2014-02-10, 14:11
Hi there!
I have choosen XBMC as it seems that it will fit my visions about the system which will be able to manage my multimedia collections... :-) But before I let it to do so, I like to play with it to better understand its priciples, logic and functions. I have also read several forum threads, wiki and other sources and here are my first notes.
First of all I have tried to understand workflows related to metadata processing, as I would like to prepare my collections for it. It seems that there are two main steps:
1. Reading/fetching/digging metadata from media files;
2. Scraping (additional) data from the web.
This workflow attitude make sense to me, as the basic metadata are mostly the same for all media, but details are „region-sensitive“ - someone would like to have descriptions in local language, some „national“ content is not well worldwide known, etc. ;-)
I don't want to go deep into proceses of creation video (movies, TV series, ...) collections (yet). It seems that XBMC just take the filename and go directly to choosen scraper to try to get more data. There is nothing complicated and no support to read metadata from video files/containers (e.g. Matroska) – that is what I have understand from several forum threads...
Afterward I have made some tests with audiofiles – I have found out several oddities. (OK – no current support for Matroska container as mentioned above...). It seems that there are several ways how to provide XBMC with audio metadata. I think that this is the order sorted by importance:
- .NFO file: XML file which can provide all needed metadata (incl. album and artist details).
- .CUE file: TXT file with basic metadata and „play-list“.
- Tags (MP3, OGG/Vorbis): As a part of each audio-file.
I am not sure whether the XBMC is going through these files in this particular order and whether there are any rules what can overwrite what data (I know, there is a possible parameter „clear“ within .NFO XML scheme). I have tested some FLAC and some MP3 files with .NFO and/or .CUE files within the album directory but I don't understand it fully...
Before I will put here my questions, some details and background:
I would like to use FLAC for my audio collections (mostly) – because it is loseless and I can have some albums as one file with .CUE list which is supported by XBMC (instead of Matroska MKA with chapters which is not supported :-( – only MKV is able to work with chapters – I have tried it...). This is because there are many albums without gaps between songs and I hate to divide it into several files (like some soundtracks, electronic music, classical records etc.).
I have played with the current XBMC 12.3 Frodo (on Win and also as XBMCbuntu instaled on USB stick). Mostly I have used fresh and clear instalation ;-)
So...
1. I have had a problems with .NFO files: I was not able to push XBMC to load any data from my .NFO file. I have tried several file names (album.nfo, [album name].nfo). I have also tried to export library data scraped from web (Queen: Innuendo) to have the proper .NFO file but without any success. Is this OK or is it a bug?
2. Import metadata from .CUE file works fine but beside the the right artist it is generating also an „Unknown“ artist where are listed all albums with .CUE file in their directory. I think that XBMC is not reading album metadata from the .CUE file – info about artist is taken from „Performer“ for each song separately...
3. I have noticed no problems with metadata directly from the media files. But these are loaded only if no .CUE file is present. (And also everytime, when I have tried provide it via .NFO file.)
4. Last question for now: During my testing I have never reached a situation where XBMC should automatically load artist and album data using scraper (similar to video). Every time I should manualy click on author/album and run the scraper. Why? :-)
If it is needed, I can provide you with more details and samples...
I have choosen XBMC as it seems that it will fit my visions about the system which will be able to manage my multimedia collections... :-) But before I let it to do so, I like to play with it to better understand its priciples, logic and functions. I have also read several forum threads, wiki and other sources and here are my first notes.
First of all I have tried to understand workflows related to metadata processing, as I would like to prepare my collections for it. It seems that there are two main steps:
1. Reading/fetching/digging metadata from media files;
2. Scraping (additional) data from the web.
This workflow attitude make sense to me, as the basic metadata are mostly the same for all media, but details are „region-sensitive“ - someone would like to have descriptions in local language, some „national“ content is not well worldwide known, etc. ;-)
I don't want to go deep into proceses of creation video (movies, TV series, ...) collections (yet). It seems that XBMC just take the filename and go directly to choosen scraper to try to get more data. There is nothing complicated and no support to read metadata from video files/containers (e.g. Matroska) – that is what I have understand from several forum threads...
Afterward I have made some tests with audiofiles – I have found out several oddities. (OK – no current support for Matroska container as mentioned above...). It seems that there are several ways how to provide XBMC with audio metadata. I think that this is the order sorted by importance:
- .NFO file: XML file which can provide all needed metadata (incl. album and artist details).
- .CUE file: TXT file with basic metadata and „play-list“.
- Tags (MP3, OGG/Vorbis): As a part of each audio-file.
I am not sure whether the XBMC is going through these files in this particular order and whether there are any rules what can overwrite what data (I know, there is a possible parameter „clear“ within .NFO XML scheme). I have tested some FLAC and some MP3 files with .NFO and/or .CUE files within the album directory but I don't understand it fully...
Before I will put here my questions, some details and background:
I would like to use FLAC for my audio collections (mostly) – because it is loseless and I can have some albums as one file with .CUE list which is supported by XBMC (instead of Matroska MKA with chapters which is not supported :-( – only MKV is able to work with chapters – I have tried it...). This is because there are many albums without gaps between songs and I hate to divide it into several files (like some soundtracks, electronic music, classical records etc.).
I have played with the current XBMC 12.3 Frodo (on Win and also as XBMCbuntu instaled on USB stick). Mostly I have used fresh and clear instalation ;-)
So...
1. I have had a problems with .NFO files: I was not able to push XBMC to load any data from my .NFO file. I have tried several file names (album.nfo, [album name].nfo). I have also tried to export library data scraped from web (Queen: Innuendo) to have the proper .NFO file but without any success. Is this OK or is it a bug?
2. Import metadata from .CUE file works fine but beside the the right artist it is generating also an „Unknown“ artist where are listed all albums with .CUE file in their directory. I think that XBMC is not reading album metadata from the .CUE file – info about artist is taken from „Performer“ for each song separately...
3. I have noticed no problems with metadata directly from the media files. But these are loaded only if no .CUE file is present. (And also everytime, when I have tried provide it via .NFO file.)
4. Last question for now: During my testing I have never reached a situation where XBMC should automatically load artist and album data using scraper (similar to video). Every time I should manualy click on author/album and run the scraper. Why? :-)
If it is needed, I can provide you with more details and samples...