Network Working Group
Request for Comments: 3625
Updates: 3555
Category: Informational
R. Gellens
H. Garudadri
Qualcomm
September 2003

The QCP File Format and Media Types for Speech Data

Status of this Memo

This memo provides information for the Internet community. It does not specify an Internet standard of any kind. Distribution of this memo is unlimited.

Copyright Notice

Copyright © The Internet Society (2003). All Rights Reserved.

Abstract

RFC 2658 specifies the streaming format for 3GPP2 13K vocoder (High Rate Speech Service Option 17 for Wideband Spread Spectrum Communications Systems, also known as QCELP 13K vocoder) data, but does not specify a storage format. Many implementations have been using the "QCP" file format (named for its file extension) for exchanging QCELP 13K data as well as Enhanced Variable Rate Coder (EVRC) and Selectable Mode Vocoders (SMV) data. (For example, Eudora(r), QuickTime(r), and cmda2000(r) handsets).

This document specifies the QCP file format and updates the audio/qcelp media registration to specify this format for storage, and registers the audio/evrc-qcp and audio/smv-qcp media types for EVRC and SMV (respectively) data stored in this format.

Table of Contents

   1.  Introduction . . . . . . . . . . . . . . . . . . . . . . . . .  2
   2.  Conventions Used in this Document. . . . . . . . . . . . . . .  2
   3.  QCP File Format. . . . . . . . . . . . . . . . . . . . . . . .  2
   4.  IANA Considerations. . . . . . . . . . . . . . . . . . . . . . 10
       4.1.  Update of Media Type Audio/qcelp . . . . . . . . . . . . 10
       4.2.  Update of Media Type Audio/vnd.qcelp . . . . . . . . . . 10
       4.3.  Registration of Audio/EVRC-QCP . . . . . . . . . . . . . 11
       4.4.  Registration of Audio/SMV-QCP. . . . . . . . . . . . . . 12
   5.  Security Considerations. . . . . . . . . . . . . . . . . . . . 13
   6.  Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . 13
   7.  References . . . . . . . . . . . . . . . . . . . . . . . . . . 13
       7.1.  Normative References . . . . . . . . . . . . . . . . . . 13
       7.2.  Informative References . . . . . . . . . . . . . . . . . 13
   8.  Intellectual Property Statement. . . . . . . . . . . . . . . . 14
   9.  Editors' Addresses . . . . . . . . . . . . . . . . . . . . . . 14
   10. Full Copyright Statement . . . . . . . . . . . . . . . . . . . 15

1. Introduction

This document specifies the QCP format for storage of [QCELP] 13K, [EVRC], and [SMV] vocoder frames in files which may reside on network elements (e.g., media servers, Multimedia Messaging System (MMS) centers, etc.) and third-generation cellular (3G) handsets, or be transmitted via email or other mechanisms.

Vocoder frames read from a QCP file may be streamed using protocols and formats outside the scope of this document, delivered to an [EVRC], [SMV], or [QCELP] 13K decoder, or otherwise processed.

The QCP format supports [QCELP] 13K as well as [EVRC] and [SMV] data. Note that this format is different from the EVRC and SMV storage format specified in [EVRC-SMV].

2. Conventions Used in this Document

The key words "REQUIRED", "MUST", "MUST NOT", "SHOULD", "SHOULD NOT", and "MAY" in this document are to be interpreted as described in BCP 14, RFC 2119 [KEYWORDS].

3. QCP File Format

The following describes the QCP file format using [ABNF].

   A qcp file contains data frames generated by the [QCELP], [EVRC], or
   [SMV] vocoders.
   qcp-file        = riff-qlcm fmt vrat [labl] [offs] data [cnfg]
                     [text]
   
   riff-qlcm       = RIFF riff-size QLCM
   
   fmt             = FMT chunk-size major minor codec-info
                     ; major and minor are set to the version of
                     ; the QCP file format used to generate the
                     ; file - currently, the following versions
                     ; are defined:
                     ;
                     ; major = 2, minor = 0
                     ;    used for SMV.
                     ; major = 1, minor = 0
                     ;    used for QCELP-13K and EVRC
   
   vrat            = VRAT chunk-size var-rate-flag size-in-packets
   
   labl            = LABL chunk-size label
   
   offs            = OFFS chunk-size step-size num-offsets *offset
                     ; number of repetitions of offset must equal
                     ; num-offsets
   
   data            = DATA chunk-size *packet [pad]
                     ; repetitions of packet contain the actual
                     ; packet data -- there should be as many
                     ; packets as indicated by the size-in-packets
                     ; element found elsewhere in this file format
   
   cnfg            = CNFG chunk-size config
   
   text            = TEXT chunk-size string [pad]
                     ; string contains any information defined
                     ; by the application
   
   riff-size       = UINT32
                     ; riff-size should equal total size of file
                     ; in bytes, excluding the size of RIFF and
                     ; riff-size
   
   chunk-size      = UINT32
                     ; chunk-size should equal the total size of
                     ; the chunk described by the rule containing
                     ; chunk-size, excluding the size of
                     ; chunk-size itself and any elements that
                     ; come before it in the rule, as well as
                     ; the pad element, if present
   
   major           = OCTET
                     ; major version number of QCP format --
                     ; currently set as "2" -- however, "1" should
                     ; be used instead for QCELP-13K and EVRC, to
                     ; maintain compatibility with older software
                     ; platforms
   
   minor           = OCTET
                     ; minor version number: currently set to "0"
   
   codec-info      = codec-guid codec-version codec-name
                             average-bps packet-size
                             block-size sampling-rate
                             sample-size variable-rate 5*UINT32
                     ; this identifies the codec used to encode
                     ; the speech in this file, and any parameters
                     ; needed in order to decode the speech

; the five UINT32 elements are reserved for ; use in future versions of this format, ; should be set to zero for now, and readers ; should allow non-zero values when reading ; files in this format

   var-rate-flag   = UINT32
                     ; if zero, sizes of packets in data chunk are
                     ; fixed at the value indicated by the
                     ; packet-size field, found elsewhere in this
                     ; file format

; if var-rate-flag is greater than zero but ; less than %xFFFF0000, packets are variable ; rate, and rate is determined by rate octet ; in each packet -- to determine the size of ; a packet, map the first octet (the rate ; octet) to the size of the rest of the ; packet, according to the rate-map-table ; information found elsewhere in this file

; values %xFFFF0001 and higher are reserved ; for use in future versions of this format

size-in-packets = UINT32

; this is the total number of packets in the ; data chunk in the file

   label           = 48OCTET
                     ; label is for generic storage for
                     ; application use
   
   step-size       = UINT32
                     ; difference in the times of sequential
                     ; application stored in this chunk, in
                     ; 100ms increments (step-size of 10 means
                     ; 1 second)

; currently only a step-size of "10" is ; guaranteed to be supported by software ; capable of reading this file format

   num-offsets     = UINT32
                     ; number of offsets in offs chunk -- must
                     ; equal number of repetitions of offset
                     ; element
   
   offset          = UINT32
                     ; absolute octet offset in this QCP file
                     ; where the beginning of the packet that is
                     ; at a time index one step past the last
                     ; offset is located -- step size is given
                     ; in step-size
   
   packet          = [rate-octet] (1*OCTET)
                     ; the rate-octet element is present only if
                     ; the var-rate-flag found elsewhere in the file
                     ; indicates that the file contains
                     ; variable-rate packets - if it is present, the
                     ; octet is used to determine the size of the
                     ; remainder of the packet (the number of octets
                     ; which follow the rate-octet)

; No rate-octet is present if the var-rate-flag ; is set to zero, indicating a fixed size ; packet file - in this case, the number of ; octets in packet is equal to the value set ; in packet-size, found elsewhere in this ; file format

; It should be noted that there is always a ; codec rate octet in a packet, even if the ; var-rate-flag is zero, indicating fixed size ; packet files - in this case, although there ; is no rate-octet element in this syntax, the ; first octet in the packet is still considered

; the "rate" for the packet.

; Two examples follow to illustrate this.

                     ; Example 1.  Variable-rate QCELP
                     ; var-rate-flag = %d1
                     ; variable-rate = %d5 %d34 %d4 %d16 %d3 %d7 %d2
                     ;                 %d3 %d1 %d0 %d0 3(%d0 %d0)
                     ; packet        = %d4 (34OCTET)
                     ; packet        = %d4 (34OCTET)
                     ; packet        = %d3 (16OCTET)
                     ; packet        = %d1 (3OCTET)
                     
                     ; Example 2.  Fixed-rate, full-rate QCELP
                     ; var-rate-flag = %d0
                     ; packet-size   = %d35
                     ; packet        = (35OCTET)  ; first octet is %d4
                     ; packet        = (35OCTET)  ; first octet is %d4
                     ; packet        = (35OCTET)  ; first octet is %d4
                     ; packet        = (35OCTET)  ; first octet is %d4
   
   pad             = %x00
                     ; pad is present only if the number of bytes in
                     ; the chunk described by the rule containing
                     ; pad would otherwise be odd - if present, its
                     ; size should NOT be included in the
                     ; calculation for any chunk-size element also
                     ; present in the chunk
   
   config          = UINT16
                     ; config is a bitmapped configuration word,
                     ; for application use
   
   string          = *(%x01-FF) %x00
                     ; this is a zero-terminated string of octets --
                     ; although not limited to it, typically the
                     ; string consists of us-ascii characters
   
   codec-guid      = UINT32 UINT16 UINT16 8OCTET
                     ; this is the unique identifier for the codec
                     ; used to encode the packets in the data chunk

; The elements of this rule match the structure ; defined for a GUID in other specifications ; and formats. The first three elements are ; stored in little-endian octet order.
; When values for a GUID are expressed, the ; first three elements are expressed as a

; sequence of hexadecimal digits in normal ; network ordering (big-endian or most ; significant digit first) while the eight ; octet element is broken up into two groups, ; the first having four hexadecimal digits, ; and the second having twelve digits, all ; expressed in network ordering.

                     ; For example, if the octets in a GUID as
                     ; stored in a file are:
                     ;   %12 %34 %56 %78 %9A %BC %DE %F0
                     ;   %0F %ED %CB %A9 %87 %65 %43 %21
                     ; then the GUID would have these values:
                     ;   UINT32 = %x78563412
                     ;   UINT16 = %xBC9A
                     ;   UINT16 = %xF0DE
                     ;   8OCTET = %x0F.ED.CB.A9.87.65.43.21
                     ; and the whole GUID would be expressed as:
                     ;   {78563412-BC9A-F0DE-0FED-CBA987654321}
                     
                     ; The following codec GUIDs are currently
                     ; defined for QCP file format:
                     ;
                     ; QCELP-13K:
                     ;   {5E7F6D41-B115-11D0-BA91-00805FB4B97E}
                     ;   {5E7F6D42-B115-11D0-BA91-00805FB4B97E}
                     ; EVRC:
                     ;   {E689D48D-9076-46B5-91EF-736A5100CEB4}
                     ; SMV:
                     ;   {8D7C2B75-A797-ED49-985E-D53C8CC75F84}
   
   codec-version   = UINT16
                     ; version number of codec used to encode the
                     ; packets in the data chunk

; This value depends on the particular codec ; used to encode the packets. The following ; versions are currently defined:

                     ; QCELP-13K:
                     ;   1 or 2
                     ; EVRC, and SMV:
                     ;   1
   
   codec-name      = 80OCTET
                     ; the proper name of the codec, in us-ascii -
                     ; unused octets after the name are set to zero
   
   average-bps     = UINT16
                     ; average data rate, in bits per second, of
                     ; the speech data represented in this file
   
   packet-size     = UINT16
                     ; the size in octets of the largest possible
                     ; packet in the data chunk
   
   block-size      = UINT16
                     ; the number of samples encoded in every packet
                     ; in the data chunk
   
   sampling-rate   = UINT16
                     ; number of speech samples per second
                     ; (typically 8000)
   
   sample-size     = UINT16
                     ; number of bits per speech sample
                     ; (typically 16)
   
   variable-rate   = num-rates rate-map-table
                     ; if num-rates is zero, and major version
                     ; number of the QCP file is 2, then the
                     ; rate-map-table is not used - instead,
                     ; it is up to the decoder to determine the
                     ; sizes of packets in the data chunk, even
                     ; if var-rate-flag indicates the file uses
                     ; variable rate packets

; otherwise, num-rates specifies how many ; different possible rate octets there are ; for the packets in the data chunk, and ; thus how many of the rate-map-entry elements ; contain valid information

   num-rates       = UINT32
                     ; this is the number of possible rate octets
                     ; used in the packets in the data chunk
   
   rate-map-table  = 8rate-map-entry
                     ; any unused entries SHOULD be filled with
                     ; %d0 %d0
   
   rate-map-entry  = rate-size rate-octet
                     ; this maps a possible rate octet for a packet
                     ; to the size of the rest of the packet having
                     ; that value for the rate octet
   
   rate-size       = OCTET
                     ; this is the size of a packet, excluding the
                     ; value for the rate octet
   
   rate-octet      = OCTET
                     ; this is the first octet of a packet in the
                     ; data chunk, when the var-rate-flag set for
                     ; the file indicates the file is variable rate
   
   UINT32          = 4OCTET
                     ; this field contains a 32-bit integer stored
                     ; as a sequence of four octets, in
                     ; little-endian order (least significant
                     ; octet first)
   
   UINT16          = 2OCTET
                     ; this field contains a 16-bit integer stored
                     ; as a sequence of two octets, in
                     ; little-endian order (least significant
                     ; octet first)
   
   OCTET           = %x00-FF
                     ; an octet, also called a byte - any possible
                     ; combination of eight bits, forming a single
                     ; integer or part of a larger integer having
                     ; more than eight bits
   
   RIFF            = %x52 %x49 %x46 %x46
   
   QLCM            = %x51 %x4C %x43 %x4D
   
   FMT             = %x66 %x6D %x74 %x20
   
   LABL            = %x6C %x61 %x62 %x6C
   
   OFFS            = %x6F %x66 %x66 %x73
   
   DATA            = %x64 %x61 %x74 %x61
   
   CNFG            = %x63 %x6E %x66 %x67
   
   TEXT            = %x74 %x65 %x78 %x74

4. IANA Considerations

IANA has updated the audio/qcelp and audio/vnd.qcelp registrations and has added the audio/evrc-qcp and audio/smv-qcp registrations as specified here.

4.1. Update of Media Type Audio/qcelp

The audio/qcelp media registration has been updated to indicate that this specification is to be used for storage.

4.2. Update of Media Type Audio/vnd.qcelp

The audio/vnd.qcelp media registration has been updated to indicate that use of this media type is deprecated and to note that the media type audio/qcelp should be used instead.

4.3. Registration of Audio/EVRC-QCP

   Media Type Name:           audio
   
   Media Subtype Name:        evrc-qcp
   
   Required Parameter:        none
   
   Optional parameters:       none

Encoding considerations:

The storage format specified in this document may be used with any transport mechanism.

Security considerations:

See Section 5 "Security Considerations" of this document.

   Public specification:       this document
   
   Additional information:     no

Magic number:

                     First four octets: RIFF
                     Octets 9-12: QLCM

(Note: octets 5-8 constitute the riff-size field, which is the size of the file minus the RIFF header. Since this could be anything, it is not fixed and thus can not be used as part of the magic number.)

File extensions: qcp, QCP
Macintosh file type code: none
Object identifier or OID: none

Intended usage:

COMMON. This file format is already in wide use in Internet email user agents, multimedia authoring and playing software, and cdma2000(r) handsets.

Person & email address to contact for further information:

       Harinath Garudadri hgarudad@qualcomm.com

Change controller:

       The IETF

4.4. Registration of Audio/SMV-QCP

   Media Type Name:           audio
   
   Media Subtype Name:        smv-qcp
   
   Required Parameter:        none
   
   Optional parameters:       none

Encoding considerations:

The storage format specified in this document may be used with any transport mechanism.

Security considerations:

See Section 5 "Security Considerations" of this document.

   Public specification:       this document
   
   Additional information:     no

Magic number:

                     First four octets: RIFF
                     Octets 9-12: QLCM

(Note: octets 5-8 constitute the riff-size field, which is the size of the file minus the RIFF header. Since this could be anything, it is not fixed and thus can not be used as part of the magic number.)

File extensions: qcp, QCP
Macintosh file type code: none
Object identifier or OID: none

Intended usage:

COMMON. This file format is already in wide use in Internet email user agents, multimedia authoring and playing software, and cdma2000(r) handsets.

Person & email address to contact for further information:

       Harinath Garudadri hgarudad@qualcomm.com

Change controller:

       The IETF

5. Security Considerations

This document specifies a file format only, not a streaming protocol payload format, nor a transfer method. As such, it introduces no security risks aside from those associated with any audio codec or media file format (for example, denial of service by transmitting a file larger than the receiver can handle). Note that those security concerns should be understood before using the file format specified here.

6. Acknowledgements

Richard Walters created the ABNF notation for this specification and proof-read the text, among other helpful tasks.

The qcp file format was originally developed by others within Qualcomm. The editor would like to thank Chuck Han and Livingstone Song for their contributions leading to this specification.

7. References

7.1. Normative References

   [ABNF]      Crocker, D., Ed. and P. Overell, "Augmented BNF for
               Syntax Specifications: ABNF", RFC 2234, November 1997.
   
   [KEYWORDS]  Bradner, S., "Key words for use in RFCs to Indicate
               Requirement Levels", BCP 14, RFC 2119, March 1997.
   
   [QCELP]     3GPP2 C.S0020 "High Rate Speech Service Option 17 for
               Wideband Spread Spectrum Communications Systems".
   
   [EVRC]      3GPP2 C.S0014 "Enhanced Variable Rate Codec, Speech
               Service Option 3 for Wideband Spread Spectrum Digital
               Systems ". (Used to be TIA/EIA/IS-127)
   
   [PureVoice] McKay, K., "RTP Payload Format for PureVoice(tm) Audio",
               RFC 2658, August 1999.
   
   [SMV]       3GPP2 C.S0030 "Selectable Mode Vocoder, Service Option
               for Wideband Spread Spectrum Communication Systems".

7.2. Informative References

   [EVRC-SMV]  Li, A., "RTP Payload Format for Enhanced Variable Rate
               Codecs (EVRC) and Selectable Mode Vocoders (SMV)", RFC
               3558, July 2003.

8. Intellectual Property Statement

The IETF takes no position regarding the validity or scope of any intellectual property or other rights that might be claimed to pertain to the implementation or use of the technology described in this document or the extent to which any license under such rights might or might not be available; neither does it represent that it has made any effort to identify any such rights. Information on the IETF's procedures with respect to rights in standards-track and standards-related documentation can be found in BCP-11. Copies of claims of rights made available for publication and any assurances of licenses to be made available, or the result of an attempt made to obtain a general license or permission for the use of such proprietary rights by implementors or users of this specification can be obtained from the IETF Secretariat.

The IETF invites any interested party to bring to its attention any copyrights, patents or patent applications, or other proprietary rights which may cover technology that may be required to practice this standard. Please address the information to the IETF Executive Director.

9. Editors' Addresses

   Randall Gellens
   QUALCOMM Incorporated
   5775 Morehouse Drive
   San Diego, CA  92121
   USA

EMail:

          randy@qualcomm.com
   
   Harinath Garudadri
   QUALCOMM Incorporated
   5775 Morehouse Drive
   San Diego, CA  92121
   USA

EMail:

          hgarudad@qualcomm.com

10. Full Copyright Statement

Copyright © The Internet Society (2003). All Rights Reserved.

This document and translations of it may be copied and furnished to others, and derivative works that comment on or otherwise explain it or assist in its implementation may be prepared, copied, published and distributed, in whole or in part, without restriction of any kind, provided that the above copyright notice and this paragraph are included on all such copies and derivative works. However, this document itself may not be modified in any way, such as by removing the copyright notice or references to the Internet Society or other Internet organizations, except as needed for the purpose of developing Internet standards in which case the procedures for copyrights defined in the Internet Standards process must be followed, or as required to translate it into languages other than English.

The limited permissions granted above are perpetual and will not be revoked by the Internet Society or its successors or assignees.

This document and the information contained herein is provided on an "AS IS" basis and THE INTERNET SOCIETY AND THE INTERNET ENGINEERING TASK FORCE DISCLAIMS ALL WARRANTIES, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO ANY WARRANTY THAT THE USE OF THE INFORMATION HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE.

Acknowledgement

Funding for the RFC Editor function is currently provided by the Internet Society.