transmuxer.js

/**
 * mux.js
 *
 * Copyright (c) 2015 Brightcove
 * All rights reserved.
 *
 * A stream-based mp2t to mp4 converter. This utility can be used to
 * deliver mp4s to a SourceBuffer on platforms that support native
 * Media Source Extensions.
 */
'use strict';

var Stream = require('../utils/stream.js');
var mp4 = require('./mp4-generator.js');
var m2ts = require('../m2ts/m2ts.js');
var AdtsStream = require('../codecs/adts.js');
var H264Stream = require('../codecs/h264').H264Stream;
var AacStream = require('../aac');

// object types
var VideoSegmentStream, AudioSegmentStream, Transmuxer, CoalesceStream;

// Helper functions
var
  createDefaultSample,
  collectDtsInfo,
  clearDtsInfo,
  calculateTrackBaseMediaDecodeTime,
  arrayEquals,
  sumFrameByteLengths;

/**
 * Default sample object
 * see ISO/IEC 14496-12:2012, section 8.6.4.3
 */
createDefaultSample = function () {
  return {
    size: 0,
    flags: {
      isLeading: 0,
      dependsOn: 1,
      isDependedOn: 0,
      hasRedundancy: 0,
      degradationPriority: 0
    }
  };
};

/**
 * Compare two arrays (even typed) for same-ness
 */
arrayEquals = function(a, b) {
  var
    i;

  if (a.length !== b.length) {
    return false;
  }

  // compare the value of each element in the array
  for (i = 0; i < a.length; i++) {
    if (a[i] !== b[i]) {
      return false;
    }
  }

  return true;
};

/**
 * Sum the `byteLength` properties of the data in each AAC frame
 */
sumFrameByteLengths = function(array) {
  var
    i,
    currentObj,
    sum = 0;

  // sum the byteLength's all each nal unit in the frame
  for (i = 0; i < array.length; i++) {
    currentObj = array[i];
    sum += currentObj.data.byteLength;
  }

  return sum;
};

/**
 * Constructs a single-track, ISO BMFF media segment from AAC data
 * events. The output of this stream can be fed to a SourceBuffer
 * configured with a suitable initialization segment.
 */
AudioSegmentStream = function(track) {
  var
    adtsFrames = [],
    sequenceNumber = 0,
    earliestAllowedDts = 0;

  AudioSegmentStream.prototype.init.call(this);

  this.push = function(data) {
    collectDtsInfo(track, data);

    if (track) {
      track.audioobjecttype = data.audioobjecttype;
      track.channelcount = data.channelcount;
      track.samplerate = data.samplerate;
      track.samplingfrequencyindex = data.samplingfrequencyindex;
      track.samplesize = data.samplesize;
    }

    // buffer audio data until end() is called
    adtsFrames.push(data);
  };

  this.setEarliestDts = function(earliestDts) {
    earliestAllowedDts = earliestDts - track.timelineStartInfo.baseMediaDecodeTime;
  };

  this.flush = function() {
    var
      frames,
      moof,
      mdat,
      boxes;

    // return early if no audio data has been observed
    if (adtsFrames.length === 0) {
      this.trigger('done');
      return;
    }

    frames = this.trimAdtsFramesByEarliestDts_(adtsFrames);

    // we have to build the index from byte locations to
    // samples (that is, adts frames) in the audio data
    track.samples = this.generateSampleTable_(frames);

    // concatenate the audio data to constuct the mdat
    mdat = mp4.mdat(this.concatenateFrameData_(frames));

    adtsFrames = [];

    calculateTrackBaseMediaDecodeTime(track);
    moof = mp4.moof(sequenceNumber, [track]);
    boxes = new Uint8Array(moof.byteLength + mdat.byteLength);

    // bump the sequence number for next time
    sequenceNumber++;

    boxes.set(moof);
    boxes.set(mdat, moof.byteLength);

    clearDtsInfo(track);

    this.trigger('data', {track: track, boxes: boxes});
    this.trigger('done');
  };

  // If the audio segment extends before the earliest allowed dts
  // value, remove AAC frames until starts at or after the earliest
  // allowed DTS so that we don't end up with a negative baseMedia-
  // DecodeTime for the audio track
  this.trimAdtsFramesByEarliestDts_ = function(adtsFrames) {
    if (track.minSegmentDts >= earliestAllowedDts) {
      return adtsFrames;
    }

    // We will need to recalculate the earliest segment Dts
    track.minSegmentDts = Infinity;

    return adtsFrames.filter(function(currentFrame) {
      // If this is an allowed frame, keep it and record it's Dts
      if (currentFrame.dts >= earliestAllowedDts) {
        track.minSegmentDts = Math.min(track.minSegmentDts, currentFrame.dts);
        track.minSegmentPts = track.minSegmentDts;
        return true;
      }
      // Otherwise, discard it
      return false;
    });
  };

  // generate the track's raw mdat data from an array of frames
  this.generateSampleTable_ = function(frames) {
    var
      i,
      currentFrame,
      samples = [];

    for (i = 0; i < frames.length; i++) {
      currentFrame = frames[i];
      samples.push({
        size: currentFrame.data.byteLength,
        duration: 1024 // For AAC audio, all samples contain 1024 samples
      });
    }
    return samples;
  };

  // generate the track's sample table from an array of frames
  this.concatenateFrameData_ = function(frames) {
    var
      i,
      currentFrame,
      dataOffset = 0,
      data = new Uint8Array(sumFrameByteLengths(frames));

    for (i = 0; i < frames.length; i++) {
      currentFrame = frames[i];

      data.set(currentFrame.data, dataOffset);
      dataOffset += currentFrame.data.byteLength;
    }
    return data;
  };
};

AudioSegmentStream.prototype = new Stream();

/**
 * Constructs a single-track, ISO BMFF media segment from H264 data
 * events. The output of this stream can be fed to a SourceBuffer
 * configured with a suitable initialization segment.
 * @param track {object} track metadata configuration
 */
VideoSegmentStream = function(track) {
  var
    sequenceNumber = 0,
    nalUnits = [],
    config,
    pps;

  VideoSegmentStream.prototype.init.call(this);

  delete track.minPTS;

  this.gopCache_ = [];

  this.push = function(nalUnit) {
    collectDtsInfo(track, nalUnit);

    // record the track config
    if (nalUnit.nalUnitType === 'seq_parameter_set_rbsp' &&
        !config) {
      config = nalUnit.config;

      track.width = config.width;
      track.height = config.height;
      track.sps = [nalUnit.data];
      track.profileIdc = config.profileIdc;
      track.levelIdc = config.levelIdc;
      track.profileCompatibility = config.profileCompatibility;
    }

    if (nalUnit.nalUnitType === 'pic_parameter_set_rbsp' &&
        !pps) {
      pps = nalUnit.data;
      track.pps = [nalUnit.data];
    }

    // buffer video until flush() is called
    nalUnits.push(nalUnit);
  };

  this.flush = function() {
    var
      frames,
      gopForFusion,
      gops,
      moof,
      mdat,
      boxes;

    // Throw away nalUnits at the start of the byte stream until
    // we find the first AUD
    while (nalUnits.length) {
      if (nalUnits[0].nalUnitType === 'access_unit_delimiter_rbsp') {
        break;
      }
      nalUnits.shift();
    }

    // Return early if no video data has been observed
    if (nalUnits.length === 0) {
      this.resetStream_();
      this.trigger('done');
      return;
    }

    // Organize the raw nal-units into arrays that represent
    // higher-level constructs such as frames and gops
    // (group-of-pictures)
    frames = this.groupNalsIntoFrames_(nalUnits);
    gops = this.groupFramesIntoGops_(frames);

    // If the first frame of this fragment is not a keyframe we have
    // a problem since MSE (on Chrome) requires a leading keyframe.
    //
    // We have two approaches to repairing this situation:
    // 1) GOP-FUSION:
    //    This is where we keep track of the GOPS (group-of-pictures)
    //    from previous fragments and attempt to find one that we can
    //    prepend to the current fragment in order to create a valid
    //    fragment.
    // 2) KEYFRAME-PULLING:
    //    Here we search for the first keyframe in the fragment and
    //    throw away all the frames between the start of the fragment
    //    and that keyframe. We then extend the duration and pull the
    //    PTS of the keyframe forward so that it covers the time range
    //    of the frames that were disposed of.
    //
    // #1 is far prefereable over #2 which can cause "stuttering" but
    // requires more things to be just right.
    if (!gops[0][0].keyFrame) {
      // Search for a gop for fusion from our gopCache
      gopForFusion = this.getGopForFusion_(nalUnits[0], track);

      if (gopForFusion) {
        gops.unshift(gopForFusion);
        // Adjust Gops' metadata to account for the inclusion of the
        // new gop at the beginning
        gops.byteLength += gopForFusion.byteLength;
        gops.nalCount += gopForFusion.nalCount;
        gops.pts = gopForFusion.pts;
        gops.dts = gopForFusion.dts;
        gops.duration += gopForFusion.duration;
      } else {
        // If we didn't find a candidate gop fall back to keyrame-pulling
        gops = this.extendFirstKeyFrame_(gops);
      }
    }
    collectDtsInfo(track, gops);

    // First, we have to build the index from byte locations to
    // samples (that is, frames) in the video data
    track.samples = this.generateSampleTable_(gops);

    // Concatenate the video data and construct the mdat
    mdat = mp4.mdat(this.concatenateNalData_(gops));

    // save all the nals in the last GOP into the gop cache
    this.gopCache_.unshift({
      gop: gops.pop(),
      pps: track.pps,
      sps: track.sps
    });

    // Keep a maximum of 6 GOPs in the cache
    this.gopCache_.length = Math.min(6, this.gopCache_.length);

    // Clear nalUnits
    nalUnits = [];

    calculateTrackBaseMediaDecodeTime(track);

    this.trigger('timelineStartInfo', track.timelineStartInfo);

    moof = mp4.moof(sequenceNumber, [track]);

    // it would be great to allocate this array up front instead of
    // throwing away hundreds of media segment fragments
    boxes = new Uint8Array(moof.byteLength + mdat.byteLength);

    // Bump the sequence number for next time
    sequenceNumber++;

    boxes.set(moof);
    boxes.set(mdat, moof.byteLength);

    this.trigger('data', {track: track, boxes: boxes});

    this.resetStream_();

    // Continue with the flush process now
    this.trigger('done');
  };

  this.resetStream_ = function() {
    clearDtsInfo(track);

    // reset config and pps because they may differ across segments
    // for instance, when we are rendition switching
    config = undefined;
    pps = undefined;
  };

  // Search for a candidate Gop for gop-fusion from the gop cache and
  // return it or return null if no good candidate was found
  this.getGopForFusion_ = function (nalUnit) {
    var
      halfSecond = 45000, // Half-a-second in a 90khz clock
      allowableOverlap = 10000, // About 3 frames @ 30fps
      nearestDistance = Infinity,
      dtsDistance,
      nearestGopObj,
      currentGop,
      currentGopObj,
      i;

    // Search for the GOP nearest to the beginning of this nal unit
    for (i = 0; i < this.gopCache_.length; i++) {
      currentGopObj = this.gopCache_[i];
      currentGop = currentGopObj.gop;

      // Reject Gops with different SPS or PPS
      if (!(track.pps && arrayEquals(track.pps[0], currentGopObj.pps[0])) ||
          !(track.sps && arrayEquals(track.sps[0], currentGopObj.sps[0]))) {
        continue;
      }

      // Reject Gops that would require a negative baseMediaDecodeTime
      if (currentGop.dts < track.timelineStartInfo.dts) {
        continue;
      }

      // The distance between the end of the gop and the start of the nalUnit
      dtsDistance = (nalUnit.dts - currentGop.dts) - currentGop.duration;

      // Only consider GOPS that start before the nal unit and end within
      // a half-second of the nal unit
      if (dtsDistance >= -allowableOverlap &&
          dtsDistance <= halfSecond) {

        // Always use the closest GOP we found if there is more than
        // one candidate
        if (!nearestGopObj ||
            nearestDistance > dtsDistance) {
          nearestGopObj = currentGopObj;
          nearestDistance = dtsDistance;
        }
      }
    }

    if (nearestGopObj) {
      return nearestGopObj.gop;
    }
    return null;
  };

  this.extendFirstKeyFrame_ = function(gops) {
    var
      h, i,
      currentGop,
      newGops;

    if (!gops[0][0].keyFrame) {
      // Remove the first GOP
      currentGop = gops.shift();

      gops.byteLength -=  currentGop.byteLength;
      gops.nalCount -= currentGop.nalCount;

      // Extend the first frame of what is now the
      // first gop to cover the time period of the
      // frames we just removed
      gops[0][0].dts = currentGop.dts;
      gops[0][0].pts = currentGop.pts;
      gops[0][0].duration += currentGop.duration;
    }

    return gops;
  };

  // Convert an array of nal units into an array of frames with each frame being
  // composed of the nal units that make up that frame
  // Also keep track of cummulative data about the frame from the nal units such
  // as the frame duration, starting pts, etc.
  this.groupNalsIntoFrames_ = function(nalUnits) {
    var
      i,
      currentNal,
      startPts,
      startDts,
      currentFrame = [],
      frames = [];

    currentFrame.byteLength = 0;

    for (i = 0; i < nalUnits.length; i++) {
      currentNal = nalUnits[i];

      // Split on 'aud'-type nal units
      if (currentNal.nalUnitType === 'access_unit_delimiter_rbsp') {
        // Since the very first nal unit is expected to be an AUD
        // only push to the frames array when currentFrame is not empty
        if (currentFrame.length) {
          currentFrame.duration = currentNal.dts - currentFrame.dts;
          frames.push(currentFrame);
        }
        currentFrame = [currentNal];
        currentFrame.byteLength = currentNal.data.byteLength;
        currentFrame.pts = currentNal.pts;
        currentFrame.dts = currentNal.dts;
      } else {
        // Specifically flag key frames for ease of use later
        if (currentNal.nalUnitType === 'slice_layer_without_partitioning_rbsp_idr') {
          currentFrame.keyFrame = true;
        }
        currentFrame.duration = currentNal.dts - currentFrame.dts;
        currentFrame.byteLength += currentNal.data.byteLength;
        currentFrame.push(currentNal);
      }
    }

    // For the last frame, use the duration of the previous frame if we
    // have nothing better to go on
    if (frames.length &&
        !currentFrame.duration ||
        currentFrame.duration <= 0) {
      currentFrame.duration = frames[frames.length - 1].duration;
    }

    // Push the final frame
    frames.push(currentFrame);
    return frames;
  };

  // Convert an array of frames into an array of Gop with each Gop being composed
  // of the frames that make up that Gop
  // Also keep track of cummulative data about the Gop from the frames such as the
  // Gop duration, starting pts, etc.
  this.groupFramesIntoGops_ = function(frames) {
    var
      i,
      currentFrame,
      currentGop = [],
      gops = [];

    // We must pre-set some of the values on the Gop since we
    // keep running totals of these values
    currentGop.byteLength = 0;
    currentGop.nalCount = 0;
    currentGop.duration = 0;
    currentGop.pts = frames[0].pts;
    currentGop.dts = frames[0].dts;

    // store some metadata about all the Gops
    gops.byteLength = 0;
    gops.nalCount = 0;
    gops.duration = 0;
    gops.pts = frames[0].pts;
    gops.dts = frames[0].dts;

    for (i = 0; i < frames.length; i++) {
      currentFrame = frames[i];

      if (currentFrame.keyFrame) {
        // Since the very first frame is expected to be an keyframe
        // only push to the gops array when currentGop is not empty
        if (currentGop.length) {
          gops.push(currentGop);
          gops.byteLength += currentGop.byteLength;
          gops.nalCount += currentGop.nalCount;
          gops.duration += currentGop.duration;
        }

        currentGop = [currentFrame];
        currentGop.nalCount = currentFrame.length;
        currentGop.byteLength = currentFrame.byteLength;
        currentGop.pts = currentFrame.pts;
        currentGop.dts = currentFrame.dts;
        currentGop.duration = currentFrame.duration;
      } else {
        currentGop.duration += currentFrame.duration;
        currentGop.nalCount += currentFrame.length;
        currentGop.byteLength += currentFrame.byteLength;
        currentGop.push(currentFrame);
      }
    }

    if (gops.length && currentGop.duration <= 0) {
      currentGop.duration = gops[gops.length - 1].duration;
    }
    gops.byteLength += currentGop.byteLength;
    gops.nalCount += currentGop.nalCount;
    gops.duration += currentGop.duration;

    // push the final Gop
    gops.push(currentGop);
    return gops;
  };

  // generate the track's sample table from an array of gops
  this.generateSampleTable_ = function(gops, baseDataOffset) {
    var
      h, i,
      sample,
      currentGop,
      currentFrame,
      currentSample,
      dataOffset = baseDataOffset || 0,
      samples = [];

    for (h = 0; h < gops.length; h++) {
      currentGop = gops[h];

      for (i = 0; i < currentGop.length; i++) {
        currentFrame = currentGop[i];

        sample = createDefaultSample();

        sample.dataOffset = dataOffset;
        sample.compositionTimeOffset = currentFrame.pts - currentFrame.dts;
        sample.duration = currentFrame.duration;
        sample.size = 4 * currentFrame.length; // Space for nal unit size
        sample.size += currentFrame.byteLength;

        if (currentFrame.keyFrame) {
          sample.flags.dependsOn = 2;
        }

        dataOffset += sample.size;

        samples.push(sample);
      }
    }
    return samples;
  };

  // generate the track's raw mdat data from an array of gops
  this.concatenateNalData_ = function (gops) {
    var
      h, i, j,
      currentGop,
      currentFrame,
      currentNal,
      dataOffset = 0,
      nalsByteLength = gops.byteLength,
      numberOfNals = gops.nalCount,
      totalByteLength = nalsByteLength + 4 * numberOfNals,
      data = new Uint8Array(totalByteLength),
      view = new DataView(data.buffer);

    // For each Gop..
    for (h = 0; h < gops.length; h++) {
      currentGop = gops[h];

      // For each Frame..
      for (i = 0; i < currentGop.length; i++) {
        currentFrame = currentGop[i];

        // For each NAL..
        for (j = 0; j < currentFrame.length; j++) {
          currentNal = currentFrame[j];

          view.setUint32(dataOffset, currentNal.data.byteLength);
          dataOffset += 4;
          data.set(currentNal.data, dataOffset);
          dataOffset += currentNal.data.byteLength;
        }
      }
    }
    return data;
  };
};

VideoSegmentStream.prototype = new Stream();

/**
 * Store information about the start and end of the track and the
 * duration for each frame/sample we process in order to calculate
 * the baseMediaDecodeTime
 */
collectDtsInfo = function (track, data) {
  if (typeof data.pts === 'number') {
    if (track.timelineStartInfo.pts === undefined) {
      track.timelineStartInfo.pts = data.pts;
    }

    if (track.minSegmentPts === undefined) {
      track.minSegmentPts = data.pts;
    } else {
      track.minSegmentPts = Math.min(track.minSegmentPts, data.pts);
    }

    if (track.maxSegmentPts === undefined) {
      track.maxSegmentPts = data.pts;
    } else {
      track.maxSegmentPts = Math.max(track.maxSegmentPts, data.pts);
    }
  }

  if (typeof data.dts === 'number') {
    if (track.timelineStartInfo.dts === undefined) {
      track.timelineStartInfo.dts = data.dts;
    }

    if (track.minSegmentDts === undefined) {
      track.minSegmentDts = data.dts;
    } else {
      track.minSegmentDts = Math.min(track.minSegmentDts, data.dts);
    }

    if (track.maxSegmentDts === undefined) {
      track.maxSegmentDts = data.dts;
    } else {
      track.maxSegmentDts = Math.max(track.maxSegmentDts, data.dts);
    }
  }
};

/**
 * Clear values used to calculate the baseMediaDecodeTime between
 * tracks
 */
clearDtsInfo = function (track) {
  delete track.minSegmentDts;
  delete track.maxSegmentDts;
  delete track.minSegmentPts;
  delete track.maxSegmentPts;
};

/**
 * Calculate the track's baseMediaDecodeTime based on the earliest
 * DTS the transmuxer has ever seen and the minimum DTS for the
 * current track
 */
calculateTrackBaseMediaDecodeTime = function (track) {
  var
    oneSecondInPTS = 90000, // 90kHz clock
    scale,
    // Calculate the distance, in time, that this segment starts from the start
    // of the timeline (earliest time seen since the transmuxer initialized)
    timeSinceStartOfTimeline = track.minSegmentDts - track.timelineStartInfo.dts,
    // Calculate the first sample's effective compositionTimeOffset
    firstSampleCompositionOffset = track.minSegmentPts - track.minSegmentDts;

  // track.timelineStartInfo.baseMediaDecodeTime is the location, in time, where
  // we want the start of the first segment to be placed
  track.baseMediaDecodeTime = track.timelineStartInfo.baseMediaDecodeTime;

  // Add to that the distance this segment is from the very first
  track.baseMediaDecodeTime += timeSinceStartOfTimeline;

  // Subtract this segment's "compositionTimeOffset" so that the first frame of
  // this segment is displayed exactly at the `baseMediaDecodeTime` or at the
  // end of the previous segment
  track.baseMediaDecodeTime -= firstSampleCompositionOffset;

  // baseMediaDecodeTime must not become negative
  track.baseMediaDecodeTime = Math.max(0, track.baseMediaDecodeTime);

  if (track.type === 'audio') {
    // Audio has a different clock equal to the sampling_rate so we need to
    // scale the PTS values into the clock rate of the track
    scale = track.samplerate / oneSecondInPTS;
    track.baseMediaDecodeTime *= scale;
    track.baseMediaDecodeTime = Math.floor(track.baseMediaDecodeTime);
  }
};

/**
 * A Stream that can combine multiple streams (ie. audio & video)
 * into a single output segment for MSE. Also supports audio-only
 * and video-only streams.
 */
CoalesceStream = function(options) {
  // Number of Tracks per output segment
  // If greater than 1, we combine multiple
  // tracks into a single segment
  this.numberOfTracks = 0;
  this.metadataStream = options.metadataStream;

  if (typeof options.remux !== 'undefined') {
    this.remuxTracks = !!options.remux;
  } else {
    this.remuxTracks = true;
  }

  this.pendingTracks = [];
  this.videoTrack = null;
  this.pendingBoxes = [];
  this.pendingCaptions = [];
  this.pendingMetadata = [];
  this.pendingBytes = 0;
  this.emittedTracks = 0;

  CoalesceStream.prototype.init.call(this);

  // Take output from multiple
  this.push = function(output) {
    // buffer incoming captions until the associated video segment
    // finishes
    if (output.text) {
      return this.pendingCaptions.push(output);
    }
    // buffer incoming id3 tags until the final flush
    if (output.frames) {
      return this.pendingMetadata.push(output);
    }

    // Add this track to the list of pending tracks and store
    // important information required for the construction of
    // the final segment
    this.pendingTracks.push(output.track);
    this.pendingBoxes.push(output.boxes);
    this.pendingBytes += output.boxes.byteLength;

    if (output.track.type === 'video') {
      this.videoTrack = output.track;
    }
    if (output.track.type === 'audio') {
      this.audioTrack = output.track;
    }
  };
};

CoalesceStream.prototype = new Stream();
CoalesceStream.prototype.flush = function() {
  var
    offset = 0,
    event = {
      captions: [],
      metadata: []
    },
    caption,
    id3,
    initSegment,
    timelineStartPts = 0,
    i;

  // Return until we have enough tracks from the pipeline to remux
  if (this.pendingTracks.length === 0 ||
     (this.remuxTracks && this.pendingTracks.length < this.numberOfTracks)) {
    return;
  }

  if (this.videoTrack) {
    timelineStartPts = this.videoTrack.timelineStartInfo.pts;
  } else if (this.audioTrack) {
    timelineStartPts = this.audioTrack.timelineStartInfo.pts;
  }

  if (this.pendingTracks.length === 1) {
    event.type = this.pendingTracks[0].type;
  } else {
    event.type = 'combined';
  }
  this.emittedTracks += this.pendingTracks.length;

  initSegment = mp4.initSegment(this.pendingTracks);

  // Create a new typed array large enough to hold the init
  // segment and all tracks
  event.data = new Uint8Array(initSegment.byteLength +
                              this.pendingBytes);

  // Create an init segment containing a moov
  // and track definitions
  event.data.set(initSegment);
  offset += initSegment.byteLength;

  // Append each moof+mdat (one per track) after the init segment
  for (i = 0; i < this.pendingBoxes.length; i++) {
    event.data.set(this.pendingBoxes[i], offset);
    offset += this.pendingBoxes[i].byteLength;
  }

  // Translate caption PTS times into second offsets into the
  // video timeline for the segment
  for (i = 0; i < this.pendingCaptions.length; i++) {
    caption = this.pendingCaptions[i];
    caption.startTime = (caption.startPts - timelineStartPts);
    caption.startTime /= 90e3;
    caption.endTime = (caption.endPts - timelineStartPts);
    caption.endTime /= 90e3;
    event.captions.push(caption);
  }

  // Translate ID3 frame PTS times into second offsets into the
  // video timeline for the segment
  for (i = 0; i < this.pendingMetadata.length; i++) {
    id3 = this.pendingMetadata[i];
    id3.cueTime = (id3.pts - timelineStartPts);
    id3.cueTime /= 90e3;
    event.metadata.push(id3);
  }
  // We add this to every single emitted segment even though we only need
  // it for the first
  event.metadata.dispatchType = this.metadataStream.dispatchType;

  // Reset stream state
  this.pendingTracks.length = 0;
  this.videoTrack = null;
  this.pendingBoxes.length = 0;
  this.pendingCaptions.length = 0;
  this.pendingBytes = 0;
  this.pendingMetadata.length = 0;

  // Emit the built segment
  this.trigger('data', event);

  // Only emit `done` if all tracks have been flushed and emitted
  if (this.emittedTracks >= this.numberOfTracks) {
    this.trigger('done');
    this.emittedTracks = 0;
  }
};
/**
 * A Stream that expects MP2T binary data as input and produces
 * corresponding media segments, suitable for use with Media Source
 * Extension (MSE) implementations that support the ISO BMFF byte
 * stream format, like Chrome.
 */
Transmuxer = function(options) {
  var
    self = this,
    videoTrack,
    audioTrack,
    packetStream, parseStream, elementaryStream,
    adtsStream, h264Stream,aacStream,
    videoSegmentStream, audioSegmentStream, captionStream,
    coalesceStream,
    headOfPipeline;

  this.setupAacPipeline = function() {
    this.metadataStream = new m2ts.MetadataStream();
    options.metadataStream = this.metadataStream;

    // set up the parsing pipeline
    aacStream = new AacStream();
    adtsStream = new AdtsStream();
    coalesceStream = new CoalesceStream(options);
    headOfPipeline = aacStream;

    aacStream.pipe(adtsStream);
    aacStream.pipe(this.metadataStream);
    this.metadataStream.pipe(coalesceStream);

    this.metadataStream.on('timestamp', function(frame) {
      aacStream.setTimestamp(frame.timeStamp);
    });
    this.addAacListener();
  };


  this.addAacListener = function() {
    aacStream.on('data', function(data) {
      var i;

      if (data.type === 'timed-metadata') {
        var track = {
            timelineStartInfo: {
            baseMediaDecodeTime: 0 },
            codec: 'adts',
            type: 'audio' };

        if (track && !audioSegmentStream) {
          // hook up the audio segment stream to the first track with aac data
          coalesceStream.numberOfTracks++;
          audioSegmentStream = new AudioSegmentStream(track);
          // Set up the final part of the audio pipeline
          adtsStream
            .pipe(audioSegmentStream)
            .pipe(coalesceStream);
        }
      }
    });
  };

  this.setupTsPipeline = function() {
    this.metadataStream = new m2ts.MetadataStream();

    options.metadataStream = this.metadataStream;

    // set up the parsing pipeline
    packetStream = new m2ts.TransportPacketStream();
    parseStream = new m2ts.TransportParseStream();
    elementaryStream = new m2ts.ElementaryStream();
    adtsStream = new AdtsStream();
    h264Stream = new H264Stream();
    captionStream = new m2ts.CaptionStream();
    coalesceStream = new CoalesceStream(options);
    headOfPipeline = packetStream;

    // disassemble MPEG2-TS packets into elementary streams
    packetStream
      .pipe(parseStream)
      .pipe(elementaryStream);

    // !!THIS ORDER IS IMPORTANT!!
    // demux the streams
    elementaryStream
      .pipe(h264Stream);
    elementaryStream
      .pipe(adtsStream);

    elementaryStream
      .pipe(this.metadataStream)
      .pipe(coalesceStream);

    // Hook up CEA-608/708 caption stream
    h264Stream.pipe(captionStream)
      .pipe(coalesceStream);
    this.addTsListener();
  };

  this.addTsListener = function() {

    elementaryStream.on('data', function(data) {
      var i;

      if (data.type === 'metadata') {
        i = data.tracks.length;

        // scan the tracks listed in the metadata
        while (i--) {
          if (!videoTrack && data.tracks[i].type === 'video') {
            videoTrack = data.tracks[i];
            videoTrack.timelineStartInfo.baseMediaDecodeTime = self.baseMediaDecodeTime;
          } else if (!audioTrack && data.tracks[i].type === 'audio') {
            audioTrack = data.tracks[i];
            audioTrack.timelineStartInfo.baseMediaDecodeTime = self.baseMediaDecodeTime;
          }
        }

        // hook up the video segment stream to the first track with h264 data
        if (videoTrack && !videoSegmentStream) {
          coalesceStream.numberOfTracks++;
          videoSegmentStream = new VideoSegmentStream(videoTrack);

          videoSegmentStream.on('timelineStartInfo', function(timelineStartInfo){
          // When video emits timelineStartInfo data after a flush, we forward that
          // info to the AudioSegmentStream, if it exists, because video timeline
          // data takes precedence.
            if (audioTrack) {
              audioTrack.timelineStartInfo = timelineStartInfo;
              // On the first segment we trim AAC frames that exist before the
              // very earliest DTS we have seen in video because Chrome will
              // interpret any video track with a baseMediaDecodeTime that is
              // non-zero as a gap.
              audioSegmentStream.setEarliestDts(timelineStartInfo.dts);
            }
          });

          // Set up the final part of the video pipeline
          h264Stream
            .pipe(videoSegmentStream)
            .pipe(coalesceStream);
        }

        if (audioTrack && !audioSegmentStream) {
          // hook up the audio segment stream to the first track with aac data
          coalesceStream.numberOfTracks++;
          audioSegmentStream = new AudioSegmentStream(audioTrack);

          // Set up the final part of the audio pipeline
          adtsStream
            .pipe(audioSegmentStream)
            .pipe(coalesceStream);
        }
      }
    });
  };
  Transmuxer.prototype.init.call(this);
  options = options || {};

  this.baseMediaDecodeTime = options.baseMediaDecodeTime || 0;

  // expose the metadata stream
  if (options.aacfile === undefined) {
    this.setupTsPipeline();
  } else {
    this.setupAacPipeline();
  }

  // hook up the segment streams once track metadata is delivered
  this.setBaseMediaDecodeTime = function (baseMediaDecodeTime) {
    this.baseMediaDecodeTime = baseMediaDecodeTime;
    if (audioTrack) {
      audioTrack.timelineStartInfo.dts = undefined;
      audioTrack.timelineStartInfo.pts = undefined;
      clearDtsInfo(audioTrack);
      audioTrack.timelineStartInfo.baseMediaDecodeTime = baseMediaDecodeTime;
    }
    if (videoTrack) {
      videoSegmentStream.gopCache_ = [];
      videoTrack.timelineStartInfo.dts = undefined;
      videoTrack.timelineStartInfo.pts = undefined;
      clearDtsInfo(videoTrack);
      videoTrack.timelineStartInfo.baseMediaDecodeTime = baseMediaDecodeTime;
    }
  };

  // feed incoming data to the front of the parsing pipeline
  this.push = function(data) {
    headOfPipeline.push(data);
  };

  // flush any buffered data
  this.flush = function() {
    // Start at the top of the pipeline and flush all pending work
    headOfPipeline.flush();
  };

  // Re-emit any data coming from the coalesce stream to the outside world
  coalesceStream.on('data', function (data) {
    self.trigger('data', data);
  });
  // Let the consumer know we have finished flushing the entire pipeline
  coalesceStream.on('done', function () {
    self.trigger('done');
  });
};
Transmuxer.prototype = new Stream();

module.exports = {
  Transmuxer: Transmuxer,
  VideoSegmentStream: VideoSegmentStream,
  AudioSegmentStream: AudioSegmentStream,
};