[Busta Uppa]

I am using a view for albums that includes an infopanel with "style," "type," and "mood" listed at the top.

I figured "type" should be taken care of, given that my files (alac) are all tagged with MusicBrainz Picard and include the "Release Type" tag (e.g. "single," "album," sometimes multi-value like "album; remix" etc.). However the "type" info rarely shows up. Seems like only 1 out of 20 actually list it.

The "style" and "mood" fields seem to be populated more frequently, however I cannot see any tag that this info is associated with. I assume the info has been scraped and is not actually embedded in the files.

And an odd thing I have noticed - I have yet to find an album that has all three fields filled in! Albums with "type" data never have anything listed for the other categories. "Style" seems to be the most common and is paired with "mood" data about half of the time.

My questions:

1) Are any of these data types actually embedded in the files? In the case of Release Type, I am not sure if my the sporadic data is from inconsistent tag reading or just inconsistent scraping.

2) If the data is all being scraped, is there a way to get more complete results? I already have MBID identifiers in my tags. I do not have "prefer online information" enabled since I have already used Picard and don't wish to have anything overwritten by Kodi. Are there other tags, maybe in an older format or something, that might interfere with proper scraping?

For what it's worth, I am using the "shift" albums view in the MadNox skin. I love leaving that infopanel on, but it gets annoying seeing "No info available" on like 80% of the content. If that could be remedied I would be 100% sold on these view settings. Any insight appreciated, thanks!

[jjd-uk]

The tags that Kodi currently reads can be found at:

http://kodi.wiki/view/Music_tagging#Tags_Kodi_reads

Note that page is slightly out of date as Dave has added the sort tags for v18.

From that I can see only Mood that can gets defintely read from tags, not sure if Type is related to RELEASETYPE as I seem to remember a discussion about there being two different things named Type, as for Style hadn't even heard of that so guess that could be scraped only.

[DaveBlake]

Album "style" and "mood" are not from tags, but are scraped from either album.NFO files or online sources such as TADB. Having mbid tags will increase the accuracy of your scraping, but not all albums have these values available.

The mood tag is used for song mood, but not for album.

In v17 onwrards album type comes from the RELEASETYPE tag, but can be overwritten by scraped values.

The sporadic completion of these fields you experience is just the nature of the online data sources. Maybe update TADB, it is a community effort. Or put the values you want into NFO files.

[Powerhouse]

If you want Style and Mood tags, there is an add-on for Picard that gives you these.

https://picard.musicbrainz.org/plugins/

I'ts the Last.fm.Plus

[Busta Uppa]

Thanks for the tips everybody. Sounds like that Picard plugin will come in handy. Plus I may contribute to TADB or start editing my NFO files... hooray for options! Thanks!

[meowmoo]

Isn't last.FM plus broken since the api change?

You need last.fm NG plugin

[Powerhouse]

Here is the fixed version of Last.FM.Plus

https://community.metabrainz.org/t/last-...on/61041/3

[docwra]

Filling in this data on TheAudioDB helps everyone.

Its pretty cool to sit down in the evening and select some "relaxed" music

Let me know if you want any extra moods or styles added to the TADB lists.

[DaveBlake]

Using Confluence I can see the type, style and mood etc., so this is most likely a skin issue, ask on the skin forum.

EDIT: So I have split and moved the post to http://forum.kodi.tv/showthread.php?tid=315482

[Busta Uppa]

Thanks for sharing this on the Aeon skin forum, DaveBlake!

However I have reason to believe this is an issue that persists across skins.

I did an experiment with "Aglio e Olio" (EP) by Beastie Boys. Here is the album as it appeared for me in Madnox. It exhibits my usual problem - type is properly filled in, with no additional album info:




Now here is the album info as it appeared in Confluence. Same as what I got in Madnox:




I did a manual refresh of album data - now the rest of the album data is shown, but type has been removed!




And now we return to my original Madnox view - I have the rest of the album info that was recently scraped, but "ep" is no longer listed under type (same as where I left off in Confluence):





So even in Confluence skin, it seems like album type info cannot coexist with the mood/style/description.

[DaveBlake]

Co-existence is no problem (in Confluence at least), look all the fields here



What is probably happening is that refreshing (and thus scraping again) has simply removed the "type" value. Although it is initially derived from the album (release) type tag, it is also a value that can be scraped.

[Busta Uppa]

Point taken on co-existence. The more likely conflict, now that I think about it, is between the scraped info and my own embedded tags. While my files do not contain any mood or style info, they have all been tagged with "Release Type" in Picard.

So in my examples it looks like Kodi initially displays only what it finds in tags (release type). Then when I force a refresh, it displays only the online scraped info (style, mood, details/review; omitting release type).

What are the proper settings for Kodi to show embedded info and then "fill in the gaps" with anything else scraped online? I really thought I had all my scraping set up this way, but I must have misunderstood some of the options. While I do wish to eventually tag everything with mood and style info (I prefer to have as much metadata stored locally as possible), I'd still want to see some online scraped info, to account for album descriptions and such.

[jjd-uk]

My guess would have been that just filling in the gaps from tags or doing a complete overwrite in the db would have been controlled by the "prefer online information" but since you don't have that enabled (I assume what you wrote in the first post still applies) then maybe a fill in the gaps only is not possible.

[Busta Uppa]

My guess would have been that just filling in the gaps from tags or doing a complete overwrite in the db would have been controlled by the "prefer online information" but since you don't have that enabled (I assume what you wrote in the first post still applies) then maybe a fill in the gaps only is not possible.

I'm starting to wonder if I should just go with "prefer online information". How aggressively does it overwrite tags? Is embedded album art overwritten too? I've put a lot of effort into my tagging so I'm hesitant to just hand it all over.

[DaveBlake]

Point taken on co-existence. The more likely conflict, now that I think about it, is between the scraped info and my own embedded tags. While my files do not contain any mood or style info, they have all been tagged with "Release Type" in Picard.

Let's be very clear (again):
1) the mood tag is taken as the song mood, the album mood is not taken from any tag values, it is only set by scraping either an NFO file or from an online source that provides album mood.
2) There is no "style" tag, songs do not have a style. Album style is set by scraping either an NFO file or from an online source that provides album style.
3) The "Release Type" tag in Picard is initially applied to the album as the album type. However it can also be scraped, and any scraped value will replace the one that was derived form the tags.

So in my examples it looks like Kodi initially displays only what it finds in tags (release type). Then when I force a refresh, it displays only the online scraped info (style, mood, details/review; omitting release type).

No. On scanning files into the library of album style, mood and type the only values you have are those for type. Album style and mood are blank because they do not come from tags.
On scraping (from NFO or online sources) you can gain values for album style, mood and type.

What are the proper settings for Kodi to show embedded info and then "fill in the gaps" with anything else scraped online?

"Prefer online information" disabled, "Fetch additional info on update" disabled, and then "query info for all" from context menu to scrape and additional data. "Server unavailable" errors my mean you do this many times e.g. last time I tried it took 13 tries to get all the info for 850 albums.

I really thought I had all my scraping set up this way, but I must have misunderstood some of the options. While I do wish to eventually tag everything with mood and style info (I prefer to have as much metadata stored locally as possible), I'd still want to see some online scraped info, to account for album descriptions and such.

You can tag songs with mood if you like, it will show for the song. There is no "style" tag. Whan these things for album locally then you will need to use NFO files.

I'm starting to wonder if I should just go with "prefer online information". How aggressively does it overwrite tags? Is embedded album art overwritten too? I've put a lot of effort into my tagging so I'm hesitant to just hand it all over.

"Prefer online information" enabled does not overwrite tags or change your music files in any way, it simply replaces some of the album data in the music library that was derived from tags. Namely album genres (just the text not the values used in navugation), Year, Compilation and Label, but it will also change the artists if the scraper mis-identifies the album.

I have recently discovered that this setting from originally intended for those with all music fully tagged with mbids to be able to keep updated with any Musicbrainz database changes without having to use Picard again. This isn't quite how it works, but it will do in v18. Unless you want total automation (no personal control), and have tagged everything with mbids, I suggest this is left disabled.