I'm looking for an easy way to catalog my mp3 files which are in an organised directory structure. I want to write a script that uses id3info to read each file in turn then output the tags to a .csv file but I can't see how to do this recursively so that it looks at each directory in turn.

find /mnt/mp3s -name *.mp3 > ~/mp3list.txt

will generate a list of filenames and paths but I can't work out how to pipe this list line by line to id3info (if this is the right approach!).

Any clues please ?

You could try the following:-
The 'tr " " "^"' and 'tr "^" " "' is in case the are any file names with spaces in, as the 'for i in' uses spaces as a delimiter.

Of course, if you are likely to encounter a "^" in a file name you should change it for something more obscure.

You might also look at an available shell program named 'mp3info'

... script worked a treat, thanks a lot.

Chas.

Glad it helped, but this one is shorter, easier and quicker:-
>Note to self - Do not post when tired.

Thanks.

I didn't really think this through enough :-)

I assumed the output would be regular and parsable but unfortunately it only returns the tags where there is a value and you cannot request particular tags. Further it (id3info) does not always return the tags in the same order !

The script plus id3info currently returns a large file comprising something like as follows ....


*** Tag information for /mnt/mp3s/blur/13/01-tender.mp3
=== TIT2 (Title/songname/content description): Tender
=== TPE1 (Lead performer(s)/Soloist(s)): Blur
=== TALB (Album/Movie/Show title): 13
=== TYER (Year): 1999
=== COMM (Comments): ()[]:
=== TCON (Content type): (17)
=== TRCK (Track number/Position in set): 1
*** mp3 info
MPEG1/layer III
Bitrate: 320KBps
Frequency: 44KHz

*** Tag information for /mnt/mp3s/blur/13/02-bugman.mp3
=== TIT2 (Title/songname/content description): Bugman
=== TPE1 (Lead performer(s)/Soloist(s)): Blur
=== TALB (Album/Movie/Show title): 13
=== TYER (Year): 1999
=== COMM (Comments): ()[]:
=== TCON (Content type): (17)
=== TRCK (Track number/Position in set): 2
*** mp3 info
MPEG1/layer III
Bitrate: 320KBps
Frequency: 44KHz

.... but if the record length is not always consistent .... the "field" order not fixed ..... and the fields not fixed (!) ..... then I'm struggling.


In truth what I really want is a nicely formated row per record :-

/mnt/mp3s/blur/13|02-bugman.mp3|Bugman|Blur|13|1999|2|MPEG1/layer III|320KBps|44KHz

i.e. PATH|FILE|TRACKNAME|ARTIST|CDTITLE|YEAR|TRACKNUMBER|FORMAT|BITRATE|FREQUENCY

... or delimeted with "," or ";" or "~" or whatever.


AWK and SED just make my brain liquify and run out of my ears, but it should be doable I think. Is this really, really hard or am I just a dunce ?

Did you check out mp3info? ... Here's the man page ... http://www.die.net/doc/linux/man/man1/mp3info.1.html

It could be done, see http://www.tgunkel.de/software/files/mp32ogg for an idea how.

However rickh is right, the best tool for the job is mp3info. If you haven't got it, its less than 50Kb. http://www.ibiblio.org/mp3info/

Thanks for the script. I knew it would work, but hadn't gotten around to figuring out the syntax yet.

Well had to figure out syntax anyway. Many of my files are ripped from vinyl, or I deleted the id3 tag for some reason. The information I want is Artist, Title, Size, Duration, and Bitrate. mp3info needs to compute it from scanning the file rather than getting it from a tag.

Works good, tho. Gives me a nicely delimited file with the info I want.

Also ... check out a program called Tellico for a library ... Near as I'm able to tell tho, you need this formatting tool to prepare data for import.

... It just works (tm) :-)

Thanks for all.