Network Working Group A. Gupta Internet-Draft Qualcomm Technologies Intended status: Standards Track Incorporated Expires: November 3, 2013 May 2, 2013 RTP payload format for Enhanced Variable Rate Narrowband-Wideband plus 2kbps Codec (EVRC-NW2K) draft-agupta-payload-rtp-evrc-nw2k-00 Abstract This document specifies real-time transport protocol (RTP) payload formats to be used for the Enhanced Variable Rate Narrowband-Wideband plus 2kbps Codec (EVRC-NW2K). Three media type registrations are included for EVRC-NW2K RTP payload formats. In addition, a file format is specified for transport of EVRC-NW2K speech data in storage mode applications such as e-mail. Status of this Memo This Internet-Draft is submitted in full conformance with the provisions of BCP 78 and BCP 79. Internet-Drafts are working documents of the Internet Engineering Task Force (IETF). Note that other groups may also distribute working documents as Internet-Drafts. The list of current Internet- Drafts is at http://datatracker.ietf.org/drafts/current/. Internet-Drafts are draft documents valid for a maximum of six months and may be updated, replaced, or obsoleted by other documents at any time. It is inappropriate to use Internet-Drafts as reference material or to cite them other than as "work in progress." This Internet-Draft will expire on November 3, 2013. Copyright Notice Copyright (c) 2013 IETF Trust and the persons identified as the document authors. All rights reserved. This document is subject to BCP 78 and the IETF Trust's Legal Provisions Relating to IETF Documents (http://trustee.ietf.org/license-info) in effect on the date of publication of this document. Please review these documents carefully, as they describe your rights and restrictions with respect to this document. Code Components extracted from this document must include Simplified BSD License text as described in Section 4.e of Gupta Expires November 3, 2013 [Page 1] Internet-Draft EVRC-NW2K RTP payload format May 2013 the Trust Legal Provisions and are provided without warranty as described in the Simplified BSD License. Table of Contents 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 3 2. Conventions . . . . . . . . . . . . . . . . . . . . . . . . . 4 3. Background . . . . . . . . . . . . . . . . . . . . . . . . . . 5 4. EVRC-NW2K codec . . . . . . . . . . . . . . . . . . . . . . . 6 5. RTP header usage . . . . . . . . . . . . . . . . . . . . . . . 7 6. Payload format . . . . . . . . . . . . . . . . . . . . . . . . 8 6.1. Encoding capability identification in EVRC-NW2K interleaved/bundled format . . . . . . . . . . . . . . . . 8 7. Congestion Control Considerations . . . . . . . . . . . . . . 11 8. Storage format for the EVRC-NW2K Codec . . . . . . . . . . . . 12 9. IANA considerations . . . . . . . . . . . . . . . . . . . . . 13 9.1. Media Type Registrations . . . . . . . . . . . . . . . . . 13 9.1.1. Registration of Media Type audio/EVRCNW2K . . . . . . 13 9.1.2. Registration of Media Type audio/EVRCNW2K0 . . . . . . 15 9.1.3. Registration of Media Type audio/EVRCNW2K1 . . . . . . 16 10. SDP mode attributes for EVRC-NW2K . . . . . . . . . . . . . . 19 11. Mode Change Request/Response Considerations . . . . . . . . . 20 12. Mapping EVRC-NW2K media type parameters into SDP . . . . . . . 22 13. Offer-Answer Model Considerations for EVRC-NW2K . . . . . . . 23 14. Declarative SDP Considerations . . . . . . . . . . . . . . . . 25 15. Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 16. Security Considerations . . . . . . . . . . . . . . . . . . . 30 17. References . . . . . . . . . . . . . . . . . . . . . . . . . . 31 17.1. Normative References . . . . . . . . . . . . . . . . . . . 31 17.2. Informative References . . . . . . . . . . . . . . . . . . 32 Author's Address . . . . . . . . . . . . . . . . . . . . . . . . . 33 Gupta Expires November 3, 2013 [Page 2] Internet-Draft EVRC-NW2K RTP payload format May 2013 1. Introduction This document specifies the payload formats for packetization of EVRC-NW2K encoded speech signals into the real-time transport protocol (RTP). It defines support for the header-free, interleaved/ bundled, and compact bundle packet formats for the EVRC-NW2K codec as well as discontinuous transmission (DTX) support for EVRC-NW2K encoded speech transported via RTP. The EVRC-NW2K codec offers better speech quality than the EVRC and EVRC-B codecs and better capacity than EVRC-WB codec. The EVRC-NW2K codec adds a new 2kbps low rate codec to EVRC-NW. EVRC-NW2K belongs to the EVRC family of codecs. Gupta Expires November 3, 2013 [Page 3] Internet-Draft EVRC-NW2K RTP payload format May 2013 2. Conventions The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in RFC 2119 [1]. Gupta Expires November 3, 2013 [Page 4] Internet-Draft EVRC-NW2K RTP payload format May 2013 3. Background EVRC-NW2K is an extension of EVRC-B [2] EVRC-WB [3] and EVRC-NW [4] speech codecs developed in 3GPP2 with support for discontinuous transmission (DTX). It provides enhanced voice quality and high spectral efficiency. The EVRC-NW2K codec operates on 20 ms frames, and the default sampling rate is 16 kHz (wideband. Input and output at 8 kHz sampling rate (narrowband) is also supported. The EVRC-NW2K codec can operate in eight modes (0 to 7) defined in [5]. EVRC-NW2K modes 0, 1, and 7 are interoperable with EVRC-WB. EVRC-NW2K modes 1 to 2 and 4 to 7 are interoperable with EVRC-B. EVRC-NW2K modes 0 to 2 and 4 to 6 use the full set or a subset of full rate, 1/2 rate, 1/4 rate and 1/8 rate frames. EVRC-NW2K replaces mode 3 of EVRC-B with a new low rate 2kbps mode that uses only 1/4 rate and 1/8 rate frames. EVRC-NW2K mode 7 uses only 1/2 rate and 1/8 rate frames. By default, EVRC-NW2K supports all narrowband modes of EVRC-NW except 3(modes 1 to 7 except 3)which is replaced by a new narrowband 2kbps mode. Mode change among modes 1 to 7 excluding 3 (or among modes 0 to 7 excluding 3 if the receiver supports wideband mode) results in codec output bit-rate change but does not cause any decoding problems at the receiver. EVRC-NW2K provides a standardized solution for packetized voice applications that allow transitions between enhanced quality and increased capacity. The most important service addressed is IP telephony. Target devices can be IP phones or VoIP handsets, media gateways, voice messaging servers, etc. Gupta Expires November 3, 2013 [Page 5] Internet-Draft EVRC-NW2K RTP payload format May 2013 4. EVRC-NW2K codec The EVRC-NW2K codec operates on 20 ms frames. It produces output frames of one of the four different sizes: 171 bits (Rate 1), 80 bits (Rate 1/2), 40 bits (Rate 1/4), or 16 bits (Rate 1/8). In addition, there are two zero-bit codec frame types: blank (null) frames and erasure frames. The default sampling rate is 16 kHz. Input and output at 8 kHz sampling rate is also supported. The frame type values and sizes of the associated codec data frames are listed in the table below: Value Rate Total codec data frame size in bytes (and in bits) -------------------------------------------------------------------- 0 Blank(Null) 0 (0 bits) 1 1/8 2 (16 bits) 2 1/4 5 (40 bits) 3 1/2 10 (80 bits) 4 1 22 (171 bits; 5 bits padded at the end) 5 Erasure 0 (SHOULD NOT be transmitted by sender) Gupta Expires November 3, 2013 [Page 6] Internet-Draft EVRC-NW2K RTP payload format May 2013 5. RTP header usage The format of the RTP header is specified in RFC 3550 [6]. The EVRC- NW2K payload formats (Section 6) use the fields of the RTP header as specified in RFC 3550 [6]. EVRC-NW2K has also the capability to operate with 8 kHz sampled input/ output signals. The decoder does not require a priori knowledge about the sampling rate of the original signal at the input of the encoder. The decoder output can be at 8kHz or 16kHz regardless of the sampling rate used at the encoder. Therefore, depending on the implementation and the electroacoustic audio capabilities of the devices, the input of the encoder and/or the output of the decoder can be configured at 8 kHz; however, a 16 kHz RTP clock rate MUST always be used. The RTP timestamp is increased by 320 for each 20 milliseconds. The RTP header marker bit (M) SHALL be set to 1 if the first frame carried in the packet contains a speech frame which is the first in a talkspurt. For all other packets the marker bit SHALL be set to zero (M=0). Gupta Expires November 3, 2013 [Page 7] Internet-Draft EVRC-NW2K RTP payload format May 2013 6. Payload format Three RTP packet formats are supported for the EVRC-NW2K codec - the interleaved/bundled packet format, the header-free packet format, and the compact bundled packet format. For all these formats, the operational details and capabilities, such as ToC, interleaving, DTX, and bundling, of EVRC-NW2K are exactly the same as those defined in EVRC [7], EVRC-B [2], EVRC-WB [3], and EVRC-NW [4], except that 1. the mode change request field in the interleaved/bundled packet format MUST be interpreted according to the definition of the RATE_REDUC parameter as defined in EVRC-NW2K [5]. 2. the mode change request field in the interleaved/bundled packet format SHOULD be honored by an EVRCNW2K encoding end point in an one-to-one session with a dedicated EVRCNW2K decoding end point such as in a two-party call or in a conference leg. 3. the reserved bit field in the first octet of the interleaved/ bundled format has only one bit. Bit 1 of the first octet is an EVRC-NW2K wideband/narrowband encoding capability identification flag. The media type audio/EVRCNW2K maps to the interleaved/bundled packet format, audio/EVRCNW2K0 maps to the header-free packet format, and audio/EVRCNW2K1 maps to the compact bundled packet format. 6.1. Encoding capability identification in EVRC-NW2K interleaved/ bundled format The EVRC-NW2K interleaved/bundled format defines an encoding capability identification flag, which is used to signal the local EVRC-NW2K wideband/narrowband encoding capability at the time of construction of an RTP packet to the far end of a communication session. This capability identification flag allows the far end to use the MMM field in its out-going (returning) EVRC-NW2K interleaved/ bundled format packets to request the desired EVRC-NW2K wideband or narrowband encoding mode in accordance with the dynamic/instantaneous encoding capability information. See RFC 3558 [7] for the definition of MMM field. The following examples illustrate a few scenarios where the encoding capability information is used: o An end-to-end wideband communication is established first between two communication end points using EVRC-NW2K interleaved/bundled format. The called end point becomes wideband encoding incapable during the call and makes the other end aware of this change using the encoding capability identification flag. Based on the new information the calling end point could change the MMM value in Gupta Expires November 3, 2013 [Page 8] Internet-Draft EVRC-NW2K RTP payload format May 2013 its outgoing EVRC-NW2K packets from Mode-0 to Mode-4 to request narrowband encoded traffic for bandwidth efficiency or from Mode-0 to Mode-1 for best perceptual quality. o An end-to-end narrowband communication is established between an EVRC-NW2K wideband encoding capable calling end point and an EVRC- NW2K wideband encoding incapable called end point. The called end point becomes EVRC-NW2K wideband encoding capable during the call and makes the other end aware of this change using the encoding capability identification flag. Based on the new information the calling end point could change the MMM value in its outgoing EVRC- NW2K packets from non-Mode-0 to Mode-0 to request wideband traffic. EVRC-NW2K interleaved/bundled format defines the encoding capability identification flag in bit 1 of the first octet, as illustrated in the figure below. The flag shall be set to zero (C=0) when the local EVRC-NW2K encoder is capable of Mode-0 wideband encoding. The flag shall be set to one (C=1) when the local EVRC-NW2K encoder is capable of non-Mode-0 narrowband encoding only. See RFC 3558 [6] for original definitions of other fields in the interleaved/bundled format. 0 1 2 3 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | RTP Header | +=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+ |R|C| LLL | NNN | MMM | Count | TOC | ... | TOC |padding| +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | one or more codec data frames, one per TOC entry | | .... | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ Reserved (R): 1 bit Reserved bit. MUST be set to zero by sender, SHOULD be ignored by receiver. Encoding capability identification (C): 1 bit Must be set to zero by sender to indicate wideband encoding capable or set to one to indicate narrowband encoding capable only. Gupta Expires November 3, 2013 [Page 9] Internet-Draft EVRC-NW2K RTP payload format May 2013 C = 0 : Mode-0 wideband encoding capable = 1 : Mode-0 wideband encoding incapable, i.e. narrowband encoding only. Gupta Expires November 3, 2013 [Page 10] Internet-Draft EVRC-NW2K RTP payload format May 2013 7. Congestion Control Considerations Congestion control for RTP is discussed in RFC 3550 [6], and in applicable RTP profiles, e.g., RFC3551 [8]. This document does not change those considerations. Due to the header overhead, the number of frames encapsulated in each RTP packet influences the overall bandwidth of the RTP stream. Packing more frames in each RTP packet can reduce the number of packets sent and hence the header overhead, at the expense of increased delay and reduced error robustness. Gupta Expires November 3, 2013 [Page 11] Internet-Draft EVRC-NW2K RTP payload format May 2013 8. Storage format for the EVRC-NW2K Codec The storage format is used for storing EVRC-NW2K encoded speech frames, e.g., as a file or e-mail attachment. The file begins with a magic number to identify the vocoder that is used. The magic number for EVRC-NW2K corresponds to the ASCII character string "#!EVRCNW2K\n", i.e., "0x23 0x21 0x45 0x56 0x52 0x43 0x4E 0x57 0x32 0x4B 0x0A". The codec data frames are stored in consecutive order, with a single ToC entry field, extended to one octet, prefixing each codec data frame. The ToC field is extended to one octet by setting the four most significant bits of the octet to zero. For example, a ToC value of 4 (a full-rate frame) is stored as 0x04. The Value column in the table in Section 4 provides the TOC values for corresponding frame types. Speech frames lost in transmission and non-received frames MUST be stored as erasure frames (ToC value of 5) to maintain synchronization with the original media. Gupta Expires November 3, 2013 [Page 12] Internet-Draft EVRC-NW2K RTP payload format May 2013 9. IANA considerations This document introduces a new EVRC-NW2K 'audio' media subtype. 9.1. Media Type Registrations Following the guidelines in RFC 4855 [9] and RFC 4288 [10], this section registers new 'audio' media subtypes for EVRC-NW2K. 9.1.1. Registration of Media Type audio/EVRCNW2K Type name: audio Subtype names: EVRCNW2K Required parameters: None Optional parameters: These parameters apply to RTP transfer only. mode-set-recv: A subset of EVRC-NW2K modes. Possible values are a comma separated list of modes from the set {0,1,2,3,4,5,6,7} (see Table 2.7.1.2-1 in 3GPP2 C.S0014-E). A decoder can use this attribute to inform an encoder of its preference to operate in a specified subset of modes. Absence of this parameter signals the mode set {1,2,3,4,5,6,7}. ptime: see RFC 4566 [11]. maxptime: see RFC 4566. maxinterleave: Maximum number for interleaving length (field LLL in the Interleaving Octet)[0..7]. The interleaving lengths used in the entire session MUST NOT exceed this maximum value. If not signaled, the maxinterleave length MUST be 5. silencesupp: see Section 6.1 in RFC 4788. dtxmax: see Section 6.1 in RFC 4788. dtxmin: see Section 6.1 in RFC 4788. hangover: see Section 6.1 in RFC 4788. Encoding considerations: This media type is framed binary data (see RFC 4288, Section 4.8) and Gupta Expires November 3, 2013 [Page 13] Internet-Draft EVRC-NW2K RTP payload format May 2013 is defined for transfer of EVRC-NW2K encoded data via RTP using the interleaved/bundled packet format specified in RFC 3558 [7]. Security considerations: See Section 16. Interoperability considerations: None Published specification: The EVRC-NW2K vocoder is specified in 3GPP2 C.S0014-E. The transfer method with the interleaved/bundled packet format via RTP is specified in RFC 3558 [7]. See Section 6 of RFC XXXX for details for EVRC-NW2K. [Note to the RFC editor: please replace XXXX with the RFC number of this document.] Applications that use this media type: It is expected that many VoIP applications (as well as mobile applications) will use this type. Additional information: The following applies to stored-file transfer methods: Magic number: #!EVRCNW2K\n (see Section 8) File extensions: enw2k, ENW2K Macintosh file type code: None Object identifier or OID: None EVRC-NW2K speech frames may also be stored in the file format "3g2" defined in 3GPP2 C.S0050-B, which is identified using the media types "audio/3gpp2" or "video/3gpp2" registered by RFC 4393 [12]. Person & email address to contact for further information: Alok Gupta Intended usage: COMMON Restrictions on usage: This media type can be used with the file format defined in Section 8 of RFC XXXX in contexts other than RTP. In context of transfers over RTP, the RTP payload format specified in Section 4.1 of RFC 3558 [7] is used for this media type. [Note to the RFC editor: please replace Gupta Expires November 3, 2013 [Page 14] Internet-Draft EVRC-NW2K RTP payload format May 2013 XXXX with the RFC number of this document.] Author: Alok Gupta Change controller: IETF Payload working group delegated from the IESG. 9.1.2. Registration of Media Type audio/EVRCNW2K0 Type name: audio Subtype names: EVRCNW2K0 Required parameters: None Optional parameters: These parameters apply to RTP transfer only. mode-set-recv: A subset of EVRC-NW2K modes. Possible values are a comma separated list of modes from the set {0,1,2,3,4,5,6,7} (see Table 2.7.1.2-1 in 3GPP2 C.S0014-E). A decoder can use this attribute to inform an encoder of its preference to operate in a specified subset of modes. Absence of this parameter signals the mode set {1,2,3,4,5,6,7}. ptime: see RFC 4566. silencesupp: see Section 6.1 in RFC 4788. dtxmax: see Section 6.1 in RFC 4788. dtxmin: see Section 6.1 in RFC 4788. hangover: see Section 6.1 in RFC 4788. Encoding considerations: This media type is framed binary data (see RFC 4288, Section 4.8) and is defined for transfer of EVRC-NW2K encoded data via RTP using the header-free packet format specified in RFC 3558 [7]. Security considerations: See Section 16. Interoperability considerataions: None Gupta Expires November 3, 2013 [Page 15] Internet-Draft EVRC-NW2K RTP payload format May 2013 Published specification: The EVRC-NW2K vocoder is specified in 3GPP2 C.S0014-E. The transfer method with the header-free packet format via RTP is specified in RFC 3558 [7]. Applications that use this media type: It is expected that many VoIP applications (as well as mobile applications) will use this type. Additional information: None Person & email address to contact for further information: Alok Gupta Intended usage: COMMON Restrictions on usage: This media type depends on RTP framing, and hence is only defined for transfer via RTP [6], the RTP payload format specified in Section 4.2 of RFC 3558 [7] SHALL be used. This media type SHALL NOT be used for storage or file transfer, instead audio/EVRCNW SHALL be used. Author: Alok Gupta Change controller: IETF Payload working group delegated from the IESG. 9.1.3. Registration of Media Type audio/EVRCNW2K1 Type name: audio Subtype names: EVRCNW2K1 Required parameters: None Optional parameters: These parameters apply to RTP transfer only. mode-set-recv: A subset of EVRC-NW2K modes. Possible values are a comma separated list of modes from the set {0,1} (see Table 2.7.1.2-1 Gupta Expires November 3, 2013 [Page 16] Internet-Draft EVRC-NW2K RTP payload format May 2013 in 3GPP2 C.S0014-E). A decoder can use this attribute to inform an encoder of its preference to operate in a specified subset of modes. A value of 0 signals the support for wideband fixed rate (full or half rate, depending on the value of 'fixedrate' parameter). A value of 1 signals narrowband fixed rate (full or half rate, depending on the value of 'fixedrate' parameter). Absence of this parameter signals the mode 1. ptime: see RFC 4566. maxptime: see RFC 4566. fixedrate: Indicates the EVRC-NW2K rate of the session while in single rate operation. Valid values include: 0.25, 0.5, and 1, where a value of 0.25 indicates the 1/4 rate, 0.5 indicates the 1/2 rate while a value of 1 indicates the full rate. If this parameter is not present, 1/4 rate is assumed. silencesupp: see Section 6.1 in RFC 4788. dtxmax: see Section 6.1 in RFC 4788. dtxmin: see Section 6.1 in RFC 4788. hangover: see Section 6.1 in RFC 4788. Encoding considerations: This media type is framed binary data (see RFC 4288, Section 4.8) and is defined for transfer of EVRC-NW2K encoded data via RTP using the compact bundled packet format specified in RFC 4788. Security considerations: See Section 16 Interoperability considerataions: None Published specification: The EVRC-NW2K vocoder is specified in 3GPP2 C.S0014-E. The transfer method with the compact bundled packet format via RTP is specified in RFC 4788. Applications that use this media type: It is expected that many VoIP applications (as well as mobile applications) will use this type. Additional information: None Gupta Expires November 3, 2013 [Page 17] Internet-Draft EVRC-NW2K RTP payload format May 2013 Person & email address to contact for further information: Alok Gupta Intended usage: COMMON Restrictions on usage: This media type depends on RTP framing, and hence is only defined for transfer via RTP [6], the RTP payload format specified in Section 4 of RFC 4788 SHALL be used. This media type SHALL NOT be used for storage or file transfer, instead audio/EVRCNW2K SHALL be used. Author: Alok Gupta Change controller: IETF Payload working group delegated from the IESG. Gupta Expires November 3, 2013 [Page 18] Internet-Draft EVRC-NW2K RTP payload format May 2013 10. SDP mode attributes for EVRC-NW2K 'mode-set-recv' can be used by a decoder to inform an encoder of its preference to operate in a specified subset of modes. Note that indicating a preference implicitly indicates support for that capability. If mode 0 is not preferred for media type EVRCNW2K0 or EVRCNW2K1, then there is no indication that mode 0 is supported. However absence of this parameter or absence of mode 0 in this parameter for media type EVRCNW2K shall not preclude mode 0 support during a call where mode 0 may be requested via the MMM field. To inform the capability for wideband mode support, a decoder can always decode all the narrowband modes (modes 1 to 7). Unless the decoder indicates the support of mode 0 (i.e., preference) in this parameter or in the MMM mode request field in interleaved/bundled payload format, an encoder at the other side shall not operate in mode 0. To indicate a preference to operate in a subset of modes, a set has been defined so that several modes can be expressed as a preference in one attempt. For instance, the set {4,5,6,7} signals that the receiver prefers the sender to operate in bandwidth-efficient narrowband modes of EVRC-NW2K. Note, during an active call session using the interleaved/bundled packet format, the MMM mode request received from a communication partner can contain a mode request different than the values in the last mode-set-recv attribute. The partner's EVRC-NW2K wideband decoding capability is determined by the latest mode-set-recv attribute or MMM mode request field. For example, a mode request with MMM=0 from a communication partner is an implicit indication of the partner's EVRCNW2K wideband decoding capability and preference. An EVRCNW2K wideband capable node receiving the request can operate in wideband mode. A mode request with MMM=1, 2, ..., or 7 from a communication partner is an implicit indication of the partner's EVRCNW2K narrowband decoding preference. The encoder of an EVRCNW2K node receiving the request shall honor the request and operate in narrowband mode. 'sendmode' is used as a SDP mode attribute in EVRC [7], EVRC-B [2] and EVRC-WB [3]. However it is deprecated in EVRC-NW and EVRC-NW2k. Gupta Expires November 3, 2013 [Page 19] Internet-Draft EVRC-NW2K RTP payload format May 2013 11. Mode Change Request/Response Considerations The interleaved/bundled packet format for the EVRC family of vocoders supports a 3-bit field (MMM) that a communication node can use to indicate its preferred compression mode to an opposite node. The concept of the compression mode (also known as Capacity Operating Point) was introduced to allow a controlled trade-off between voice quality and channel capacity. The notion makes it possible to exercise vocoders at the highest possible (average) bit-rate (hence, highest voice quality) when the network is lightly loaded. Conversely, once the network load increases the vocoders can be requested to operate at lower average bit-rates so as to absorb the additional network load without causing an undue increase in the frame-erasure rates; the underlying premise is that while a higher bit-rate improves the vocoder performance, it also increases the network loading, risking a sharp decline in voice quality should the frame-erasure rate be too high. By contrast, a lower bit-rate mode of operation can result in accommodation of the additional network load without causing unduly high frame-erasure rates, resulting in better overall quality despite the inherently lower voice quality of the lower bit-rate mode of the vocoder. Accordingly, the MMM field should be used to request the far-end to transmit compressed-speech using a mode that provides the best balance between voice quality and capacity. However, in the case of mobile-mobile calls, for example, there are two wireless sides involved, each with a potentially different network load level and hence a different preferred mode. In such cases, achieving optimal end-to-end performance depends on coherent management of the operative mode by the two sides. This requires that even if the local node prefers a higher bit-rate vocoder mode, it should adjust to a lower bit-rate mode if requested by the far end, in order to avoid potentially high frame erasure rates due to heavy load at the far end network. For similar reasons, in cases where a mode requested by the far end should not be supported, it might still be beneficial to consider switching to a supported vocoder mode corresponding to a lower average bit-rate than requested. It is recommended that the next lower average bit-rate supported vocoder mode be used for encoding when a mode requested by the far end is not supported. A wideband-capable endpoint can use the information conveyed by the C-bit of the RTP payload header to determine the optimal mode to request of the far end. If the far end cannot provide Mode0 packets (C-bit=1), then the choice of MMM can be based strictly on the local network load. If the C-bit indicates remote end's Mode0 encoding capability (C-bit=0), then even if the local network load is not light, Mode0 can be requested knowing definitively that it will be Gupta Expires November 3, 2013 [Page 20] Internet-Draft EVRC-NW2K RTP payload format May 2013 supported. This will permit operators to treat wideband-capable mobiles preferentially, should they wish to adopt such policy. Gupta Expires November 3, 2013 [Page 21] Internet-Draft EVRC-NW2K RTP payload format May 2013 12. Mapping EVRC-NW2K media type parameters into SDP Information carried in the media type specification has a specific mapping to fields in the Session Description Protocol (SDP) [11], which is commonly used to describe RTP sessions. When SDP is used to specify sessions employing EVRC-NW2K encoded speech, the mapping is as follows. o The media type ("audio") goes in SDP "m=" as the media name. o The media subtype ("EVRCNW2K", "EVRCNW2K0" or "EVRCNW2K1") goes in SDP "a=rtpmap" as the encoding name. o The optional parameters 'ptime and 'maxptime' (for subtypes EVRCNW2K, EVRCNW2K1) go in the SDP "a=ptime" and "a=maxptime" attributes, respectively. o Any remaining parameters (for subtypes EVRCNW2K, EVRCNW2K0 and EVRCNW2K1) go in the SDP "a=fmtp" attribute by copying them from the media type string as a semicolon separated list of parameter=value pairs. Gupta Expires November 3, 2013 [Page 22] Internet-Draft EVRC-NW2K RTP payload format May 2013 13. Offer-Answer Model Considerations for EVRC-NW2K The following considerations apply when using the SDP offer-answer procedures of RFC 3264 [13] to negotiate the use of EVRC-NW2K payload in RTP: o Since EVRC-NW2K is an extension of EVRC-B, EVRC-WB and EVRC-NW, the offerer SHOULD also announce EVRC-B, EVRC-WB and EVRC-NW support in its "m=audio" lines, with EVRC-NW2K as the preferred codec. This will allow interoperability with an answerer which supports only EVRC-B, EVRC-WB and/or EVRC-NW. Below is an example of such an offer: m=audio 55954 RTP/AVP 97 98 99 100 a=rtpmap:97 EVRCNW2K0/16000 a=rtpmap:98 EVRCNW0/16000 a=rtpmap:99 EVRCWB0/16000 a=rtpmap:100 EVRCB0/8000 a=fmtp:97 mode-set-recv=0,1,2,3,4,5,6 a=fmtp:98 mode-set-recv=0,1,2,3,4,5,6 a=fmtp:99 mode-set-recv=0,4 a=fmtp:100 recvmode=0 If the answerer supports EVRC-NW2K then the answerer can keep the payload type 97 in its answer and the conversation can be done using EVRC-NW2K. Else, if the answerer supports only EVRC-NW, EVRC-WB and/or EVRC-B then the answerer will leave only the payload type 98, 99 and/or 100 respectively in its answer and the conversation will be done using EVRC-NW, EVRC-WB and/or EVRC-B respectively. An example answer for the above offer: m=audio 55954 RTP/AVP 97 a=rtpmap:97 EVRCNW2K0/16000 a=fmtp:97 mode-set-recv=4 o 'mode-set-recv' is a uni-directional receive only parameter. o An offerer can use 'mode-set-recv' to request that the remote sender's encoder be limited to the list of modes signaled in 'mode-set-recv'. A remote sender MAY ignore 'mode-set-recv' requests. However, a remote sender shall not assume the other side can support mode 0, unless the offer includes mode 0 explicitly in 'mode-set-recv' or the remote sender receives mode requests with MMM = 0 from the communication partner during an Gupta Expires November 3, 2013 [Page 23] Internet-Draft EVRC-NW2K RTP payload format May 2013 active call using EVRC-NW2K interleaved/bundled format. o The parameters 'maxptime' and 'ptime' will in most cases not affect interoperability, however the setting of the parameters can affect the performance of the application. The SDP offer-answer handling of the 'ptime' parameter is described in RFC 3264 [13]. The 'maxptime' parameter MUST be handled in the same way. o For a sendonly stream, the 'mode-set-recv' parameter is not useful and SHOULD NOT be used. o When using EVRCNW2K1, the entire session MUST use the same fixed rate and mode (0-Wideband or 1-Narrowband). o For additional rules which MUST be followed while negotiating DTX parameters, see Section 6.8 in RFC 4788 [2]. o Any unknown parameter in an SDP offer MUST be ignored by the receiver and MUST NOT be included in the SDP answer. Gupta Expires November 3, 2013 [Page 24] Internet-Draft EVRC-NW2K RTP payload format May 2013 14. Declarative SDP Considerations For declarative use of SDP in SAP [16] and RTSP [17], the following considerations apply: o Any 'maxptime' and 'ptime' values should be selected with care to ensure that the session's participants can achieve reasonable performance. o The payload format configuration parameters are all declarative and a participant MUST use the configuration(s) that is provided for the session. More than one configuration MAY be provided if necessary by declaring multiple RTP payload types, however the number of types SHOULD be kept small. For declarative examples, see Section 15. o The usage of unidirectional receive-only parameters, such as 'mode-set-recv', should be excluded in any declarations, since these parameters are meaningless in one-way streaming applications. Gupta Expires November 3, 2013 [Page 25] Internet-Draft EVRC-NW2K RTP payload format May 2013 15. Examples Some example SDP session descriptions utilizing EVRC-NW2K encodings follow. In these examples, long a=fmtp lines are folded to meet the column width constraints of this document. The backslash ("\") at the end of a line and the carriage return that follows it should be ignored. Note that media subtype names are case-insensitive. Parameter names are case-insensitive both in media types and in the mapping to the SDP a=fmtp attribute. Example usage of EVRCNW2K if wideband mode is supported: m=audio 49120 RTP/AVP 96 97 98 99 a=rtpmap:96 EVRCNW2K/16000 a=rtpmap:97 EVRCNW/16000 a=rtpmap:98 EVRCWB/16000 a=rtpmap:99 EVRCB/8000 a=fmtp:96 mode-set-recv=0,1,2,3,4,5,6 a=fmtp:97 mode-set-recv=0,1,2,3,4,5,6 a=fmtp:98 mode-set-recv=0,4 a=fmtp:99 recvmode=0 a=maxptime:120 Example usage of EVRCNW2K if wideband mode is not supported: m=audio 49120 RTP/AVP 96 97 98 99 a=rtpmap:96 EVRCNW2K/16000 a=rtpmap:97 EVRCNW/16000 a=rtpmap:98 EVRCWB/16000 a=rtpmap:99 EVRCB/8000 a=fmtp:96 mode-set-recv=1,2,3,4,5,6 a=fmtp:97 mode-set-recv=1,2,3,4,5,6 a=fmtp:98 mode-set-recv=4 a=fmtp:99 recvmode=0 a=maxptime:120 Example usage of EVRCNW2K0: m=audio 49120 RTP/AVP 96 97 98 99 a=rtpmap:96 EVRCNW2K0/16000 a=rtpmap:97 EVRCNW0/16000 a=rtpmap:98 EVRCWB0/16000 a=rtpmap:99 EVRCB0/8000 a=fmtp:96 mode-set-recv=0,1,2,3,4,5,6 a=fmtp:97 mode-set-recv=0,1,2,3,4,5,6 a=fmtp:98 mode-set-recv=0,4 a=fmtp:99 recvmode=0 Gupta Expires November 3, 2013 [Page 26] Internet-Draft EVRC-NW2K RTP payload format May 2013 Example SDP answer from a media gateway requesting a terminal to limit its encoder operation to EVRC-NW2K mode 3. m=audio 49120 RTP/AVP 96 a=rtpmap:96 EVRCNW2K0/16000 a=fmtp:96 mode-set-recv=3 Example usage of EVRCNW2K1: m=audio 49120 RTP/AVP 96 97 98 99 a=rtpmap:96 EVRCNW2K1/16000 a=rtpmap:97 EVRCNW1/16000 a=rtpmap:98 EVRCWB1/16000 a=rtpmap:99 EVRCB1/8000 a=fmtp:96 fixedrate=0.5 a=fmtp:97 fixedrate=0.5 a=fmtp:98 fixedrate=0.5 a=fmtp:99 fixedrate=0.5 a=maxptime:100 Example usage of EVRCNW2K with DTX with silencesupp=1: m=audio 49120 RTP/AVP 96 97 98 99 a=rtpmap:96 EVRCNW2K/16000 a=rtpmap:97 EVRCNW/16000 a=rtpmap:98 EVRCWB/16000 a=rtpmap:99 EVRCB/8000 a=fmtp:96 silencesupp=1;dtxmax=32;dtxmin=12;hangover=1; \ mode-set-recv=0,1,2,3,4,5,6 a=fmtp:97 silencesupp=1;dtxmax=32;dtxmin=12;hangover=1; \ mode-set-recv=0,1,2,3,4,5,6 a=fmtp:98 silencesupp=1;dtxmax=32;dtxmin=12;hangover=1; \ mode-set-recv=0,4 a=fmtp:99 recvmode=0 a=maxptime:120 Examples usage of EVRCNW2K with DTX with silencesupp=0: Gupta Expires November 3, 2013 [Page 27] Internet-Draft EVRC-NW2K RTP payload format May 2013 m=audio 49120 RTP/AVP 96 97 98 99 a=rtpmap:96 EVRCNW2K/16000 a=rtpmap:97 EVRCNW/16000 a=rtpmap:98 EVRCWB/16000 a=rtpmap:99 EVRCB/8000 a=fmtp:96 silencesupp=0;dtxmax=32;dtxmin=12;hangover=1; \ mode-set-recv=0,1,2,3,4,5,6 a=fmtp:97 silencesupp=0;dtxmax=32;dtxmin=12;hangover=1; \ mode-set-recv=0,1,2,3,4,5,6 a=fmtp:98 silencesupp=0;dtxmax=32;dtxmin=12;hangover=1; \ mode-set-recv=0,4 a=fmtp:99 recvmode=0 a=maxptime:120 Example offer answer exchange between EVRC-NW2K and legacy EVRC-B (RFC 4788): Offer: m=audio 55954 RTP/AVP 96 97 98 99 a=rtpmap:96 EVRCNW2K0/16000 a=rtpmap:97 EVRCNW0/16000 a=rtpmap:98 EVRCWB0/16000 a=rtpmap:99 EVRCB0/8000 a=rtpmap:96 mode-set-recv=0,1,2,3,4,5,6 a=rtpmap:97 mode-set-recv=0,1,2,3,4,5,6 a=fmtp:98 mode-set-recv=0,4 a=fmtp:99 recvmode=0 Answer: m=audio 55954 RTP/AVP 99 a=rtpmap:99 EVRCB0/8000 Example offer answer exchange between EVRC-NW2K and legacy EVRC-WB (RFC 5188): Gupta Expires November 3, 2013 [Page 28] Internet-Draft EVRC-NW2K RTP payload format May 2013 Offer: m=audio 55954 RTP/AVP 96 97 98 99 a=rtpmap:96 EVRCNW2K0/16000 a=rtpmap:97 EVRCNW0/16000 a=rtpmap:98 EVRCWB0/16000 a=rtpmap:99 EVRCB0/8000 a=rtpmap:96 mode-set-recv=0,1,2,3,4,5,6 a=rtpmap:97 mode-set-recv=0,1,2,3,4,5,6 a=fmtp:98 mode-set-recv=0,4 a=fmtp:99 recvmode=0 Answer: m=audio 55954 RTP/AVP 98 99 a=rtpmap:98 EVRCWB0/16000 Example offer answer exchange between EVRC-NW2K and legacy EVRC-NW (RFC ????):