[vorbis-dev] RECORDNUMBER

[Beni Cherniavsky]

On Sat, 25 Aug 2001, Ralph Giles wrote:

> I'd probably use VOLUMENUMBER instead. Or better yet just include
> it in the title. Or, depending on how the tracks were listed,
> use the same album title, but increment the TRACKNUMBER across
> each of the disks.
>
One related point to think of: I'm soon going to encode 4 CDs, each
containing 2 original albums (which probably were on vynil).  Now I'm
interested in both of the volume/track breakdowns: CD 1-4 / track 1-24
(for example)  vs.  album 1-8 (name actually) / song 1-12.

I want the breakdown onto the CDs to locate the song on the CDs and I want
the original albom breakdown for oranizing my songs (e.g. generate a
playlist for each albom).

Now how do I represent this?  VOLUMENUMBER/DISCNUMBER/RECORDNUMBER should
probably refer to the 1-4 CD number.  The 1 of 8 albums will be refered to
by name in ALBUM.  Then I also need to refer to the whole 4 CD compilation
name - what tag do I put it in?  Then there are the two concepts of track
numbers.  Which is more appropriate for TRACKNUMBER and how should I name
the tag for the other one?

> For the DATE header, we do suggest that you use the ISO format, or
> at least spell out the month (or the day of week) to remove
> ambiguity. But to demand standardized, machine parseable data
> misses the point; the spec clearly indicates the vorbis comment
> header is intended for simple human-readable data only.
>
I think machine parsabily is meaningless to demand here but the guidelines
should request that the user applies common sense to format the date in
a way that is unambiguos to all humans - e.g. don't do 1-3-99 since it can
also mean 3-1-99; add timezone (or clear location) for precise times but
probably not needed for if hour is unknown anyway...  And add ISODATE if
you need it.

As a sidenote about different calendars, partial information can be very
unconvenient for _exact_ translation into e.g. an ISO date.  For example
hebrew years begin somewhere about Sep. (floating from year to year).  So
if I know only the hebrew year, it can't be exaclty mapped onto a
gregorian date that I know only the year of - it must become a range,
something like 1980-09-13 - 1981-09-23.  Similarly, hebrew days change at
the sunset, so If I know a hebrew day, the exact mapping would be e.g.
1981-03-05 18:38 - 1981-03-06 18:39 (plus timezone info).

>
> Our standard answer to this has been that we plan to have a structured
> format for rich (aka 'kitchen sink') metadata that's intended for
> machine comparison and searching. One of the reasons I've not been
> in a hurry with this is it's more a nicety for distribution. You
> can always use a catalog program with an external database to manage
> your music collection, and/or use the comment metadata to look up
> details in an internet db.
>
> Of course, if your application does make more sense with machinable
> fields in the comment header, go ahead and define your own tags. We
> might not think it's the best idea, but it's none of our business,
> and that's why the format is extensible.
>
Agreed.  That's why if the suggestion to make oggenc reject non-iso dates
(somewhere in this thread) would be accepted, I would get the sources,
remove this and recompile ;).  Freedom is certainly the way to go.


-- 
Beni Cherniavsky <cben at crosswinds.net>
                 (also scben at t2,cben at tx in Technion)

--- >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-dev-request at xiph.org'
containing only the word 'unsubscribe' in the body.  No subject is needed.
Unsubscribe messages sent to the list will be ignored/filtered.