diff --git a/dom/html/HTMLMediaElement.cpp b/dom/html/HTMLMediaElement.cpp
index ab0cb924a3..23222d0ffa 100644
--- a/dom/html/HTMLMediaElement.cpp
+++ b/dom/html/HTMLMediaElement.cpp
@@ -1874,11 +1874,11 @@ HTMLMediaElement::CaptureStreamInternal(bool aFinishWhenEnded)
     if (mReadyState >= HAVE_METADATA) {
       // Expose the tracks to JS directly.
       if (HasAudio()) {
-        TrackID audioTrackId = mMediaInfo.mAudio.mTrackInfo.mOutputId;
+        TrackID audioTrackId = mMediaInfo.mAudio.mTrackId;
         out->mStream->CreateDOMTrack(audioTrackId, MediaSegment::AUDIO);
       }
       if (HasVideo()) {
-        TrackID videoTrackId = mMediaInfo.mVideo.mTrackInfo.mOutputId;
+        TrackID videoTrackId = mMediaInfo.mVideo.mTrackId;
         out->mStream->CreateDOMTrack(videoTrackId, MediaSegment::VIDEO);
       }
     }
@@ -3172,11 +3172,11 @@ void HTMLMediaElement::MetadataLoaded(const MediaInfo* aInfo,
   // Expose the tracks to JS directly.
   for (OutputMediaStream& out : mOutputStreams) {
     if (aInfo->HasAudio()) {
-      TrackID audioTrackId = aInfo->mAudio.mTrackInfo.mOutputId;
+      TrackID audioTrackId = aInfo->mAudio.mTrackId;
       out.mStream->CreateDOMTrack(audioTrackId, MediaSegment::AUDIO);
     }
     if (aInfo->HasVideo()) {
-      TrackID videoTrackId = aInfo->mVideo.mTrackInfo.mOutputId;
+      TrackID videoTrackId = aInfo->mVideo.mTrackId;
       out.mStream->CreateDOMTrack(videoTrackId, MediaSegment::VIDEO);
     }
   }
diff --git a/dom/media/MediaDecoder.cpp b/dom/media/MediaDecoder.cpp
index 239df3bb3e..3694c88f25 100644
--- a/dom/media/MediaDecoder.cpp
+++ b/dom/media/MediaDecoder.cpp
@@ -1908,7 +1908,7 @@ MediaDecoder::ConstructMediaTracks()
 
   AudioTrackList* audioList = element->AudioTracks();
   if (audioList && mInfo->HasAudio()) {
-    TrackInfo info = mInfo->mAudio.mTrackInfo;
+    const TrackInfo& info = mInfo->mAudio;
     nsRefPtr<AudioTrack> track = MediaTrackList::CreateAudioTrack(
     info.mId, info.mKind, info.mLabel, info.mLanguage, info.mEnabled);
 
@@ -1917,7 +1917,7 @@ MediaDecoder::ConstructMediaTracks()
 
   VideoTrackList* videoList = element->VideoTracks();
   if (videoList && mInfo->HasVideo()) {
-    TrackInfo info = mInfo->mVideo.mTrackInfo;
+    const TrackInfo& info = mInfo->mVideo;
     nsRefPtr<VideoTrack> track = MediaTrackList::CreateVideoTrack(
     info.mId, info.mKind, info.mLabel, info.mLanguage);
 
diff --git a/dom/media/MediaDecoderStateMachine.cpp b/dom/media/MediaDecoderStateMachine.cpp
index 3659782443..4009f1fad9 100644
--- a/dom/media/MediaDecoderStateMachine.cpp
+++ b/dom/media/MediaDecoderStateMachine.cpp
@@ -410,7 +410,7 @@ void MediaDecoderStateMachine::SendStreamData()
 
     if (!stream->mStreamInitialized) {
       if (mInfo.HasAudio()) {
-        TrackID audioTrackId = mInfo.mAudio.mTrackInfo.mOutputId;
+        TrackID audioTrackId = mInfo.mAudio.mTrackId;
         AudioSegment* audio = new AudioSegment();
         mediaStream->AddAudioTrack(audioTrackId, mInfo.mAudio.mRate, 0, audio,
                                    SourceMediaStream::ADDTRACK_QUEUED);
@@ -419,7 +419,7 @@ void MediaDecoderStateMachine::SendStreamData()
         stream->mNextAudioTime = mStartTime + stream->mInitialTime;
       }
       if (mInfo.HasVideo()) {
-        TrackID videoTrackId = mInfo.mVideo.mTrackInfo.mOutputId;
+        TrackID videoTrackId = mInfo.mVideo.mTrackId;
         VideoSegment* video = new VideoSegment();
         mediaStream->AddTrack(videoTrackId, 0, video,
                               SourceMediaStream::ADDTRACK_QUEUED);
@@ -438,7 +438,7 @@ void MediaDecoderStateMachine::SendStreamData()
 
     if (mInfo.HasAudio()) {
       MOZ_ASSERT(stream->mNextAudioTime != -1, "Should've been initialized");
-      TrackID audioTrackId = mInfo.mAudio.mTrackInfo.mOutputId;
+      TrackID audioTrackId = mInfo.mAudio.mTrackId;
       nsAutoTArray<nsRefPtr<AudioData>,10> audio;
       // It's OK to hold references to the AudioData because AudioData
       // is ref-counted.
@@ -464,7 +464,7 @@ void MediaDecoderStateMachine::SendStreamData()
 
     if (mInfo.HasVideo()) {
       MOZ_ASSERT(stream->mNextVideoTime != -1, "Should've been initialized");
-      TrackID videoTrackId = mInfo.mVideo.mTrackInfo.mOutputId;
+      TrackID videoTrackId = mInfo.mVideo.mTrackId;
       nsAutoTArray<nsRefPtr<VideoData>,10> video;
       // It's OK to hold references to the VideoData because VideoData
       // is ref-counted.
@@ -575,7 +575,7 @@ bool MediaDecoderStateMachine::HaveEnoughDecodedAudio(int64_t aAmpleAudioUSecs)
 
   if (stream && stream->mStreamInitialized && !stream->mHaveSentFinishAudio) {
     MOZ_ASSERT(mInfo.HasAudio());
-    TrackID audioTrackId = mInfo.mAudio.mTrackInfo.mOutputId;
+    TrackID audioTrackId = mInfo.mAudio.mTrackId;
     if (!stream->mStream->HaveEnoughBuffered(audioTrackId)) {
       return false;
     }
@@ -598,7 +598,7 @@ bool MediaDecoderStateMachine::HaveEnoughDecodedVideo()
 
   if (stream && stream->mStreamInitialized && !stream->mHaveSentFinishVideo) {
     MOZ_ASSERT(mInfo.HasVideo());
-    TrackID videoTrackId = mInfo.mVideo.mTrackInfo.mOutputId;
+    TrackID videoTrackId = mInfo.mVideo.mTrackId;
     if (!stream->mStream->HaveEnoughBuffered(videoTrackId)) {
       return false;
     }
diff --git a/dom/media/MediaInfo.h b/dom/media/MediaInfo.h
index 990d43eab7..8ed16ece59 100644
--- a/dom/media/MediaInfo.h
+++ b/dom/media/MediaInfo.h
@@ -6,64 +6,125 @@
 #if !defined(MediaInfo_h)
 #define MediaInfo_h
 
-#include "nsSize.h"
 #include "nsRect.h"
-#include "ImageTypes.h"
+#include "nsRefPtr.h"
+#include "nsSize.h"
 #include "nsString.h"
 #include "nsTArray.h"
+#include "ImageTypes.h"
+#include "MediaData.h"
 #include "StreamBuffer.h" // for TrackID
 
 namespace mozilla {
 
-struct TrackInfo {
-  void Init(const nsAString& aId,
+class TrackInfo {
+public:
+  enum TrackType {
+    kUndefinedTrack,
+    kAudioTrack,
+    kVideoTrack,
+    kTextTrack
+  };
+  TrackInfo(TrackType aType,
+            const nsAString& aId,
             const nsAString& aKind,
             const nsAString& aLabel,
             const nsAString& aLanguage,
             bool aEnabled,
-            TrackID aOutputId = TRACK_INVALID)
+            TrackID aTrackId = TRACK_INVALID)
+    : mId(aId)
+    , mKind(aKind)
+    , mLabel(aLabel)
+    , mLanguage(aLanguage)
+    , mEnabled(aEnabled)
+    , mTrackId(aTrackId)
+    , mDuration(0)
+    , mMediaTime(0)
+    , mType(aType)
+  {
+  }
+
+  // Only used for backward compatibility. Do not use in new code.
+  void Init(TrackType aType,
+            const nsAString& aId,
+            const nsAString& aKind,
+            const nsAString& aLabel,
+            const nsAString& aLanguage,
+            bool aEnabled,
+            TrackID aTrackId = TRACK_INVALID)
   {
     mId = aId;
     mKind = aKind;
     mLabel = aLabel;
     mLanguage = aLanguage;
     mEnabled = aEnabled;
-    mOutputId = aOutputId;
+    mTrackId = aTrackId;
+    mType = aType;
   }
 
+  // Fields common with MediaTrack object.
   nsString mId;
   nsString mKind;
   nsString mLabel;
   nsString mLanguage;
   bool mEnabled;
-  TrackID mOutputId;
+
+  TrackID mTrackId;
+
+  nsAutoCString mMimeType;
+  int64_t mDuration;
+  int64_t mMediaTime;
+  CryptoTrack mCrypto;
+
+  bool IsAudio() const
+  {
+    return mType == kAudioTrack;
+  }
+  bool IsVideo() const
+  {
+    return mType == kVideoTrack;
+  }
+  bool IsText() const
+  {
+    return mType == kTextTrack;
+  }
+  TrackType GetType() const
+  {
+    return mType;
+  }
+  bool virtual IsValid() const = 0;
+
+private:
+  TrackType mType;
 };
 
 // Stores info relevant to presenting media frames.
-class VideoInfo {
-private:
-  void Init(int32_t aWidth, int32_t aHeight, bool aHasVideo)
-  {
-    mDisplay = nsIntSize(aWidth, aHeight);
-    mStereoMode = StereoMode::MONO;
-    mHasVideo = aHasVideo;
-    mIsHardwareAccelerated = false;
-
-    // TODO: TrackInfo should be initialized by its specific codec decoder.
-    // This following call should be removed once we have that implemented.
-    mTrackInfo.Init(NS_LITERAL_STRING("2"), NS_LITERAL_STRING("main"),
-                    EmptyString(), EmptyString(), true, 2);
-  }
-
+class VideoInfo : public TrackInfo {
 public:
   VideoInfo()
+    : VideoInfo(0, 0, false)
   {
-    Init(0, 0, false);
   }
 
   VideoInfo(int32_t aWidth, int32_t aHeight)
+    : VideoInfo(aWidth, aHeight, true)
   {
-    Init(aWidth, aHeight, true);
+  }
+
+  VideoInfo(int32_t aWidth, int32_t aHeight, bool aHasVideo)
+    : TrackInfo(kVideoTrack, NS_LITERAL_STRING("2"), NS_LITERAL_STRING("main"),
+                EmptyString(), EmptyString(), true, 2)
+    , mDisplay(nsIntSize(aWidth, aHeight))
+    , mStereoMode(StereoMode::MONO)
+    , mImage(nsIntSize(aWidth, aHeight))
+    , mExtraData(new DataBuffer)
+    , mHasVideo(aHasVideo)
+  {
+  }
+
+  virtual bool IsValid() const override
+  {
+    return mDisplay.width > 0 && mDisplay.height > 0;
   }
 
   // Size in pixels at which the video is rendered. This is after it has
@@ -73,25 +134,31 @@ public:
   // Indicates the frame layout for single track stereo videos.
   StereoMode mStereoMode;
 
+  // Size in pixels of decoded video's image.
+  nsIntSize mImage;
+
+  nsRefPtr<DataBuffer> mExtraData;
+
   // True if we have an active video bitstream.
   bool mHasVideo;
 
-  TrackInfo mTrackInfo;
-
   bool mIsHardwareAccelerated;
 };
 
-class AudioInfo {
+class AudioInfo : public TrackInfo {
 public:
   AudioInfo()
-    : mRate(44100)
-    , mChannels(2)
+    : TrackInfo(kAudioTrack, NS_LITERAL_STRING("1"), NS_LITERAL_STRING("main"),
+                EmptyString(), EmptyString(), true, 1)
+    , mRate(0)
+    , mChannels(0)
+    , mBitDepth(0)
+    , mProfile(0)
+    , mExtendedProfile(0)
+    , mCodecSpecificConfig(new DataBuffer)
+    , mExtraData(new DataBuffer)
     , mHasAudio(false)
   {
-    // TODO: TrackInfo should be initialized by its specific codec decoder.
-    // This following call should be removed once we have that implemented.
-    mTrackInfo.Init(NS_LITERAL_STRING("1"), NS_LITERAL_STRING("main"),
-                    EmptyString(), EmptyString(), true, 1);
   }
 
   // Sample rate.
@@ -100,10 +167,25 @@ public:
   // Number of audio channels.
   uint32_t mChannels;
 
+  // Bits per sample.
+  uint32_t mBitDepth;
+
+  // Codec profile.
+  int8_t mProfile;
+
+  // Extended codec profile.
+  int8_t mExtendedProfile;
+
+  nsRefPtr<DataBuffer> mCodecSpecificConfig;
+  nsRefPtr<DataBuffer> mExtraData;
+
+  virtual bool IsValid() const override
+  {
+    return mChannels > 0 && mRate > 0;
+  }
+
   // True if we have an active audio bitstream.
   bool mHasAudio;
-
-  TrackInfo mTrackInfo;
 };
 
 class EncryptionInfo {
diff --git a/dom/media/ogg/OggReader.cpp b/dom/media/ogg/OggReader.cpp
index cb7b0329d3..8b18712413 100644
--- a/dom/media/ogg/OggReader.cpp
+++ b/dom/media/ogg/OggReader.cpp
@@ -111,7 +111,10 @@ static const nsString GetKind(const nsCString& aRole)
   return EmptyString();
 }
 
-static void InitTrack(MessageField* aMsgInfo, TrackInfo* aInfo, bool aEnable)
+static void InitTrack(TrackInfo::TrackType aTrackType,
+                      MessageField* aMsgInfo,
+                      TrackInfo* aInfo,
+                      bool aEnable)
 {
   MOZ_ASSERT(aMsgInfo);
   MOZ_ASSERT(aInfo);
@@ -120,7 +123,8 @@ static void InitTrack(MessageField* aMsgInfo, TrackInfo* aInfo, bool aEnable)
   nsCString* sRole = aMsgInfo->mValuesStore.Get(eRole);
   nsCString* sTitle = aMsgInfo->mValuesStore.Get(eTitle);
   nsCString* sLanguage = aMsgInfo->mValuesStore.Get(eLanguage);
-  aInfo->Init(sName? NS_ConvertUTF8toUTF16(*sName):EmptyString(),
+  aInfo->Init(aTrackType,
+              sName? NS_ConvertUTF8toUTF16(*sName):EmptyString(),
               sRole? GetKind(*sRole):EmptyString(),
               sTitle? NS_ConvertUTF8toUTF16(*sTitle):EmptyString(),
               sLanguage? NS_ConvertUTF8toUTF16(*sLanguage):EmptyString(),
@@ -322,7 +326,10 @@ void OggReader::SetupMediaTracksInfo(const nsTArray<uint32_t>& aSerials)
       }
 
       if (msgInfo) {
-        InitTrack(msgInfo, &mInfo.mVideo.mTrackInfo, mTheoraState == theoraState);
+        InitTrack(TrackInfo::kVideoTrack,
+                  msgInfo,
+                  &mInfo.mVideo,
+                  mTheoraState == theoraState);
       }
 
       nsIntRect picture = nsIntRect(theoraState->mInfo.pic_x,
@@ -343,7 +350,10 @@ void OggReader::SetupMediaTracksInfo(const nsTArray<uint32_t>& aSerials)
       }
 
       if (msgInfo) {
-        InitTrack(msgInfo, &mInfo.mAudio.mTrackInfo, mVorbisState == vorbisState);
+        InitTrack(TrackInfo::kAudioTrack,
+                  msgInfo,
+                  &mInfo.mAudio,
+                  mVorbisState == vorbisState);
       }
 
       mInfo.mAudio.mHasAudio = true;
@@ -356,7 +366,10 @@ void OggReader::SetupMediaTracksInfo(const nsTArray<uint32_t>& aSerials)
       }
 
       if (msgInfo) {
-        InitTrack(msgInfo, &mInfo.mAudio.mTrackInfo, mOpusState == opusState);
+        InitTrack(TrackInfo::kAudioTrack,
+                  msgInfo,
+                  &mInfo.mAudio,
+                  mOpusState == opusState);
       }
 
       mInfo.mAudio.mHasAudio = true;
@@ -784,7 +797,7 @@ bool OggReader::ReadOggChain()
     LOG(PR_LOG_DEBUG, ("New vorbis ogg link, serial=%d\n", mVorbisSerial));
 
     if (msgInfo) {
-      InitTrack(msgInfo, &mInfo.mAudio.mTrackInfo, true);
+      InitTrack(TrackInfo::kAudioTrack, msgInfo, &mInfo.mAudio, true);
     }
     mInfo.mAudio.mRate = newVorbisState->mInfo.rate;
     mInfo.mAudio.mChannels = newVorbisState->mInfo.channels;
@@ -800,7 +813,7 @@ bool OggReader::ReadOggChain()
     SetupTargetOpus(newOpusState);
 
     if (msgInfo) {
-      InitTrack(msgInfo, &mInfo.mAudio.mTrackInfo, true);
+      InitTrack(TrackInfo::kAudioTrack, msgInfo, &mInfo.mAudio, true);
     }
     mInfo.mAudio.mRate = newOpusState->mRate;
     mInfo.mAudio.mChannels = newOpusState->mChannels;