Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[EpF]

I realise that as J River are trying to get MC 11 out the door, this is probably not a good time, but I just thought of it, so maybe it'll go in the 'in-tray' for a while...

Would it be possible to have an 'IsDuplicate' expression, such that I could have:

If((IsDuplicate([artist], 1), IsDuplicate([artist rating], 0)), Not a Duplicate, [artist])

To the effect, if the artists are the same, but the aritst ratings are not, then return the artist name...

??

I live in hope!!   

Listening to 'Black Lines' by 1 Mile North, from 'Minor Shadows' (2003)

[StiX]

Not sure how can be implemented the IsDuplicate(), but another Function, which is definitely necessary to add to the list of functions is the IsContain() function. This function could be used in same frequence as the IsEqual() function, so I don't see even a sense to bring an example. Currently it is possible to particularly simulate this function with the Mid(), but that is very limited.

Another problem is that there's no way to add spaces in conditionalities. Ex. If(True,Text,  Text after two space characters).

Fundamental problem with DBExpressions is that there's no way to use them to perform operations on Field Data. Currently it is possible to do such operations only through a plug-in (developing it by using the SDK) and performing operations on files which was added to Playing Now (ex. the array of slow VBSparta plug-ins). Support for DBExpressions should be added to "Find and Replace" dialog with possibility to use Find/Replace presets and presets to perform Find/Replace Batch operations.

Second problem with MC itself, is that there's no way to create linking between songs.

A simple example on song "Ace Of Base - Life Is a Flower". This song appears in 4 albums:

Artist   =Name   Duration   Bitrate   =Alb.Year   =Album/Compilation/Soundtrack
Ace of Base   Life Is a Flower      3:43   192   '98|'99   [Compilation] Singles of the 90's (15/16)
Ace of Base   Life Is a Flower      3:46   192   '98   [Single] Life Is a Flower (1/5)
Ace of Base   Life Is a Flower      3:44   192   '98   Flowers (1/14)
Ace of Base   Life Is a Flower (Extended Version)      5:46   192   '98|'99   [Compilation] Top Remix Album '99 (17/18)
Ace of Base   Life Is a Flower (Extended Version)      5:44   192   '98   [Single] Life Is a Flower (3/5)
Ace of Base   Life Is a Flower (Reggae Version)      3:34   192   '98|'99   [Compilation] Top Remix Album '99 (8/18)
Ace of Base   Life Is a Flower (Reggae Version)      3:32   192   '98   [Single] Life Is a Flower (2/5)
Ace of Base   Life Is A Flower (Soul Poets Night Club Mix)      5:21   192   '98|'99   [Compilation] Top Remix Album '99 (2/18)
Ace of Base   Life Is a Flower (Soul Poets Night Club Mix)      5:21   192   '98   [Single] Life Is a Flower (4/5)

I like to assign following fields to each song (in alphabetical order):

Code: [Select]

[Album Notes], [Album Compilation], [Album Soundtrack], [Artist Bios], [Artist Genre], [Artist Nationality], [Chart], [Cues], [Dominant Instruments], [Genre], [Genre Style], [Image], [Language], [Lyrics], [Mood], [Popularity], [Preference], [Preference %], [Rating], [Similarity (artist)], [Similarity (name)], [Similarity (soundtrack]), [Situation], [Subject], [Tempo]
Another problem with expressions is that they should be written as single line of code, instead of a regular multi-line code. The Column "=Album/Compilation/Soundtrack" have following expression:

Code: [Select]

Clean(If(IsEqual([Collection Type],Assorted), If(IsEqual([Album Soundtrack],), If(IsEqual([Album Compilation],),If(IsEqual([Album],),,[!a] [Album]),[!cIf(IsEqual([Album],),,!a)] [Album Compilation]), [!sIf(IsEqual([Album Soundtrack Type],),-?,)If(IsEqual([Album Soundtrack Type],Movie),,-FixCase(Mid([Album Soundtrack Type],0,3),4))If(IsEqual([Album],),,!a)If(IsEqual([Album Compilation],),,!c)] [Album Soundtrack]), If(IsEqual([Collection Type],Album),If(IsEqual([Tracks Total],),[a],) [Album],)If(IsEqual([Collection Type],Compilation),If(IsEqual([Tracks Total],),[c],[C]) [Album Compilation],)If(IsEqual([Collection Type],DJ Set),[DJ] [Album],)If(IsEqual([Collection Type],Karaoke),[K] [Album],)If(IsEqual([Collection Type],Mix),[Mix] [Album],)If(IsEqual([Collection Type],Genre Preview),[GP] [Album Compilation],)If(IsEqual([Collection Type],Humour),[H] [Album],)If(IsEqual([Collection Type],Recording),[Rec] [Album],)If(IsEqual([Collection Type],SmartSound),[SS] [Album],)If(IsEqual(Mid([Collection Type],0,10),Soundtrack),[If(IsEqual([Tracks Total],),s,S)If(IsEqual([Collection Type],Soundtrack Compilation),If(IsEqual([Tracks Total],),c,C),)If(IsEqual([Album Soundtrack Type],),-?,)If(IsEqual([Album Soundtrack Type],Movie),,-FixCase(Mid([Album Soundtrack Type],0,3),4))] [Album Soundtrack] If(IsEqual([Collection Type],Soundtrack Compilation),[[Album Compilation]] ,),) ) ) Clean( If(IsEqual([Disc #],),,°[Disc #]) /([Track #]If(IsEqual([Tracks Total],),,//[Tracks Total])/) )As you can see it's hard to follow the expression. That's why I always have to write the expression in other editor, which support syntax highlighting or in a Word-like text editor and manually colorize parts of expression, so that I can read and understand it after long time. Also, MC Expressions doesn't support Variables, there's no way to add a statement.

For more than 4 years I'm using Media Center (it was my decision from tens of other tested Playlisters). Primary reasons was because of possibility to add Custom Fields to songs and to search inside the main UI (without using dialogs). I had few thousand of songs. Now, when I have tens thousands, and almost every day I add new music to the library, I constantly have the problem of lack of time to tag each song.

Simple example: I got a new album "[Compilation] Greatest Hits of Ace of Base 1991-2005". Now I have to perform following operations:

• 1. Fill New Unique Album related data: [Album], [Album Notes], [Artist], [Date (year)], [Image], [Record Label]
• 2. Right Click on one of the Songs of the New Album: Context Menu > Locate > Artist > Fill Redundant/Same Artist related data: [Artist Bios], [Artist Genre], [Artist Nationality]
• 3. And now comes the most boring repetitive and monotonous job. Consequently Right Click on Each Song of the New Album: Context Menu > Locate > Advanced > Name > Fill  Redundant/Same Song related data: [Album Compilation], [Album Soundtrack], [Chart], [Cues], [Dominant Instruments], [Genre], [Genre Style], [Language], [Lyrics], [Mood], [Popularity], [Preference], [Preference %], [Rating], [Similarity (artist)], [Similarity (name)], [Similarity (soundtrack]), [Situation], [Subject], [Tempo]
  

And there's even no available Entry for these Commands (Locate > ...) in Resource.xml.

The problem is an opposite of the non-existing function IsDuplicate(). Instead of applying data to each similar song, it would be more convenient to create a link between songs. When I have 10 new albums added to the library I have to spend several hours to just fill same data to each song. A way to share Field Values between songs is critical.

[StiX]

Now, how to realize this through expressions?
Currently there's no way to get access to other songs data from expressions. If we would had this possibility, then the problem (and I assured others have this problem too) could be solved particularly only for New Songs in "Find and Replace" Dialog by writing an Expression with specific functions that allow to gather data of other songs in library. Because for each field the Find/Replace operation should be applied separately, a possibility to perform Batch Find/Replace operations (and certainly a possibility to save Batch operations as a preset) is necessary. Example of Find/Replace operations:

• Gather all songs of currently processed song's [Artist] and apply following Artist-related fields data: [Artist Bios], [Artist Genre], [Artist Nationality]
[/li]

• Gather all songs with same [Artist]-[Name] Values compared to current song's [Artist]-[Name] and apply following Song-related Field Values: [Genre], [Genre Style], [Language], [Lyrics], [Mood].

But, we can have a problem when other "same" songs already have different [Lyrics], [Genre] or any other [Field] that we want to apply to the New Songs. In that case a Dialog should appear with a list of matched songs to allow the user to preview any of listed songs and select appropriate Value and an option to apply this selection to all of the listed songs, to put All these songs to right/same Values.

But, what if while listening to a song, we want to change/add data of one of Song-related fields ([Rating], [Tempo], [Mood], [Similarity], etc.)? We can't do this instantly, we can't change the Field data without performing additional tasks:

• Currently we have to first perform the "Context Menu > Locate > Advanced > Name" command and Fill the necessary data (Ex. [Rating]) to all related songs.
• Even by having a possibility to gather and apply data from similar songs of the Library, we can't do this instantly. We would have to Fill the Data to for Current Song, then perform the "Context Menu > Locate > Advanced > Name" command, then select all other files and perform a "Find and Replace" Batch operation to apply the data automatically to all related songs. As you can see this way is even more tedious than previous in current case (that's not true for Newly added Music).

Now, after reviewing the problem and the possible solutions, here is a Total Solution to this problem: An Option in "Options" dialog, to create groups of Fields, that share Field Values among same Fields. I propose to create following Factory Default groups (when this functionality will be realized):

• Same Album Artist (auto): [Artist Bios], [Artist Genre], [Artist Nationality]
• Same [Album Artist (auto)]-[Album]: [Album Notes], [Album Type], [Album Rating]
• Same [Artist]-[Name]: [Chart], [Genre], [Genre Style], [Language], [Lyrics], [Mood], [Preference], [Rating], [Situation], [Subject], [Tempo]

Of course there should be examples to each of the Grouping in Help File. I could do this if time will come.

By linking/grouping Fields, there would be no need to perform any of represented tasks. It would be enough to easily change Field Values of currently playing song without worrying about duplicate/similar songs as changes would as well reflect to other songs Field Values. Some concluding examples:

• After changing [Artist Bios], all songs of Same [Album Artist (auto)] will get this Artist-related data automatically
• After changing [Lyrics] or [Tempo] of a song, all songs of Same [Artist]-[Name] will have this same Song-related data automatically
• After changing [Album Notes] or [Album Rating] of a song, all songs of Same [Album Artist (auto)]-[Album] will gather this same Album-related Data automatically

As a result, independently of how and where you are browsing your library, you will not have to worry about duplicates. Of course in case if you have same [Artist]-[Name] songs but with different content, you have to add another [Version] Field to particular songs, to differentiate them. Values examples for such Field: "Remix", "Club Mix", "Long", "Jazzy", etc. So that you can create additional Field groups. Ex.

• Same [Artist]-[Name]: [Language], [Lyrics]
• Same [Artist]-[Name]-[Version]: [Chart], [Genre], [Genre Style], [Mood]

Also, by doing this, you have to change your "[Name]" Column with an expression like

Code: [Select]

"Clean([Name] /([Version]/) If(IsEqual([Album Soundtrack],),,/(from [Album Soundtrack Type] "[Album Soundtrack]"/) ) )"
Of course a Field can't be added simultaneously to several Field Groups. Design of "Customize Toolbar" Dialog can be used for this.

Perhaps this functionality should be added to MC12 alpha 

[StiX]

I want to extend my opinions about Fields Grouping and what else should be introduced in MC12. I wont create new topics for these feature requests, as everything now should spin around DBExpressions (they should be available everywhere).

Let's start from Top of Designing the Structure of New Major Version of the MC. First of all Fields should be separated to several groups: Artist, Album and Track-related data,

• For [Artist]-related Data Fields Group, or let's call it Artist-related Database a separate internal database should exist. Because of possible Identically Named Artists, each "Different" Artist should have some additional unique characteristics in Artist-related Group of Fields, like [Artist Birthday], [Artist Bios], [Artist Country], [Artist Native Language], [Artist Genre], etc. These Fields should by default not be saved into mp3 Tags (there's not much sense in doing this, as no other software can read Metadata written by MediaCenter), and all Artist-related fields should behave as a whole (as One Field) for Each Different Artist. When a respectable user will want to separate two artists in it's Library, he will need to fill some of the Artist-related fields and a special so called [Artist Alias] field, which will be used for Viewing purposes, so that artists can be distinguished in Comboboxes, Views. Some examples of Identically named Bands:
• [Artist] "Aurora" Mid 00's German Trance
• [Artist] "Aurora" Late 90's US Breakbeat
• [Artist] "Aurora" Early 90's UK Hardcore Adrenalin Records
• [Artist] "Aurora" Late 80's US Deathrock
• For [Album]-related Data Fields Group or let's call it Album-related Database also a separate internal database should exist: a Group of Album-related Fields for each [Artist]-[Album] ([Album Artist (auto)]-[Album]). The [Artist] here is also a Live Linked Tag to [Artist] from Artist-related Database. Fields like [Album Notes], [Album Rating], [Album Artist], [Album Year], etc.

By doing this Field Grouping, there'll be no need to *EVER* search, select all songs of same particular Artist or Album when it is needed to change some Artist/Album-related data, and most importantly, after these additions, the main Database will have some more freedom and we should receive some more speed improvements when working with Large Libraries. Of course there should be also a possibility to perform editing in classic (current) track-by-track manner (in case of mistypings or fraud data).


When Music of some particular Artist or Album is Deleted from Library, all the information should be retained in Library and by will it could be reviewed at any time, even if the corresponding music files aren't in the Library, because both Artist and Album-related fields will be in separate Additional Databases.


Additionally to Field Grouping, finally some new Field Data Types should be necessarily added:

• [URL]
• [Expression] (this one can be realized as a replacement to Calculated Data Template)
• [HTML Notes] or [Notes (HTML)]
  A simple HTML Editor should be included with following abilities:
• Ability to format Text, import Images, add URLs to HTML Notes.
• Ability to put Live Links to Artists/Songs/Albums/Documents/Images/Video (relative to current Artist/Song/Album) in Database. Links should have a form kinda mc://songcode which will automatically turn to Live [Artist]-[Name]-[Album]-[Year] Data from Library (defaults could be configured in Options), and additionally a link should have a Rule or Expression property for resulting String (iow. Link Text) similarly to what is now used in View Schemes and plus some popup text should be generated to this HTML page with beneficial information (which could be configured in Options)
• Ability to perform regular operations from context menu on these links.
• Relational [Fields] Currently there is no easy way to Build a Real Relational Data Type Tree. (perhaps I use not the best naming convention for what it is intended to do).  Will explain the purpose of this Field Data on example
• [Genre] > [Genre Style] > [Genre SubStyle]
  Currently to see only the [Styles] of some particular [Genre], we would need to create for each genre an additional Field with following predefined values:
• [Genre Style (Easy Listening)]: Acoustic;Bachelor Pad;Crooners/Vocals;General Easy Listening;Lounge;Love Songs;Mood Music;Musicals/Broadway;New Age;Soft Rock;Spiritual
• [Genre Style (Electronica)]: Ambient;Ambient > Abstract;Breakbeat/Breaks;Breakbeat/Breaks;Dance;Dance;Down Tempo;Drum & Bass;Electronic Cover Songs;Electronica;Ethnic;Experimental;Hardcore;House;...
• [Genre SubStyle (Electronica > Breakbeat/Breaks)]: Big Beat (Chemical beat);Funky Breaks;Hardhop;UK Breakbeat/Nu Skool
• [Genre SubStyle (Electronica > Dance)]: Club;Euro;Hard;Electro Freestyle
• [Genre SubStyle (Electronica > ...)]: ...
• ... and so on, until we will have additional, over 100 Brand New Genre-related Fields. That's not right.
• [Similarity (Soundtrack)] > [Similarity (Artist)] > [Similarity (Name)]
• Linked [Fields] Values. Comboboxes for ALL of following fields should contain same list of Values:
• [Artist]; [Album Artist]; [Track Featured Artist]; [Track Modification Artist]; [Similarity (Artist)]; [Similarity (Name)]
• [Preference]; [Preference (artist)]; [Preference (album)]
• [Tempo]; [Tempo (artist)]; [Tempo (album)]
• [Date (year)]; [Album Year]
• [Nationality]; [Language]

[EpF]

I find that hard to believe..!

 

[StiX]

I find that hard to believe..!

 

Oh  , well, on this board, yes

Last two months I spent on performing a "4th generation structurization" (in my terms). This long time was used to standardize all Fields: their Naming, Values, Relations. Many MC Expressions was developed. Main target with Expressions in Columns was to simultaneously view on screen Field Values of almost 30 Fields. Because currently impossible to put icons or characters from other fonts into output line, I had to use acronyms for designations. Here is a random tracks example:

Code: [Select]

Artist =Name Duration Bitrate =Alb.Year =General =Extended
3T Anything   [*] 4:21 256 '96 [Eng]  CHAN Ball [A#+C! AA Eng Ball] *15 LyrMoSimSitSuT
Celine Dion Because You Loved Me  [S] 4:32 256 '96 []  CHAN Ball [A#+C!S AA Eng Ball] *01¤4°4 LyrMoPuSitSuT
Captain Hollywood Flying High [C]  4:40 218 [] (VBR) ELEC [C] 
Francesco Napoli Stella [C]  3:07 256 04|''' [Eng]  ELEC Elec>Dance > Euro [C! Eng Elec]  Cm
Francis Goya Gipsy    2:34 160 [Eng]  INST Inst>Folk   ISu°
Nicolas de Angelis La valsa    2:33 112 [Lat] ('Low B') INST Inst   ISu°
Scorpions Send Me an Angel    4:31 192 '91|00 [Eng]  ROCK Rock [A#+] *44 DupLyr
The Bangles Eternal Flame [C]  3:40 192 '89 [Eng]  CHAN Ball [C] *01 LyrMoSitSuT
Êðàñíàÿ ïëåñåíü Àðèÿ àãåíòà 007    3:53 192 00 [Rus] ◊ HUMO Humo>Parodies [A#+ N]  Comp
Ñìûñëîâûå Ãàëëþöèíàöèè Çà÷åì òîïòàòü ìîþ ëþáîâü    4:21 256 00|''' []  POP Pop [A#+] ¤4°4 MoSitSuT
[Karaoke] Ñîôèÿ Ðîòàðó - ×åðâîíà ðóòà    2:52 256 01|''' [Rus]  CHAN Inst [A#+ AA Oldi] 

Also a lot of time was spent in MusicBrainz http://wiki.musicbrainz.org/wiki.pl?HomePage which is definitely a "way to the future" in music Tag (Field) Values categorization standards. They have a Database of near-consistently typed 3144626 Track Names, which is a replacement to other databases like FreeDB, Amazon, etc. Authors of MC should use MusicBrainz service in MediaCenter. The service is totally free.

2004-07-16
French computer magazine Clubic has an [review of the MusicBrainzTagger] as an alternative to MoodLogic. (En Français)

[MoodLogic] is a project that is similar to MusicBrainz in some ways. They have created a gigantic database with song information that can be used to correct ID3 tags and also to create dynamic playlists.

The company has been slowly but surely falling into disrepute with its users due to a perceived neglect of their requests. --PetervanHardenberg

Which is a point where MusicBrainz could come in and fill the gap: MusicBrainz is open source and made by the users. And the AdvancedRelationships model would allow for IntelligentPlaylistGeneration. --DonRedman

Following URL is for MediaCenter Developers:
http://wiki.musicbrainz.org\wiki.pl?ClientLibrary


Some of my proposals on bearing with fields, which could be applied to a Database Software, such as MC, are already invented as a web service at MusicBrainz. Others was born after reading hundreds of pages at MusicBrainz.

AdvancedRelationshipType:
http://wiki.musicbrainz.org/wiki.pl?AdvancedRelationshipType

Code: [Select]

Artist  Album             Track 
Artist  [Artist--Artist]  -              - 
Album   [Album--Artist]  [Album--Album]  - 
Track   [Artist--Track]  [Album--Track]  [Track--Track] 
Fields relationships above, are different from what I propose. I propose Field Grouping and a Tree-like parent-child-sibling Fields Relationships (iow. Explorer-like).

Music Data Relationships at MusicBrainz are historical-biographical-cover-genre-instrument relationships between various Artists, Albums and Tracks. Surely, this model and the database should be realized and used in MC, because there's no chance that MoodLogic or any other service will be *ever* free... This data should be accessible in what I called [HTML Notes]. Surely, there's needed much more details to create an outline of how everything should cooperate with each other, but currently I don't see *any* concernment from J River about new features.


Current version of Picard Tagger v0.4.0 https://helixcommunity.org/download.php/653/picard-setup-0.4.0.exe overwrites all MediaCenter IDs with TRM IDs (TRM Recognizes Music), that represent the audio signature of a musical piece, so currently the service can't be easily used in cooperation with MC.

[JimH]

Also a lot of time was spent in MusicBrainz http://wiki.musicbrainz.org/wiki.pl?HomePage which is definitely a "way to the future" in music Tag (Field) Values categorization standards. They have a Database of near-consistently typed 3144626 Track Names, which is a replacement to other databases like FreeDB, Amazon, etc. Authors of MC should use MusicBrainz service in MediaCenter. The service is totally free.Following URL is for MediaCenter Developers:
http://wiki.musicbrainz.org\wiki.pl?ClientLibrary

Musicbrainz has 3,144,626 tracks.

Our database, YADB, has 935,610 CD's or about 9,356,100 tracks.

http://www.yadb.com/stats.html

If both were entered by users, they're likely to be similar in quality.

Stix,
Keep posting.  We're reading.  Thanks for your interest.

[StiX]

It's my pleasure.

MusicBrainz
1. Advanced Moderation of Data.
http://wiki.musicbrainz.org\wiki.pl?AdvancedModeration
Perhaps from 10 million entries in YADB less than half is in correct state.
2. Artist Aliases.
http://wiki.musicbrainz.org/wiki.pl?ArtistAlias
3. Foreign Language Support
http://wiki.musicbrainz.org/wiki.pl?InterNationalization
I couldn't retrieve information for my Russian/Ukrainian music from YADB.

4. Creates TRMs: The Universal Barcode for Music and Media from Relatable®
http://www.relatable.com/tech/trm.html

TRM is an audio fingerprinting technology that generates a unique fingerprint for an audio file based on an analysis of the acoustic properties of the audio itself. Each audio fingerprint is unique and can be used to identify a track precisely, regardless of whether any associated text identifiers are present or accurate.

YADB is even not close to this.

And, finally...
5. Intelligent Playlist Generation (ala MoodLogic) by using TRMs
http://wiki.musicbrainz.org\wiki.pl?IntelligentPlaylistGeneration

Any rejections?

[JimH]

TRM is an audio fingerprinting technology that generates a unique fingerprint for an audio file based on an analysis of the acoustic properties of the audio itself. Each audio fingerprint is unique and can be used to identify a track precisely, regardless of whether any associated text identifiers are present or accurate.

YADB has used a similar technique for track look-up for several years.

What you say about Musicbrainz is interesting, but we think that YADB is a very good solution.  It has been more reliable and accurate than our previous one.

[StiX]

Sorry for my ignorance of the YADB info.

But, the disadvantages of YADB are clear, and you can't say nothing:
1. Doesn't use anymore a similar technique in purpose to recognize Music.
2. Doesn't use anymore a similar technique in purpose to create MoodLogic playlists.
3. Not International
4. Nobody moderates the Data entered by users
5. No Artist Aliases (misspelled, alternative Artist/Band Titles)


MC-related disadvantages
6. MC have, or more precisely, doesn't have an UI to check the differences between imported YADB data and perhaps the more correct manually filled data (Only by pressing Ctrl+Z, Ctrl+Y repeatedly).
7. MC doesn't have an option for Plug-in developers to perform tasks on selected files in any place of the Library. You have to put the music into the Playing Now to perform tasks, such as Import Lyrics, Cover Art from Internet.

I don't want to hurt anybody, J River team made a great work on MC11, but some things should be improved. I don't mean now, just write it to your ToDo, please.