LISTSERV mailing list manager LISTSERV 16.0

Help for METS Archives


METS Archives

METS Archives


METS@LISTSERV.LOC.GOV


View:

Message:

[

First

|

Previous

|

Next

|

Last

]

By Topic:

[

First

|

Previous

|

Next

|

Last

]

By Author:

[

First

|

Previous

|

Next

|

Last

]

Font:

Proportional Font

LISTSERV Archives

LISTSERV Archives

METS Home

METS Home

METS  December 2001

METS December 2001

Subject:

LONG: Thank you Dave Ackerman, comments

From:

Carl Fleischhauer <[log in to unmask]>

Reply-To:

Metadata Encoding and Transmission Standard <[log in to unmask]>

Date:

Tue, 11 Dec 2001 16:34:16 -0500

Content-Type:

TEXT/PLAIN

Parts/Attachments:

Parts/Attachments

TEXT/PLAIN (243 lines)

This document is framed as a <dialog> with David Ackerman
([log in to unmask]) in response to his useful email of Thu Dec
6.  I have extracted the core statements from his email, and have
CATEGORIZED them below.  The sections are marked up as <dave></dave> and
<carl></carl>.  Notes that pertain exclusively to our relational database
(and are of local interest to my colleagues) are marked <LCdb></LCdb>.

CHANNELS:

<dave> We (the AES) propose something like the following for showing
channel relationships:
<sound_channel_map>
        <channel_assignment sound_channel="0" sound_map_loc="left front
speaker" />
        <channel_assignment sound_channel="1" sound_map_loc="right front
speaker"/>
</sound_channel_map>

in addition, an attribute of  num_channels ="2"

and a sound_field="multitrack" attribute that describes the aural space
the channels are meant to be presented as. For example, 6 channels might
be a surround mix in 5.1, or they might be multi-track.  At the AES
conference last weekend I  heard of a label that is releasing their
material in a non-standard surround format utilizing the standard 6
channels available but mapping the sub channel to another full range
speaker that is meant to give a sense of height when placed in their
non-standard speaker arrangement. </dave>

<carl> Comment: Am I correct to understand Dave to be saying that there
are four possible attributes to the element
channel_assignment: (1) sound_channel, (2) sound_map_loc,
(3) num_channels, and (4) sound_field.  If so, I wonder, why does the
example above not include "num_channels?"   BTW: we will concoct a similar
but separate element/attributes for analog source items where we are
sorting out tracks and not "channels."

Anyway, proceeding on the assumption that these are four attributes of
"channel_assignment," here is a comparison to our database back end.

(1) sound_channel: this seems to mean "if there are more than one channel
in this multiplexed signal, which channel [by number] is this?"  For mono,
do we leave this empty/not used?  If we had to state a value for mono,
using the example above, it looks like the value would be
sound_channel="0" which seems counterintuitive to me.  Why isn't the first
channel called "1" ?

<LCdb>LC AV current database does not have this data unit ("which channel
am I, by number?") unless it is implied in some way.  We need something we
might call "sound_channel_num," no?</LCdb>

(2) sound_map_loc: the location (destination in playback) of this
channel: which loudspeaker are we heading for.

<LCdb>LC AV current database field "channel_track_information" which is
the same.  Proposal: adopt Dave's name.   Our instances so far are limited
to mono and stereo so if we take his example to heart, our enumerated list
would be like this: if stereo, then left front speaker and right front
speaker.  If mono, can we just leave this out/blank/omit?  In the Builder
pulldown, in instances like this, do you have a selection for data-entry
staff called "none" that does not actually put anything in back end?<LCdb>

(3) num_channels.  LC AV current database field has
"channel_track_quantity" which is the same.

<LCdb>Proposal: adopt Dave's name or a variation in the db.  The
enumerated list for this would just positive real integers, no?</LCdb>

(4) sound_field: our same-name field is just like his.

<LCdb>We have an enumerated list but I bet that the AES enumerated list
(if there is one) will be longer or more detailed.  Nothing to do about
that right now. </LCdb></carl>

SYNCHRONIZATION:

<dave>Have a look at the AES-31 Standard.  The proposed Audio Decision
List performs the synchronization function for multi channel audio. You
could even do this with the METS Structure map by including time stamp
information in the audio metadata that indicates when each track starts
against a common time line [assuming no edits]. Processing information can
be included in the digi-prov bucket. From the AES core audio metadata
proposal:

<!ELEMENT timestamp EMPTY>
      <!ATTLIST timestamp
             timecode_start CDATA #REQUIRED
          timecode_end CDATA #IMPLIED
             timcode_format CDATA #REQUIRED> </dave>

<carl>I do not really understand this, although I deduce that it applies
to instances of multi-files for multi-channel (as contrasted with
multiplexed multi-channel in one file). I suspect that is doesn't apply in
our LC reformatting work, although we might receive some born-digital
content in this form.  We have NO timestamp/timecode fields in the
database.  Even if we don't foresee using them in the short or medium
term, is it better to leave them out or put them in?</carl>

THE STRUCTURE OF THE FILE ITSELF:

<carl>This has been very instructive to me in the sense of making clear
how poorly I grasp what is going on! Yikes!!  I am inclined to assume that
Dave and the AES are authoritative.  But so far no one has called my
attention to an LC need to capture this kind of data in our reformatting
work.  And further, in reformatting, there is probably a reasonable
business rule to normalize data so we don't ever keep "raw" files.  But
maybe we already have files with hidden digital workstation data in them
and I don't know it.

<LCdb>Once again, I ask, if there are elements and attributes in the AES
schema that we cannot foresee using at this time, is it still a good
practice to put the fields in the database back end?</LCdb></carl>

<dave> [re MSU response]  I would not recommend leaving these details to
be "inferred from the basic type/format relationship" as you
suggest. There is a lot of goofy software out there. Sonic solutions has
no less that 4 ways of packing their audio into an AIFF. Some Apple
software (like OS X) writes AIFF little-endian despite the fact that the
Macintosh is traditionally thought of as a big-endian platform [and the
fact that the specification of AIFF [published by Apple] says that data
should be packed big-endian]. It is hard enough to find out what is going
on now, with the tools available to us.  What will someone 50 years from
now be able to infer about these data formats?

So, these issues are dealt with in 3 blocks of elements in the AES
draft.

<dave block 1>  Any data compression/bit rate reduction is described in
the following:

<!ELEMENT bit_rate_reduction EMPTY>
     <!ATTLIST bit_rate_reduction
            codec_name CDATA #REQUIRED
              codec_creator_app CDATA #REQUIRED
                codec_quality (lossy | code-regenerating) #REQUIRED
                 data_rate CDATA #IMPLIED
                data_rate_mode (fixed | variable) #IMPLIED > </dave block
1>

<carl>I am inclined to accept Dave's ideas.  Dave has three "units" for
codec: codec_name, codec_creator_app, codec_quality (lossy |
code-regenerating).

<LCdb>We used to have two database fields for audio and video, although
the October 31 version of the Excel data dictionary has these in video
only: compression_method and compression_amount.  Proposal: we accept
Dave's trio and use them in BOTH audio and video.</LCdb>

Dave has two "units" for data rate: data_rate and data_rate_mode (fixed |
variable).  We have these in both audio and video. Hooray.</carl>

<dave block 2>The method of encoding the audio is described in an
attribute of the format_att element: audio_data_encoding="pcm" DSD is
another likely value for this field.</dave block 2>

<carl>I am not sure I am know where Dave's "format_att" element lives,
which he refers to as though it were richly described in some other place
that he did not copy for us.  Does this then relate to the ongoing
discussion we have been having with MSU about "format" related
stuff?  What follows is my attempt to catalog what we have on deck in the
way of file-ish data:

In the fileGrp part of METS
1.  internet_media_type [MIMEtype]
2.  filesize
3.  datetime_ create, datetime_ modify, datetime_ access

From our ALLFILES set, now orphaned but likely to stay in techMD:
4.  Byte_order
5, 6, 7.  Checksum trio: (1) checksum_type (2) checksum_value
(3) checksum_datetime
8. Security (encryption)
9. Watermark.
10. Encapsulation.
11. Owner.
12.  Use.

And waiting in the wings, inspired by the MSU comments:
13. format_name [which may be used to identify the class, and thus be used
in data-administration system to "get" format-level documentation]
14. filename_extension

Finally, Dave's specific-to-audio example:
audio_data_encoding

PROPOSAL: For now, I am inclined to add audio_data_encoding (understood to
mean "digital sampling method") to audioMD, and hope that if it is an
attribute of something (format_att?) that we have in the right place in
our database tables.</carl>

<dave block 3> And then the actual nitty gritty details of how the bits
are arranged:
<!ELEMENT data_format EMPTY>
    <!ATTLIST data_format
           word_size CDATA #REQUIRED
                audio_data_orientation (big_endian |
little_endian) #REQUIRED
           first_sample_offset CDATA #IMPLIED
              audio_block_size CDATA #IMPLIED
        first_valid_byte_block CDATA #IMPLIED
           last_valid_byte_block CDATA #IMPLIED>

"The data_format element is used to describe the organization of digital
audio data at the byte level. The word_size attribute describes the number
of bytes that comprise a single sample of audio data. Generally this will
map directly to the format_att elements bit_depth attribute. Note that 24
bit will usually be expressed as a 3 byte word_size; though some
applications may store 24 bit audio in a 4 byte word. The
audio_data_orientation attribute indicates whether the sound data is
written in big endian format or little endian format. The last 4
attributes may be useful when describing sound files that implement a data
blocking scheme. The first_sample_offset is used to record the location of
the first valid sound byte in the file. The audio_block_size records the
size of the blocks in bytes. Additionally, some formats interleave
non-audio data with the sound data. The first_valid_byte_block indicates
where the firs t byte of valid sound within a block is located while
last_valid_byte_block indicates the last valid byte of sound data within
the block, forming a range of valid audio data. All other data in the
block is non-audio data." </dave block 3>

<carl>Dick, Peter Alyea, and I struggled with these data units months ago.
Now we have a bit more explanation and see that they will be in the AES
schema.

<LCdb>The argument for putting these in the db back end is not that we
will reformat in a manner that requires them (I think) but rather that we
may receive born digital content structured in this way, e.g., from a
sister archive like Harvard.  I have no earthly idea if these reappear in
the world of video and am inclined just now to keep this to audio.

Proposal: add these six fields to audioMD table:
(1) word_size. Looks like our default value will be 3 bytes.  Or 2 bytes
when we get 16-bit content, less frequent.
(2) audio_data_orientation (big_endian | little_endian).  Redundant with
byte_order, which we may still be getting in our ALLFILES or in fileGrp,
depending on where Jerry ends up.
(3) first_sample_offset, (4) audio_block_size, (5) first_valid_byte_block,
and (6) last_valid_byte_block.  Peter, do we do any audio work in formats
that have a "data blocking scheme?"  If not, these will just sit there
empty.</LCdb></carl>

WHEW!  Carl

Top of Message | Previous Page | Permalink

Advanced Options


Options

Log In

Log In

Get Password

Get Password


Search Archives

Search Archives


Subscribe or Unsubscribe

Subscribe or Unsubscribe


Archives

January 2020
December 2019
November 2019
October 2019
September 2019
August 2019
May 2019
April 2019
March 2019
November 2018
October 2018
September 2018
July 2018
June 2018
April 2018
February 2018
January 2018
December 2017
November 2017
October 2017
September 2017
July 2017
June 2017
May 2017
April 2017
January 2017
October 2016
August 2016
July 2016
June 2016
May 2016
April 2016
January 2016
September 2015
August 2015
June 2015
May 2015
April 2015
March 2015
February 2015
October 2014
September 2014
August 2014
July 2014
June 2014
May 2014
April 2014
January 2014
December 2013
October 2013
September 2013
August 2013
June 2013
May 2013
April 2013
February 2013
November 2012
October 2012
September 2012
August 2012
June 2012
May 2012
March 2012
February 2012
January 2012
November 2011
October 2011
September 2011
August 2011
July 2011
June 2011
May 2011
April 2011
March 2011
February 2011
January 2011
December 2010
November 2010
October 2010
September 2010
August 2010
July 2010
June 2010
May 2010
April 2010
February 2010
January 2010
December 2009
November 2009
October 2009
September 2009
August 2009
July 2009
June 2009
May 2009
April 2009
March 2009
February 2009
January 2009
December 2008
November 2008
October 2008
September 2008
August 2008
July 2008
June 2008
May 2008
April 2008
March 2008
February 2008
January 2008
December 2007
November 2007
October 2007
September 2007
August 2007
July 2007
June 2007
May 2007
April 2007
March 2007
February 2007
January 2007
December 2006
November 2006
October 2006
September 2006
August 2006
July 2006
June 2006
May 2006
April 2006
March 2006
February 2006
January 2006
December 2005
November 2005
October 2005
September 2005
August 2005
July 2005
June 2005
May 2005
April 2005
March 2005
February 2005
January 2005
December 2004
November 2004
October 2004
September 2004
August 2004
July 2004
June 2004
May 2004
April 2004
March 2004
February 2004
January 2004
December 2003
November 2003
October 2003
September 2003
August 2003
June 2003
May 2003
April 2003
March 2003
February 2003
January 2003
December 2002
November 2002
October 2002
September 2002
August 2002
July 2002
June 2002
May 2002
April 2002
March 2002
February 2002
January 2002
December 2001
November 2001
October 2001
September 2001
August 2001
July 2001
June 2001
May 2001
April 2001

ATOM RSS1 RSS2



LISTSERV.LOC.GOV

CataList Email List Search Powered by the LISTSERV Email List Manager