably the best catalog out there for latin/salsa, folkloric and latin jazz. It would be very cool if the schema supported their catalog numbers. (I have some Python code to parse Descarga's web pages, but I'm not sure if it's kosher to do this on a massive scale... I used the code only on my large collection of Cuban music CDs.). --MartinBlais


Support for Subdirectories in the MusicBrainzTagger

It would be cool if it could support sub folders when you choose the dir... --anonymous

You can use drag and drop with the tagger. That should help --DonRedman


Refresh folder view automatically after albums were saved to improve usability. --stupor


Lyrics option tied to songs in database. --anonymous



Subscribe to this release

It would be nice if you could subsribe to a special release. So it is possible to get a mail when changes to your favourite releases are made. This is could for Compilations by Various Artist and if you only want to subsribe one release by an artist and not the whole "discography". --anonymous




Status: This page describes a failed proposal. It is not official, and should only be used, if at all, as the basis for a new proposal.



Proposal number: RFC-Unassigned
Champion: None
Status: Failed, due to Officially closed as Abandoned, March 24, 2010
This proposal was not tracked in Trac.


Support for Hidden Bonus Tracks

If a CD has extra-long tracks (usually the last track) which contain a large amount of silence and then a hidden / bonus track, I usually split them into 2 files: the track named on the sleeve, and the bonus track. (For example, http://musicbrainz.org/showalbum.html?albumid=39118) It would be useful if MusicBrainz could:

  • identify these derived tracks, so that the appropriate tagging info could be used. Currently, they are unidentified as they are a completely different length to the originals. I am hesitant to add the TRM for the trimmed version as an alternative for the last track, as I'm not sure whether it should count as belonging to that release or not.
  • this is a bit more BlueSky, I know, but it might be worth considering a metadata representation for the positions of multiple songs in releases of this sort, which could be used as a guide for splitting the track. - IMSoP

I too would like this issue resolved --anonymous

I would love to see a more complex way of dealing with Hidden Tracks. It seems logical to me that parts of a track could be named. Imagine an artist releases a CD with, say, track 15 as the final track on the CD, but that track contains two songs separated by 10 minutes silence. Wouldn't it be logical for MusicBrainz to do something more intellegent with this track and split it into the two components: Track 15 from 0:00 - 3:15 and Track 15 from 13:15 - 15:00. This could easily be displayed on the website by having this indented beneath the last track. It could be 'extra information' that isn't used if the program using the query doesn't understand it, but would provide ripping tools etc. with an invaluable tool.

It would also allow users (like myself) that split the track into two songs to reference both to (different) records in the MB database. It would allow MB to be focused on tracks, but capture the song information in the common situation when track information isn't enough for the user. --anonymous


Resolve Redundant/Duplicate Releases

MusicBrainz stores a unique ID for TRM, track, artist and release. Do these IDs have a date code built into them/or are the obseleted when changes are made in the database? My thinking is: the Tagger clearly uses these codes to confirm that an audio file is already saved. It does not appear to check the server before moving these files to the saved category. If it did and if there were built-in obsolescence codes when a change is made in the server database then previously tagged files would be automatically updated. For example, currently there are redundant releases in the database (i.e. the Beatles White Album or Elton John's Yellow Brick Road). Songs from one disc are identified as belonging to several different releases (as evident by the different release names, some with disc numbers in their name). Currently I don't tag these because, though the artist and title are correct, there is confusion as to the release naming. I would be willing to tag these if I knew that the tags would be automatically updated one day when someone corrects the redundancies in the database. --anonymous

(moved here from below because it covers the same topic --DonRedman): The user should have the *option* of selecting an release where there are multiple release matches for a given track by an artist. For example, MusicBrainz frequently selects a greatest hits release (not necessarily official) or even obscure compilations of multiple artists. The user could be more accurate in identifying source release from a group of potential release sources. In my case, the user may wish to be deliberately inaccurate--selecting the original release rather than the source greatest hits release. --anonymous

If I understand this right, then this issue will be resolved by the PicardTagger. Can anybody confirm this and then move this to OldFeatureSuggestions? --DonRedman



Set Preferred Geographic Region in the Taggger

There should be a preference in the tagger for a geographic region (for example, if all my releases were bought in the US, then disregard editions of the release published in Australia). This would lead to more accurate tagging and perhaps reduce the likelihood of TrmCollisions. --anonymous



Support the replaygain tag

I think Musicbrainz should write replaygain tags. I have no idea how or what you do to write them but im sure you can google it. For those who dont know, replaygain is a standard to fix the variable volumes in mp3's, I know foobar2k uses it, and it works real well. --anonymous

This might not be too smart since the volume is defined by the encoder-settings. Besides foobar2k recalculates the replaygain every time you play the song by default (if I remember correctly), it sure do calculate the values if the files doesn't have 'em. And thirdly, only foobar2k uses these tags yet. The replaygainplugin for winamp use its own tags. So there is no need to implement a field which wouldn't be of any use to most people, and most probably would be wrong for the few who uses it!

  • That is quite incorrect. ReplayGain was designed to be a common standard - to work in the same way on all platforms and all players. And currently, the only way to scan for ReplayGain data in Foobar2000 is to do it manually - it does not recalculate ReplayGain every time when playing the song. MusicBrainz should leave the ReplayGain tag intact, since the actual sound stream is not modified. Currently, the ReplayGain tags are always removed. --intgr

I support the feature request of anonymous, and I think it is entirely reasonable and desirable. I can see two benefits: First, reading the replaygain tag from the database is quicker than scanning the audio data locally, so if for example a new CD is inserted into a player or computer, the correct playback gain can be applied right away rather than having to scan the disk first. Second, one can query the database for the replaygain tags of discs one does not (yet) own. This can be used to find out about the relative loudness of the tracks, which may be interesting for statistics, for purchase decisions (a track that is very loud may be overcompressed and hence sound bad), or for comparing different versions of the same song. --Stefan

The value calculated by ReplayGain can vary greatly based on normalization, file format (lossless v. lossy), encoder software, encoder settings, etc. A value don't really apply to a song in general but only to the specific file that was scanned. --dkg



Present a List of Choices in the Tagger UI when TrmCollisions Occur

It would be a major time-saver in the MusicBrainzTagger if, when a TRM matches more than one track, I could pick the correct one without having to send it to the unidentified tab. I assume it's a user-interface safeguard not that they have to be moved at all. What would be better would be if ALL exact matches appeared in the entry, so i could radio-button the correct one. --ChozCunningham

i second this one.. major head ache... even if a 4th tab is needed or something for files whose trm matches more than 1 song... --anonymous

As it is, it would appear that no further automated tagging occurs following a trm collision (even though, for there to be a trm collision, it would seem that there are at least two matches). A context menu in the Tagger with the ability to chose one of the two trms (or none) would be very nice. --anonymous




Status: This page describes a failed proposal. It is not official, and should only be used, if at all, as the basis for a new proposal.



Proposal number: RFC-Unassigned
Champion: None
Status: Failed, due to Officially closed as Abandoned, March 24, 2010
This proposal was not tracked in Trac.


Enhance the RelatedArtists Feature

PLEASE PLASE PLASE! add a feature similar to "people who liked this song/artist also liked..." --anonymous

  • This could probably be quite easily implemented using the XML interface of www.echocloud.com --Pelpet
    • That's what the RelatedArtists are. --RjMunro
      • Try last.fm - you may be able to link up with their system to impliment? -- Kim
        • this seems to be well underway and already running on the testserver. last.fm specifies what kind of other music listen to as well, if you share similar tastes --Keschte
          • RobertKaye said that currently, contract issues with last.fm do not allow us to publish the related artist data, yet. --Keschte

Song recognition (using a bit of song)

I just wanted to suggest a new function to add to your wonderful program: the possibility to know a song title and artist only by using 20-30 seconds of the song (as the "Shazam" service for mobile phones). For example, if I'm listening to a beautiful song at the radio and I'd like to know the song title I could record a piece of song and submit it to you to receive the title. Is it feasible with your architecture?






Add the ability to collapse tracks from the same artist that appear on multiple releases into one single track with only one TrackID

If the above was done then one side-effect would be the ability to use the tracklists of CDs you don't own as readymade playlists. For example, if you owned all of the individual releases released by a band such as Simon & Garfunkel you would be able to use one of their many Greatest Hits as a playlist to save you creating your own. Or if you owned a lot of songs from the 60s (volumes 1 through X of "Hits of the 60s" perhaps) then you may be interested in listening to "Girl Groups of the 60s" or "#1 hits of 1967" or "The Psychedelic Sounds of the 60s" or "The Songs of Burt Bacharach" etc.

Analyzing this relationship data (artist/song x appears on n compilations with artist/song z) would also probably give (yet another) metric of relatedness for songs. See RelatedArtists.

The trick with this though is that there ARE some alternate versions of songs by the same artist with identical titles (perhaps alternate mixes on different releases) and also, that the TrackIDs as they stand right now also help to identify what release a track is from, which can be handy in some cases. Perhaps an additional "General Song UUID" would be better to track songs (as opposed to tracks)? -- Pipian

  • I thought that would be a good idea too. But it seems it would just be a mess to implement. People would collapse tracks which aren't neccesarily the same, or at least try. Sometimes they might have very similar times and all, so how is a voter to tell without having both tracks. However, the idea might be good for relationships: e.g. you could "lock" tracks so that if, for example, someone adds a composer to the track on Release A, that relationship will carry through to the track on Release B. Being able to group multidisc releases in a better way seems on a similar line of though, yet far easier to implement and verify. That way certain universally true data can be automatically copied - a given release with two discs will have the same release events, ASIN etc. -- DaveAndrews


Suggestion: I've noticed my version of the Tagger, 0.8.5, compares both the TRM ID and the current tag info when calculating the degree of similarity between my track and the ones on the server. Maybe my track info is all screwed up, and the program tells me the song is only a little similar to the one in the server, and only if it gets the right song. This could be reduced if the comparison involved the TRM ID only. Maybe only after that has failed, the current tag info would be used to look for a match.

Daniel CSF, 20030329



Add an option to control placement of "The" in titles. For instance, "The Who" versus "Who, The" and some may prefer just "Rolling Stones", for example.

  • I assume you mean into the tagger... It's easily possible given that MusicBrainz has a sortname field. The last I heard this was being added into the MusicBrainz library, so it's only a matter of time before it should get into the tagger. - ArchieB
  • You may also want to harass the makers of your favourite mp3 player to support the sort the TSOP tag in ID3v2 which is designed to contain the sortname of the performer. This means you can have "Bob Marley", "Bob Marley & the Wailers" and "Ziggy Marley" listed together under M, without having to display them as "Marley, Bob", "Marley, Bob & the Wailers" and "Marley, Ziggy".


Some singles have the same name as the release. Making the metadata visible (and usable in the tagger) in some way may reduce the urge for people to tack "single" on to the end of the title. The same goes for live and instrumental releases. If people want this style for their own data then it should be an option in the tagger.

Also, there is a great singles database available from www.discogs.com and I believe that it could be accessed in much the same way as freedb.com is used now. - ObedFernandez



What about the ability to search on artist AND release? This would allow much greater accuracy when searching the site.

- You can do this using the "Advanced search" part of the search page (basically it's the Tagger lookup, but without the Tagger) - DaveEvans



I assume (from the donate button) that you need money to keep the project going. Have you considered a 'sponsor an improvement' scheme. I'd be willing to donate money to see the two improvements above get done. People could pledge money to particular improvements, and once a set limit was hit, they would get done. This would also let you know which improvements were the most popular.



The interface could use some tweaking. The ability to select multiple releases and do a batch command on them from the normal, non-expanded release list view has disappeared. This forces the user to go to the compact view just to do batch commands. :(

Also, we could consolidate the info/edit lines in the release details view. Right now there are way too many links when you're looking at a single release.

Finally, a really useful improvement would be to list the date of release for the release in the normal, non-expanded release list view. It would be great to have these automatically sorted by date, too.

~Tenebrous~



Another good thing would be for the interface to remember the last set state of the release listing for a users when they are logged in. So if you set an artist to a compact listing it will remain that way as long as you are logged in. This would prevent the user from having to keep setting the release listing when it's different from what the "view compact when more than X releases are listed" preference.

~Tenebrous~


Also when you are importing an release, if we could add a release at that time, it would save another step.  

~Tenebrous

FreeDB stores year info for many releases, and it is surprisingly accurate. --bawjaws



But FreeDB doesn't have release date for many releases and when it does, MusicMoz doesn't get that piece of info on input. Also, we need a way to denote which record company/label released the release.

~Tenebrous~



It's important to distinguish between types of releases in terms of how available a release is. Contrast a major-label release and an indie release with a specialty release done by the band themselves in which only 1000 copies are ever made. I propose creating an extra tag or category (could be in addition to existing structure or just add this to the release/ep/etc list) called "Limited" for these specialty releases.

~Tenebrous~


It also would be nice to choose an abbreviated release selection page when importing releases from freedb. If you've done it more than 40 times, you know what each item is, yet you still have to scroll the whole page down to get to the submit button.  

~Tenebrous~



I second everyting that has been said about release years. Also it should be possible to mass-import release years for the releases already in the database from freedb or amazon (but they should probably be marked as "inaccurate" or similar so that future moderators won't be confused if they see a bad release year).

/Beliuxxx


I have 2 ideas, both of which are for the tagger. First, I would like for there to be a way to not just remove a track from being analyzed by the tagger, but to totally delete it from my computer. Too often, either a downloaded mp3 or one that i have ripped myself will have errors in it, and it would be beneficial to be able to remove the track from within the tagger. Secondly, since all the tracks have a unique track ID, would it be possible to store this ID in the comment field of an ID3v1 or ID3v2 tag? As a corrallary, this info could be used to refresh my own collection to any changes that may have occured since the last time the tracks were tagged without needing to completely reanalyze the track.

--If you were to use this just for personal use, it might be beneficial to create a table (read: XML) with the TRM and a file's corresponding checksum, so that the process of refreshing a collection occurs quicker. I think putting the TRM in the comment field would be necessary only for the convenience of other users who download from you. I'm not sure if either way is better / easier, but just thought I'd throw my comment out there. -evileyetmc



BetterReleaseRecognition



A suggestion for a new search on the Advanced page: moderations in which you added a note. It's difficult to keep track of what you agreed to do when you can't search by this and the search doesn't let you plop down in the middle of all your moderations, but forces you to go to the a number and hit > to get the next seven entries or so. ~Tene



Dont know if these features have already been suggested (or already exist!)- i don't want to read the complete page as it is my first visit to this wiki... - so please drop me a short comment if they are. - (1) I'd like to start browsing by entering the artist name so i haven't to click letter by letter. - And (2): Could the .cddb file for an release be made available for download?? - FloK, 04-01-20



Don't know if this is possible or not, but can there be release art integrated into the database. I would really like to completely get rid of MusicMatch but this is the only thing holding me back.

Lyrics would be nice too. /Pelpet



I think it would be nice if the tagger could open playlist files. Then I could easily select poorly tagged files in winamp/foobar2k and export a list to the tagger.

/Pelpet


SubscribeToEditor

The following is largely addressed by the recent suggestion of SubscribeToEditor ("watch" another editor's edits/votes):

This is more of a social / website suggestion, but when I was going through some random moderations, I saw that somebody as anal as I am had gone and done some thorough updates (fixes, not just freedb imports). It's easy to click through the user link to see all their mods, but actually what I wanted was a list of all their open mods so that I could vote for them. It's a bit of a pain to set up an advanced search for one moderator's open mods, but I was able to do so manually, and I voted for pretty much every one of them. I had some confidence in their correctness - they weren't all perfect, but all were improvements over what was already there. It's a nice reward for somebody who has obviously put more than a few minutes into correcting data, and actually takes much less time than voting on an equivalent number of random mods (most of which are release adds/imports that have already been entered anyhow).

It would be nice if individual moderator pages had a quick link to "vote on this person's moderations" that would pull up an advanced search for that person's open mods - then when you click on submit, the page refreshes with the remaining open mods. It might also be helpful to add something to the moderation faq suggesting this sort of "buddy mod".



The Edit Releases pages are awkward - it requires a minimum of three extra pages and four clicks to add a single release date to an release that has no release information. And the first time I entered release data, I didn't realized that I had to "Enter Moderations" for them to take effect (I thought that the "Submit" button had already done it). It would be much nicer if "Edit Releases" for an release with no releast info took you directly to a page where you could enter a release without having to click on "Add Release" and "Submit" - it would also be easier if the "Submit" button actually submitted any changes. Dupuy


Have more Server Access Paths for cgi-bin/mq_2_1.pl: In particular, allow fuzzy searching rather than requiring an ID. Why? Say I want to link to something in RDF but I don't want to spend exhorbitant amounts of time looking for a TRM/MBID - or I have an automated service which *cannot* do it. This would mean that simple RDF clients like Foaf Explorer could explore MusicBrainz without having to know SOAP to discover fuzzy IDs. If (when) this is implemented, please contact me -- DanielOconnor (mailto:daniel.oconnor@gmail.com) 

Have you seen the distributed moderation system in place at urbandictionary.com ? If you add or use their data, it has a very quick and fun feature that shows you previous moderation decisions, and lets you say good, bad, indifferent to them. 

It's a way of getting time-poor types like me to contribute stuff, as it becomes a nice, gentle background task and quite compelling.

Kim



When the tagger can't find the release based on the current ID3 info, could it look to file name/path? Although my collection is in rough shape wrt ID3s, it is very accurately sorted in the file structure, which is definitely info accessible to the tagger.



DataTrivia




Status: This page describes a failed proposal. It is not official, and should only be used, if at all, as the basis for a new proposal.



Proposal number: RFC-Unassigned
Champion: None
Status: Failed, due to Officially closed as Abandoned, March 24, 2010
This proposal was not tracked in Trac.


Replace AccurateRip

There's a database and SDK called AccurateRip which allows users to submit checksums of tracks so that it is possible to see if one got the same result as everyone else, in which case one can be very sure that the rip was perfect. AccurateRip is not free/open software and can therefor not be integrated into a free ripper like cdparanoia. I suggest that for each discid, one can submit md5 checksums for each track (excluding pre-gap) together with the CD-drive used and the offset used. The offset thing is because most cd-drives miss by a fixed number of samples when asked to seek to a particular position. Therefor that offset must first be known, which requires comparing against a known valid rip. AccurateRip does this by having many reference discs which one can calibrate with, and I think they actually save some sectors of audio data to be able to do this. If that constitutes copyright infringement or not I don't know, but perhaps some clever method of doing the same can be found without saving anything that could be used to reconstruct meaningful audio data (e.g. the first sample of the first 10 sectors or similar).


Save Album Art to Album Folder

Picard is already showing us the Art in the the corner, can we make it so when we save the album and it does the tagging/renaming/moving is puts the artwork in the folder too? I know this isn't exactly tagging related, but it'd be incredibly easy to implement.

  • In Picard, go to Options -> Options -> Cover Art -> Check "Save cover images as separate files", change the name for those files as you please, and you're done. --HumHumXX 17:18, 25 February 2010 (UTC)