live555 源码分析： PLAY 的处理

在 SETUP 请求之后，客户端会发起 PLAY 请求，以请求服务器开始传输音视频数据。在 PLAY 请求执行时，一定是已经执行过 SETUP 请求，建立好了客户端会话，因而会与其它要求客户端会话已经建立的请求一起，通过 clientSession->handleCmd_withinSession() 执行：

      } else if (strcmp(cmdName, "TEARDOWN") == 0          || strcmp(cmdName, "PLAY") == 0 || strcmp(cmdName, "PAUSE") == 0          || strcmp(cmdName, "GET_PARAMETER") == 0          || strcmp(cmdName, "SET_PARAMETER") == 0) {        if (clientSession != NULL) {          clientSession->handleCmd_withinSession(this, cmdName, urlPreSuffix,              urlSuffix, (char const*) fRequestBuffer);        } else {          handleCmd_sessionNotFound();        }      }

RTSPServer::RTSPClientSession::handleCmd_withinSession() 的定义如下：

void RTSPServer::RTSPClientSession::handleCmd_withinSession(RTSPServer::RTSPClientConnection* ourClientConnection,    char const* cmdName,    char const* urlPreSuffix, char const* urlSuffix,    char const* fullRequestStr) {  // This will either be:  // - a non-aggregated operation, if "urlPreSuffix" is the session (stream)  //   name and "urlSuffix" is the subsession (track) name, or  // - an aggregated operation, if "urlSuffix" is the session (stream) name,  //   or "urlPreSuffix" is the session (stream) name, and "urlSuffix" is empty,  //   or "urlPreSuffix" and "urlSuffix" are both nonempty, but when concatenated, (with "/") form the session (stream) name.  // Begin by figuring out which of these it is:  ServerMediaSubsession* subsession;  if (fOurServerMediaSession == NULL) { // There wasn't a previous SETUP!    ourClientConnection->handleCmd_notSupported();    return;  } else if (urlSuffix[0] != '\0' && strcmp(fOurServerMediaSession->streamName(), urlPreSuffix) == 0) {    // Non-aggregated operation.    // Look up the media subsession whose track id is "urlSuffix":    ServerMediaSubsessionIterator iter(*fOurServerMediaSession);    while ((subsession = iter.next()) != NULL) {      if (strcmp(subsession->trackId(), urlSuffix) == 0) break; // success    }    if (subsession == NULL) { // no such track!      ourClientConnection->handleCmd_notFound();      return;    }  } else if (strcmp(fOurServerMediaSession->streamName(), urlSuffix) == 0 ||      (urlSuffix[0] == '\0' && strcmp(fOurServerMediaSession->streamName(), urlPreSuffix) == 0)) {    // Aggregated operation    subsession = NULL;  } else if (urlPreSuffix[0] != '\0' && urlSuffix[0] != '\0') {    // Aggregated operation, if <urlPreSuffix>/<urlSuffix> is the session (stream) name:    unsigned const urlPreSuffixLen = strlen(urlPreSuffix);    if (strncmp(fOurServerMediaSession->streamName(), urlPreSuffix, urlPreSuffixLen) == 0        && fOurServerMediaSession->streamName()[urlPreSuffixLen] == '/'        && strcmp(&(fOurServerMediaSession->streamName())[urlPreSuffixLen + 1], urlSuffix) == 0) {      subsession = NULL;    } else {      ourClientConnection->handleCmd_notFound();      return;    }  } else { // the request doesn't match a known stream and/or track at all!    ourClientConnection->handleCmd_notFound();    return;  }  if (strcmp(cmdName, "TEARDOWN") == 0) {    handleCmd_TEARDOWN(ourClientConnection, subsession);  } else if (strcmp(cmdName, "PLAY") == 0) {    handleCmd_PLAY(ourClientConnection, subsession, fullRequestStr);  } else if (strcmp(cmdName, "PAUSE") == 0) {    handleCmd_PAUSE(ourClientConnection, subsession);  } else if (strcmp(cmdName, "GET_PARAMETER") == 0) {    handleCmd_GET_PARAMETER(ourClientConnection, subsession, fullRequestStr);  } else if (strcmp(cmdName, "SET_PARAMETER") == 0) {    handleCmd_SET_PARAMETER(ourClientConnection, subsession, fullRequestStr);  }}

在这个函数中，首先会找到请求的 URL 所相应的 ServerMediaSubsession，如果查找失败，会直接返回 404 并结束处理；然后即是根据请求的具体类型，通过不同的方法进行处理。

找到的 ServerMediaSubsession 存储于 subsession 中。subsession 值为空并不表示查找失败。这里通过值为 NULL 表示操作应用于 fOurServerMediaSession 描述的整个流媒体会话。请求的资源的路径可以分为两部分，分别是资源路径前缀 urlPreSuffix 和标识具体 track 的后缀 urlSuffix。live555 中通过 ServerMediaSession 和 ServerMediaSubsession 两级对象管理流媒体会话。

根据请求的资源的路径与流媒体会话状态的对应关系的不同，判断操作是否可以支持，以及操作应用的范围。

查找 ServerMediaSubsession 之后，PLAY 请求通过RTSPServer::RTSPClientSession::handleCmd_PLAY() 函数处理：

void RTSPServer::RTSPClientSession::handleCmd_PLAY(RTSPServer::RTSPClientConnection* ourClientConnection,    ServerMediaSubsession* subsession, char const* fullRequestStr) {  char* rtspURL = fOurRTSPServer.rtspURL(fOurServerMediaSession,      ourClientConnection->fClientInputSocket);  unsigned rtspURLSize = strlen(rtspURL);  // Parse the client's "Scale:" header, if any:  float scale;  Boolean sawScaleHeader = parseScaleHeader(fullRequestStr, scale);  // Try to set the stream's scale factor to this value:  if (subsession == NULL /*aggregate op*/) {    fOurServerMediaSession->testScaleFactor(scale);  } else {    subsession->testScaleFactor(scale);  }  char buf[100];  char* scaleHeader;  if (!sawScaleHeader) {    buf[0] = '\0'; // Because we didn't see a Scale: header, don't send one back  } else {    sprintf(buf, "Scale: %f\r\n", scale);  }  scaleHeader = strDup(buf);  // Parse the client's "Range:" header, if any:  float duration = 0.0;  double rangeStart = 0.0, rangeEnd = 0.0;  char* absStart = NULL; char* absEnd = NULL;  Boolean startTimeIsNow;  Boolean sawRangeHeader    = parseRangeHeader(fullRequestStr, rangeStart, rangeEnd, absStart, absEnd, startTimeIsNow);  if (sawRangeHeader && absStart == NULL/*not seeking by 'absolute' time*/) {    // Use this information, plus the stream's duration (if known), to create our own "Range:" header, for the response:    duration = subsession == NULL /*aggregate op*/      ? fOurServerMediaSession->duration() : subsession->duration();    if (duration < 0.0) {      // We're an aggregate PLAY, but the subsessions have different durations.      // Use the largest of these durations in our header      duration = -duration;    }    // Make sure that "rangeStart" and "rangeEnd" (from the client's "Range:" header)    // have sane values, before we send back our own "Range:" header in our response:    if (rangeStart < 0.0) rangeStart = 0.0;    else if (rangeStart > duration) rangeStart = duration;    if (rangeEnd < 0.0) rangeEnd = 0.0;    else if (rangeEnd > duration) rangeEnd = duration;    if ((scale > 0.0 && rangeStart > rangeEnd && rangeEnd > 0.0)        || (scale < 0.0 && rangeStart < rangeEnd)) {      // "rangeStart" and "rangeEnd" were the wrong way around; swap them:      double tmp = rangeStart;      rangeStart = rangeEnd;      rangeEnd = tmp;    }  }  // Create a "RTP-Info:" line.  It will get filled in from each subsession's state:  char const* rtpInfoFmt =    "%s" // "RTP-Info:", plus any preceding rtpInfo items    "%s" // comma separator, if needed    "url=%s/%s"    ";seq=%d"    ";rtptime=%u"    ;  unsigned rtpInfoFmtSize = strlen(rtpInfoFmt);  char* rtpInfo = strDup("RTP-Info: ");  unsigned i, numRTPInfoItems = 0;  // Do any required seeking/scaling on each subsession, before starting streaming.  // (However, we don't do this if the "PLAY" request was for just a single subsession  // of a multiple-subsession stream; for such streams, seeking/scaling can be done  // only with an aggregate "PLAY".)  for (i = 0; i < fNumStreamStates; ++i) {    if (subsession == NULL /* means: aggregated operation */ || fNumStreamStates == 1) {      if (fStreamStates[i].subsession != NULL) {        if (sawScaleHeader) {          fStreamStates[i].subsession->setStreamScale(fOurSessionId, fStreamStates[i].streamToken, scale);        }        if (absStart != NULL) {          // Special case handling for seeking by 'absolute' time:          fStreamStates[i].subsession->seekStream(fOurSessionId, fStreamStates[i].streamToken, absStart, absEnd);        } else {          // Seeking by relative (NPT) time:          u_int64_t numBytes;          if (!sawRangeHeader || startTimeIsNow) {            // We're resuming streaming without seeking, so we just do a 'null' seek            // (to get our NPT, and to specify when to end streaming):            fStreamStates[i].subsession->nullSeekStream(fOurSessionId, fStreamStates[i].streamToken,                rangeEnd, numBytes);          } else {            // We do a real 'seek':            double streamDuration = 0.0; // by default; means: stream until the end of the media            if (rangeEnd > 0.0 && (rangeEnd + 0.001) < duration) {              // the 0.001 is because we limited the values to 3 decimal places              // We want the stream to end early.  Set the duration we want:              streamDuration = rangeEnd - rangeStart;              if (streamDuration < 0.0) streamDuration = -streamDuration;              // should happen only if scale < 0.0            }            fStreamStates[i].subsession->seekStream(fOurSessionId, fStreamStates[i].streamToken,                rangeStart, streamDuration, numBytes);          }        }      }    }  }  // Create the "Range:" header that we'll send back in our response.  // (Note that we do this after seeking, in case the seeking operation changed the range start time.)  if (absStart != NULL) {    // We're seeking by 'absolute' time:    if (absEnd == NULL) {      sprintf(buf, "Range: clock=%s-\r\n", absStart);    } else {      sprintf(buf, "Range: clock=%s-%s\r\n", absStart, absEnd);    }    delete[] absStart; delete[] absEnd;  } else {    // We're seeking by relative (NPT) time:    if (!sawRangeHeader || startTimeIsNow) {      // We didn't seek, so in our response, begin the range with the current NPT (normal play time):      float curNPT = 0.0;      for (i = 0; i < fNumStreamStates; ++i) {        if (subsession == NULL /* means: aggregated operation */        || subsession == fStreamStates[i].subsession) {          if (fStreamStates[i].subsession == NULL) continue;          float npt = fStreamStates[i].subsession->getCurrentNPT(fStreamStates[i].streamToken);          if (npt > curNPT)            curNPT = npt;          // Note: If this is an aggregate "PLAY" on a multi-subsession stream,          // then it's conceivable that the NPTs of each subsession may differ          // (if there has been a previous seek on just one subsession).          // In this (unusual) case, we just return the largest NPT; I hope that turns out OK...        }      }      rangeStart = curNPT;    }    if (rangeEnd == 0.0 && scale >= 0.0) {      sprintf(buf, "Range: npt=%.3f-\r\n", rangeStart);    } else {      sprintf(buf, "Range: npt=%.3f-%.3f\r\n", rangeStart, rangeEnd);    }  }  char* rangeHeader = strDup(buf);  // Now, start streaming:  for (i = 0; i < fNumStreamStates; ++i) {    if (subsession == NULL /* means: aggregated operation */    || subsession == fStreamStates[i].subsession) {      unsigned short rtpSeqNum = 0;      unsigned rtpTimestamp = 0;      if (fStreamStates[i].subsession == NULL) continue;      fStreamStates[i].subsession->startStream(fOurSessionId,          fStreamStates[i].streamToken,          (TaskFunc*) noteClientLiveness, this,          rtpSeqNum, rtpTimestamp,          RTSPServer::RTSPClientConnection::handleAlternativeRequestByte, ourClientConnection);      const char *urlSuffix = fStreamStates[i].subsession->trackId();      char* prevRTPInfo = rtpInfo;      unsigned rtpInfoSize = rtpInfoFmtSize          + strlen(prevRTPInfo)          + 1          + rtspURLSize + strlen(urlSuffix)          + 5 /*max unsigned short len*/          + 10 /*max unsigned (32-bit) len*/          + 2 /*allows for trailing \r\n at final end of string*/;      rtpInfo = new char[rtpInfoSize];      sprintf(rtpInfo, rtpInfoFmt, prevRTPInfo,          numRTPInfoItems++ == 0 ? "" : ",", rtspURL, urlSuffix, rtpSeqNum,          rtpTimestamp);      delete[] prevRTPInfo;    }  }  if (numRTPInfoItems == 0) {    rtpInfo[0] = '\0';  } else {    unsigned rtpInfoLen = strlen(rtpInfo);    rtpInfo[rtpInfoLen] = '\r';    rtpInfo[rtpInfoLen + 1] = '\n';    rtpInfo[rtpInfoLen + 2] = '\0';  }  // Fill in the response:  snprintf((char*) ourClientConnection->fResponseBuffer,      sizeof ourClientConnection->fResponseBuffer, "RTSP/1.0 200 OK\r\n"          "CSeq: %s\r\n"          "%s"          "%s"          "%s"          "Session: %08X\r\n"          "%s\r\n",          ourClientConnection->fCurrentCSeq,          dateHeader(),          scaleHeader,          rangeHeader,          fOurSessionId,          rtpInfo);  delete[] rtpInfo;  delete[] rangeHeader;  delete[] scaleHeader;  delete[] rtspURL;}

在具体分析这个函数之前，先来看一个 PLAY 请求/响应的示例。PLAY 请求示例：

PLAY rtsp://10.240.248.20:8554/video/raw_h264_stream.264/ RTSP/1.0Range: npt=0.000-CSeq: 4User-Agent: Lavf56.40.101Session: D10C8C71

PLAY 响应的示例：

RTSP/1.0 200 OKCSeq: 4Date: Sat, Sep 02 2017 08:54:03 GMTRange: npt=0.000-Session: D10C8C71RTP-Info: url=rtsp://10.240.248.20:8554/video/raw_h264_stream.264/track1;seq=12647;rtptime=2457491257

然后来看 RTSPServer::RTSPClientSession::handleCmd_PLAY() 的具体处理过程。

第一步，解析 Scale: 头部，为流媒体会话的各个子会话设置 Scale，并构造要返回的 Scale: 头部。

  // Parse the client's "Scale:" header, if any:  float scale;  Boolean sawScaleHeader = parseScaleHeader(fullRequestStr, scale);  // Try to set the stream's scale factor to this value:  if (subsession == NULL /*aggregate op*/) {    fOurServerMediaSession->testScaleFactor(scale);  } else {    subsession->testScaleFactor(scale);  }  char buf[100];  char* scaleHeader;  if (!sawScaleHeader) {    buf[0] = '\0'; // Because we didn't see a Scale: header, don't send one back  } else {    sprintf(buf, "Scale: %f\r\n", scale);  }  scaleHeader = strDup(buf);

用于解析 Scale: 头部的 parseScaleHeader() 定义如下：

Boolean parseScaleHeader(char const* buf, float& scale) {  // Initialize the result parameter to a default value:  scale = 1.0;  // First, find "Scale:"  while (1) {    if (*buf == '\0') return False; // not found    if (_strncasecmp(buf, "Scale:", 6) == 0) break;    ++buf;  }  char const* fields = buf + 6;  while (*fields == ' ') ++fields;  float sc;  if (sscanf(fields, "%f", &sc) == 1) {    scale = sc;  } else {    return False; // The header is malformed  }  return True;}

scale 默认情况下会被设置为1.0。

解释一下 scale 值的作用。根据 RTSP 的规范 RFC 2326 12.34 Scale，scale 值是用于控制播放速度的。scale 值为 1，表示以正常的向前观看速率正常播放或记录。如果不是 1，值对应于相对于正常观看速率的比率。比如，值为 2 的比率表示两倍的正常观看速率（“快进”），而值为 0.5 的比率表示一半的观看速率。换句话说，值为 2 的比率播放的时间增长速率是墙上时钟的两倍。对于流逝的每一秒时间，将有 2 秒的内容被发送。负值表示相反的方向。

除非 Speed 参数另有要求，否则数据速率不得更改。Scale
变化的实现取决于服务器和媒体类型。对于视频，服务器可以，比如，传送一个关键帧或选择的关键帧。对于音频，它可以在保持音调的同时缩放音频，或者更不希望地传送音频片段。服务器应尝试提供近似的观看速率，但可以限制其支持的比例值范围。响应要包含服务器选择的实际的 scale。如果请求包含了 Range 参数，新的 scale 值将在此时生效。

应用于整个流媒体会话的 PLAY 请求，通过 ServerMediaSession::testScaleFactor(float& scale) 设置 scale，应用于具体子会话的 PLAY 请求，则通过 ServerMediaSubsession 的 testScaleFactor(float& scale) 设置。

对于我们的 H264VideoFileServerMediaSubsession，testScaleFactor(float& scale) 的实现还是在 ServerMediaSubsession 中：

void ServerMediaSubsession::testScaleFactor(float& scale) {  // default implementation: Support scale = 1 only  scale = 1;}

ServerMediaSession::testScaleFactor(float& scale) 的定义如下：

void ServerMediaSession::testScaleFactor(float& scale) {  // First, try setting all subsessions to the desired scale.  // If the subsessions' actual scales differ from each other, choose the  // value that's closest to 1, and then try re-setting all subsessions to that  // value.  If the subsessions' actual scales still differ, re-set them all to 1.  float minSSScale = 1.0;  float maxSSScale = 1.0;  float bestSSScale = 1.0;  float bestDistanceTo1 = 0.0;  ServerMediaSubsession* subsession;  for (subsession = fSubsessionsHead; subsession != NULL;       subsession = subsession->fNext) {    float ssscale = scale;    subsession->testScaleFactor(ssscale);    if (subsession == fSubsessionsHead) { // this is the first subsession      minSSScale = maxSSScale = bestSSScale = ssscale;      bestDistanceTo1 = (float)fabs(ssscale - 1.0f);    } else {      if (ssscale < minSSScale) {        minSSScale = ssscale;      } else if (ssscale > maxSSScale) {        maxSSScale = ssscale;      }      float distanceTo1 = (float) fabs(ssscale - 1.0f);      if (distanceTo1 < bestDistanceTo1) {        bestSSScale = ssscale;        bestDistanceTo1 = distanceTo1;      }    }  }  if (minSSScale == maxSSScale) {    // All subsessions are at the same scale: minSSScale == bestSSScale == maxSSScale    scale = minSSScale;    return;  }  // The scales for each subsession differ.  Try to set each one to the value  // that's closest to 1:  for (subsession = fSubsessionsHead; subsession != NULL;       subsession = subsession->fNext) {    float ssscale = bestSSScale;    subsession->testScaleFactor(ssscale);    if (ssscale != bestSSScale) break; // no luck  }  if (subsession == NULL) {    // All subsessions are at the same scale: bestSSScale    scale = bestSSScale;    return;  }  // Still no luck.  Set each subsession's scale to 1:  for (subsession = fSubsessionsHead; subsession != NULL;       subsession = subsession->fNext) {    float ssscale = 1;    subsession->testScaleFactor(ssscale);  }  scale = 1;}

在这个函数中，执行的步骤如下：
Step 1：尝试将所有子会话的 scale 设置客户端请求的值，并记录下子会话选择的 scale 的最大值，最小值，以及与 1 最接近的值。
Step 2：第 1 步执行过程中，所有子会话选择的 scale 值相同，则将该 scale 返回给调用者。
Step 3：第 1 步执行过程中，所有子会话选择的 scale 值不同，则尝试将与 1 最接近的 scale 值设置给所有的。
Step 4：第 3 步可以成功将与 1 最接近的 scale 值设置给所有子会话，则将该 scale 返回给调用者。
Step 5：第 3 步无法将与 1 最接近的 scale 值设置给所有子会话，则将每个子会话的 scale 值设置为 1。

回到 RTSPServer::RTSPClientSession::handleCmd_PLAY()。

第二步，解析客户端的 Range: 头部，并根据 scale 值更新 range 值。

  // Parse the client's "Range:" header, if any:  float duration = 0.0;  double rangeStart = 0.0, rangeEnd = 0.0;  char* absStart = NULL; char* absEnd = NULL;  Boolean startTimeIsNow;  Boolean sawRangeHeader    = parseRangeHeader(fullRequestStr, rangeStart, rangeEnd, absStart, absEnd, startTimeIsNow);  if (sawRangeHeader && absStart == NULL/*not seeking by 'absolute' time*/) {    // Use this information, plus the stream's duration (if known), to create our own "Range:" header, for the response:    duration = subsession == NULL /*aggregate op*/      ? fOurServerMediaSession->duration() : subsession->duration();    if (duration < 0.0) {      // We're an aggregate PLAY, but the subsessions have different durations.      // Use the largest of these durations in our header      duration = -duration;    }    // Make sure that "rangeStart" and "rangeEnd" (from the client's "Range:" header)    // have sane values, before we send back our own "Range:" header in our response:    if (rangeStart < 0.0) rangeStart = 0.0;    else if (rangeStart > duration) rangeStart = duration;    if (rangeEnd < 0.0) rangeEnd = 0.0;    else if (rangeEnd > duration) rangeEnd = duration;    if ((scale > 0.0 && rangeStart > rangeEnd && rangeEnd > 0.0)        || (scale < 0.0 && rangeStart < rangeEnd)) {      // "rangeStart" and "rangeEnd" were the wrong way around; swap them:      double tmp = rangeStart;      rangeStart = rangeEnd;      rangeEnd = tmp;    }  }

第四步，创建 RTP-Info: 行：

  // Create a "RTP-Info:" line.  It will get filled in from each subsession's state:  char const* rtpInfoFmt =    "%s" // "RTP-Info:", plus any preceding rtpInfo items    "%s" // comma separator, if needed    "url=%s/%s"    ";seq=%d"    ";rtptime=%u"    ;  unsigned rtpInfoFmtSize = strlen(rtpInfoFmt);  char* rtpInfo = strDup("RTP-Info: ");  unsigned i, numRTPInfoItems = 0;

第五步，在开始流式传输之前，对每个子会话进行所需的 seeking/scaling，即设置播放进度和播放速率。

  for (i = 0; i < fNumStreamStates; ++i) {    if (subsession == NULL /* means: aggregated operation */ || fNumStreamStates == 1) {      if (fStreamStates[i].subsession != NULL) {        if (sawScaleHeader) {          fStreamStates[i].subsession->setStreamScale(fOurSessionId, fStreamStates[i].streamToken, scale);        }        if (absStart != NULL) {          // Special case handling for seeking by 'absolute' time:          fStreamStates[i].subsession->seekStream(fOurSessionId, fStreamStates[i].streamToken, absStart, absEnd);        } else {          // Seeking by relative (NPT) time:          u_int64_t numBytes;          if (!sawRangeHeader || startTimeIsNow) {            // We're resuming streaming without seeking, so we just do a 'null' seek            // (to get our NPT, and to specify when to end streaming):            fStreamStates[i].subsession->nullSeekStream(fOurSessionId, fStreamStates[i].streamToken,                rangeEnd, numBytes);          } else {            // We do a real 'seek':            double streamDuration = 0.0; // by default; means: stream until the end of the media            if (rangeEnd > 0.0 && (rangeEnd + 0.001) < duration) {              // the 0.001 is because we limited the values to 3 decimal places              // We want the stream to end early.  Set the duration we want:              streamDuration = rangeEnd - rangeStart;              if (streamDuration < 0.0) streamDuration = -streamDuration;              // should happen only if scale < 0.0            }            fStreamStates[i].subsession->seekStream(fOurSessionId, fStreamStates[i].streamToken,                rangeStart, streamDuration, numBytes);          }        }      }    }  }

第六步，创建将会在响应中发送的 Range: 头部。

  if (absStart != NULL) {    // We're seeking by 'absolute' time:    if (absEnd == NULL) {      sprintf(buf, "Range: clock=%s-\r\n", absStart);    } else {      sprintf(buf, "Range: clock=%s-%s\r\n", absStart, absEnd);    }    delete[] absStart; delete[] absEnd;  } else {    // We're seeking by relative (NPT) time:    if (!sawRangeHeader || startTimeIsNow) {      // We didn't seek, so in our response, begin the range with the current NPT (normal play time):      float curNPT = 0.0;      for (i = 0; i < fNumStreamStates; ++i) {        if (subsession == NULL /* means: aggregated operation */        || subsession == fStreamStates[i].subsession) {          if (fStreamStates[i].subsession == NULL) continue;          float npt = fStreamStates[i].subsession->getCurrentNPT(fStreamStates[i].streamToken);          if (npt > curNPT)            curNPT = npt;          // Note: If this is an aggregate "PLAY" on a multi-subsession stream,          // then it's conceivable that the NPTs of each subsession may differ          // (if there has been a previous seek on just one subsession).          // In this (unusual) case, we just return the largest NPT; I hope that turns out OK...        }      }      rangeStart = curNPT;    }    if (rangeEnd == 0.0 && scale >= 0.0) {      sprintf(buf, "Range: npt=%.3f-\r\n", rangeStart);    } else {      sprintf(buf, "Range: npt=%.3f-%.3f\r\n", rangeStart, rangeEnd);    }  }  char* rangeHeader = strDup(buf);

第七步，启动流媒体传输，并产生最终的 RTP-Info: 行。

  // Now, start streaming:  for (i = 0; i < fNumStreamStates; ++i) {    if (subsession == NULL /* means: aggregated operation */    || subsession == fStreamStates[i].subsession) {      unsigned short rtpSeqNum = 0;      unsigned rtpTimestamp = 0;      if (fStreamStates[i].subsession == NULL) continue;      fStreamStates[i].subsession->startStream(fOurSessionId,          fStreamStates[i].streamToken,          (TaskFunc*) noteClientLiveness, this,          rtpSeqNum, rtpTimestamp,          RTSPServer::RTSPClientConnection::handleAlternativeRequestByte, ourClientConnection);      const char *urlSuffix = fStreamStates[i].subsession->trackId();      char* prevRTPInfo = rtpInfo;      unsigned rtpInfoSize = rtpInfoFmtSize          + strlen(prevRTPInfo)          + 1          + rtspURLSize + strlen(urlSuffix)          + 5 /*max unsigned short len*/          + 10 /*max unsigned (32-bit) len*/          + 2 /*allows for trailing \r\n at final end of string*/;      rtpInfo = new char[rtpInfoSize];      sprintf(rtpInfo, rtpInfoFmt, prevRTPInfo,          numRTPInfoItems++ == 0 ? "" : ",", rtspURL, urlSuffix, rtpSeqNum,          rtpTimestamp);      delete[] prevRTPInfo;    }  }  if (numRTPInfoItems == 0) {    rtpInfo[0] = '\0';  } else {    unsigned rtpInfoLen = strlen(rtpInfo);    rtpInfo[rtpInfoLen] = '\r';    rtpInfo[rtpInfoLen + 1] = '\n';    rtpInfo[rtpInfoLen + 2] = '\0';  }

第八步，产生最终的响应，并释放临时分配的内存。

  // Fill in the response:  snprintf((char*) ourClientConnection->fResponseBuffer,      sizeof ourClientConnection->fResponseBuffer, "RTSP/1.0 200 OK\r\n"          "CSeq: %s\r\n"          "%s"          "%s"          "%s"          "Session: %08X\r\n"          "%s\r\n",          ourClientConnection->fCurrentCSeq,          dateHeader(),          scaleHeader,          rangeHeader,          fOurSessionId,          rtpInfo);  delete[] rtpInfo;  delete[] rangeHeader;  delete[] scaleHeader;  delete[] rtspURL;

Done。

live555 源码分析系列文章

live555 源码分析：简介
live555 源码分析：基础设施
live555 源码分析：MediaSever
Wireshark 抓包分析 RTSP/RTP/RTCP 基本工作过程
live555 源码分析：RTSPServer
live555 源码分析： DESCRIBE 的处理
live555 源码分析： SETUP 的处理
live555 源码分析： PLAY 的处理