Vorbis - Dec 2001 - (Classical) Request for Standardization of expanded TAGS

A month or two ago I sent an email to this list proposing to expand
the list of "standard" tags for Ogg Vorbis.  No tag would be required,
but if you wanted to encode certain types of information about a file,
you could use a standard tag.

I went through the whole discussion, and revised my proposal in light
of all the comments from everyone.  Here is the updated proposal.

This document is written in the context of
     http://www.xiph.org/ogg/vorbis/doc/v-comment.html
which hasn't been updated since Feb 19, 2001.If a tag can denote
a list of items, and not just a singleton, repeat the tag for each
item, one item for each usage of the tag.  After the list of tags
and what they mean, I will give an example that attempts to use
each one, then a more "normal" example, then a minimal example, and
what the result would look like when played by ogg123.

At first, I just meant for these tags to help out the Classical music
side of things.  But I've been assured that (other than Opus
and Conductor) these tags are equally as useful in Pop music.
Although with Metallicas "Enter Sandman" you might even need the
CONDUCTOR tag, for the guy who conducted the cello players.

--------------------------------------------------------
First, a few tags are now DEPRECATED:

VERSION
        version information is typically included in the title,
        such as "Doctoring the Tardis - 12" Remix"

DESCRIPTION
        role subsumed by ADDENDUM tag

ORGANIZATION
        role fulfilled by PUBLISHER, PRODUCER, LABEL,
        and COPYRIGHT tags.

ARTIST
        role fulfilled by COMPOSER, LYRICIST, PERFORMER, ENSEMBLE,
        CONDUCTOR, AUTHOR, PRODUCER, and ARRANGER tags.

--------------------------------------------------------
We are then left with the following RECOMMENDED tags:

OPUS
        the number of the work; ie, Opus 10, BVW 81, K6
        for a regular opus, just type the number, eg "10".
        For corner cases, type in the full designation,
        such as "BVW 231".

COMPOSER
        composer of the work. ie, Gustav Mahler

ARRANGER
        the person who arranged the piece.

PRODUCER
        the person who produced the recording

PUBLISHER
        webmaster from Hyperion will explain this
        one; its necessary for radio stations. if it
        turns out to be the same as the COPYRIGHT tag,
        I'll recommend we drop it entirely.

LYRICIST
        the person who wrote the lyrics
        
AUTHOR
        author of spoken text, eg, JRR Tolkien

CONDUCTOR
        conductor of the work; ie Herbert von Karajan

PERFORMER
        individual performers singled out for mention;
        ie, Yoyo Ma (violinist)

ENSEMBLE
        the group playing the peice, whether orchestra, choir,
        quartet, or rock group. ie, Berlin Philharmonic

LABEL
        the record label or imprint of the CD.

ISBN
        if found, put ISBN or EIN of the compact disc here.

ISRC
        documentation isn't available, but if you need it...

DISCID
        since the EIN, ISBN, etc numbers aren't to be reliably found
        on the CD, nor is the catalog number reliable, the FREEDB
        index hash should go here

TRACKNUMBER
        the track number on the CD

ALBUM
        if appropriate, an album name

TITLE
        "the work", whether a symphony, or a pop song

PART
        a division within a work; movement of a symphony, eg.

TYPE
        type of work, ie, symphony, concerto, song, speech

GENRE
        id3 type classification (classical, pop, jazz, blues, etc)

DATE
        date of recording, if known

LOCATION
        location of recording, if known

COPYRIGHT
        who holds copyright to the track

ADDENDUM
        additional comments pertaining to the piece.

--------------------------------------------------------
Here is a ``minimal'' example.

ALBUM=Joyride
ENSEMBLE=Roxette
TITLE=Joyride

Here is what you might see if you played it with ogg123:

Album: Joyride
Ensemble: Roxette
Title: Joyride

<p>Bitstream is 2 channel, 44100Hz

--------------------------------------------------------
Here is a ``typical'' example.

LABEL=Deutsche Grammophon
COMPOSER=Gustav Mahler
CONDUCTOR=Herbert von Karajan
ENSEMBLE=Berliner Philharmoniker
PERFORMER=Liza don Getti (soprano)
PERFORMER=Joe Barr (piano)
OPUS=8
PART="movement 1. Allegreto"
TITLE="Symphony no. 4"
ADDENDUM=I was present when this recording was made; met my wife there.
ADDENDUM=The flautist ate a green pickle with a purple egg.

And here is what you might see if you played it with ogg123:

Playing from file ogg/mystery_file_I_got_from_gnutella.ogg
Label: Deutsche Grammophon
Composer: Gustav Mahler
Conductor: Herbert von Karajan
Ensemble: Berliner Philharmoniker
Performers:  Liza don Getti (soprano)
             Joe Barr (piano)
Title: Symphony no. 4 (Opus 8), movement 1. Allegreto
Addenda: I was present when this recording was made; met
          my wife there.  The flautist ate a green pickle
          with a purple egg.

<p>Bitstream is 2 channel, 44100Hz
--------------------------------------------------------
Here is a complete example.
OPUS=1
COMPOSER=Stravinsky
ARRANGER=Ravel
PRODUCER=Fat Fred
PUBLISHER=Unknown
LYRICIST=Puccini
AUTHOR=Reverend Lovejoy
CONDUCTOR=Zubin Meta
PERFORMER=Nix Nax (piccolo)
ENSEMBLE=Upper East German Woodwind Band
LABEL=BIZ
ISBN=11111111111
ISRC=format unknown
DISCID=al;dkjf;ajsdl;fkjasl;kj
TRACKNUMBER=8
ALBUM=Tootles and Tunes
TITLE=The Life and Times of Friar Tux
PART=first minuet (allegro con molto)
TYPE=ballet
GENRE=classical (avant garde)
DATE=Sept 3, 1974
LOCATION=Easthampton Cathedral, Chesterwickshire
COPYRIGHT=Unclaimed
ADDENDUM=This is the wierdest, ugliest piece of music I've ever heard.  Its
the ugly duckling of the classical music world.

And here is what it would like like from ogg123

Album: Tootles and Tunes
Track: 8
Title: The Life and Times of Friar Tux (Opus 1), first minuet (allegro con
molto)
Composer:  Stravinksy
Conductor: Zubin Meta
Performer: Nix Nax (piccolo)
Ensemble:  Upper East German Woodwind Band
Arranger:  Ravel
Lyricist:  Puccini
Author:    Reverend Lovejoy
Label:     BIZ
Publisher: Unknown
Copyright: Unclaimed
ISBN: 111111111
ISRC: format unknwon
Type: ballet
Genre: classical (avant garde)
DiscID: al;dkjf;ajsdl;fkjasl;kj
Producer: Fat Fred
Recorded on: Sept 3, 1974
Recorded at: Easthampton Cathedral, Chesterwickshire
Addenda:  This is the wierdest, ugliest piece of music I've ever heard.
          Its the ugly duckling of the classical music world.  It has a
          monotone reading throughout, of that famous sermon by
          Reverend Lovejoy, "Life and Times of Friar Tux" 

<p>Bitstream is 2 channel, 44100Hz

-------------- next part --------------
A non-text attachment was scrubbed...
Name: part
Type: application/pgp-signature
Size: 797 bytes
Desc: not available
Url :
http://lists.xiph.org/pipermail/vorbis/attachments/20011206/c4cfc694/part-0001.pgp

Jonathan Walther (krooger@debian.org) wrote:
> First, a few tags are now DEPRECATED:
> ARTIST
> 	role fulfilled by COMPOSER, LYRICIST, PERFORMER, ENSEMBLE,
> 	CONDUCTOR, AUTHOR, PRODUCER, and ARRANGER tags.
I can see why this is necessary, but it's going to break backward
compatibility with existing Ogg files and players.  For that reason alone,
it's going to be extremely hard to get this adopted.

You might want to mention how players that adopt your standard should
deal with old tags.  It shouldn't be terribly hard -- just something
like this (pseudocode):

    if (exists(ARTIST)) {
        old-style tag parsing
    } else {
            new-style tag parsing
    }

Apart from that issue, I like this proposal.


-- 
Greg Wooledge                  |   "Truth belongs to everybody."
greg@wooledge.org              |    - The Red Hot Chili Peppers
http://wooledge.org/~greg/     |


-------------- next part --------------
A non-text attachment was scrubbed...
Name: part
Type: application/pgp-signature
Size: 241 bytes
Desc: not available
Url :
http://lists.xiph.org/pipermail/vorbis/attachments/20011206/a42be456/part-0001.pgp

Sgrifennodd Jonathan Walther:
> First, a few tags are now DEPRECATED:
> 
> ORGANIZATION
> 	role fulfilled by PUBLISHER, PRODUCER, LABEL,
> 	and COPYRIGHT tags.
Hmm, are radio broadcasts considered publishing?

..ciaran

-- 
Ciaran Anscomb, BBC Internet Services
"Energy goes very fast - it's the essence of all the creations!"

--- >8 ----
List archives:  http://www.xiph.org/archives/
Ogg project homepage: http://www.xiph.org/ogg/
To unsubscribe from this list, send a message to
'vorbis-request@xiph.org'
containing only the word 'unsubscribe' in the body.  No subject is
needed.
Unsubscribe messages sent to the list will be ignored/filtered.

On Thursday, December 06, 2001 9:42 AM, Jonathan Walther 
[SMTP:krooger@debian.org] wrote:
> First, a few tags are now DEPRECATED:
>
> VERSION
> 	version information is typically included in the title,
> 	such as "Doctoring the Tardis - 12" Remix"
I've not encoded many files to ogg yet, but I imagine that this tag is 
something I'd have used a lot (especially with live tracks, and 
occasionally with remixes).  Personally I prefer a TITLE/VERSION setup to 
one where all the info appears in TITLE.

I'd be interested to hear the reasoning behind this tag's deprecation --
the archived thread from early October doesn't seem to go into it.  Also, 
will deprecation ultimately mean that players won't support it as a 
standard tag?  Will it have any repurcussions at all for files encoded by 
people who thought it was a good idea?

In the absence of an argument to convince me that lumping everything into 
TITLE is a good idea, consider this a vote against VERSION's deprecation 
:-)

Cheers,
Steve


--
Steve Tregidgo

<p><p>--- >8 ----
List archives:  http://www.xiph.org/archives/
Ogg project homepage: http://www.xiph.org/ogg/
To unsubscribe from this list, send a message to
'vorbis-request@xiph.org'
containing only the word 'unsubscribe' in the body.  No subject is
needed.
Unsubscribe messages sent to the list will be ignored/filtered.

> ENSEMBLE
> 	the group playing the peice, whether orchestra, choir,
> 	quartet, or rock group. ie, Berlin Philharmonic
Are solo artists (ie. no group) also considered an 'ensemble' ?


-- 
Trick
__________________________________________
Linux User #229006 * http://counter.li.org

"There is no magic." - Nakor, magic user.

--- >8 ----
List archives:  http://www.xiph.org/archives/
Ogg project homepage: http://www.xiph.org/ogg/
To unsubscribe from this list, send a message to
'vorbis-request@xiph.org'
containing only the word 'unsubscribe' in the body.  No subject is
needed.
Unsubscribe messages sent to the list will be ignored/filtered.

On 2001-12-06, Jonathan Walther wrote:
> I went through the whole discussion, and revised my proposal in light
> of all the comments from everyone.  Here is the updated proposal.
>
I generally like it.
> First, a few tags are now DEPRECATED:
>
> VERSION
> 	version information is typically included in the title,
> 	such as "Doctoring the Tardis - 12" Remix"
>
> DESCRIPTION
> 	role subsumed by ADDENDUM tag
>
> ORGANIZATION
> 	role fulfilled by PUBLISHER, PRODUCER, LABEL,
> 	and COPYRIGHT tags.
>
> ARTIST
> 	role fulfilled by COMPOSER, LYRICIST, PERFORMER, ENSEMBLE,
> 	CONDUCTOR, AUTHOR, PRODUCER, and ARRANGER tags.
>
Great!  ARTIST was never enough for me.  There is a minor problem of
backward compatibility - the two solutions I see are:
1) Use ARTIST for PERFORMER since it's the de-facto use in music metadata.
2) Switch to PERFORMER; ARTIST from now on means "one of the above, I
   don't know which one".
I prefer the second, since lot's of metadata with single artist field
(e.g. ID3) floats around and one should have a way to express this
common uncertainty it in vorbis comments.

P.S.  To all vorbis GUI tag editor writers - implement a simple way to
rename a comment (like making the name and not only the value editable
when you double click a comment) if you didn't already.
> --------------------------------------------------------
> We are then left with the following RECOMMENDED tags:
>
> OPUS
> COMPOSER
> ARRANGER
I like these.
> PRODUCER
> 	the person who produced the recording
>
> PUBLISHER
> 	webmaster from Hyperion will explain this
> 	one; its necessary for radio stations. if it
> 	turns out to be the same as the COPYRIGHT tag,
> 	I'll recommend we drop it entirely.
>
Could you explain this better?
> LYRICIST
> 	the person who wrote the lyrics
>
> AUTHOR
> 	author of spoken text, eg, JRR Tolkien
>
Where is the border?  If it's a poem, is it a lyricist or an author?  What
if it's half sung?
> CONDUCTOR
> PERFORMER
> ENSEMBLE
I like these.
> LABEL
> 	the record label or imprint of the CD.
>
> ISBN
>         if found, put ISBN or EIN of the compact disc here.
>
Isn't sound specifiaclly excluded from ISBN's scope?  See:
http://www.isbn.org/standards/home/isbn/international/html/usm3.htm
http://www.isbn.org/standards/home/isbn/international/html/usmnote.htm#cdrom
There are some exceptions: books on cassetes, sound on CD/Video whose
main purpose is educational, etc. - did you mean these cases?
> ISRC
> 	documentation isn't available, but if you need it...
>
It's probably temporarily down, since it's in google's cache but you
need
to abort download before it tries to run a script (or disable javascript
in the browser) to see it.  In particular try
http://www.google.com/search?q=cache:http://www.ifpi.org/online/isrc_sites.html
for related sites (that are on-line now :-).
Some info appears in http://www.riaa.org/Audio-Standards-3.cfm

It has shortcomings.  IIRC most people on this ML agreed that it will run
out of numbers in a quite finite time (number contains only 3 digits for
registrant (company) and this assumes that big record companies will
eliminate all small producers).  And it has 2 digits per year!

Net search appears to show it as almost dead, compared to ISWC.  It's a
different thing - identifies a musical work disregarding performance.
See:  www.iswc.org  and  http://www.mpa.org/press/Press2_8.html
At least ISWC seems better designed - single 9-digit space.

Does anybody want an ISWC tag?

Lot's of new IDs are beeing developed, they will get anywhere IMHO if they
change the standards every two years.  These standards should've been
designed to last until the fall of the Galactic Empire...

I'm afraid after 10 years the vorbis tag namespace will contain more
standard IDs than useful tags...  How about moving all IDs into one tag:

ID=ISRC=...
ID=ISBN=...
ID=ISAN=...

and so on.  URN goes this way.  Maybe adopt URNs?  They don't have
anything registered for sound yet...  See:
http://www.uri.net/urn-nid-status.html.  This will make the vorbis tag
space cleaner.  The second separator doesn't have to be an `=' - `:'
or
anything else is equally good.  Also notice that the namespace name part
is case-sensitive for vorbis tools since it's part of the tag value.
> DISCID
> 	since the EIN, ISBN, etc numbers aren't to be reliably found
> 	on the CD, nor is the catalog number reliable, the FREEDB
> 	index hash should go here
>
Is the freedb index a Good one?  Never mind, it's widely used so it should
be put in but again, ID=FREEDB=.... is better.
> TRACKNUMBER
> 	the track number on the CD
>
A disknumber for multi-disk recorings was requested once.  Maybe `2-7'
in TRACKNUMBER meaning 2nd disk, 7th track is better.
> ALBUM
> 	if appropriate, an album name
>
> TITLE
> 	"the work", whether a symphony, or a pop song
>
> PART
> 	a division within a work; movement of a symphony, eg.
>
> TYPE
> 	type of work, ie, symphony, concerto, song, speech
>
> GENRE
> 	id3 type classification (classical, pop, jazz, blues, etc)
>
> DATE
> 	date of recording, if known
>
> LOCATION
> 	location of recording, if known
>
> COPYRIGHT
> 	who holds copyright to the track
>
> ADDENDUM
> 	additional comments pertaining to the piece.
>
A little too latin for me.  I like DESCRIPTION more.  If I undersatnd
correctly this exactly replaces description but is a more exact name.  If
so I think it's not worth the compatibility change.
> --------------------------------------------------------
> Here is a ``typical'' example.
>
> LABEL=Deutsche Grammophon
> COMPOSER=Gustav Mahler
> CONDUCTOR=Herbert von Karajan
> ENSEMBLE=Berliner Philharmoniker
> PERFORMER=Liza don Getti (soprano)
> PERFORMER=Joe Barr (piano)
> OPUS=8
> PART="movement 1. Allegreto"
> TITLE="Symphony no. 4"
> ADDENDUM=I was present when this recording was made; met my wife there.
> ADDENDUM=The flautist ate a green pickle with a purple egg.
>
> And here is what you might see if you played it with ogg123:
>
> Playing from file ogg/mystery_file_I_got_from_gnutella.ogg
> Label: Deutsche Grammophon
> Composer: Gustav Mahler
> Conductor: Herbert von Karajan
> Ensemble: Berliner Philharmoniker
> Performers:  Liza don Getti (soprano)
> 	     Joe Barr (piano)
> Title: Symphony no. 4 (Opus 8), movement 1. Allegreto
> Addenda: I was present when this recording was made; met
>           my wife there.  The flautist ate a green pickle
> 	  with a purple egg.
>
Is the player app suppossed to know latin singular/plurar grammar? Note
also that this gives no indication of line break inside the comment
(that's legal!) vs. two comments of same name.
> And here is what it would like like from ogg123
>
> [snip]
A good player (like ogg123) should allow a user to configure what comments
he wishes to see and in what order should they be sorted.


-- 
Beni Cherniavsky <cben@tx.technion.ac.il>
                 (also scben@t2 in Technion)

<p><p><p>--- >8 ----
List archives:  http://www.xiph.org/archives/
Ogg project homepage: http://www.xiph.org/ogg/
To unsubscribe from this list, send a message to
'vorbis-request@xiph.org'
containing only the word 'unsubscribe' in the body.  No subject is
needed.
Unsubscribe messages sent to the list will be ignored/filtered.

On Thursday, December 06, 2001 1:05 PM, Beni Cherniavksy 
[SMTP:cben@techunix.technion.ac.il] wrote:
> A disknumber for multi-disk recorings was requested once.  Maybe `2-7'
> in TRACKNUMBER meaning 2nd disk, 7th track is better.
That sounds good.  The alternative would seem to be to add " (disc X)"
to
the end of the ALBUM value for multi-disc recordings, which doesn't seem 
like a clean use of the fields to me.  An album that comes on several discs 
would show up as multiple albums when grouped/sorted by human-readable 
fields (ie excluding freedb ID etc).

I think the suggested format is better than the use of, say, DISKNUMBER 
(which adds, IMO, an explicit level of hierarchy to track info in the 
context of albums).  Or has this been discussed before and I've missed it 
(I subscribed to this list mid-October so that's quite likely)?

Cheers,
Steve


--
Steve Tregidgo

<p><p>--- >8 ----
List archives:  http://www.xiph.org/archives/
Ogg project homepage: http://www.xiph.org/ogg/
To unsubscribe from this list, send a message to
'vorbis-request@xiph.org'
containing only the word 'unsubscribe' in the body.  No subject is
needed.
Unsubscribe messages sent to the list will be ignored/filtered.

On Thursday 06 December 2001 22:42, Jonathan Walther wrote:
> --------------------------------------------------------
> First, a few tags are now DEPRECATED:
>
> VERSION
> 	version information is typically included in the title,
> 	such as "Doctoring the Tardis - 12" Remix"
How is having a version tag for spliting remix and cover information out 
from the title any less legitimate than separating opus and part?
> DESCRIPTION
>        role subsumed by ADDENDUM tag
Why replace a word everyone understands with one that it's so clear or 
commonly used? 
> ARTIST
> 	role fulfilled by COMPOSER, LYRICIST, PERFORMER, ENSEMBLE,
> 	CONDUCTOR, AUTHOR, PRODUCER, and ARRANGER tags.
I think this is misguide. See my notes below.
> --------------------------------------------------------
> We are then left with the following RECOMMENDED tags:
[misc tag suggestions]

These are all good, but they're also fairly ambiguous. You're going to
have to
be a lot more specific about what each tag is supposed to do and not do to
avoid that problem. 

In fact, I don't even think it's _possible_ to usefully eliminate the 
ambiguity in every tag - you're better off crafting a way to make it not 
matter.
> DISCID
> 	since the EIN, ISBN, etc numbers aren't to be reliably found
> 	on the CD, nor is the catalog number reliable, the FREEDB
> 	index hash should go here
CDDB index hashes are not unique - not really a problem if the album's entry
is in your local cache, but a hassle if you're looking it up in the server. 
At the very least, you need to include the catagory the cd has been placed 
in. 
> TRACKNUMBER
> 	the track number on the CD
Let's include a DISCNUMBER, too. I hate having to put that in the album.

<p>> GENRE
> 	id3 type classification (classical, pop, jazz, blues, etc)
While it's sensible to map id3 genre classifications into the GENRE tag if 
you're converting tags, it's more sensible to treat each genre tag as a 
fairly arbitary keyword that you can run a text search on. One man's metal
is
another man's hard rock, after all. 
> --------------------------------------------------------
> Here is a ``minimal'' example.
>
> ALBUM=Joyride
> ENSEMBLE=Roxette
> TITLE=Joyride
>
> Here is what you might see if you played it with ogg123:
>
> Album: Joyride
> Ensemble: Roxette
> Title: Joyride
Ok, speaking as a member of the vast hordes of pop kiddies who don't know
one
decomposing composer from the next, there is just _no_way_ I'm going to use 
the ENSEMBLE tag to describe a pop group. Forget it. I won't do it, and I 
won't use a tag system that stops me from using artist, either. 

Speaking as someone who's writing a library for converting audio metatdata 
between arbitary tag formats, how do I figure out what to convert
'artist'
to? Ensemble? But what if they're just a performer, or a composer, or a 
string with the composer, conductor and orchestra in it?

Leave ARTIST in. It makes automatic conversion easy, and it's good enough
for
the majority of users. But provide good implementation advice to tag editor 
authors to make it easy to break down an artist tag into more specific parts, 
and to player authors so they can assemble a sensible display from title and 
artist components if the simple tags don't exist.

Don't get me wrong - I like these tags, and look forward to using many of 
them, but they require writing more complicated code to benefit a minority of 
users, so if you want to see them be adopted, you'll have to do a lot more 
work to make the lives of the implementors easier.

John

--- >8 ----
List archives:  http://www.xiph.org/archives/
Ogg project homepage: http://www.xiph.org/ogg/
To unsubscribe from this list, send a message to
'vorbis-request@xiph.org'
containing only the word 'unsubscribe' in the body.  No subject is
needed.
Unsubscribe messages sent to the list will be ignored/filtered.

Jonathan Walther (krooger@debian.org) wrote:
> PERFORMER
> 	individual performers singled out for mention;
> 	ie, Yoyo Ma (violinist)
How do actually handle this?  Sample CD liner notes:

  PRODUCED BY CHRIS BLACKWELL for Island Records, Inc.
  Associate Producer: Robert Ash
  Engineer: Robert Ash
  Assistant Engineer: Cass Rigby
  CDD Pre-Mastering by WCI Record Group
  Art Direction: Sue Ab Surd
  Photography: George DuBose
  Hairdos by: La Verne
  Recorded at Compass Point Studios, Nassau, Bahamas
  Conseilleur: Gary Kurfist

  KATE PIERSON:
    vocals, organ, keyboard bass
    2nd guitar (52 Girls, Hero Worship)
  FRED SCHNEIDER:
    vocals, walkie talkie, toy piano
    keyboard bass (Hero Worship)
  KEITH STRICKLAND:
    drums, percussion, Claire sounds
  CINDY WILSON:
    vocals, bongoes, 2nd guitar (Moon)
    tambourine
  RICKY WILSON:
    guitar, smoke alarm

(These are the first two pages of the liner notes for the first album
by The B-52's.)  So, how do we tag this under the proposal?  Do we use
this:

  TITLE=Hero Worship
  PERFORMER=Kate Pierson (vocals)
  PERFORMER=Kate Pierson (organ)
  PERFORMER=Kate Pierson (keyboard bass)
  PERFORMER=Kate Pierson (2nd guitar)
  PERFORMER=Fred Schneider (vocals)
  PERFORMER=Fred Schneider (walkie talkie)
  ...

Or do we do this:

  TITLE=Hero Worship
  PERFORMER=Kate Pierson (vocals, organ, keyboard bass)
  PERFORMER=Fred Schneider (vocals, walkie talkie, toy piano, keyboard bass)
  ...

Where do we put things like "Assistant Engineer"?  What about
"Photography" and "Hairdos by"?

Finally, what about music that includes samples of other music?  For
example, from Madonna's "Erotica":

  (c) 1992 WB Music Corp./Bleu Disque Music Co. Inc./Webo Girl
  Publishing, Inc. Adm. by WB Music Corp./Shepsongs Adm. by MCA Music
  Publishing Inc. ASCAP

  Shep Pettibone and Tony Shimkin - Sequencing, Keyboards and Programming
  Joe Moskowitz - Additional Keyboards
  P. Dennis Mitchell and Robin Hancock - Recording Engineers
  George Karras - Mix Engineer

  Contains a sample of "Jungle Boogie" by Kool & the Gang, under
  license from Polygram Special Markets, a division of Polygram Group
  Distribution, Inc.  Written by R. Bell, C. Smith, R. Mickens, D. Boyce,
  R. Westfield, D. Thomas, R. Bell, G. Brown.  Warner-Tamerlane
  Publishing Corp./Second Decade Music Co. BMI.  All rights administered
  by Warner-Tamerlane Publishing Corp. Used by permission.  All rights
  reserved.

A lot of this is just copyright stuff that most people don't care about
an awful lot, but at the very least I'd expect to be able to include
the TITLE and ARTIST (heh) of the sampled work in some standardized way.


-- 
Greg Wooledge                  |   "Truth belongs to everybody."
greg@wooledge.org              |    - The Red Hot Chili Peppers
http://wooledge.org/~greg/     |

P.S.: Lyrics!!


-------------- next part --------------
A non-text attachment was scrubbed...
Name: part
Type: application/pgp-signature
Size: 241 bytes
Desc: not available
Url :
http://lists.xiph.org/pipermail/vorbis/attachments/20011206/541370bf/part-0001.pgp

On Friday, December 07, 2001 6:52 AM, Jonathan Walther 
[SMTP:krooger@debian.org] wrote:
> Now, if it was something like "Bon Jovi appearing live with
> Guns'n'Roses", then I'd say
>
> PERFORMER=Bon Jovi
> ENSEMBLE=Guns'n'Roses
>
> because they are thought of as two separate entities collaborating,
> instead of one group.
If we're talking about the band Bon Jovi (as opposed to the lead singer, 
Jon Bon Jovi), would you suggest two ENSEMBLE tags in your scheme?  I ask 
mainly for clarification on the use of multiple tags with the same name -- 
I know it's supported in general, but does the proposed standard seek to 
limit any of these to singletons?  (Not asking to be pedantic on the 
singer/band ambiguity :-)

Cheers,
Steve


--
Steve Tregidgo

<p><p>--- >8 ----
List archives:  http://www.xiph.org/archives/
Ogg project homepage: http://www.xiph.org/ogg/
To unsubscribe from this list, send a message to
'vorbis-request@xiph.org'
containing only the word 'unsubscribe' in the body.  No subject is
needed.
Unsubscribe messages sent to the list will be ignored/filtered.

> You noticed the extreme nuisance was only in the creating of oggs,
> right? Once they're out there, you won't get any hastle.
That is quite enough hassle for me, I encode a lot of oggs, some for
song archival, some for DJing, some for just generally testing the
formats quality. I don't want to be hassled at all about what tags I use. 
Will it hassle me if I include performer/ensemble alongside an artist tag?  
what if I'm just doing a rip to have a listen to how good a song sounds
at a particular nominal rate, and I just want to be lazy and use the
artist/title from freedb. Will it hassle me about that?
>Here is a way to make people avoid ARTIST by positive rather than negative
>means: in the tagging GUI, near every artist field, show a radio list of
>PERFORMER, COMPOSER, AUTHOR,
That there is a great idea! rather than bug me because I'm being lazy
and using the artist tag, encourage me to use something else without
annoying me! 

Mark




On Wed, Dec 12, 2001 at 03:22:06PM +0200, Beni Cherniavksy
wrote:
>I second (fifteenth? :) many people that say PERFORMER / ENSEBMLE
>separation is problematic.  I don't know how to use it _usefully_ with
I said it in email; maybe I should say it in the proposal.  ENSEMBLE is
for people that want to use it.  Other people can use PERFORMER as they
use ARTIST.
>A related question on this: I have many songs by two ensembles (one duo,
>one trio) that kept changed their composition (one man left, another one
>got in) but kept the name and style.  So I usually want to just write the
>band name, e.g. when searching for songs by them but I sometimes I want to
>know who exactly it was.  Should I indicate this in the ENSEMBLE or by
>adding the PERFORMER that made the difference?  The later seems better but
>is ambigous to someone not knowing these detail well: when you see both
>ENSEMBLE and PERFORMER, does it mean that it's an additional performer
or
>a part of this ensemble that the tagger decided to specifically note?
Ah.  Finally, a good example.  In such a case, the ENSEMBLE/PERFORMER
distinction makes a big difference.  A performer can be a soloist that
just happens to be with a group, but a performer can also be a member of
the group who is noteworthy on his own.  For instance, the first violin
of a symhony orchestra might be singled out for mention on the cover of
a CD.  By analogy, you could put the groups name in ENSEMBLE, and the
names of both performers in their own PERFORMER tags.  In my mind, the
ENSEMBLE tag is the "important" tag, and PERFORMER secondary.
>I also strongly dislike the "Any attempt to use the ARTIST tag must be
an
>extreme nuisance for the user of the program" statement!  I would like
to
You noticed the extreme nuisance was only in the creating of oggs,
right?  Once they're out there, you won't get any hastle.

Your suggestions below are good; they are what I meant by nuisance.
Design the UI to encourage use of standard tags first, but allow artist
tags after some struggle.  Was this not clear from the proposal?
>see a copy of this proposal distributed along with the different tagging
>tools, and availiable by a single click on a "Recommended tags"
button in
>the tagging dialog.  Also provide simple means to use this tags but
don't
>do anything negative.  I like very much to catalogue my music but I myself
>will use ARTIST sometimes (when I don't know more information).  I
don't
>want to be told by programs "That's deprecated, if you do it more
than
>twice a week I will remind you again, then if you don't get the point I
>will delete all files with ARTIST tags from your hard disk so that you
>learn to tag correctly..."  Of course I exaggerated but _I know what I
do
>when I use ARTIST_.  And people who don't won't learn from
punishment,
>they will only become angry.
>
>Here is a way to make people avoid ARTIST by positive rather than negative
>means: in the tagging GUI, near every artist field, show a radio list of
>PERFORMER, COMPOSER, AUTHOR, etc... that when selected mutates the ARTIST
>into that tag.  It's a good idea for some other places too, to lower the
>energy barrier of being more specific.
Thanks.

Jonathan

<p>--- >8 ----
List archives:  http://www.xiph.org/archives/
Ogg project homepage: http://www.xiph.org/ogg/
To unsubscribe from this list, send a message to
'vorbis-request@xiph.org'
containing only the word 'unsubscribe' in the body.  No subject is
needed.
Unsubscribe messages sent to the list will be ignored/filtered.

I don't like this:
"Any attempt to use the ARTIST
tag must be an extreme nuisance for the user of the program."

I'd prefer something like "Use of the "PERFORMER/ENSEMBLE"
tags instead
of  the "ARTIST" tag should be encouraged. This may be acheived
through
interface cues etc. The use of the artist tag should not be made
impossible, or necessarily harder, but the option to use the
PERFORMER/ENSEMBLE tag should be made obvious and convenient."

While silently converting freedb tags to performer would work fine, I
think it might break players looking for the "Artist" tag. (loads
xmms).
Hrmmm, it says "%p - performer/artist", how convenient!

Does winamp handle it gracefully with the vorbis plugin?

Now, let me just check grip, it has a "%a" for artist straight from
freedb..... And it's passed straight to oggenc, checking oggenc, oggenc
only allows "-a" for inputing artist tags. How does oggenc save it in
the file? I'm assuming it saves it as "Artist=...."

Well if oggenc allows a "-p" switch to input performer, I can be
compliant by changing one character in my grip config. Or if oggenc
saves it in a performer tag already I'm already compliant.

I just think alienating users by making it really hard to use an artist
tag is a bit harsh, and elitist. (<- maybe not the words I was looking
for exactly, but anyway).

<p><p>


On Thu, Dec 13, 2001 at 10:02:52AM +1100, MARK JAMES HETHERINGTON
wrote:
>>Here is a way to make people avoid ARTIST by positive rather than
negative
>>means: in the tagging GUI, near every artist field, show a radio list of
>>PERFORMER, COMPOSER, AUTHOR,
>
>That there is a great idea! rather than bug me because I'm being lazy
>and using the artist tag, encourage me to use something else without
>annoying me! 
How would you recommend the proposal read?  I am a big fan of positive
encouragement.  Here is your chance to influence the standard.  I
envisioned ARTIST tags from freedb being silently coerced into PERFORMER
tags; if you really want an ARTIST tag you would have to change it back.
That would meet the standard.

<p>--- >8 ----
List archives:  http://www.xiph.org/archives/
Ogg project homepage: http://www.xiph.org/ogg/
To unsubscribe from this list, send a message to
'vorbis-request@xiph.org'
containing only the word 'unsubscribe' in the body.  No subject is
needed.
Unsubscribe messages sent to the list will be ignored/filtered.

Excellent, I much prefer the proposal worded that way.

I didn't even know of "vorbiscomment" existing on my system till I
read
that email, and decided to take a look. 

I don't know how I can easily get grip to use it. There is no post
compression command, but I guess I could do something like "oggenc <ogg
options>; vorbiscomment <comment options>" and it would do it....




On Thu, Dec 13, 2001 at 10:53:44AM +1100, MARK JAMES HETHERINGTON
wrote:
>I don't like this:
>"Any attempt to use the ARTIST
>tag must be an extreme nuisance for the user of the program."
>
>I'd prefer something like "Use of the
"PERFORMER/ENSEMBLE" tags instead
>of  the "ARTIST" tag should be encouraged. This may be acheived
through
>interface cues etc. The use of the artist tag should not be made
>impossible, or necessarily harder, but the option to use the
>PERFORMER/ENSEMBLE tag should be made obvious and convenient."
Thanks. I've updated the proposal to just remove that troublesome ARTIST
clause.  It now reads like this:

    Compliant tag editors and rippers may support tags not in the
    standard, as long as they encourage use of standard tags over
    any non-standard tags.

Is that acceptable?
>While silently converting freedb tags to performer would work fine, I
>think it might break players looking for the "Artist" tag. (loads
xmms).
>Hrmmm, it says "%p - performer/artist", how convenient!
I have a feeling that the winamp/xmms architectures are flexible enough
that if we adopt this standard, their support will shortly follow.  It
will just involve the addition of another strcmp() call.
>Now, let me just check grip, it has a "%a" for artist straight
from
>freedb..... And it's passed straight to oggenc, checking oggenc, oggenc
>only allows "-a" for inputing artist tags. How does oggenc save it
in
>the file? I'm assuming it saves it as "Artist=...."
grip can move over to using vorbiscomment with a minimum of fuss.

Jonathan

<p>--- >8 ----
List archives:  http://www.xiph.org/archives/
Ogg project homepage: http://www.xiph.org/ogg/
To unsubscribe from this list, send a message to
'vorbis-request@xiph.org'
containing only the word 'unsubscribe' in the body.  No subject is
needed.
Unsubscribe messages sent to the list will be ignored/filtered.