rfc: http://www.ietf.org/rfc/rfc3550.txt
The RTP header has the following format:
The first twelve octets are present in every RTP packet, while the list of CSRC identifiers is present only when inserted by a mixer.

前12个字节在每个RTP包中都存在,他们是:V、P、X、CC、M、PT、sequence number、timestamp、×××C.其中V是版本号(占2个bit);P是填充位(占1个bit);X是扩展位(占1个bit);CC是CSRC的记数位(占4个bit);M是标记位(占1个bit);PT是有效载荷的类型(占7个bit);sequence number是RTP包记数位(占16个bit);timestamp是时间戳位(占4个字节);SSRC是同步标志位(占4个字节)。CSRC不是RTP必须的(占4个字节)。
                                                                                                             

The fields have the following meaning: 
   version (V): 2 bits
      This field identifies the version of RTP.  The version defined by this specification is two (2).  (The value 1 is used by the first draft version of RTP and the value 0 is used by the protocol; initially implemented in the "vat" audio tool.)

   padding (P): 1 bit
      If the padding bit is set, the packet contains one or more additional padding octets at the end which are not part of the payload.  The last octet of the padding contains a count of how many padding octets should be ignored, including itself.  Padding may be needed by some encryption algorithms with fixed block sizes or for carrying several RTP packets in a lower-layer protocol data unit.

   extension (X): 1 bit
      If the extension bit is set, the fixed header MUST be followed by exactly one header extension, with a format defined in Section 5.3.1.
 
   CSRC count (CC): 4 bits
      The CSRC count contains the number of CSRC identifiers that follow the fixed header.

   CSRC记数(CC)    表示CSRC标识的数目。CSRC标识紧跟在RTP固定头部之后,用来表示RTP数据报的来源,RTP协议允许在同一个会话中存在多个数据源,它们可以通过RTP混合器合并为一个数据源。例如,可以产生一个CSRC列表来表示一个电话会议,该会议通过一个RTP混合器将所有讲话者的语音数据组合为一个RTP数据源。

   marker (M): 1 bit
      The interpretation of the marker is defined by a profile.  It is intended to allow significant events such as frame boundaries to be marked in the packet stream.  A profile MAY define additional marker bits or specify that there is no marker bit by changing the number of bits in the payload type field (see Section 5.3).

   payload type (PT): 7 bits
      This field identifies the format of the RTP payload and determines its interpretation by the application.  A profile MAY specify a default static mapping of payload type codes to payload formats. Additional payload type codes MAY be defined dynamically through non-RTP means (see Section 3).  A set of default mappings for audio and video is specified in the companion RFC 3551 [1].  An RTP source MAY change the payload type during a session, but this field SHOULD NOT be used for multiplexing separate media streams (see Section 5.2). A receiver MUST ignore packets with payload types that it does not understand.

   负载类型(PT)    标明RTP负载的格式,包括所采用的编码算法、采样频率、承载通道等。例如,类型2表明该RTP数据包中承载的是用ITU G.721算法编码的语音数据,采样频率为8000Hz,并且采用单声道。

   sequence number: 16 bits
      The sequence number increments by one for each RTP data packet sent, and may be used by the receiver to detect packet loss and to restore packet sequence.  The initial value of the sequence number SHOULD be random (unpredictable) to make known-plaintext attacks on encryption more difficult, even if the source itself does not encrypt according to the method in Section 9.1, because the packets may flow through a translator that does.  Techniques for choosing unpredictable numbers are discussed in [17].

    序列号  用来为接收方提供探测数据丢失的方法,但如何处理丢失的数据则是应用程序自己的事情,RTP协议本身并不负责数据的重传。

   timestamp: 32 bits
      The timestamp reflects the sampling instant of the first octet in the RTP data packet.  The sampling instant MUST be derived from a clock that increments monotonically and linearly in time to allow synchronization and jitter calculations (see Section 6.4.1).  The resolution of the clock MUST be sufficient for the desired synchronization accuracy and for measuring packet arrival jitter (one tick per video frame is typically not sufficient).  The clock frequency is dependent on the format of data carried as payload and is specified statically in the profile or payload format specification that defines the format, or MAY be specified dynamically for payload formats defined through non-RTP means.  If RTP packets are generated periodically, the nominal sampling instant as determined from the sampling clock is to be used, not a reading of the system clock.  As an example, for fixed-rate audio the timestamp clock would likely increment by one for each sampling period.  If an audio application reads blocks covering     160 sampling periods from the input device, the timestamp would be increased by 160 for each such block, regardless of whether the block is transmitted in a packet or dropped as silent.

    时间戳  记录了负载中第一个字节的采样时间,接收方能够时间戳能够确定数据的到达是否受到了延迟抖动的影响,但具体如何来补偿延迟抖动则是应用程序自己的事情。

      The initial value of the timestamp SHOULD be random, as for the sequence number.  Several consecutive RTP packets will have equal timestamps if they are (logically) generated at once, e.g., belong to the same video frame.  Consecutive RTP packets MAY contain timestamps that are not monotonic if the data is not transmitted in the order it was sampled, as in the case of MPEG interpolated video frames.  (The sequence numbers of the packets as transmitted will still be monotonic.)

      RTP timestamps from different media streams may advance at different rates and usually have independent, random offsets. Therefore, although these timestamps are sufficient to reconstruct the timing of a single stream, directly comparing RTP timestamps from different media is not effective for synchronization. Instead, for each medium the RTP timestamp is related to the sampling instant by pairing it with a timestamp from a reference clock (wallclock) that represents the time when the data corresponding to the RTP timestamp was sampled.  The reference clock is shared by all media to be synchronized.  The timestamp pairs are not transmitted in every data packet, but at a lower rate in RTCP SR packets as described in Section 6.4.

      The sampling instant is chosen as the point of reference for the RTP timestamp because it is known to the transmitting endpoint and has a common definition for all media, independent of encoding delays or other processing.  The purpose is to allow synchronized presentation of all media sampled at the same time.

      Applications transmitting stored data rather than data sampled in real time typically use a virtual presentation timeline derived from wallclock time to determine when the next frame or other unit of each medium in the stored data should be presented.  In this case, the RTP timestamp would reflect the presentation time for each unit.  That is, the RTP timestamp for each unit would be related to the wallclock time at which the unit becomes current on the virtual presentation timeline.  Actual presentation occurs some time later as determined by the receiver.

      An example describing live audio narration of prerecorded video
      illustrates the significance of choosing the sampling instant as
      the reference point.  In this scenario, the video would be
      presented locally for the narrator to view and would be
      simultaneously transmitted using RTP.  The "sampling instant" of a
      video frame transmitted in RTP would be established by referencing
      its timestamp to the wallclock time when that video frame was
      presented to the narrator.  The sampling instant for the audio RTP
      packets containing the narrator's speech would be established by
      referencing the same wallclock time when the audio was sampled.
      The audio and video may even be transmitted by different hosts if
      the reference clocks on the two hosts are synchronized by some
      means such as NTP.  A receiver can then synchronize presentation
      of the audio and video packets by relating their RTP timestamps
      using the timestamp pairs in RTCP SR packets.

   ×××C: 32 bits
      The ×××C field identifies the synchronization source.  This
      identifier SHOULD be chosen randomly, with the intent that no two
      synchronization sources within the same RTP session will have the
      same ×××C identifier.  An example algorithm for generating a
      random identifier is presented in Appendix A.6.  Although the
      probability of multiple sources choosing the same identifier is
      low, all RTP implementations must be prepared to detect and
      resolve collisions.  Section 8 describes the probability of
      collision along with a mechanism for resolving collisions and
      detecting RTP-level forwarding loops based on the uniqueness of
      the ×××C identifier.  If a source changes its source transport
      address, it must also choose a new ×××C identifier to avoid being
      interpreted as a looped source (see Section 8.2).

   CSRC list: 0 to 15 items, 32 bits each
      The CSRC list identifies the contributing sources for the payload
      contained in this packet.  The number of identifiers is given by
      the CC field.  If there are more than 15 contributing sources,
      only 15 can be identified.  CSRC identifiers are inserted by
      mixers (see Section 7.1), using the ×××C identifiers of
      contributing sources.  For example, for audio packets the ×××C
      identifiers of all sources that were mixed together to create a
      packet are listed, allowing correct talker indication at the
      receiver.