Benny Prijono | ef010c5 | 2007-03-30 10:49:46 +0000 | [diff] [blame^] | 1 | <?xml version='1.0'?> |
| 2 | <!DOCTYPE rfc SYSTEM 'rfc2629.dtd'> |
| 3 | <?rfc toc="yes" ?> |
| 4 | |
| 5 | <rfc ipr="full3667" docName="RTP Payload Format for the Speex Codec"> |
| 6 | |
| 7 | <front> |
| 8 | <title>draft-herlein-speex-rtp-profile-03</title> |
| 9 | |
| 10 | <author initials="G" surname="Herlein" fullname="Greg Herlein"> |
| 11 | <organization></organization> |
| 12 | <address> |
| 13 | <email>gherlein@herlein.com</email> |
| 14 | <postal> |
| 15 | <street>2034 Filbert Street</street> |
| 16 | <city>San Francisco</city> |
| 17 | <region>California</region> |
| 18 | <code>94123</code> |
| 19 | <country>United States</country> |
| 20 | </postal> |
| 21 | </address> |
| 22 | </author> |
| 23 | |
| 24 | <author initials="S" surname="Morlat" fullname="Simon Morlat"> |
| 25 | <address> |
| 26 | <email>simon.morlat@linphone.org</email> |
| 27 | <postal> |
| 28 | <street>35, av de Vizille App 42</street> |
| 29 | <city>Grenoble</city> |
| 30 | <code>38000</code> |
| 31 | <country>France</country> |
| 32 | </postal> |
| 33 | </address> |
| 34 | </author> |
| 35 | |
| 36 | <author initials="J" surname="Jean-Marc" fullname="Jean-Marc Valin"> |
| 37 | <address> |
| 38 | <email>jean-marc.valin@hermes.usherb.ca</email> |
| 39 | <postal> |
| 40 | <street>Department of Electrical and Computer Engineering</street> |
| 41 | <street>University of Sherbrooke</street> |
| 42 | <street>2500 blvd Universite</street> |
| 43 | <city>Sherbrooke</city> |
| 44 | <region>Quebec</region> |
| 45 | <code>J1K 2R1</code> |
| 46 | <country>Canada</country> |
| 47 | </postal> |
| 48 | </address> |
| 49 | </author> |
| 50 | |
| 51 | <author initials="R" surname="Hardiman" fullname="Roger Hardiman"> |
| 52 | <address> |
| 53 | <email>roger@freebsd.org</email> |
| 54 | <postal> |
| 55 | <street>49 Nettleton Road</street> |
| 56 | <city>Cheltenham</city> |
| 57 | <region>Gloucestershire</region> |
| 58 | <code>GL51 6NR</code> |
| 59 | <country>England</country> |
| 60 | </postal> |
| 61 | </address> |
| 62 | </author> |
| 63 | |
| 64 | |
| 65 | <author initials="P" surname="Kerr" fullname="Phil Kerr"> |
| 66 | <address> |
| 67 | <email>phil@plus24.com</email> |
| 68 | <postal> |
| 69 | <country>England</country> |
| 70 | </postal> |
| 71 | </address> |
| 72 | </author> |
| 73 | |
| 74 | <date day="01" month="January" year="2005" /> |
| 75 | |
| 76 | <area>General</area> |
| 77 | <workgroup>AVT Working Group</workgroup> |
| 78 | <keyword>I-D</keyword> |
| 79 | |
| 80 | <keyword>Internet-Draft</keyword> |
| 81 | <keyword>Speex</keyword> |
| 82 | <keyword>RTP</keyword> |
| 83 | <abstract> |
| 84 | <t> |
| 85 | Speex is an open-source voice codec suitable for use in Voice over |
| 86 | IP (VoIP) type applications. This document describes the payload |
| 87 | format for Speex generated bit streams within an RTP packet. Also |
| 88 | included here are the necessary details for the use of Speex with |
| 89 | the Session Description Protocol (SDP) and a preliminary method of |
| 90 | using Speex within H.323 applications. |
| 91 | </t> |
| 92 | </abstract> |
| 93 | </front> |
| 94 | |
| 95 | <middle> |
| 96 | |
| 97 | <section anchor="Conventions used in this document" title="Conventions used in this document"> |
| 98 | <t> |
| 99 | The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", |
| 100 | "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this |
| 101 | document are to be interpreted as described in RFC 2119 <xref target="rfc2119"></xref>. |
| 102 | </t> |
| 103 | </section> |
| 104 | |
| 105 | <section anchor="Overview of the Speex Codec" title="Overview of the Speex Codec"> |
| 106 | |
| 107 | <t> |
| 108 | Speex is based on the CELP <xref target="CELP"></xref> encoding technique with support for |
| 109 | either narrowband (nominal 8kHz), wideband (nominal 16kHz) or |
| 110 | ultra-wideband (nominal 32kHz), and (non-optimal) rates up to 48 kHz |
| 111 | sampling also available. The main characteristics can be summarized |
| 112 | as follows: |
| 113 | </t> |
| 114 | |
| 115 | <t> |
| 116 | <list style="symbols"> |
| 117 | <t>Free software/open-source</t> |
| 118 | <t>Integration of wideband and narrowband in the same bit-stream</t> |
| 119 | <t>Wide range of bit-rates available</t> |
| 120 | <t>Dynamic bit-rate switching and variable bit-rate (VBR)</t> |
| 121 | <t>Voice Activity Detection (VAD, integrated with VBR)</t> |
| 122 | <t>Variable complexity</t> |
| 123 | </list> |
| 124 | </t> |
| 125 | |
| 126 | </section> |
| 127 | |
| 128 | <section anchor="RTP payload format for Speex" title="RTP payload format for Speex"> |
| 129 | |
| 130 | <t> |
| 131 | For RTP based transportation of Speex encoded audio the standard |
| 132 | RTP header [2] is followed by one or more payload data blocks. |
| 133 | An optional padding terminator may also be used. |
| 134 | </t> |
| 135 | <artwork><![CDATA[ |
| 136 | 0 1 2 3 |
| 137 | 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 |
| 138 | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |
| 139 | | RTP Header | |
| 140 | +=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+ |
| 141 | | one or more frames of Speex .... | |
| 142 | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |
| 143 | | one or more frames of Speex .... | padding | |
| 144 | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |
| 145 | ]]></artwork> |
| 146 | |
| 147 | </section> |
| 148 | |
| 149 | <section anchor="RTP Header" title="RTP Header"> |
| 150 | |
| 151 | <artwork><![CDATA[ |
| 152 | 0 1 2 3 |
| 153 | 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 |
| 154 | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |
| 155 | |V=2|P|X| CC |M| PT | sequence number | |
| 156 | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |
| 157 | | timestamp | |
| 158 | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |
| 159 | | synchronization source (SSRC) identifier | |
| 160 | +=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+ |
| 161 | | contributing source (CSRC) identifiers | |
| 162 | | ... | |
| 163 | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |
| 164 | ]]></artwork> |
| 165 | |
| 166 | <t> |
| 167 | The RTP header begins with an octet of fields (V, P, X, and CC) to |
| 168 | support specialized RTP uses (see <xref target="rfc3550"></xref> and <xref target="rfc3551"></xref> for details). For |
| 169 | Speex the following values are used. |
| 170 | </t> |
| 171 | |
| 172 | <t>Version (V): 2 bits</t><t> |
| 173 | This field identifies the version of RTP. The version |
| 174 | used by this specification is two <xref target="rfc3550"></xref>. |
| 175 | </t> |
| 176 | |
| 177 | <t>Padding (P): 1 bit</t><t> |
| 178 | If the padding bit is set, the packet contains one or more |
| 179 | additional padding octets at the end which are not part of |
| 180 | the payload. P is set if the total packet size is less than |
| 181 | the MTU. |
| 182 | </t> |
| 183 | |
| 184 | <t>Extension (X): 1 bit</t><t> |
| 185 | If the extension, X, bit is set, the fixed header MUST be |
| 186 | followed by exactly one header extension, with a format defined |
| 187 | in Section 5.3.1. of <xref target="rfc3550"></xref>. |
| 188 | </t> |
| 189 | |
| 190 | <t>CSRC count (CC): 4 bits</t><t> |
| 191 | The CSRC count contains the number of CSRC identifiers. |
| 192 | </t> |
| 193 | |
| 194 | <t>Marker (M): 1 bit</t><t> |
| 195 | The M bit indicates if the packet contains comfort noise. This |
| 196 | field is used in conjunction with the cng SDP attribute and is |
| 197 | detailed further in section 5 below. In normal usage this bit |
| 198 | is set if the packet contains comfort noise. |
| 199 | </t> |
| 200 | |
| 201 | <t>Payload Type (PT): 7 bits</t><t> |
| 202 | An RTP profile for a class of applications is expected to assign |
| 203 | a payload type for this format, or a dynamically allocated |
| 204 | payload type SHOULD be chosen which designates the payload as |
| 205 | Speex. |
| 206 | </t> |
| 207 | |
| 208 | <t>Sequence number: 16 bits</t><t> |
| 209 | The sequence number increments by one for each RTP data packet |
| 210 | sent, and may be used by the receiver to detect packet loss and |
| 211 | to restore packet sequence. This field is detailed further in |
| 212 | <xref target="rfc3550"></xref>. |
| 213 | </t> |
| 214 | |
| 215 | <t>Timestamp: 32 bits</t><t> |
| 216 | A timestamp representing the sampling time of the first sample of |
| 217 | the first Speex packet in the RTP packet. The clock frequency |
| 218 | MUST be set to the sample rate of the encoded audio data. |
| 219 | |
| 220 | Speex uses 20 msec frames and a variable sampling rate clock. |
| 221 | The RTP timestamp MUST be in units of 1/X of a second where X |
| 222 | is the sample rate used. Speex uses a nominal 8kHz sampling rate |
| 223 | for narrowband use, a nominal 16kHz sampling rate for wideband use, |
| 224 | and a nominal 32kHz sampling rate for ultra-wideband use. |
| 225 | </t> |
| 226 | |
| 227 | <t>SSRC/CSRC identifiers:</t><t> |
| 228 | These two fields, 32 bits each with one SSRC field and a maximum |
| 229 | of 16 CSRC fields, are as defined in <xref target="rfc3550"></xref>. |
| 230 | </t> |
| 231 | |
| 232 | </section> |
| 233 | |
| 234 | <section anchor="Speex payload" title="Speex payload"> |
| 235 | |
| 236 | <t> |
| 237 | For the purposes of packetizing the bit stream in RTP, it is only |
| 238 | necessary to consider the sequence of bits as output by the Speex |
| 239 | encoder <xref target="speexenc"></xref>, and present the same sequence to the decoder. The |
| 240 | payload format described here maintains this sequence. |
| 241 | </t> |
| 242 | |
| 243 | <t> |
| 244 | A typical Speex frame, encoded at the maximum bitrate, is approx. |
| 245 | 110 octets and the total number of Speex frames SHOULD be kept |
| 246 | less than the path MTU to prevent fragmentation. Speex frames MUST |
| 247 | NOT be fragmented across multiple RTP packets, |
| 248 | </t> |
| 249 | |
| 250 | <t> |
| 251 | An RTP packet MAY contain Speex frames of the same bit rate or of |
| 252 | varying bit rates, since the bit-rate for a frame is conveyed in |
| 253 | band with the signal. |
| 254 | </t> |
| 255 | |
| 256 | <t> |
| 257 | The encoding and decoding algorithm can change the bit rate at any |
| 258 | 20 msec frame boundary, with the bit rate change notification provided |
| 259 | in-band with the bit stream. Each frame contains both "mode" |
| 260 | (narrowband, wideband or ultra-wideband) and "sub-mode" (bit-rate) |
| 261 | information in the bit stream. No out-of-band notification is |
| 262 | required for the decoder to process changes in the bit rate sent |
| 263 | by the encoder. |
| 264 | </t> |
| 265 | |
| 266 | <t> |
| 267 | It is RECOMMENDED that values of 8000, 16000 and 32000 be used |
| 268 | for normal internet telephony applications, though the sample |
| 269 | rate is supported at rates as low as 6000 Hz and as high as |
| 270 | 48 kHz. |
| 271 | </t> |
| 272 | |
| 273 | <t> |
| 274 | The RTP payload MUST be padded to provide an integer number of |
| 275 | octets as the payload length. These padding bits are LSB aligned |
| 276 | in network octet order and consist of a 0 followed by all ones |
| 277 | (until the end of the octet). This padding is only required for |
| 278 | the last frame in the packet, and only to ensure the packet |
| 279 | contents ends on an octet boundary. |
| 280 | </t> |
| 281 | |
| 282 | </section> |
| 283 | |
| 284 | <section anchor="Example Speex packet" title="Example Speex packet"> |
| 285 | |
| 286 | <t> |
| 287 | In the example below we have a single Speex frame with 5 bits |
| 288 | of padding to ensure the packet size falls on an octet boundary. |
| 289 | </t> |
| 290 | |
| 291 | <artwork><![CDATA[ |
| 292 | 0 1 2 3 |
| 293 | 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 |
| 294 | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |
| 295 | |V=2|P|X| CC |M| PT | sequence number | |
| 296 | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |
| 297 | | timestamp | |
| 298 | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |
| 299 | | synchronization source (SSRC) identifier | |
| 300 | +=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+ |
| 301 | |
| 302 | 0 1 2 3 |
| 303 | 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 |
| 304 | +=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+ |
| 305 | | contributing source (CSRC) identifiers | |
| 306 | | ... | |
| 307 | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |
| 308 | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |
| 309 | | ..speex data.. | |
| 310 | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |
| 311 | | ..speex data.. |0 1 1 1 1| |
| 312 | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |
| 313 | ]]></artwork> |
| 314 | |
| 315 | </section> |
| 316 | |
| 317 | <section anchor="Multiple Speex frames in a RTP packet" title="Multiple Speex frames in a RTP packet"> |
| 318 | |
| 319 | <t> |
| 320 | Below is an example of two Speex frames contained within one RTP |
| 321 | packet. The Speex frame length in this example fall on an octet |
| 322 | boundary so there is no padding. |
| 323 | </t> |
| 324 | |
| 325 | <t> |
| 326 | Speex codecs <xref target="speexenc"></xref> are able to detect the the bitrate from the |
| 327 | payload and are responsible for detecting the 20 msec boundaries |
| 328 | between each frame. |
| 329 | </t> |
| 330 | |
| 331 | <artwork><![CDATA[ |
| 332 | 0 1 2 3 |
| 333 | 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 |
| 334 | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |
| 335 | |V=2|P|X| CC |M| PT | sequence number | |
| 336 | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |
| 337 | | timestamp | |
| 338 | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |
| 339 | | synchronization source (SSRC) identifier | |
| 340 | +=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+ |
| 341 | | contributing source (CSRC) identifiers | |
| 342 | | ... | |
| 343 | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |
| 344 | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |
| 345 | | ..speex data.. | |
| 346 | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |
| 347 | | ..speex data.. | ..speex data.. | |
| 348 | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |
| 349 | | ..speex data.. | |
| 350 | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |
| 351 | ]]></artwork> |
| 352 | |
| 353 | </section> |
| 354 | |
| 355 | <section anchor="MIME registration of Speex" title="MIME registration of Speex"> |
| 356 | |
| 357 | <t> |
| 358 | Full definition of the MIME <xref target="rfc2045"></xref> type for Speex will be part of the Ogg |
| 359 | Vorbis MIME type definition application <xref target="rfc3534"></xref>. |
| 360 | </t> |
| 361 | |
| 362 | <t>MIME media type name: audio</t> |
| 363 | |
| 364 | <t>MIME subtype: speex</t> |
| 365 | |
| 366 | <t>Optional parameters:</t> |
| 367 | |
| 368 | <t>Required parameters: to be included in the Ogg MIME specification.</t> |
| 369 | |
| 370 | <t>Encoding considerations:</t> |
| 371 | |
| 372 | <t>Security Considerations:</t> |
| 373 | <t>See Section 6 of RFC 3047.</t> |
| 374 | |
| 375 | <t>Interoperability considerations: none</t> |
| 376 | |
| 377 | <t>Published specification: </t> |
| 378 | |
| 379 | <t>Applications which use this media type:</t> |
| 380 | |
| 381 | <t>Additional information: none</t> |
| 382 | |
| 383 | <t>Person & email address to contact for further information:<vspace blankLines="1" /> |
| 384 | <list style="empty"> |
| 385 | <t>Greg Herlein <gherlein@herlein.com></t> |
| 386 | <t>Jean-Marc Valin <jean-marc.valin@hermes.usherb.ca></t> |
| 387 | </list> |
| 388 | </t> |
| 389 | |
| 390 | <t>Intended usage: COMMON</t> |
| 391 | |
| 392 | <t>Author/Change controller:</t> |
| 393 | |
| 394 | <t> |
| 395 | <list style="empty"> |
| 396 | <t>Author: Greg Herlein <gherlein@herlein.com></t> |
| 397 | <t>Change controller: Greg Herlein <gherlein@herlein.com></t> |
| 398 | <t>Change controller: IETF AVT Working Group</t> |
| 399 | </list> |
| 400 | </t> |
| 401 | |
| 402 | <t> |
| 403 | This transport type signifies that the content is to be interpreted |
| 404 | according to this document if the contents are transmitted over RTP. |
| 405 | Should this transport type appear over a lossless streaming protocol |
| 406 | such as TCP, the content encapsulation should be interpreted as an |
| 407 | Ogg Stream in accordance with <xref target="rfc3534"></xref>, with the exception that the |
| 408 | content of the Ogg Stream may be assumed to be Speex audio and |
| 409 | Speex audio only. |
| 410 | </t> |
| 411 | |
| 412 | </section> |
| 413 | |
| 414 | <section anchor="SDP usage of Speex" title="SDP usage of Speex"> |
| 415 | |
| 416 | <t> |
| 417 | When conveying information by SDP <xref target="rfc2327"></xref>, the encoding name MUST be |
| 418 | set to "speex". An example of the media representation in SDP for |
| 419 | offering a single channel of Speex at 8000 samples per second might |
| 420 | be: |
| 421 | </t> |
| 422 | |
| 423 | <vspace blankLines="1" /> |
| 424 | <list style="empty"> |
| 425 | <t>m=audio 8088 RTP/AVP 97</t> |
| 426 | <t>a=rtpmap:97 speex/8000</t> |
| 427 | </list> |
| 428 | |
| 429 | <t> |
| 430 | Note that the RTP payload type code of 97 is defined in this media |
| 431 | definition to be 'mapped' to the speex codec at an 8kHz sampling |
| 432 | frequency using the 'a=rtpmap' line. Any number from 96 to 127 |
| 433 | could have been chosen (the allowed range for dynamic types). |
| 434 | </t> |
| 435 | |
| 436 | <t> |
| 437 | The value of the sampling frequency is typically 8000 for narrow band |
| 438 | operation, 16000 for wide band operation, and 32000 for ultra-wide |
| 439 | band operation. |
| 440 | </t> |
| 441 | |
| 442 | <t> |
| 443 | If for some reason the offerer has bandwidth limitations, the client |
| 444 | may use the "b=" header, as explained in SDP <xref target="rfc2327"></xref>. The following example |
| 445 | illustrates the case where the offerer cannot receive more than |
| 446 | 10 kbit/s. |
| 447 | </t> |
| 448 | |
| 449 | <vspace blankLines="1" /> |
| 450 | <list style="empty"> |
| 451 | <t>m=audio 8088 RTP/AVP 97</t> |
| 452 | <t>b=AS:10</t> |
| 453 | <t>a=rtmap:97 speex/8000</t> |
| 454 | </list> |
| 455 | |
| 456 | <t> |
| 457 | In this case, if the remote part agrees, it should configure its |
| 458 | Speex encoder so that it does not use modes that produce more than |
| 459 | 10 kbit/s. Note that the "b=" constraint also applies on all |
| 460 | payload types that may be proposed in the media line ("m="). |
| 461 | </t> |
| 462 | |
| 463 | <t> |
| 464 | An other way to make recommendations to the remote Speex encoder |
| 465 | is to use its specific parameters via the a=fmtp: directive. The |
| 466 | following parameters are defined for use in this way: |
| 467 | </t> |
| 468 | |
| 469 | <vspace blankLines="1" /> |
| 470 | <list style="empty"> |
| 471 | <t>ptime: duration of each packet in milliseconds.<vspace blankLines="1" /></t> |
| 472 | |
| 473 | <t>sr: actual sample rate in Hz.<vspace blankLines="1" /></t> |
| 474 | |
| 475 | <t>ebw: encoding bandwidth - either 'narrow' or 'wide' or |
| 476 | 'ultra' (corresponds to nominal 8000, 16000, and |
| 477 | 32000 Hz sampling rates).<vspace blankLines="1" /></t> |
| 478 | |
| 479 | <t>vbr: variable bit rate - either 'on' 'off' or 'vad' |
| 480 | (defaults to off). If on, variable bit rate is |
| 481 | enabled. If off, disabled. If set to 'vad' then |
| 482 | constant bit rate is used but silence will be encoded |
| 483 | with special short frames to indicate a lack of voice |
| 484 | for that period.<vspace blankLines="1" /></t> |
| 485 | |
| 486 | <t>cng: comfort noise generation - either 'on' or 'off'. If |
| 487 | off then silence frames will be silent; if 'on' then |
| 488 | those frames will be filled with comfort noise.<vspace blankLines="1" /></t> |
| 489 | |
| 490 | <t>mode: Speex encoding mode. Can be {1,2,3,4,5,6,any} |
| 491 | defaults to 3 in narrowband, 6 in wide and ultra-wide.<vspace blankLines="1" /></t> |
| 492 | |
| 493 | <t>penh: use of perceptual enhancement. 1 indicates |
| 494 | to the decoder that perceptual enhancement is recommended, |
| 495 | 0 indicates that it is not. Defaults to on (1).<vspace blankLines="1" /></t> |
| 496 | </list> |
| 497 | |
| 498 | <t>Examples:</t> |
| 499 | |
| 500 | <vspace blankLines="1" /> |
| 501 | <list style="empty"> |
| 502 | <t>m=audio 8008 RTP/AVP 97</t> |
| 503 | <t>a=rtpmap:97 speex/8000</t> |
| 504 | <t>a=fmtp:97 mode=4</t> |
| 505 | </list> |
| 506 | |
| 507 | <t> |
| 508 | This examples illustrate an offerer that wishes to receive |
| 509 | a Speex stream at 8000Hz, but only using speex mode 3. |
| 510 | </t> |
| 511 | |
| 512 | <t> |
| 513 | The offerer may suggest to the remote decoder to activate |
| 514 | its perceptual enhancement filter like this: |
| 515 | </t> |
| 516 | |
| 517 | <vspace blankLines="1" /> |
| 518 | <list style="empty"> |
| 519 | <t>m=audio 8088 RTP/AVP 97</t> |
| 520 | <t>a=rtmap:97 speex/8000</t> |
| 521 | <t>a=fmtp:97 penh=1 </t> |
| 522 | </list> |
| 523 | |
| 524 | <t> |
| 525 | Several Speex specific parameters can be given in a single |
| 526 | a=fmtp line provided that they are separated by a semi-colon: |
| 527 | </t> |
| 528 | |
| 529 | <vspace blankLines="1" /> |
| 530 | <list style="empty"> |
| 531 | <t>a=fmtp:97 mode=any;penh=1</t> |
| 532 | </list> |
| 533 | |
| 534 | <t> |
| 535 | The offerer may indicate that it wishes to send variable bit rate |
| 536 | frames with comfort noise: |
| 537 | </t> |
| 538 | |
| 539 | <vspace blankLines="1" /> |
| 540 | <list style="empty"> |
| 541 | <t>m=audio 8088 RTP/AVP 97</t> |
| 542 | <t>a=rtmap:97 speex/8000</t> |
| 543 | <t>a=fmtp:97 vbr=on;cng=on</t> |
| 544 | </list> |
| 545 | |
| 546 | <t> |
| 547 | The "ptime" attribute is used to denote the packetization |
| 548 | interval (ie, how many milliseconds of audio is encoded in a |
| 549 | single RTP packet). Since Speex uses 20 msec frames, ptime values |
| 550 | of multiples of 20 denote multiple Speex frames per packet. |
| 551 | Values of ptime which are not multiples of 20 MUST be ignored |
| 552 | and clients MUST use the default value of 20 instead. |
| 553 | </t> |
| 554 | |
| 555 | <t> |
| 556 | In the example below the ptime value is set to 40, indicating that |
| 557 | there are 2 frames in each packet. |
| 558 | </t> |
| 559 | |
| 560 | <vspace blankLines="1" /> |
| 561 | <list style="empty"> |
| 562 | <t>m=audio 8008 RTP/AVP 97</t> |
| 563 | <t>a=rtpmap:97 speex/8000</t> |
| 564 | <t>a=ptime:40</t> |
| 565 | </list> |
| 566 | |
| 567 | <t> |
| 568 | Note that the ptime parameter applies to all payloads listed |
| 569 | in the media line and is not used as part of an a=fmtp directive. |
| 570 | </t> |
| 571 | |
| 572 | <t> |
| 573 | Values of ptime not multiple of 20 msec are meaningless, so the |
| 574 | receiver of such ptime values MUST ignore them. If during the |
| 575 | life of an RTP session the ptime value changes, when there are |
| 576 | multiple Speex frames for example, the SDP value must also reflect |
| 577 | the new value. |
| 578 | </t> |
| 579 | |
| 580 | <t> |
| 581 | Care must be taken when setting the value of ptime so that the |
| 582 | RTP packet size does not exceed the path MTU. |
| 583 | </t> |
| 584 | |
| 585 | </section> |
| 586 | <section anchor="ITU H.323/H.245 Use of Speex" title="ITU H.323/H.245 Use of Speex"> |
| 587 | |
| 588 | <t> |
| 589 | Application is underway to make Speex a standard ITU codec. |
| 590 | However, until that is finalized, Speex MAY be used in H.323 <xref target="H323"></xref> by |
| 591 | using a non-standard codec block definition in the H.245 <xref target="H245"></xref> codec |
| 592 | capability negotiations. |
| 593 | </t> |
| 594 | |
| 595 | </section> |
| 596 | |
| 597 | <section anchor="NonStandardMessage format" title="NonStandardMessage format"> |
| 598 | |
| 599 | <t> |
| 600 | For Speex use in H.245 <xref target="H245"></xref> based systems, the fields in the |
| 601 | NonStandardMessage should be: |
| 602 | </t> |
| 603 | |
| 604 | <vspace blankLines="1" /> |
| 605 | <list style="empty"> |
| 606 | <t>t35CountryCode = Hex: B5</t> |
| 607 | <t>t35Extension = Hex: 00</t> |
| 608 | <t>manufacturerCode = Hex: 0026</t> |
| 609 | <t>[Length of the Binary Sequence (8 bit number)]</t> |
| 610 | <t>[Binary Sequence consisting of an ASCII string, no NULL terminator]</t> |
| 611 | </list> |
| 612 | |
| 613 | <t> |
| 614 | The binary sequence is an ascii string merely for ease of use. |
| 615 | The string is not null terminated. The format of this string is |
| 616 | </t> |
| 617 | |
| 618 | <vspace blankLines="1" /> |
| 619 | <list style="empty"> |
| 620 | <t>speex [optional variables]</t> |
| 621 | </list> |
| 622 | |
| 623 | <t> |
| 624 | The optional variables are identical to those used for the SDP |
| 625 | a=fmtp strings discussed in section 5 above. The string is built |
| 626 | to be all on one line, each key-value pair separated by a |
| 627 | semi-colon. The optional variables MAY be omitted, which causes |
| 628 | the default values to be assumed. They are: |
| 629 | </t> |
| 630 | |
| 631 | <vspace blankLines="1" /> |
| 632 | <list style="empty"> |
| 633 | <t>ebw=narrow;mode=3;vbr=off;cng=off;ptime=20;sr=8000;penh=no;</t> |
| 634 | </list> |
| 635 | |
| 636 | <t> |
| 637 | The fifth octet of the block is the length of the binary sequence. |
| 638 | </t> |
| 639 | |
| 640 | <t> |
| 641 | NOTE: this method can result in the advertising of a large number |
| 642 | of Speex 'codecs' based on the number of variables possible. For |
| 643 | most VoIP applications, use of the default binary sequence of |
| 644 | 'speex' is RECOMMENDED to be used in addition to all other options. |
| 645 | This maximizes the chances that two H.323 based applications that |
| 646 | support Speex can find a mutual codec. |
| 647 | </t> |
| 648 | |
| 649 | </section> |
| 650 | |
| 651 | <section anchor="RTP Payload Types" title="RTP Payload Types"> |
| 652 | |
| 653 | <t> |
| 654 | Dynamic payload type codes MUST be negotiated 'out-of-band' |
| 655 | for the assignment of a dynamic payload type from the |
| 656 | range of 96-127. H.323 applications MUST use the H.245 |
| 657 | H2250LogicalChannelParameters encoding to accomplish this. |
| 658 | </t> |
| 659 | |
| 660 | </section> |
| 661 | |
| 662 | <section anchor="Security Considerations" title="Security Considerations"> |
| 663 | |
| 664 | <t> |
| 665 | RTP packets using the payload format defined in this specification |
| 666 | are subject to the security considerations discussed in the RTP |
| 667 | specification <xref target="rfc3550"></xref>, and any appropriate RTP profile. This implies |
| 668 | that confidentiality of the media streams is achieved by encryption. |
| 669 | Because the data compression used with this payload format is applied |
| 670 | end-to-end, encryption may be performed after compression so there is |
| 671 | no conflict between the two operations. |
| 672 | </t> |
| 673 | |
| 674 | <t> |
| 675 | A potential denial-of-service threat exists for data encodings using |
| 676 | compression techniques that have non-uniform receiver-end |
| 677 | computational load. The attacker can inject pathological datagrams |
| 678 | into the stream which are complex to decode and cause the receiver to |
| 679 | be overloaded. However, this encoding does not exhibit any |
| 680 | significant non-uniformity. |
| 681 | </t> |
| 682 | |
| 683 | <t> |
| 684 | As with any IP-based protocol, in some circumstances a receiver may |
| 685 | be overloaded simply by the receipt of too many packets, either |
| 686 | desired or undesired. Network-layer authentication may be used to |
| 687 | discard packets from undesired sources, but the processing cost of |
| 688 | the authentication itself may be too high. |
| 689 | </t> |
| 690 | |
| 691 | </section> |
| 692 | |
| 693 | <section anchor="Acknowledgments" title="Acknowledgments"> |
| 694 | |
| 695 | <t> |
| 696 | The authors would like to thank Equivalence Pty Ltd of Australia |
| 697 | for their assistance in attempting to standardize the use of Speex |
| 698 | in H.323 applications, and for implementing Speex in their open |
| 699 | source OpenH323 stack. The authors would also like to thank Brian |
| 700 | C. Wiles <brian@streamcomm.com> of StreamComm for his assistance in |
| 701 | developing the proposed standard for Speex use in H.323 |
| 702 | applications. |
| 703 | </t> |
| 704 | |
| 705 | <t> |
| 706 | The authors would also like to thank the following members of the |
| 707 | Speex and AVT communities for their input: Ross Finlayson, |
| 708 | Federico Montesino Pouzols, Henning Schulzrinne, Magnus Westerlund. |
| 709 | </t> |
| 710 | </section> |
| 711 | |
| 712 | </middle> |
| 713 | |
| 714 | <back> |
| 715 | |
| 716 | <references title="Normative References"> |
| 717 | |
| 718 | <reference anchor="rfc2119"> |
| 719 | <front> |
| 720 | <title>Key words for use in RFCs to Indicate Requirement Levels </title> |
| 721 | <author initials="S." surname="Bradner" fullname="Scott Bradner"></author> |
| 722 | </front> |
| 723 | <seriesInfo name="RFC" value="2119" /> |
| 724 | </reference> |
| 725 | |
| 726 | <reference anchor="rfc3550"> |
| 727 | <front> |
| 728 | <title>RTP: A Transport Protocol for real-time applications</title> |
| 729 | <author initials="H." surname="Schulzrinne" fullname=""></author> |
| 730 | <author initials="S." surname="Casner" fullname=""></author> |
| 731 | <author initials="R." surname="Frederick" fullname=""></author> |
| 732 | <author initials="V." surname="Jacobson" fullname=""></author> |
| 733 | </front> |
| 734 | <seriesInfo name="RFC" value="3550" /> |
| 735 | </reference> |
| 736 | |
| 737 | <reference anchor="rfc2045"> |
| 738 | <front> |
| 739 | <title>Multipurpose Internet Mail Extensions (MIME) Part One: Format of Internet Message Bodies</title> |
| 740 | <author initials="" surname="" fullname=""></author> |
| 741 | </front> |
| 742 | <date month="November" year="1998" /> |
| 743 | <seriesInfo name="RFC" value="2045" /> |
| 744 | </reference> |
| 745 | |
| 746 | <reference anchor="rfc2327"> |
| 747 | <front> |
| 748 | <title>SDP: Session Description Protocol</title> |
| 749 | <author initials="V." surname="Jacobson" fullname=""></author> |
| 750 | <author initials="M." surname="Handley" fullname=""></author> |
| 751 | </front> |
| 752 | <date month="April" year="1998" /> |
| 753 | <seriesInfo name="RFC" value="2327" /> |
| 754 | </reference> |
| 755 | |
| 756 | <reference anchor="H323"> |
| 757 | <front> |
| 758 | <title>Packet-based Multimedia Communications Systems</title> |
| 759 | <author initials="" surname="" fullname=""></author> |
| 760 | </front> |
| 761 | <date month="" year="1998" /> |
| 762 | <seriesInfo name="ITU-T Recommendation" value="H.323" /> |
| 763 | </reference> |
| 764 | |
| 765 | <reference anchor="H245"> |
| 766 | <front> |
| 767 | <title>Control of communications between Visual Telephone Systems and Terminal Equipment</title> |
| 768 | <author initials="" surname="" fullname=""></author> |
| 769 | </front> |
| 770 | <date month="" year="1998" /> |
| 771 | <seriesInfo name="ITU-T Recommendation" value="H.245" /> |
| 772 | </reference> |
| 773 | |
| 774 | <reference anchor="rfc3551"> |
| 775 | <front> |
| 776 | <title>RTP Profile for Audio and Video Conferences with Minimal Control.</title> |
| 777 | <author initials="H." surname="Schulzrinne" fullname=""></author> |
| 778 | <author initials="S." surname="Casner" fullname=""></author> |
| 779 | </front> |
| 780 | <date month="July" year="2003" /> |
| 781 | <seriesInfo name="RFC" value="3551" /> |
| 782 | </reference> |
| 783 | |
| 784 | <reference anchor="rfc3534"> |
| 785 | <front> |
| 786 | <title>The application/ogg Media Type</title> |
| 787 | <author initials="L." surname="Walleij" fullname=""></author> |
| 788 | </front> |
| 789 | <date month="May" year="2003" /> |
| 790 | <seriesInfo name="RFC" value="3534" /> |
| 791 | </reference> |
| 792 | |
| 793 | </references> |
| 794 | |
| 795 | <references title="Informative References"> |
| 796 | |
| 797 | <reference anchor="speexenc"> |
| 798 | <front> |
| 799 | <title>Speexenc/speexdec, reference command-line encoder/decoder</title> |
| 800 | </front> |
| 801 | <seriesInfo name="Speex website" value="http://www.speex.org/" /> |
| 802 | </reference> |
| 803 | |
| 804 | <reference anchor="CELP"> |
| 805 | <front> |
| 806 | <title>CELP, U.S. Federal Standard 1016.</title> |
| 807 | <author initials="" surname="" fullname=""></author> |
| 808 | </front> |
| 809 | <seriesInfo name="National Technical Information Service (NTIS) website" value="http://www.ntis.gov/" /> |
| 810 | </reference> |
| 811 | |
| 812 | </references> |
| 813 | |
| 814 | </back> |
| 815 | </rfc> |