Main title

Embedding Metadata in Digital Audio Files

Guideline for Federal Agency Use of Broadcast WAVE Files

By the Federal Agencies Audio-Visual Working Group

http://www.digitizationguidelines.gov/audio-visual/

Version 1.0. Approved version, September 15, 2009 (clarification wording November 24, 2009)

TABLE OF CONTENTS

I. THE BEXT CHUNK

I.A. Strongly recommended elements (page 2)

Originator, OriginatorReference, Description (for additional identifiers),

OriginationDate, Version, Reserved

I.B. Recommended element (page 5)

TimeReference

I.C. Optional elements (page 5)

OriginationTime, CodingHistory

II. THE INFO CHUNK

II.A. Recommended element (page 7)

IARL (archival location)

II.B. Optional elements (page 7)

INAM (name, title), ICMT (comment for identifiers), ICRD (creation date), ICOP

(copyright and other restrictions)

III. SPECIFICATIONS AND OTHER REFERENCES (page 9)

What is this document?

This is one of three docum

ents pertaining to the embedding of metadata in digital audio files

prepared by the Federal Agencies Audio-Visual Working Group in 2009. The two companion

documents are:

 Introductory Discussion for the Proposed Federal Agencies Guideline

http://www.digitizationguidelines.gov/audio-visual/documents/Embed_Intro_090915.pdf

 Consultant’s report on embedding options in digital audio files

http://www.digitizationguidelines.gov/audio-visual/documents/AVPS_Audio_Metadata_Overview_090612.pdf

PART I. THE BEXT CHUNK

I.A. Strongly recommended elements for bext chunk. If the Working Group had the

authority to do so, these would be "required."

Originator

BWF Spec ASCII string (maximum 32 characters) containing the name of the originator/producer of the

audio file. If the length of the string is less than 32 characters, the field is ended by a null

character.

Fed Agencies

Application

This element contains the entity responsible for the "archiving" (creation, maintenance,

preservation) of this digital item. Entity designations should be as specific as possible

including a two-character county code to avoid the potential for conflict in the archiving

entity’s name.

If space permits within the 32 character limit, the archival entity should be identified at the

most specific level within the institution.

Use a standard abbreviation of entity names such as those found in the Guide to Government

Acronyms & Abbreviations. If an entity is not on this list, use a familiar abbreviation. Use

the standard two-character ISO 3166 alpha 2 country code list

Char limit 32

Mandatory/optional Strongly recommended (if the Working Group had authority: "required")

Values [Country code]comma space[Entity name]

Example LC US, LOC/RSS

[RSS = Recorded Sound Section]

Example NARA US, NARA

Example EPA US, EPA

OriginatorReference

BWF Spec ASCII string (maximum 32 characters) containing a non ambiguous reference allocated by

the originating organization. If the length of the string is less than 32 characters, the field is

ended by a null character.

Fed Agencies

Application

This element contains the principal identifier or the "best" identifier which uniquely

differentiates one object from another, preferably at the file level.

If the principal identifier string is less than 32 characters, enter the entire identifier string.

If the principal identifier string is longer than 32 characters, enter this text: "See Description

for identifiers."

Do not embed identifiers which could pose a possible security risk, e.g., by exposing exact

pathnames.

NOTE: The Working Group perceived value in the practice of repeating the principal

identifier as the first identifier in the BEXT Description element (which has 256 available

characters) but did not make this detail a strong recommendation. Comments from readers

are welcome.

Char limit 32

Mandatory/optional Strongly recommended (if the Working Group had authority: "required")

Values Under 32 characters: Identifier string

Over 32 characters: See Description

Example NARA See Description for identifiers

[Explanation: actual identifiers extend more than 32 characters.]

Example LC 1201566-2-1

[Explanation: unique number generated by the MAVIS collections management

database. Preferred identifier for the LC Recorded Sound Section. Generally for

local use.]

Description

BWF Spec ASCII string (maximum 256 characters) containing a free description of the sequence. To

help applications which only display a short description, it is recommended that a résumé of

the description is contained in the first 64 characters, and the last 192 characters are use for

details. If the length of the string is less than 256 characters, the last one is followed by a

null character. (00)

Fed Agencies

Application

COMMENT: This element is recommended as a container for identifiers for the work at

hand and/or as pointers to additional, non-embedded (externally maintained) metadata.

Members of the Working Group have repeatedly encountered the need to provide multiple

identifiers for a given item. The resulting extent of data cannot be accommodated in the

OriginatorReference element. For these reasons, the Working Group's recommendations for

the Description element deviate from the EBU specification.

In some cases, the 256-character limit will prevent an agency from listing all of its

identifiers; the most important or helpful should be provided.

Do not embed identifiers which could pose a possible security risk, e.g., by exposing exact

pathnames.

NOTE: The Working Group perceived value in two practices but wished to leave these as

optional. The first is the tagging of identifiers (see examples, typically URLs) to permit

them to be properly understood. The second is the repetition of the principal identifier (as

provided without tagging) in OriginatorReference as the first identifier in Description, where

labeling as to its origin or purpose can be provided. Comments from readers are welcome.

Char limit 256

Mandatory/optional Strongly recommended (if the Working Group had authority: "required")

Values If labeled: Identifier [comma space] type [comma space] comment [semicolon-space if more

than one identifier]

If no labeling: Identifier

Example NARA [Two labeled identifiers]

58979818, local, principal ID original filename; 306-MUSA-9658B, local, RG-Series-Item

Number

Example LC [One labeled identifier]

http://hdl.loc.gov/loc.mbrsmi/westhpp.2033, URL, principal ID handle

Example LC [One unlabeled identifier]

http://hdl.loc.gov/loc.mbrsmi/westhpp.2033

Example LC [One unlabeled identifier]

1201566-2-1

[Explanation: unique number generated by the MAVIS collections management

database. Generally for local use.]

Example LC [One unlabeled identifier]

RYI_6039

[Explanation: Recorded Sound Section shelf number for the original physical item

that has been digitally reformatted. For local use.]

Example LC [One unlabeled identifier]

Harmonia Mundi France HM 957

[Explanation: Label information for a phonodisc in the Recorded Sound Section

collection]

Example California

Digital Library

[One labeled identifier]

ark.cdlib.org.org/ark:/13030/tf5p30086k, URL, ARK

Example LC [One labeled identifier]

http://lccn.loc.gov/mp76000002, URL, Permalink

OriginationDate

BWF Spec Ten ASCII characters containing the date of creation of the audio sequence.

Fed Agencies

Application

This element contains the file creation date. This is understood to mean the local date in the

timezone for the archival entity; the structure of the bext chunk does not permit ISO 8601

datetime indication, which unambiguously indicates date and time in terms of UTC

(Coordinated Universal Time or Temps Universel Coordonné).

Char limit 10

Mandatory/optional Strongly recommended (if the Working Group had authority: "required")

Values ISO 8601 YYYY-MM-DD. Year is defined from 0000 to 9999; Month is defined from 01 to

12 (use leading zeroes if less than 10); Day is defined from 01 to 28, 29, 30 or 31 (use

leading zeroes if less than 10). The separator between the items is a hyphen

[-] with no spaces. Note: the Working Group adheres to the ISO 8601 truncation

convention: the year should always be given but, if unknown, values for month and day are

omitted. Thus the string length may be four, seven, or ten. If the string entered is less than

ten characters, end it with a null character.

Example 2005-11-09

Version

BWF Spec An unsigned binary number giving the version of the BWF, particularly the contents of the

Reserved field.

UMID 64 bytes containing a UMID (Unique Material Identifier) to the SPMTE 330M

standard. If only a 32 byte .basic UMID. is used, the last 32 bytes should be set to zero. (The

length of the UMID is given internally.) Note: The EBU intends to publish guidance on the

use of the UMID in audio files.

This element contains the BWF version. The current version is Version 1 which replaced

Version 0. For Version 1, this element is set to 0001h.

Fed Agencies

Application

Follow EBU specification.

NOTE regarding the UMID identifier. The Working Group’s understanding is that when

file-reading software recognizes an instance of BWF version 1, it will look in the Version

field to see if there is a UMID and report it. Readers that recognize an instance of version 0

will not look for a UMID. Meanwhile, no federal agencies are using UMID for audio files

at this time and the Working Group has no particular recommendation regarding this

identifier. NARA reports that their digitization systems produce version 1 files with the

"slot" for the UMID left blank.

Char limit 64 bytes reserved for UMID

Mandatory/optional Required by EBU specification

Values

Example 0001h

Reserved

BWF Spec 190 bytes reserved for extensions. If the Version field is set to 0001h, these 190 bytes must

be set to a NULL (zero) value.

Fed Agencies

Application

Follow EBU specification.

Char limit 190 bytes

Mandatory/optional Required by EBU specification

Values

Example NULL

I.B. Recommended element for bext chunk

TimeReference

BWF Spec This field contains the timecode of the sequence. It is a 64-bit value which contains the first

sample count since midnight. The number of samples per second depends on the sample

frequency which is defined in the field <nSamplesPerSec> from the <format chunk>.

Fed Agencies

Application

The Working Group believes that TimeReference can be used in a manner that supports the

management of files that are parts of multitrack or multisegment works and plans to provide

recommendations in the future.

Char limit 64-bit value

Mandatory/optional Optional but recommended for complex recordings

Values

Example 2374

I.C. Optional elements for bext chunk

OriginationTime

BWF Spec Eight ASCII characters containing the time of creation of the audio sequence.

Fed Agencies

Application

This element contains the file creation time. This is understood to mean the local time in the

timezone for the archival entity; the structure of the bext chunk does not permit ISO 8601

datetime indication, which unambiguously indicates date and time in terms of UTC

(Coordinated Universal Time or Temps Universel Coordonné).

Char limit 8

Mandatory/optional Optional

Values ISO 8601 HH:MM:SS. Hour is defined from 00 to 23 (use leading zeroes if less than 10).

Minute and second are defined from 00 to 59 (use leading zeroes if less than 10). The

separator between the items is a colon [:] with no spaces. Note: the Working Group adheres

to the ISO 8601 truncation convention: the hour should always be given but, if unknown,

values for minutes and second are omitted. Thus the string length may be two, five, or eight.

If the string entered is less than eight characters, end the data with a null character.

Example 01:45:25

CodingHistory

BWF Spec Non-restricted ASCII characters, containing a collection of strings terminated by CR/LF.

Each string contains a description of a coding process applied to the audio data. Each new

coding application is required to add a new string with the appropriate information. This

information must contain the type of sound (PCM or MPEG) with its specific parameters:

PCM : mode (mono, stereo), size of the sample (8, 16 bits) and sample frequency: MPEG :

sample frequency, bit-rate, layer (I or II) and the mode (mono, stereo, joint stereo or dual

channel). It is recommended that the manufacturers of the coders provide an ASCII string

for use in the coding history.

Fed Agencies

Application

This element is designed to hold data on the digitizing process including signal chain

specifics, sample rate and bit depth, and other elements. It is defined as a collection of

strings, each presented on a separate line, containing a history of the coding processes

applied to the file. The first line documents the analog source recording, the second line

contains data on the capture process, the third line of data records information on the storage

of the file. A new line is added when the coding history related to the file is changed.

Char limit Non-restricted

Mandatory/optional Optional

Values Each variable within a string is separated by a comma-space and each line should end with a

carriage return and line feed.

Summary of subelements:

A=coding algorithm

F=sampling frequency

B=bit rate (only for MPEG)

W=word length

M=mode

T=free ASCII text string; contains no commas but semicolons may be used

Detail on subelement syntax:

A = Coding Algorithm <ANALOG, PCM, MPEG1L1, MPEG1L2, MPEG1L3,

MPEG2L1, MPEG2L2, MPEG2L3>

F=Sampling frequency <11000, 22050, 24000, 32000, 44100, 48000> Implied unit

of measure [Hz]

B (ONLY FOR MPEG ENCODING) = Bit-rate <any bit-rate allowed in MPEG 2

(ISO/IEC13818-3)>, Implied unit of measure [kbit/s per channel]

W= Word Length <8, 12, 14, 16, 18, 20, 22, 24> Implied unit of measure [bits]

M=Mode <mono, stereo, dual-mono, joint-stereo>

T=Text, free string <a free ASCII-text string for in house use. This string should

contain no commas (ASCII 2Chex). Examples of the contents: ID-No; codec type;

A/D type>

Example [Each line is another “chapter” in one history:]

A=ANALOG,M=stereo,T=Otari MX5050; SN3690; 15 ips; open reel tape

A=PCM,F=48000,W=18,M=stereo,T=NVision; NV1000; A/D

Example A=PCM,F=48000,W=16,M=stereo,T=PCX9; DIO

Example

SoundDirections

A=ANALOG,M=mono,T=Studer816; SN1007; 15 ips; open reel tape,

A=PCM,F=96000,W=24,M=mono,T=Pyramix1; SN16986,

A=PCM,F=96000,W=24,M=mono,T=Lynx; AES16; DIO,

[Explanation:

Line 1 reads: an analog, mono, open-reel tape played back on a Studer 816 tape

machine with serial number 1007 at tape speed 15 ips.

Line 2 reads: tape was digitized to PCM coding in mono mode at 96 kHz sampling

frequency and 24 bits per sample on a Pyramix 1 DAW with serial number 16986.

Line 3 reads: the audio was stored as a BWF file with PCM coding in mono mode

at 96 kHz sampling frequency and 24 bits per sample using a Lynx AES16 digital

input/output interface]

PART II. THE INFO CHUNK (“LIST info chunk”)

II.A. Recommended element for info chunk

IARL

RIFF Spec Archival Location: Indicates where the subject of the file is archived.

Fed Agencies

Application

This element is an approximate equivalent to the Originator element in the bext chunk which

records the entity responsible for the "archiving" (creation, maintenance, preservation) of

this digital item. If used, this element could repeat the data from the Originator element in

the same structured format. In some applications, it is automatically carried over to the MP3

ID3 element "ArchivalLocation." Of all the info list chunk elements, this is the most

important for Federal Agencies because it is vital to document the archival responsibility for

a content item.

Entity designations should be as specific as possible including a two-character county code

to avoid the potential for conflict in the archiving entity’s name. The archival entity should

be identified at the most specific level within the institution.

Use a standard abbreviation of entity names such as those found in the Guide to Government

Acronyms & Abbreviations. If an entity is not on this list, use a familiar abbreviation. Use

the standard two-character ISO 3166 alpha 2 country code list

Mandatory/optional Strongly recommended (“required”) if using info list chunk

Values [Country code]comma space[Entity name]

Example LC US, LOC/RSS

[RSS = Recorded Sound Section]

Example NARA US, NARA

II.B. Optional elements for info chunk

INAM

RIFF Spec Name: Stores the title of the subject of the file, such as, Seattle from Above.

Fed Agencies

Application

If used, this element should contain a working title. This phrase is intended to convey the

Working Group’s understanding that metadata may be updated over time. Users who refer

to elements like titles in an embedded block of metadata should understand that this may not

represent the latest and best information, and they should be counseled to follow one of the

identifiers in order to obtain later and/or better data.

Mandatory/optional Optional

Values Free text

Example LC Symphony no. 3 in A minor, op. 56

Example LC Interview with Bo Dollis at WWOZ, New Orleans, 1999-03-24 [title assigned by cataloger]

Example LC Adventures in research. No. 587, Sawing off Manhattan Island [episode in a radio series]

ICMT

RIFF Spec Comment: Provides general comments about the file or the subject of the file. If the

comment is several sentences long, end each sentence with a period. Do not include newline

characters (such as CR/LF/EOL).

Fed Agencies

Application

If used, this element should repeat the principal and other identifiers from the

OriginatorReference and/or Description element in the bext chunk. Each identifier will be

labeled as to its origin or purpose using the “type” and “comment” qualifiers. Using the tag

“local” in the "type" qualifier implies the entity identified in the bext Originator element. In

some applications, this info list element may carry forward automatically to the MP3 ID3

tagged element "Comment."

NOTE: The Working Group perceived value in the practices of labeling identifiers (see

examples) but wished to leave this as an optional practice.

Mandatory/optional Optional

Values If labeled: Identifier [comma space] type [comma space] comment [semicolon-space if more

than one identifier]

If no labeling: Identifier

Example NARA [Tagged identifier]

58979818, local, principal ID original filename

Example NARA [Tagged identifier]

306-MUSA-9658B, local, RG-Series-Item Number

Example LC [Two tagged identifiers, from LC, RSS]

1201566-2-1, local, system-generated number; RYI 6039, local, source location number

Example LC [Tagged identifier]

http://hdl.loc.gov/loc.mbrsmi/westhpp.2033, URL, principal ID Handle

Example California

Digital Library

[Tagged identifier]

ark.cdlib.org.org/ark:/13030/tf5p30086k, URL, ARK

Example LC [Unlabeled identifier]

http://lccn.loc.gov/mp76000002

ICRD

RIFF Spec Creation date: Specifies the date the subject of the file was created. List dates in year-month-

day format, padding one-digit months and days with a zero on the left. For example, 1999-

05-03 for May 3, 1999.

Fed Agencies

Application

This element is an equivalent to the OriginationDate in the bext chunk. If used, this element

could repeat the information from the OriginationDate element in the bext chunk using the

same structured format. In some applications, this info list element may carry forward

automatically to MP3 ID3 tagged element "CreationDate."

Mandatory/optional Optional

Values ISO 8601 YYYY-MM-DD. Year is defined from 0000 to 9999; Month is defined from 1 to

12; Day is defined from 1 to 28,29,30 or 31. The separator between the items is a hyphen [-]

with no spaces.

Example 2005-11-30

ICOP

RIFF Spec Copyright: Records the copyright information for the file. For example, Copyright

Encyclopedia International 1991. If there are multiple copyrights, separate them by a

semicolon followed by a space.

Fed Agencies

Application

Information about copyright and other restrictions (donor, privacy, etc.). Usage by federal

agencies will often refer to the documentation of restrictions provided by other, non-

embedded metadata. This element has no equivalent in the bext chunk. In some

applications, this info list element may carry forward automatically to MP3 ID3 tagged

element "Copyright."

If used, this element may contain the information as known at the time of embedding. The

Working Group understands that metadata may be updated over time. Users who refer to an

embedded block of metadata should understand that this may not represent the latest and

best information, and they should be counseled to follow one of the identifiers in order to

obtain later and/or better data.

Mandatory/optional Optional

Values Free text. If there are multiple copyrights or other restriction statements, separate them by a

semicolon followed by a space.

Example LC Publication and other forms of distribution may be restricted. For details, contact the

Recorded Sound Section of the Library of Congress. [Preferred wording from LC, RSS]

Example LC (p) Rhino Records 2002.

Example LC See Copyright Restriction Statement. [Used by LC, American Folklife Center]

III. SPECIFICATIONS AND REFERENCES

III.A. Specifications relevant to the preceding recommendations.

European Broadcast Union. Technical Recommendation R98-1999. Format for the

<CodingHistory> field in Broadcast Wave Format files, BWF. Online:

http://www.ebu.ch/CMSimages/en/tec_text_r98-1999_tcm6-4709.pdf. Accessed: March

31, 2009.

European Broadcast Union. Broadcast Wave Format (BWF) user guide. Online:

http://www.ebu.ch/en/technical/publications/userguides/bwf_user_guide.php. Accessed:

January 6, 2009.

European Broadcast Union. BWF: a format for audio data files in broadcasting. Version 1. July

2001. Online: http://www.ebu.ch/CMSimages/en/tec_doc_t3285_tcm6-10544.pdf.

Accessed: April 1, 2009.

IBM Corporation and Microsoft Corporation. Multimedia Programming Interface and Data

Specifications 1.0. Online: http://www.kk.iij4u.or.jp/~kondo/wave/mpidata.txt. Accessed:

April 1, 2009.

International Organization for Standardization. ISO 3166 Country Codes. Online:

http://www.iso.org/iso/country_codes/iso_3166_code_lists.htm or

http://www.iso.org/iso/english_country_names_and_code_elements. Accessed: June 5, 2009.

III.B. Other References

Casey, M. and B. Gordon. 2007. Sound Directions: Best Practices For Audio Preservation.

Trustees of Indiana University and President and Fellows of Harvard University. Online:

http://www.dlib.indiana.edu/projects/sounddirections/bestpractices2007/. Accessed:

January 6, 2009.

Harvard University Library. Administrative Metadata for Digital Audio Files. Online:

http://hul.harvard.edu/ois/systems/drs/audiometadata.pdf

. Accessed: January 13, 2009.

Indiana University Archives of Traditional Music. An Introduction to Technical and Structural

Metadata Elements in ATMC. Online:

http://www.dlib.indiana.edu/projects/sounddirections/papersP

resent/sd_app1_v1.pdf.

Accessed: January 13, 2009.

Smith, Kelly L. GovSpeak: A Guide to Government Acronyms & Abbreviations. Online:

http://members.cox.net/govdocs/govspeak.html

. Accessed: April 15, 2009.