Parsing of encoded duration for unencrypted opus streams.

BUG=396634

Review URL: https://codereview.chromium.org/883403002

Cr-Commit-Position: refs/heads/master@{#315395}

media/BUILD.gn

@@ -554,6 +554,8 @@ test("media_unittests") {
     "formats/common/offset_byte_queue_unittest.cc",
     "formats/webm/cluster_builder.cc",
     "formats/webm/cluster_builder.h",
+    "formats/webm/opus_packet_builder.cc",
+    "formats/webm/opus_packet_builder.h",
     "formats/webm/tracks_builder.cc",
     "formats/webm/tracks_builder.h",
     "formats/webm/webm_cluster_parser_unittest.cc",

media/base/media_log.h

@@ -8,6 +8,7 @@
 #include <sstream>
 #include <string>
 
+#include "base/logging.h"
 #include "base/memory/ref_counted.h"
 #include "base/memory/scoped_ptr.h"
 #include "media/base/media_export.h"
@@ -36,6 +37,11 @@ class LogHelper {
 
 #define MEDIA_LOG(log_cb) LogHelper(log_cb).stream()
 
+// Logs only while count < max. Increments count for each log. Use LAZY_STREAM
+// to avoid wasteful evaluation of subsequent stream arguments.
+#define LIMITED_MEDIA_LOG(log_cb, count, max) \
+  LAZY_STREAM(MEDIA_LOG(log_cb), (count) < (max) && ((count)++ || true))
+
 class MEDIA_EXPORT MediaLog : public base::RefCountedThreadSafe<MediaLog> {
  public:
   // Convert various enums to strings.

media/formats/webm/opus_packet_builder.cc

+// Copyright (c) 2015 The Chromium Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style license that can be
+// found in the LICENSE file.
+
+#include "base/logging.h"
+#include "media/formats/webm/opus_packet_builder.h"
+#include "media/formats/webm/webm_cluster_parser.h"
+
+namespace media {
+
+OpusPacket::OpusPacket(uint8_t config, uint8_t frame_count, bool is_VBR) {
+  DCHECK_GE(config, 0);
+  DCHECK_LT(config, kNumPossibleOpusConfigs);
+  DCHECK_GE(frame_count, kMinOpusPacketFrameCount);
+  DCHECK_LE(frame_count, kMaxOpusPacketFrameCount);
+
+  duration_ms_ = frame_count *
+                 WebMClusterParser::kOpusFrameDurationsMu[config] /
+                 static_cast<float>(1000);
+
+  uint8_t frame_count_code;
+  uint8_t frame_count_byte;
+
+  if (frame_count == 1) {
+    frame_count_code = 0;
+  } else if (frame_count == 2) {
+    frame_count_code = is_VBR ? 2 : 1;
+  } else {
+    frame_count_code = 3;
+    frame_count_byte = (is_VBR ? 1 << 7 : 0) | frame_count;
+  }
+
+  // All opus packets must have TOC byte.
+  uint8_t opus_toc_byte = (config << 3) | frame_count_code;
+  data_.push_back(opus_toc_byte);
+
+  // For code 3 packets, the number of frames is signaled in the "frame
+  // count byte".
+  if (frame_count_code == 3) {
+    data_.push_back(frame_count_byte);
+  }
+
+  // Packet will only conform to layout specification for the TOC byte
+  // and optional frame count bytes appended above. This last byte
+  // is purely dummy padding where frame size data or encoded data might
+  // otherwise start.
+  data_.push_back(static_cast<uint8_t>(0));
+}
+
+OpusPacket::~OpusPacket() {
+}
+
+const uint8_t* OpusPacket::data() const {
+  return &(data_[0]);
+}
+
+int OpusPacket::size() const {
+  return data_.size();
+}
+
+double OpusPacket::duration_ms() const {
+  return duration_ms_;
+}
+
+ScopedVector<OpusPacket> BuildAllOpusPackets() {
+  ScopedVector<OpusPacket> opus_packets;
+
+  for (int frame_count = kMinOpusPacketFrameCount;
+       frame_count <= kMaxOpusPacketFrameCount; frame_count++) {
+    for (int opus_config_num = 0; opus_config_num < kNumPossibleOpusConfigs;
+         opus_config_num++) {
+      bool is_VBR = false;
+      opus_packets.push_back(
+          new OpusPacket(opus_config_num, frame_count, is_VBR));
+
+      if (frame_count >= 2) {
+        // Add another packet with VBR flag toggled. For frame counts >= 2,
+        // VBR triggers changes to packet framing.
+        is_VBR = true;
+        opus_packets.push_back(
+            new OpusPacket(opus_config_num, frame_count, is_VBR));
+      }
+    }
+  }
+
+  return opus_packets.Pass();
+}
+
+}  // namespace media

media/formats/webm/opus_packet_builder.h

+// Copyright (c) 2015 The Chromium Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style license that can be
+// found in the LICENSE file.
+
+#ifndef MEDIA_FORMATS_WEBM_OPUS_PACKET_BUILDER_H_
+#define MEDIA_FORMATS_WEBM_OPUS_PACKET_BUILDER_H_
+
+#include <vector>
+
+#include "base/memory/scoped_ptr.h"
+#include "base/memory/scoped_vector.h"
+
+namespace media {
+
+// From Opus RFC. See https://tools.ietf.org/html/rfc6716#page-14
+enum OpusConstants {
+  kNumPossibleOpusConfigs = 32,
+  kMinOpusPacketFrameCount = 1,
+  kMaxOpusPacketFrameCount = 48
+};
+
+class OpusPacket {
+ public:
+  OpusPacket(uint8_t config, uint8_t frame_count, bool is_VBR);
+  ~OpusPacket();
+
+  const uint8_t* data() const;
+  int size() const;
+  double duration_ms() const;
+
+ private:
+  std::vector<uint8_t> data_;
+  double duration_ms_;
+
+  DISALLOW_COPY_AND_ASSIGN(OpusPacket);
+};
+
+// Builds an exhaustive collection of Opus packet configurations.
+ScopedVector<OpusPacket> BuildAllOpusPackets();
+
+}  // namespace media
+
+#endif  // MEDIA_FORMATS_WEBM_OPUS_PACKET_BUILDER_H_

media/formats/webm/webm_cluster_parser.h

@@ -11,6 +11,7 @@
 #include <string>
 
 #include "base/memory/scoped_ptr.h"
+#include "media/base/audio_decoder_config.h"
 #include "media/base/media_export.h"
 #include "media/base/media_log.h"
 #include "media/base/stream_parser.h"
@@ -28,13 +29,17 @@ class MEDIA_EXPORT WebMClusterParser : public WebMParserClient {
 
   // Arbitrarily-chosen numbers to estimate the duration of a buffer if none is
   // set and there is not enough information to get a better estimate.
-  // TODO(wolenetz/acolwell): Parse audio codebook to determine missing audio
-  // frame durations. See http://crbug.com/351166.
   enum {
     kDefaultAudioBufferDurationInMs = 23,  // Common 1k samples @44.1kHz
     kDefaultVideoBufferDurationInMs = 42  // Low 24fps to reduce stalls
   };
 
+  // Opus packets encode the duration and other parameters in the 5 most
+  // significant bits of the first byte. The index in this array corresponds
+  // to the duration of each frame of the packet in microseconds. See
+  // https://tools.ietf.org/html/rfc6716#page-14
+  static const uint16_t kOpusFrameDurationsMu[];
+
  private:
   // Helper class that manages per-track state.
   class Track {
@@ -87,7 +92,7 @@ class MEDIA_EXPORT WebMClusterParser : public WebMParserClient {
     // block is a keyframe.
     // |data| contains the bytes in the block.
     // |size| indicates the number of bytes in |data|.
-    bool IsKeyframe(const uint8* data, int size) const;
+    bool IsKeyframe(const uint8_t* data, int size) const;
 
     base::TimeDelta default_duration() const { return default_duration_; }
 
@@ -143,6 +148,7 @@ class MEDIA_EXPORT WebMClusterParser : public WebMParserClient {
                     const std::set<int64>& ignored_tracks,
                     const std::string& audio_encryption_key_id,
                     const std::string& video_encryption_key_id,
+                    const AudioCodec audio_codec_,
                     const LogCB& log_cb);
   ~WebMClusterParser() override;
 
@@ -154,7 +160,7 @@ class MEDIA_EXPORT WebMClusterParser : public WebMParserClient {
   // Returns -1 if the parse fails.
   // Returns 0 if more data is needed.
   // Returns the number of bytes parsed on success.
-  int Parse(const uint8* buf, int size);
+  int Parse(const uint8_t* buf, int size);
 
   base::TimeDelta cluster_start_time() const { return cluster_start_time_; }
 
@@ -194,14 +200,24 @@ class MEDIA_EXPORT WebMClusterParser : public WebMParserClient {
   WebMParserClient* OnListStart(int id) override;
   bool OnListEnd(int id) override;
   bool OnUInt(int id, int64 val) override;
-  bool OnBinary(int id, const uint8* data, int size) override;
-
-  bool ParseBlock(bool is_simple_block, const uint8* buf, int size,
-                  const uint8* additional, int additional_size, int duration,
+  bool OnBinary(int id, const uint8_t* data, int size) override;
+
+  bool ParseBlock(bool is_simple_block,
+                  const uint8_t* buf,
+                  int size,
+                  const uint8_t* additional,
+                  int additional_size,
+                  int duration,
                   int64 discard_padding);
-  bool OnBlock(bool is_simple_block, int track_num, int timecode, int duration,
-               int flags, const uint8* data, int size,
-               const uint8* additional, int additional_size,
+  bool OnBlock(bool is_simple_block,
+               int track_num,
+               int timecode,
+               int duration,
+               int flags,
+               const uint8_t* data,
+               int size,
+               const uint8_t* additional,
+               int additional_size,
                int64 discard_padding);
 
   // Resets the Track objects associated with each text track.
@@ -227,21 +243,40 @@ class MEDIA_EXPORT WebMClusterParser : public WebMParserClient {
   // if that track num is not a text track.
   Track* FindTextTrack(int track_num);
 
+  // Attempts to read the duration from the encoded audio data, returning as
+  // TimeDelta or kNoTimestamp() if duration cannot be retrieved. This obviously
+  // violates layering rules, but is useful for MSE to know duration in cases
+  // where it isn't explicitly given and cannot be calculated for Blocks at the
+  // end of a Cluster (the next Cluster in playback-order may not be the next
+  // Cluster we parse, so we can't simply use the delta of the first Block in
+  // the next Cluster). Avoid calling if encrypted; may produce unexpected
+  // output. See implementation for supported codecs.
+  base::TimeDelta TryGetEncodedAudioDuration(const uint8_t* data, int size);
+
+  // Reads Opus packet header to determine packet duration. Duration returned
+  // as TimeDelta or kNoTimestamp() upon failure to read duration from packet.
+  base::TimeDelta ReadOpusDuration(const uint8_t* data, int size);
+
+  // Tracks the number of MEDIA_LOGs made in process of reading encoded
+  // duration. Useful to prevent log spam.
+  int num_duration_errors_;
+
   double timecode_multiplier_;  // Multiplier used to convert timecodes into
                                 // microseconds.
   std::set<int64> ignored_tracks_;
   std::string audio_encryption_key_id_;
   std::string video_encryption_key_id_;
+  const AudioCodec audio_codec_;
 
   WebMListParser parser_;
 
   int64 last_block_timecode_;
-  scoped_ptr<uint8[]> block_data_;
+  scoped_ptr<uint8_t[]> block_data_;
   int block_data_size_;
   int64 block_duration_;
   int64 block_add_id_;
 
-  scoped_ptr<uint8[]> block_additional_data_;
+  scoped_ptr<uint8_t[]> block_additional_data_;
   // Must be 0 if |block_additional_data_| is null. Must be > 0 if
   // |block_additional_data_| is NOT null.
   int block_additional_data_size_;

media/formats/webm/webm_stream_parser.cc

@@ -240,6 +240,7 @@ int WebMStreamParser::ParseInfoAndTracks(const uint8* data, int size) {
       tracks_parser.ignored_tracks(),
       tracks_parser.audio_encryption_key_id(),
       tracks_parser.video_encryption_key_id(),
+      audio_config.codec(),
       log_cb_));
 
   if (!init_cb_.is_null())

media/media.gyp

@@ -1257,6 +1257,8 @@
         'formats/common/offset_byte_queue_unittest.cc',
         'formats/webm/cluster_builder.cc',
         'formats/webm/cluster_builder.h',
+        'formats/webm/opus_packet_builder.cc',
+        'formats/webm/opus_packet_builder.h',
         'formats/webm/tracks_builder.cc',
         'formats/webm/tracks_builder.h',
         'formats/webm/webm_cluster_parser_unittest.cc',