Bug Fix: CMAF DASH playback works

- syntax error fixed
- removed unnecessary track id simplification

lib/cmaf.cpp

@@ -13,14 +13,6 @@ namespace CMAF{
     return payloadSize;
   }
 
-  size_t simplifiedTrackId(const DTSC::Meta & M, size_t idx) {
-    std::string type = M.getType(idx);
-    if (type == "video") {return 1;}
-    if (type == "audio") {return 2;}
-    if (type == "meta") {return 3;}
-    return idx;
-  }
-
   std::string trackHeader(const DTSC::Meta &M, size_t track, bool simplifyTrackIds){
     std::string tType = M.getType(track);
 
@@ -43,9 +35,6 @@ namespace CMAF{
     MP4::TRAK trakBox;
 
     MP4::TKHD tkhdBox(M, track);
-    if (simplifyTrackIds){
-      tkhdBox.setTrackID(simplifiedTrackId(M, track));
-    }
     tkhdBox.setDuration(0);
     trakBox.setContent(tkhdBox, 0);
 
@@ -131,9 +120,6 @@ namespace CMAF{
     }
 
     MP4::TREX trexBox(track + 1);
-    if (simplifyTrackIds){
-      trexBox.setTrackID(simplifiedTrackId(M, track));
-    }
     trexBox.setDefaultSampleDuration(1000);
     mvexBox.setContent(trexBox, M.getVod() ? 1 : 0);
 
@@ -148,7 +134,8 @@ namespace CMAF{
       MP4::SIDX sidxBox;
       sidxBox.setReferenceID(track + 1);
       sidxBox.setTimescale(1000);
-      sidxBox.setEarliestPresentationTime(keys.getTime(0) + parts.getOffset(0) - M.getFirstms(track));
+      sidxBox.setEarliestPresentationTime(keys.getTime(0) + parts.getOffset(0) -
+                                          M.getFirstms(track));
 
       for (size_t i = 0; i < fragments.getEndValid(); i++){
         size_t firstKey = fragments.getFirstKey(i);
@@ -156,9 +143,12 @@ namespace CMAF{
             ((i + 1 < fragments.getEndValid()) ? fragments.getFirstKey(i + 1) : keys.getEndValid());
 
         MP4::sidxReference refItem;
-        refItem.referencedSize = payloadSize(M, track, keys.getTime(firstKey), keys.getTime(endKey)) + keyHeaderSize(M, track, i) + 8;
+        refItem.referencedSize =
+            payloadSize(M, track, keys.getTime(firstKey), keys.getTime(endKey)) +
+            keyHeaderSize(M, track, i) + 8;
         refItem.subSegmentDuration =
-            (endKey == keys.getEndValid() ? M.getLastms(track) : keys.getTime(endKey)) - keys.getTime(firstKey);
+            (endKey == keys.getEndValid() ? M.getLastms(track) : keys.getTime(endKey)) -
+            keys.getTime(firstKey);
         refItem.sapStart = true;
         refItem.sapType = 16;
         refItem.sapDeltaTime = 0;
@@ -209,7 +199,8 @@ namespace CMAF{
   }
 
   /// Generates the 'moof' box for a DTSC::Key based CMAF fragment.
-  std::string keyHeader(const DTSC::Meta &M, size_t track, uint64_t startTime, uint64_t endTime, uint64_t segmentNum, bool simplifyTrackIds, bool UTCTime){
+  std::string keyHeader(const DTSC::Meta &M, size_t track, uint64_t startTime, uint64_t endTime,
+                        uint64_t segmentNum, bool simplifyTrackIds, bool UTCTime){
 
     size_t firstPart = M.getPartIndex(startTime, track);
     size_t endPart = M.getPartIndex(endTime, track);
@@ -218,7 +209,6 @@ namespace CMAF{
     MP4::MFHD mfhdBox(segmentNum);
     moofBox.setContent(mfhdBox, 0);
 
-
     std::set<sortPart> trunOrder;
 
     // We use keyHeaderSize here to determine the relative offsets of the data in the 'mdat' box.
@@ -242,12 +232,10 @@ namespace CMAF{
 
     tfhdBox.setFlags(MP4::tfhdSampleFlag | MP4::tfhdBaseIsMoof | MP4::tfhdSampleDesc);
     tfhdBox.setTrackID(track + 1);
-    if (simplifyTrackIds){
-      tfhdBox.setTrackID(simplifiedTrackId(M, track));
-    }
     tfhdBox.setDefaultSampleDuration(444);
     tfhdBox.setDefaultSampleSize(444);
-    tfhdBox.setDefaultSampleFlags((M.getType(track) == "video") ? (MP4::noIPicture | MP4::noKeySample)
+    tfhdBox.setDefaultSampleFlags((M.getType(track) == "video")
+                                      ? (MP4::noIPicture | MP4::noKeySample)
                                       : (MP4::isIPicture | MP4::isKeySample));
     tfhdBox.setSampleDescriptionIndex(1);
     trafBox.setContent(tfhdBox, 0);
@@ -256,7 +244,8 @@ namespace CMAF{
     if (M.getVod()){
       tfdtBox.setBaseMediaDecodeTime(startTime - M.getFirstms(track));
     }else{
-      tfdtBox.setBaseMediaDecodeTime((UTCTime ? startTime + M.getBootMsOffset() + unixBootDiff : startTime));
+      tfdtBox.setBaseMediaDecodeTime(
+          (UTCTime ? startTime + M.getBootMsOffset() + unixBootDiff : startTime));
     }
     trafBox.setContent(tfdtBox, 1);
 
@@ -277,15 +266,14 @@ namespace CMAF{
         MP4::trunSampleInformation sampleInfo;
         sampleInfo.sampleSize = parts.getSize(it->partIndex);
         sampleInfo.sampleDuration = parts.getDuration(it->partIndex);
-        if (it == lastOne){
+        if (it == lastOne){sampleInfo.sampleDuration = endTime - it->time;}
-          sampleInfo.sampleDuration = endTime - it->time;
-        }
         sampleInfo.sampleOffset = parts.getOffset(it->partIndex);
         trunBox.setSampleInformation(sampleInfo, trunOffset++);
       }
     }else{
-      WARN_MSG("Empty CMAF header for track %zu: %" PRIu64 "-%" PRIu64 " contains no packets (first: %" PRIu64
+      WARN_MSG("Empty CMAF header for track %zu: %" PRIu64 "-%" PRIu64
-               ", last: %" PRIu64 "), firstPart=%zu, lastPart=%zu",
+               " contains no packets (first: %" PRIu64 ", last: %" PRIu64
+               "), firstPart=%zu, lastPart=%zu",
                track, startTime, endTime, M.getFirstms(track), M.getLastms(track), firstPart,
                endPart);
     }

src/output/output_cmaf.cpp

@@ -489,32 +489,32 @@ namespace Mist{
   }
 
   void OutCMAF::generateSegmentlist(size_t idx, std::stringstream &s,
-                                    void callBack(uint64_t, uint64_t, std::stringstream &, bool)){
+                                    void dashSegmentCallBack(uint64_t, uint64_t,
-    DTSC::Fragments fragments(M.fragments(idx));
+                                                             std::stringstream &, bool)){
+    // NOTE: Weirdly making the 0th track as the reference track fixed everything.
+    // Looks like a nomenclature issue.
+    // TODO: Investigate with spec and refactor stuff appropriately.
+
+    size_t mainTrack = *M.getValidTracks().begin(); // M.mainTrack();
+
+    if (mainTrack == INVALID_TRACK_ID){return;}
+    DTSC::Fragments fragments(M.fragments(mainTrack));
     uint32_t firstFragment = fragments.getFirstValid();
     uint32_t lastFragment = fragments.getEndValid();
     bool first = true;
     // skip the first two fragments if live
     if (M.getLive() && (lastFragment - firstFragment) > 6){firstFragment += 2;}
 
-    if (M.getType(idx) == "audio"){
+    DTSC::Keys keys(M.keys(mainTrack));
-      uint32_t mainTrack = M.mainTrack();
-      if (mainTrack == INVALID_TRACK_ID){return;}
-      DTSC::Fragments f(M.fragments(mainTrack));
-      uint64_t firstVidTime = M.getTimeForFragmentIndex(mainTrack, f.getFirstValid());
-      firstFragment = M.getFragmentIndexForTime(idx, firstVidTime);
-    }
-
-    DTSC::Keys keys(M.keys(idx));
     for (; firstFragment < lastFragment; ++firstFragment){
       uint32_t duration = fragments.getDuration(firstFragment);
       uint64_t starttime = keys.getTime(fragments.getFirstKey(firstFragment));
       if (!duration){
         if (M.getLive()){continue;}// skip last fragment when live
-        duration = M.getLastms(idx) - starttime;
+        duration = M.getLastms(mainTrack) - starttime;
       }
-      if (M.getVod()){starttime -= M.getFirstms(idx);}
+      if (M.getVod()){starttime -= M.getFirstms(mainTrack);}
-      callBack(starttime, duration, s, first);
+      dashSegmentCallBack(starttime, duration, s, first);
       first = false;
     }
 
@@ -662,7 +662,7 @@ namespace Mist{
          it++){
       if (M.getType(it->first) == "video"){vTracks.insert(it->first);}
       if (M.getType(it->first) == "audio"){aTracks.insert(it->first);}
-      if (M.getType(it->first) == "subtitle"){sTracks.insert(it->first);}
+      if (M.getCodec(it->first) == "subtitle"){sTracks.insert(it->first);}
     }
 
     if (!vTracks.size() && !aTracks.size()){return "";}
@@ -685,12 +685,18 @@ namespace Mist{
         << "\" suggestedPresentationDelay=\"PT5.0S\" minBufferTime=\"PT2.0S\" publishTime=\""
         << Util::getUTCString(Util::epoch()) << "\" ";
     }
+
+    r << "xmlns:xsi=\"http://www.w3.org/2001/XMLSchema-instance\" ";
+    r << "xmlns:xlink=\"http://www.w3.org/1999/xlink\" ";
+    r << "xsi:schemaLocation=\"urn:mpeg:DASH:schema:MPD:2011 "
+         "http://standards.iso.org/ittf/PubliclyAvailableStandards/MPEG-DASH_schema_files/"
+         "DASH-MPD.xsd\" ";
     r << "profiles=\"urn:mpeg:dash:profile:isoff-live:2011\" "
          "xmlns=\"urn:mpeg:dash:schema:mpd:2011\" >"
       << std::endl;
     r << "<ProgramInformation><Title>" << streamName << "</Title></ProgramInformation>"
       << std::endl;
-    r << "<Period " << (M.getLive() ? "start=\"0\"" : "") << ">" << std::endl;
+    r << "<Period " << (M.getLive() ? "start=\"PT0.0S\"" : "") << ">" << std::endl;
 
     dashAdaptation(1, vTracks, videoAligned, r);
     dashAdaptation(2, aTracks, audioAligned, r);