Trying to untangle ID3-tags for classical music
Ilja Nieuwland
First of all, I did a google search to see whether anyone had tackled
this problem before; I may have simply missed it, so please correct me
if I did.
Here goed: I use an Apple iPod to listen to classical music on the road.
I frequently rip CD's (or parts thereof) using Apple's iTunes software.
It does the job admirably, with only a little sound deformity caused by
the MP3 encoding (when I use 192Kbps encoding). However, there is a
problem. When I insert a CD, it will connect to the CDDB database to
collect its track information. Great service, but more often it stinks
in practice.
What's the problem? Well, the main one is that no-one seems to have
seriously thought about which information to present where in the ID3-
tag (the tag containing the MP3's information). As a consequence, the
'artist' field, for instance, sometimes contains the name of the
performers, sometimes that of the piece, sometimes that of the movement,
and sometimes that of the composer (even though newer incarnations of
the ID3 tag have a 'composer' field). To further complicate matters,
different formats are used as well as different languages.
To make a long story short, most of the time I find myself typing the
CD's data just in order to be able to find the piece I want on the iPod.
Wouldn't it be convenient if someone was to devise a standard format for
classical ID3 tagging, using the same language (probably English) and
the same way of spelling names, symphonies, etc. I know that this all
sounds very administratively dull and will be difficult in the light of
the internet's anarchic nature, but I'd be willing to give it a try. If
anyone else thinks this is a good idea, please contact me over e-mail (
or here).
Regards,
Ilja Nieuwland
Groningen
Netherlands
p.s. typed while listening to Gunzenhauser's excellent Taneyev 4th.
Tony Movshon
> When I insert a CD, it will connect to the CDDB database to
> collect its track information. Great service, but more often it stinks
> in practice.
>
> What's the problem? Well, the main one is that no-one seems to have
> seriously thought about which information to present where in the ID3-
> tag (the tag containing the MP3's information). As a consequence, the
> 'artist' field, for instance, sometimes contains the name of the
> performers, sometimes that of the piece, sometimes that of the movement,
> and sometimes that of the composer (even though newer incarnations of
> the ID3 tag have a 'composer' field). To further complicate matters,
> different formats are used as well as different languages.
>
> To make a long story short, most of the time I find myself typing the
> CD's data just in order to be able to find the piece I want on the iPod.
> Wouldn't it be convenient if someone was to devise a standard format for
> classical ID3 tagging, using the same language (probably English) and
> the same way of spelling names, symphonies, etc. I know that this all
> sounds very administratively dull and will be difficult in the light of
> the internet's anarchic nature, but I'd be willing to give it a try. If
> anyone else thinks this is a good idea, please contact me over e-mail (
> or here).
Agreed, but the problem is completely intractable. CDDB is a miraculously
clever thing, but its database is entirely user-created and the structure of
the database fields is ill-defined for classical music. It is immune to
centralization.
Usually some approximation to the information you want is there in a CDDB
entry, though randomly placed in different fields. I sometimes do a thing that
you can emulate -- when I reorganize or add to an entry, I upload it back to
CDDB. So at least any disk you play that has passed through my (digital) hands
is likely to have a more rational structure.
I'm especially fond of CDDB entries for Japanese disks. They're -really-
helpful ...
Tony Movshon
mov...@nyu.edu
Matthew B. Tepper (posts from uswest.net are forged)
letters to be typed in news:3E9C008...@nyu.edu:
> Ilja Nieuwland wrote:
>> When I insert a CD, it will connect to the CDDB database to collect its
>> track information. Great service, but more often it stinks in practice.
>>
>> What's the problem? Well, the main one is that no-one seems to have
>> seriously thought about which information to present where in the ID3-
>> tag (the tag containing the MP3's information). As a consequence, the
>> 'artist' field, for instance, sometimes contains the name of the
>> performers, sometimes that of the piece, sometimes that of the
>> movement, and sometimes that of the composer (even though newer
>> incarnations of the ID3 tag have a 'composer' field). To further
>> complicate matters, different formats are used as well as different
>> languages.
>>
>> To make a long story short, most of the time I find myself typing the
>> CD's data just in order to be able to find the piece I want on the iPod.
>> Wouldn't it be convenient if someone was to devise a standard format for
>> classical ID3 tagging, using the same language (probably English) and
>> the same way of spelling names, symphonies, etc. I know that this all
>> sounds very administratively dull and will be difficult in the light of
>> the internet's anarchic nature, but I'd be willing to give it a try. If
>> anyone else thinks this is a good idea, please contact me over e-mail
>> (or here).
>
> Agreed, but the problem is completely intractable. CDDB is a miraculously
> clever thing, but its database is entirely user-created and the structure
> of the database fields is ill-defined for classical music. It is immune
> to centralization.
>
> Usually some approximation to the information you want is there in a
> CDDB entry, though randomly placed in different fields. I sometimes do a
> thing that you can emulate -- when I reorganize or add to an entry, I
> upload it back to CDDB. So at least any disk you play that has passed
> through my (digital) hands is likely to have a more rational structure.
>
> I'm especially fond of CDDB entries for Japanese disks. They're -really-
> helpful ...
Agreed. It would be like trying to herd kittens.
--
Matthew B. Tepper: WWW, science fiction, classical music, ducks!
My personal home page -- http://home.earthlink.net/~oy/index.html
My main music page --- http://home.earthlink.net/~oy/berlioz.html
To write to me, do for my address what Androcles did for the lion
Mark Coy tossed off eBay? http://makeashorterlink.com/?M2B734C02
RMCR's most pointless, dumb and laughable chowderhead: Mark Coy.
Ilja Nieuwland
Granted. But the idea would rather be to do exactly what you are doing,
but collectively, and according to some type of standard - which might
be indicated by a marker (a prefix, for instance) for recognition. Maybe
the impact would be limited, but it'd certainly be an improvement on the
existing situation.
--
Ilja
Grammarian
I retag almost everything. I haven't found any way to make it much
easier than that. Byt he way, the problem isn't limitied to classical.
Everything else is messed up, too. After all, "If you make something
so simple only a idiot can screw it up, only an idiot will use it."
ulvi
FWIW, I use the "Genre" field for composer (last name only, with the
first name initials only if necessary: e.g. "R. Strauss").
The "Artist" field for the performer (again last-name only unless more
is necessary, in which case first and last name fully spelled out
as in "Sviatoslav Richter, Haitink, ..." etc.), only conductor name
for orch. works, with the orchestra acronyms added in case of possible
confusion ("Haitink/ACO. Haitink/VPO, ..." etc.). Chamber works, lieder,
etc. get performer names deliimted by slashes: ("Kremer/Afanassiev,
Hotter/Moore, ..." etc.) unless the group has a name ("Janacek
Quartet."). Similarly for concertos and orch. songs etc
(soloist/conductor). Year of performance added with another slash if
necessary("Karajan/BPO/1963").
The track name is the trickiest field. I usually accept the CDDB info as
given as long as it is sufficient to identify the movement/piece
unambigously in isolation. Unfortunately, often this field is marked
in CDDB in clueless ways such as: "Allegro Molto" etc. Then I use a
minimalist approach (the principle of least action on my part) to mark
the movement with opus number followed by sequence: e.g. "Op.31/2 3"
for the Beethoven Tempest sonata last movement. This leaves out
tonality, tempo indications and a whole bunch of interesting things, but
at least I know what I am listening to without doubt (I can then look up
the other info if I really wanted to know).
I think the genre field is completely useless for classical (do you
really need to be told that you are listening to a string quartet?), so
composer name fits very nicely in there. Artist = Performer(s) is more
or less axiomatic, so the only standardization that needs to be decided
is the Track name. I would be happy if everyone using CDDB adopted my
minimalist approach ("Op.31/2 3, Hob.XIII:48 2, K.331 1, ..." etc.),
but I'd be happy to comply with any other reasonable consensus.
Ripping is time-consuming (even with CDDB), but it's worth
it to have 500 hours of music on my Nomad Jukebox which I
can take anywhere with me :)
Ulvi
Tony Movshon
> Granted. But the idea would rather be to do exactly what you are doing,
> but collectively, and according to some type of standard - which might
> be indicated by a marker (a prefix, for instance) for recognition. Maybe
> the impact would be limited, but it'd certainly be an improvement on the
> existing situation.
All I do is put the composer's name in the "composer" field, the
performer[s] in the "artist" field, the work name in the "album" field,
and (if I have the time and inclination) the movement markings or (for
operas) aria or text marking in the "song name" field. For lieder, I
actually get to put the song name in the "song name" field!
Tony Movshon
mov...@nyu.edu
Matthew B. Tepper (posts from uswest.net are forged)
be typed in news:Xns935E9ADBFACE4...@209.204.42.170:
> Ripping is time-consuming (even with CDDB), but it's worth it to have 500
> hours of music on my Nomad Jukebox which I can take anywhere with me :)
AFTER I bought my Tungsten T, I discovered that it would function as an MP3
by the simple addition of software (my final choice was AeroPlayer, which I
paid $9 to register) and a Secure Digital card to hold the files (a 128MB
Lexar card, $45 + tax at Fry's before rebate). Add a $15 cassette adapter,
and it plays through my car stereo. Or I can listen on headphones.