What scraper are you using - or, specifically, which site is providing the metadata?
What happens if you go directly to that site and try searching?
What happens if you switch debugging on, what does the log say? You should see the query (or the response from the scraper, anyway) which may give you some pointers.
(2014-02-21, 18:06)Prof Yaffle Wrote: [ -> ]What scraper are you using - or, specifically, which site is providing the metadata?
What happens if you go directly to that site and try searching?
What happens if you switch debugging on, what does the log say? You should see the query (or the response from the scraper, anyway) which may give you some pointers.
The problem is not in the scraping part. I can manually search and detect the right album (because for xbmc my cd doesn't have an artist so it can't set it automatically). But xbmc doesn't recognize the id3 tag. xbmc can read easily the 2nd audio file artist without to scrape the info, but it can't for the dido one even if they're both tagged with musicbrainz and they have the same tags (for mediainfo and mp3tag)
Since you don't provide a debug log, it is hard to tell what the problem is.
I did scan the flac file (note flac does not have ID3, rather Vorbis comment) using XBMC 13 20140214 (so a couple day older build) and looking at my debug log, I see that the musicinfoscanner did find both the album and alub artist musicbrainz id entries, and provided those to Universal Album Scraper and Universal Artist Scraper which then found scrape data correctly, so I can't duplicate your problem.
Really need a full debug log uploaded to pastebin or xbmclogs that shows scanning that album. Here's some pertinent data from my log:
Code:
15:33:02 T:7124 DEBUG: MUSIC_INFO::CMusicInfoScanner::DoScan Scanning dir 'F:\testmusic\artist test\album test\' as not in the database
15:33:02 T:7124 DEBUG: unrecognized XipComment name: ACOUSTID_ID
15:33:02 T:7124 DEBUG: unrecognized XipComment name: ALBUMARTISTSORT
15:33:02 T:7124 DEBUG: unrecognized XipComment name: ARTISTSORT
15:33:02 T:7124 DEBUG: unrecognized XipComment name: ASIN
15:33:02 T:7124 DEBUG: unrecognized XipComment name: BARCODE
15:33:02 T:7124 DEBUG: unrecognized XipComment name: CATALOGNUMBER
15:33:02 T:7124 DEBUG: unrecognized XipComment name: DISCTOTAL
15:33:02 T:7124 DEBUG: unrecognized XipComment name: ISRC
15:33:02 T:7124 DEBUG: unrecognized XipComment name: LABEL
15:33:02 T:7124 DEBUG: unrecognized XipComment name: MEDIA
15:33:02 T:7124 DEBUG: unrecognized XipComment name: MUSICBRAINZ_RELEASEGROUPID
15:33:02 T:7124 DEBUG: unrecognized XipComment name: NOTES
15:33:02 T:7124 DEBUG: unrecognized XipComment name: ORIGINALDATE
15:33:02 T:7124 DEBUG: unrecognized XipComment name: RELEASECOUNTRY
15:33:02 T:7124 DEBUG: unrecognized XipComment name: RELEASESTATUS
15:33:02 T:7124 DEBUG: unrecognized XipComment name: RELEASETYPE
15:33:02 T:7124 DEBUG: unrecognized XipComment name: SCRIPT
15:33:02 T:7124 DEBUG: unrecognized XipComment name: TOTALDISCS
15:33:02 T:7124 DEBUG: unrecognized XipComment name: TOTALTRACKS
15:33:02 T:7124 DEBUG: unrecognized XipComment name: TRACKTOTAL
15:33:02 T:7124 DEBUG: CAnnouncementManager - Announcement: OnUpdate from xbmc
This shows file is found and some tags aren't read (normal).
Code:
15:33:02 T:7124 DEBUG: GOT ANNOUNCEMENT, type: 32, from xbmc, message OnUpdate
15:33:02 T:7124 DEBUG: MUSIC_INFO::CMusicInfoScanner::UpdateDatabaseAlbumInfo downloading info for: Life for Rent
15:33:02 T:7124 DEBUG: scraper: ResolveIDToUrl returned <url>http://musicbrainz.org/ws/2/release/62bef034-b779-30dd-bbfd-a2035d0e0f74?inc=recordings+release-groups+artists+labels+ratings</url>
15:33:04 T:7124 DEBUG: -- nfo-scraper: Universal Album Scraper
15:33:04 T:7124 DEBUG: -- nfo url: http://musicbrainz.org/ws/2/release/62bef034-b779-30dd-bbfd-a2035d0e0f74?inc=recordings+release-groups+artists+labels+ratings
This shows album MBID was read and used for the Universal Album Scraper.
Code:
15:33:06 T:7124 DEBUG: MUSIC_INFO::CMusicInfoScanner::UpdateDatabaseArtistInfo downloading info for: Dido
15:33:06 T:7124 DEBUG: scraper: ResolveIDToUrl returned <url>http://musicbrainz.org/ws/2/artist/d1353a0c-26fb-4318-a116-defde9c7c9ad?inc=url-rels</url>
15:33:08 T:7124 DEBUG: -- nfo-scraper: Universal Artist Scraper
15:33:08 T:7124 DEBUG: -- nfo url: http://musicbrainz.org/ws/2/artist/d1353a0c-26fb-4318-a116-defde9c7c9ad?inc=url-rels
This shows artist MBID was read and used for Universal Artist Scraper.
scott s.
.
Track works fine in current nightly, so I've probably already fixed it.
I found out the problem. In the folder there was also this cue file:
https://dl.dropboxusercontent.com/u/2219...202003.cue
I didn't know that xbmc looks to cue before and ignore the files in it if it finds it (even musicbrainz tags). Also when you open the folder it uses the cue data.
I opened the file and i found what was the problem.
My cue had this:
But without quotes it fails to read the artist name because it wants this:
Is this a bug? All my other cue files were with quotes but xbmc should be able to handle this easily
Never seen a .cue sheet like that before.
I doubt anyone will take the time to code up support for that. A quick look shows it's not completely trivial (i.e. we depend on the quotes being there atm).
See
https://github.com/xbmc/xbmc/tree/master...cument.cpp if you want to play. You'll want to mess around with the calls to ExtractQuoteInfo to first knock the bit off the front of the string, then trim for whitespace then trim for quotes. This will likely be less reliable, so I'd suggest first parsing quotes if available.
(2014-02-22, 04:35)jmarshall Wrote: [ -> ]Never seen a .cue sheet like that before.
I doubt anyone will take the time to code up support for that. A quick look shows it's not completely trivial (i.e. we depend on the quotes being there atm).
See https://github.com/xbmc/xbmc/tree/master...cument.cpp if you want to play. You'll want to mess around with the calls to ExtractQuoteInfo to first knock the bit off the front of the string, then trim for whitespace then trim for quotes. This will likely be less reliable, so I'd suggest first parsing quotes if available.
I don't know if it's ok with your coding methods but i could rename the function from "ExtractQuoteInfo" to "ExtractInfo" and do the parsing in the function. If the string contains 2 quotes i can return true with the string between quotes (exactly like now), if it doesn't have any i can return true with the full string and if it only have one i can return false...
The cue file always calls ExtractQuoteInfo so it should work for all the tags in the cue file
If this is okay i can do it..
That would be OK, but note that ExtractQuoteInfo currently takes the full line, so you may need to supply a bit more info to the function.
I miss that.. I can pass the string with the tag name already removed or add a new variable to the function containing the string to remove or the starting index. Probably the first is the most logical solution. What do you think?
I'd just run a .substr() on the path you pass in to ExtractQuoteInfo, similar to what is done for ExtractNumericInfo.
(2014-02-22, 05:41)jmarshall Wrote: [ -> ]I'd just run a .substr() on the path you pass in to ExtractQuoteInfo, similar to what is done for ExtractNumericInfo.
Exactly what i was thinking.
Here's my result
https://github.com/phate89/xbmc/commit/8...f6d4585cc8
I also found that some tag was already working without quotes with some check. I removeD that checks and move the handling in the function (i also changed the type to string because before it was bool only to check if there were quotes before and after).
I've made some comments. Fix it up and get it working and feel free to fire off a PR
(2014-02-22, 23:09)jmarshall Wrote: [ -> ]I've made some comments. Fix it up and get it working and feel free to fire off a PR
Done. Thanks for your help
Hi guys!
Some of your posts are to technical for me (or better I am lazy to try to understand ;-).
Anyway, I have described similar "problems" with .CUE files in my post here:
http://forum.xbmc.org/showthread.php?tid=185872
From my point of view it is/was not a problem of quotes. XBMC during the metadata scan from audiofiles alway add "Unknown" artist to albums with .CUE file present in its directory. My (temporary) solution is to rename all .CUE files to e.g. .TXT extension, import metadata and afterward rename it back to original... :-)
Will this be solved with Gotham?