为Chromium实现MediaConfig API - 过程分享

464次阅读  |  发布于2年以前

背景

在实现实现任何功能前,一般都需要有数据统计支撑,作为收益验证的手段以及测试 Case 收集的手段。因此,在尝试实现 HEVC 硬解前,首先需要统计有多少视频是 HEVC 视频,并拿到视频的解码参数。

很不幸,浏览器层面并未给我们暴露出视频的解码信息,比如分辨率,编码,Profile,色彩空间。这些一系列的视频参数,对于前端都是无感知且无法感知的。

因此,在往常,如果需要获取视频信息做数据统计,首先需要拿到视频的二进制数据,在前端解析视频的元数据(比如 HEVC VPS、SPS、PPS),来提取视频参数。这样做,缺点是显而易见的,这种方式并不支持原生的 Video 标签播放的内容,且需要额外二次发起请求,且需要针对每种编码均实现一遍。

因此,如果能从 Renderer 层,以浏览器视角,监听视频的解码信息,那或许我们则不需要做任何 Trick 的操作?我在这个过程尝试探索了一下,并成功实现了这个诉求。

Media Internals 实现原理

Media Internals 简介

Chrome(包括 Chromium 内核的 Edge)实现了很多用于排查问题的 Debug 页面,比如我们常用的chrome://flags等等,那么对于音视频,是否也有这样的页面?答案是,有的,它叫作“Media Internal”,在你的浏览器输入chrome://media-internal,并随意打开一个有视频播放的页面,例如 youtube。

则你会在这个页面看到视频的一些内部解码信息,以及MediaLog, 经过一番观察可以看到,红色画圈部分是我们需要的,如果我们能在视频播放时直接从在渲染进程获取到这部分信息,则我们即可轻松实现数据收集功能。

media-internal 支持所有平台,且默认开启:

src/chrome/browser/about_flags.cc

在 url 输入 chrome://media-internals,背后发生了什么?

路由判断,并加载对应模板的代码

Chrome 内部,会根据 url.host_piece()是否为 kChromeUIMediaInternalsHost 且开启了 media Internal 功能,创建 MediaInternalsUI 实例:

src/content/browser/webui/content_web_ui_controller_factory.cc

Web-UI 的主进程模块加载

在主进程创建MediaInternalsMessageHandler实例监听后,执行CreateMediaInternalsHTMLSource(),即通过由 c++代码并渲染前端展示的 html

src/content/browser/media/media-internals-ui.cc

// 入口
WebUIDataSource* CreateMediaInternalsHTMLSource() {
  // 创建一个叫`media-internal`的WebUIDataSource
  WebUIDataSource* source =
      WebUIDataSource::Create(kChromeUIMediaInternalsHost);

  // 将资源的路径添加到source
  source->UseStringsJs();
  source->AddResourcePaths(
      base::make_span(kMediaInternalsResources, kMediaInternalsResourcesSize));
  // 设置默认入口为media_internals.html
  source->SetDefaultResource(IDR_MEDIA_INTERNALS_MEDIA_INTERNALS_HTML);
  return source;
}

继续观察代码实现,可以看到,我们在 media-internals 页面拿到的内容,其实就是由这部分代码生成的:

media-internals 本质上会收集所有 renderer 进程产生的媒体信息事件(专业叫法叫做:MediaLog,Chrome media 模块单独定义的一种 log 类型)其中我们要的视频编码信息,在这里被定义为了一个叫做VideoDecoderConfig的类,因此我们从 media-internals 实现得到了一个重要信息,那就是:只要能获取到VideoDecoderConfig,就可以实现我们要的功能。

// 拿到了这个,我们就拿到了想要的一切
struct VideoDecoderConfig {
  VideoCodec codec;
  VideoCodecProfile profile;
  uint32 level;
  bool has_alpha;
  VideoTransformation transformation;
  gfx.mojom.Size coded_size;
  gfx.mojom.Rect visible_rect;
  gfx.mojom.Size natural_size;
  array<uint8> extra_data;
  EncryptionScheme encryption_scheme;
  VideoColorSpace color_space_info;
  gfx.mojom.HDRMetadata? hdr_metadata;
};

类似的,其实音频的信息,在 Chromium 内同样有一个类似的类,叫做AudioDecoderConfig。因此实现目标变为:拿到**VideoDecoderConfig**以及**AudioDecoderConfig**

媒体编码信息获取 API 实现

通过搜索 Chromium 的代码,我们发现,对于普通视频,媒体信息的获取是在主进程通过 FFMpegDemuxer 解析得到,MSE 视频则是利用 ChunkDemuxer 解析得到,并最终利用 Mojo 进行 IPC 通信传输到 Renderer 进程,OK,这说明理论上该操作是可行的。

设计 MediaConfig API

在实现前,首先需要思考,应该以何种形式拿到视频的信息,并优雅暴露到前端,这里我能想到的最好办法是直接在 HTMLMediaElement 类实现我们要的 API,这样最终不管是 video 标签还是 audio 标签,都可以拿到音视频编码信息,这里先放上最终设计好后的 IDL:

interface MediaDecoderConfig {
  // 总比特率(暂时不支持ObjectURL)
  readonly totalBitrate: number;
  // 视频流总比特率(暂时不支持ObjectURL)
  readonly videoBitrate: number;
  // 视频编码
  readonly videoCodec: string;
  // 视频配置
  readonly videoProfile: string;
  // 视频级别
  readonly videoLevel: string;
  // 视频色彩空间原色
  readonly videoColorSpacePrimaries: string;
  // 视频色彩空间传递函数
  readonly videoColorSpaceTransfer: string;
  // 视频色彩空间矩阵
  readonly videoColorSpaceMatrix: string;
  // 视频色彩空间范围
  readonly videoColorSpaceRange: string;
  // 视频编码宽度
  readonly videoCodedSizeWidth: number;
  // 视频编码高度
  readonly videoCodedSizeHeight: number;
  // 视频原始宽度
  readonly videoNaturalSizeWidth: number;
  // 视频原始高度
  readonly videoNaturalSizeHeight: number;
  // 视频可使区X点
  readonly videoVisibleRectX: number;
  // 视频可使区X点
  readonly videoVisibleRectY: number;
  // 视频可使区宽度
  readonly videoVisibleRectWidth: number;
  // 视频可使区高度
  readonly videoVisibleRectHeight: number;
  // 视频是否加密
  readonly videoEncryption: string;
  // 视频HDR元数据-亮度范围
  readonly videoHdrMetadataLuminanceRange: string;
  // 视频HDR元数据-主阵列
  readonly videoHdrMetadataPrimaries: string;
  // 视频HDR元数据-MaxContentLightLevel
  readonly videoHdrMetadataMaxContentLightLevel: number;
  // 视频HDR元数据-MaxFrameAverageLightLevel
  readonly videoHdrMetadataMaxFrameAverageLightLevel: number;
  // 视频旋转角度
  readonly videoTransformationRotation: string;
  // 视频是否翻转
  readonly videoTransformationFlipped: boolean;
  // 视频是否有Alpha通道
  readonly videoHasAlpha: boolean;
  // 视频是否有(内部)额外数据,可忽略
  readonly videoHasExtraData: boolean;
  // 音频比特率
  readonly audioBitrate: number;
  // 音频编码
  readonly audioCodec: string;
  // 音频配置
  readonly audioProfile: string;
  // 音频每通道比特数
  readonly audioBytesPerChannel: number;
  // 音频通道格式
  readonly audioChannelLayout: string;
  // 音频通道数
  readonly audioChannels: number;
  // 音频采样速率
  readonly audioSamplesPerSecond: number;
  // 音频采样格式
  readonly audioSampleFormat: string;
  // 音频每帧字节数
  readonly audioBytesPerFrame: number;
  // 音频预滚
  readonly audioSeekPreroll: string;
  // 音频延迟
  readonly audioCodecDelay: number;
  // 音频是否加密
  readonly audioEncryption: string;
  // 音频输出通道格式
  readonly audioTargetOutputChannelLayout: string;
  // 音频是否应该忽略解码器延迟
  readonly audioShouldDiscardDecoderDelay: boolean;
  // 音频是否有(内部)额外数据,可忽略
  readonly audioHasExtraData: boolean;
}

interface HTMLMediaElementEventMap extends HTMLElementEventMap {
  mediaconfigchange: Event;
}

interface HTMLMediaElement extends HTMLElement {
  // 视频配置属性,可在DOM查看
  readonly videoConfig: string;
  // 音频配置属性,可在DOM读取
  readonly audioConfig: string;
  // 当该事件触发时,可获取到视频配置
  onmediaconfigchange: ((this: HTMLMediaElement, ev: Event) => any) | null;
  // 获取音视频配置
  getMediaConfig(): MediaDecoderConfig;
}

监听音视频信息变化

通过观测发现,Chromium 在 Blink 的 WebMediaPlayerImpl 模块实现了 onMetadata 方法,当音视频信息变化时,会触发该方法,因此能实现我们要的 API 的本质实际上就是利用了 onMetadata 方法的触发

diff --git a/third_party/blink/renderer/platform/media/web_media_player_impl.cc b/third_party/blink/renderer/platform/media/web_media_player_impl.cc
index 18b3ef5fe312d..003b26e1d123f 100644
--- a/third_party/blink/renderer/platform/media/web_media_player_impl.cc
+++ b/third_party/blink/renderer/platform/media/web_media_player_impl.cc
@@ -1919,6 +1919,9 @@ void WebMediaPlayerImpl::OnMetadata(const media::PipelineMetadata& metadata) {
     media_metrics_provider_->SetHasAudio(metadata.audio_decoder_config.codec());
     RecordEncryptionScheme("Audio",
                            metadata.audio_decoder_config.encryption_scheme());
// 实现一个onAudioConfigChange方法,获取音频变化
+
+    if (client_)
+      client_->OnAudioConfigChange(metadata.audio_decoder_config);
   }

   if (HasVideo()) {
@@ -1926,6 +1929,9 @@ void WebMediaPlayerImpl::OnMetadata(const media::PipelineMetadata& metadata) {
     RecordEncryptionScheme("Video",
                            metadata.video_decoder_config.encryption_scheme());
// 实现一个OnVideoConfigChange方法,获取视频变化
+    if (client_)
+      client_->OnVideoConfigChange(metadata.video_decoder_config);
+
     if (overlay_enabled_) {
       // SurfaceView doesn't support rotated video, so transition back if
       // the video is now rotated.  If `always_enable_overlays_`, we keep the
@@ -1948,6 +1954,9 @@ void WebMediaPlayerImpl::OnMetadata(const media::PipelineMetadata& metadata) {
     }
   }

 // 实现最终的OnMediaConfigChange方法,用于记录是否含有音频或视频
+  if (client_)
+    client_->OnMediaConfigChange(HasVideo(), HasAudio());
+
   if (observer_)
     observer_->OnMetadataChanged(pipeline_metadata_);
diff --git a/third_party/blink/public/platform/web_media_player_client.h b/third_party/blink/public/platform/web_media_player_client.h
index 55e80c1e54281..c6ef56e36c614 100644
--- a/third_party/blink/public/platform/web_media_player_client.h
+++ b/third_party/blink/public/platform/web_media_player_client.h
@@ -218,6 +218,13 @@ class BLINK_PLATFORM_EXPORT WebMediaPlayerClient {
   // See https://wicg.github.io/video-rvfc/.
   virtual void OnRequestVideoFrameCallback() {}

+  // On metadata ready and video config aquired
+  virtual void OnVideoConfigChange(const media::VideoDecoderConfig& video_config) = 0;
+  // On metadata ready and audio config aquired
+  virtual void OnAudioConfigChange(const media::AudioDecoderConfig& audio_config) = 0;
+  // On meteadate ready and audio/config aquired then fire event
+  virtual void OnMediaConfigChange(bool has_video, bool has_audio) = 0;
+
   struct Features {
     WebString id;
     WebString width;

在 HTMLMediaElement 实现 API

拿到音视频信息变化后,我们需要在 HTMLMediaElement 实现我们要的方法。

diff --git a/third_party/blink/renderer/core/html/media/html_media_element.cc b/third_party/blink/renderer/core/html/media/html_media_element.cc
index cfcb19002a16a..cde5b8b8cc4ac 100644
--- a/third_party/blink/renderer/core/html/media/html_media_element.cc
+++ b/third_party/blink/renderer/core/html/media/html_media_element.cc
@@ -29,6 +29,7 @@
 #include <algorithm>
 #include <limits>

+#include <string>
 #include "base/auto_reset.h"
 #include "base/debug/crash_logging.h"
 #include "base/feature_list.h"
@@ -56,6 +57,7 @@
 #include "third_party/blink/public/platform/web_media_player_source.h"
 #include "third_party/blink/renderer/bindings/core/v8/script_controller.h"
 #include "third_party/blink/renderer/bindings/core/v8/script_promise_resolver.h"
+#include "third_party/blink/renderer/bindings/core/v8/v8_binding_for_core.h"
 #include "third_party/blink/renderer/core/core_initializer.h"
 #include "third_party/blink/renderer/core/core_probes_inl.h"
 #include "third_party/blink/renderer/core/css/media_list.h"
@@ -83,6 +85,7 @@
 #include "third_party/blink/renderer/core/html/media/autoplay_policy.h"
 #include "third_party/blink/renderer/core/html/media/html_media_element_controls_list.h"
 #include "third_party/blink/renderer/core/html/media/media_controls.h"
+#include "third_party/blink/renderer/core/html/media/media_decoder_config.h"
 #include "third_party/blink/renderer/core/html/media/media_error.h"
 #include "third_party/blink/renderer/core/html/media/media_fragment_uri_parser.h"
 #include "third_party/blink/renderer/core/html/media/media_source_attachment.h"
@@ -1539,6 +1542,24 @@ void HTMLMediaElement::DidAudioOutputSinkChanged(
     observer->OnAudioOutputSinkChanged(hashed_device_id);
 }

+void HTMLMediaElement::OnMediaConfigChange(bool has_video, bool has_audio) {
// 记录是否有音、视频配置
+  has_video_ = has_video;
+  has_audio_ = has_audio;
// 在HTMLMediaElement上Dispatch一个事件,通知前端
+  ScheduleEvent(event_type_names::kMediaconfigchange);
+}
+
+void HTMLMediaElement::OnVideoConfigChange(const media::VideoDecoderConfig& config) {
// 存储VideoDecoderConfig
+  video_config_ = config;
+  const std::string video_config = config.AsHumanReadableString();
// 在HTMLMediaElement上设置属性
+  setAttribute(html_names::kVideoconfigAttr, AtomicString(video_config.c_str()));
+}
+
+void HTMLMediaElement::OnAudioConfigChange(const media::AudioDecoderConfig& config) {
// 存储AudioDecoderConfig
+  audio_config_ = config;
+  const std::string audio_config = config.AsHumanReadableString();
// 在HTMLMediaElement上设置属性
+  setAttribute(html_names::kAudioconfigAttr, AtomicString(audio_config.c_str()));
+}
+
 void HTMLMediaElement::SetMediaPlayerHostForTesting(
     mojo::PendingAssociatedRemote<media::mojom::blink::MediaPlayerHost> host) {
   media_player_host_remote_->Value().Bind(
@@ -2528,6 +2549,115 @@ bool HTMLMediaElement::Autoplay() const {
   return FastHasAttribute(html_names::kAutoplayAttr);
 }

// 为HTMLMediaElement实现的videoconfig属性,只读
+String HTMLMediaElement::videoConfig() const {
+  LOG(INFO) << __FUNCTION__ << "getVideoConfig";
+  return FastGetAttribute(html_names::kVideoconfigAttr).GetString();
+}
+
+ // 为HTMLMediaElement实现的videoconfig属性,只读
+String HTMLMediaElement::audioConfig() const {
+  LOG(INFO) << __FUNCTION__ << "getAudioConfig";
+  return FastGetAttribute(html_names::kAudioconfigAttr).GetString();
+}
+ // 这里是我们最终要实现的getMediaConfig API的实现函数,大体逻辑即
+ // 将音视频配置转换为Blink IDL支持的类型
+MediaDecoderConfig* HTMLMediaElement::getMediaConfig() {
+  MediaDecoderConfig* config = MakeGarbageCollected<MediaDecoderConfig>();
+
+  if (has_video_) {
+    const int total_bitrate = video_config_.GetHumanReadableTotalBitrate();
+    const int video_bitrate = video_config_.GetHumanReadableVideoBitrate();
+    const AtomicString video_codec = AtomicString(video_config_.GetHumanReadableCodecName().c_str());
+    const AtomicString video_profile = AtomicString(video_config_.GetHumanReadableProfile().c_str());
+    const AtomicString video_level = AtomicString(video_config_.GetHumanReadableLevel().c_str());
+    const AtomicString video_color_space_primaries = AtomicString(video_config_.GetHumanReadableColorSpacePrimaries().c_str());
+    const AtomicString video_color_space_transfer = AtomicString(video_config_.GetHumanReadableColorSpaceTransfer().c_str());
+    const AtomicString video_color_space_matrix = AtomicString(video_config_.GetHumanReadableColorSpaceMatrix().c_str());
+    const AtomicString video_color_space_range = AtomicString(video_config_.GetHumanReadableColorSpaceRange().c_str());
+    const float video_coded_size_width = video_config_.GetHumanReadableCodedSizeWidth();
+    const float video_coded_size_height = video_config_.GetHumanReadableCodedSizeHeight();
+    const float video_natural_size_width = video_config_.GetHumanReadableNaturalSizeWidth();
+    const float video_natural_size_height = video_config_.GetHumanReadableNaturalSizeHeight();
+    const float video_visible_rect_x = video_config_.GetHumanReadableVisibleRectX();
+    const float video_visible_rect_y = video_config_.GetHumanReadableVisibleRectY();
+    const float video_visible_rect_width = video_config_.GetHumanReadableVisibleRectWidth();
+    const float video_visible_rect_height = video_config_.GetHumanReadableVisibleRectHeight();
+    const AtomicString video_encryption = AtomicString(video_config_.GetHumanReadableEncryption().c_str());
+    const AtomicString video_hdr_metadata_luminance_range = AtomicString(video_config_.GetHumanReadableHDRMetadataLuminanceRange().c_str());
+    const AtomicString video_hdr_metadata_primaries = AtomicString(video_config_.GetHumanReadableHDRMetadataPrimaries().c_str());
+    const unsigned video_hdr_metadata_max_content_light_level = video_config_.GetHumanReadableHDRMetadataMaxContentLightLevel();
+    const unsigned video_hdr_metadata_max_frame_average_light_level = video_config_.GetHumanReadableHDRMetadataMaxFrameAverageLightLevel();
+    const AtomicString video_transformation_rotation = AtomicString(video_config_.GetHumanReadableTransformationRotation().c_str());
+    const bool video_transformation_flipped = video_config_.GetHumanReadableTransformationFlipped();
+    const bool video_has_alpha = video_config_.GetHumanReadableHasAlpha();
+    const bool video_has_extra_data = video_config_.GetHumanReadableHasExtraData();
+
+    config->SetVideoConfig(total_bitrate,
+                           video_bitrate,
+                           video_codec,
+                           video_profile,
+                           video_level,
+                           video_color_space_primaries,
+                           video_color_space_transfer,
+                           video_color_space_matrix,
+                           video_color_space_range,
+                           video_coded_size_width,
+                           video_coded_size_height,
+                           video_natural_size_width,
+                           video_natural_size_height,
+                           video_visible_rect_x,
+                           video_visible_rect_y,
+                           video_visible_rect_width,
+                           video_visible_rect_height,
+                           video_encryption,
+                           video_hdr_metadata_luminance_range,
+                           video_hdr_metadata_primaries,
+                           video_hdr_metadata_max_content_light_level,
+                           video_hdr_metadata_max_frame_average_light_level,
+                           video_transformation_rotation,
+                           video_transformation_flipped,
+                           video_has_alpha,
+                           video_has_extra_data);
+  }
+
+  if (has_audio_) {
+    const int total_bitrate = audio_config_.GetHumanReadableTotalBitrate();
+    const int audio_bitrate = audio_config_.GetHumanReadableAudioBitrate();
、   // 知识点,std:string需要转换为AtomicString,才可暴露给前端
+    const AtomicString audio_codec = AtomicString(audio_config_.GetHumanReadableCodecName().c_str());
+    const AtomicString audio_profile = AtomicString(audio_config_.GetHumanReadableProfile().c_str());
+    const int audio_bytes_per_channel = audio_config_.GetHumanReadableBytesPerChannel();
+    const AtomicString audio_channel_layout = AtomicString(audio_config_.GetHumanReadableChannelLayout().c_str());
+    const int audio_channels = audio_config_.GetHumanReadableChannels();
+    const int audio_samples_per_second = audio_config_.GetHumanReadableSamplesPerSecond();
+    const AtomicString audio_sample_format = AtomicString(audio_config_.GetHumanReadableSampleFormat().c_str());
+    const int audio_bytes_per_frame = audio_config_.GetHumanReadableBytesPerFrame();
+    const AtomicString audio_seek_preroll = AtomicString(audio_config_.GetHumanReadableSeekPreroll().c_str());
+    const int audio_codec_delay = audio_config_.GetHumanReadableCodecDelay();
+    const AtomicString audio_encryption = AtomicString(audio_config_.GetHumanReadableEncryption().c_str());
+    const AtomicString audio_target_output_channel_layout = AtomicString(audio_config_.GetHumanReadableTargetOutputChannelLayout().c_str());
+    const bool audio_should_discard_decoder_delay = audio_config_.GetHumanReadableShouldDiscardDecoderDelay();
+    const bool audio_has_extra_data = audio_config_.GetHumanReadableHasExtraData();
+
+    config->SetAudioConfig(total_bitrate,
+                           audio_bitrate,
+                           audio_codec,
+                           audio_profile,
+                           audio_bytes_per_channel,
+                           audio_channel_layout,
+                           audio_channels,
+                           audio_samples_per_second,
+                           audio_sample_format,
+                           audio_bytes_per_frame,
+                           audio_seek_preroll,
+                           audio_codec_delay,
+                           audio_encryption,
+                           audio_target_output_channel_layout,
+                           audio_should_discard_decoder_delay,
+                           audio_has_extra_data);
+  }
+
+  return config;
+}
+
 String HTMLMediaElement::preload() const {
   if (GetLoadType() == WebMediaPlayer::kLoadTypeMediaStream)
     return PreloadTypeToString(WebMediaPlayer::kPreloadNone);
diff --git a/third_party/blink/renderer/core/html/media/html_media_element.h b/third_party/blink/renderer/core/html/media/html_media_element.h
index 4af2fe7c1c04b..eb5d1956196c2 100644
--- a/third_party/blink/renderer/core/html/media/html_media_element.h
+++ b/third_party/blink/renderer/core/html/media/html_media_element.h
@@ -31,6 +31,7 @@

 #include "base/time/time.h"
 #include "base/timer/elapsed_timer.h"
+#include "media/base/audio_decoder.h"
 #include "media/mojo/mojom/media_player.mojom-blink.h"
 #include "third_party/abseil-cpp/absl/types/optional.h"
 #include "third_party/blink/public/common/media/display_type.h"
@@ -56,6 +57,7 @@
 #include "third_party/blink/renderer/platform/timer.h"
 #include "third_party/blink/renderer/platform/weborigin/kurl.h"
 #include "third_party/blink/renderer/platform/wtf/threading_primitives.h"
+#include "third_party/blink/renderer/core/html/media/media_decoder_config.h"

 namespace cc {
 class Layer;
@@ -93,6 +95,7 @@ class VideoTrack;
 class VideoTrackList;
 class WebInbandTextTrack;
 class WebRemotePlaybackClient;
+class MediaDecoderConfig;

 class CORE_EXPORT HTMLMediaElement
     : public HTMLElement,
@@ -190,6 +193,10 @@ class CORE_EXPORT HTMLMediaElement
   String EffectivePreload() const;
   WebMediaPlayer::Preload EffectivePreloadType() const;

+  String videoConfig() const;
+  String audioConfig() const;
+  MediaDecoderConfig* getMediaConfig();
+
   WebTimeRanges BufferedInternal() const;
   TimeRanges* buffered() const;
   void load();
@@ -254,6 +261,11 @@ class CORE_EXPORT HTMLMediaElement
   void SetUserWantsControlsVisible(bool visible);
   bool UserWantsControlsVisible() const;

+  void OnMediaConfigChange(bool has_video, bool has_audio) override;
+  void OnVideoConfigChange(const media::VideoDecoderConfig& video_config) override;
+  void OnAudioConfigChange(const media::AudioDecoderConfig& audio_config) override;
+
   void TogglePlayState();

   AudioTrackList& audioTracks();
@@ -663,6 +675,11 @@ class CORE_EXPORT HTMLMediaElement
   KURL current_src_after_redirects_;
   Member<MediaStreamDescriptor> src_object_;

+  bool has_video_ = false;
+  bool has_audio_ = false;
+  media::VideoDecoderConfig video_config_;
+  media::AudioDecoderConfig audio_config_;
+
   // To prevent potential regression when extended by the MSE API, do not set
   // |error_| outside of constructor and SetError().
   Member<MediaError> error_;

接着需要为我们新增的 API 定义 IDL:

diff --git a/third_party/blink/renderer/core/html/media/html_media_element.idl b/third_party/blink/renderer/core/html/media/html_media_element.idl
index a041c328f1dec..ed2d93a8f403f 100644
--- a/third_party/blink/renderer/core/html/media/html_media_element.idl
+++ b/third_party/blink/renderer/core/html/media/html_media_element.idl
@@ -46,6 +46,9 @@ enum CanPlayTypeResult { "" /* empty string */, "maybe", "probably" };
     const unsigned short NETWORK_NO_SOURCE = 3;
     [ImplementedAs=getNetworkState] readonly attribute unsigned short networkState;
     [CEReactions] attribute DOMString preload;
+    readonly attribute DOMString videoConfig;
+    readonly attribute DOMString audioConfig;
+    MediaDecoderConfig getMediaConfig();
     readonly attribute TimeRanges buffered;
     void load();
     [CallWith=ExecutionContext, HighEntropy, Measure] CanPlayTypeResult canPlayType(DOMString type);

计算音视频的比特率

由于VideoDecoderConfig缺失了比特率这个重要的参数,为了收集到更详细的视频信息,我们还需要在 Demux 阶段计算一下音视频轨道的比特率。

diff --git a/third_party/blink/public/platform/web_media_player.h b/third_party/blink/public/platform/web_media_player.h
index 306f13ae8af00..da703e6e7bf1d 100644
--- a/third_party/blink/public/platform/web_media_player.h
+++ b/third_party/blink/public/platform/web_media_player.h
@@ -186,6 +186,8 @@ class WebMediaPlayer {
   virtual void FlingingStarted() {}
   virtual void FlingingStopped() {}
   virtual void SetPreload(Preload) {}
+  virtual void SetVideoConfig(const AtomicString&) {}
+  virtual void SetAudioConfig(const AtomicString&) {}
   virtual WebTimeRanges Buffered() const = 0;
   virtual WebTimeRanges Seekable() const = 0;
diff --git a/media/filters/ffmpeg_demuxer.h b/media/filters/ffmpeg_demuxer.h
index 8d4c66e7611ac..ec8893131bb66 100644
--- a/media/filters/ffmpeg_demuxer.h
+++ b/media/filters/ffmpeg_demuxer.h
@@ -123,6 +123,9 @@ class MEDIA_EXPORT FFmpegDemuxerStream : public DemuxerStream {
   AudioDecoderConfig audio_decoder_config() override;
   VideoDecoderConfig video_decoder_config() override;

+  void SetAudioBitrate(int audio_bitrate, int bitrate);
+  void SetVideoBitrate(int video_bitrate, int bitrate);
+
   bool IsEnabled() const;
   void SetEnabled(bool enabled, base::TimeDelta timestamp);

由于音视频可能有多个流,因此需要将流的比特率加和。

diff --git a/media/filters/ffmpeg_demuxer.cc b/media/filters/ffmpeg_demuxer.cc
index a76be9a01c5be..47ded334541aa 100644
--- a/media/filters/ffmpeg_demuxer.cc
+++ b/media/filters/ffmpeg_demuxer.cc
@@ -830,6 +830,20 @@ VideoDecoderConfig FFmpegDemuxerStream::video_decoder_config() {
   return *video_config_;
 }

 // 记录音频总比特率和音视频总比特率
+void FFmpegDemuxerStream::SetAudioBitrate(int audio_bitrate, int bitrate) {
+  if (audio_config_->IsValidConfig()) {
+    audio_config_->SetAudioStreamBitrate(audio_bitrate);
+    audio_config_->SetTotalStreamBitrate(bitrate);
+  }
+}
+ // 记录视频总比特率和音视频总比特率
+void FFmpegDemuxerStream::SetVideoBitrate(int video_bitrate, int bitrate) {
+  if (video_config_->IsValidConfig()) {
+    video_config_->SetVideoStreamBitrate(video_bitrate);
+    video_config_->SetTotalStreamBitrate(bitrate);
+  }
+}
+
 bool FFmpegDemuxerStream::IsEnabled() const {
   DCHECK(task_runner_->RunsTasksInCurrentSequence());
   return is_enabled_;
@@ -1236,6 +1250,40 @@ static int CalculateBitrate(AVFormatContext* format_context,
   return base::ClampRound(filesize_in_bytes * duration.ToHz() * 8);
 }
 // 计算音频总比特率的逻辑
+static int CalculateAudioBitrate(AVFormatContext* format_context) {
+  int audio_stream_bitrate = 0;
+
+  for (size_t i = 0; i < format_context->nb_streams; ++i) {
+    AVCodecParameters* codec_parameters = format_context->streams[i]->codecpar;
+    AVMediaType codec_type = codec_parameters->codec_type;
+    if (codec_type == AVMEDIA_TYPE_AUDIO) {
+      audio_stream_bitrate += codec_parameters->bit_rate;
+    }
+  }
+
+  if (audio_stream_bitrate > 0)
+    return audio_stream_bitrate;
+
+  return 0;
+}
+ // 计算视频总比特率的逻辑
+static int CalculateVideoBitrate(AVFormatContext* format_context) {
+  int video_stream_bitrate = 0;
+
+  for (size_t i = 0; i < format_context->nb_streams; ++i) {
+    AVCodecParameters* codec_parameters = format_context->streams[i]->codecpar;
+    AVMediaType codec_type = codec_parameters->codec_type;
+    if (codec_type == AVMEDIA_TYPE_VIDEO) {
+      video_stream_bitrate += codec_parameters->bit_rate;
+    }
+  }
+
+  if (video_stream_bitrate > 0)
+    return video_stream_bitrate;
+
+  return 0;
+}
+
 void FFmpegDemuxer::OnOpenContextDone(bool result) {
   DCHECK(task_runner_->RunsTasksInCurrentSequence());
   if (stopped_) {
@@ -1563,6 +1611,20 @@ void FFmpegDemuxer::OnFindStreamInfoDone(int result) {
   if (bitrate_ > 0)
     data_source_->SetBitrate(bitrate_);

+  int video_stream_bitrate = CalculateVideoBitrate(format_context);
+  int audio_stream_bitrate = CalculateAudioBitrate(format_context);
+  int total_stream_bitrate = video_stream_bitrate + audio_stream_bitrate;
+
+  for (const auto& stream : streams_) {
+    if (stream) {
+      if (stream->type() == DemuxerStream::AUDIO) {
+        stream->SetAudioBitrate(audio_stream_bitrate, total_stream_bitrate);
+      } else if (stream->type() == DemuxerStream::VIDEO) {
+        stream->SetVideoBitrate(video_stream_bitrate, total_stream_bitrate);
+      }
+    }
+  }
+
   LogMetadata(format_context, max_duration);
   media_tracks_updated_cb_.Run(std::move(media_tracks));

扩展 VideoDecoderConfig

我们需要将VideoDecoderConfig的属性提取成我们需要的类型,且能支持我们保存比特率,因此这里需要的额外扩展一下 DecoderConfig:

diff --git a/media/base/video_decoder_config.cc b/media/base/video_decoder_config.cc
index 40c9707fae4d8..d1b311003b975 100644
--- a/media/base/video_decoder_config.cc
+++ b/media/base/video_decoder_config.cc
// 将我们要的参数转换为我们要的格式
@@ -106,6 +106,11 @@ std::string VideoDecoderConfig::AsHumanReadableString() const {
     << ", rotation: " << VideoRotationToString(video_transformation().rotation)
     << ", flipped: " << video_transformation().mirrored
     << ", color space: " << color_space_info().ToGfxColorSpace().ToString();
+
+  if (totalStreamBitrate() > 0) {
+    s << ", total_bitrate: " << base::NumberToString(totalStreamBitrate())
+      << ", video_bitrate: " << base::NumberToString(videoStreamBitrate());
+  }

   if (hdr_metadata().has_value()) {
     s << std::setprecision(4) << ", luminance range: "
@@ -129,10 +134,158 @@ std::string VideoDecoderConfig::AsHumanReadableString() const {
   return s.str();
 }

+int VideoDecoderConfig::GetHumanReadableTotalBitrate() const {
+  return totalStreamBitrate();
+}
+
+int VideoDecoderConfig::GetHumanReadableVideoBitrate() const {
+  return videoStreamBitrate();
+}
+
 std::string VideoDecoderConfig::GetHumanReadableCodecName() const {
   return GetCodecName(codec());
 }

+std::string VideoDecoderConfig::GetHumanReadableProfile() const {
+  return GetProfileName(profile());
+}
+
+std::string VideoDecoderConfig::GetHumanReadableLevel() const {
+  return level() > kNoVideoCodecLevel ? base::NumberToString(level()) : "not available";
+}
+
+std::string VideoDecoderConfig::GetHumanReadableColorSpacePrimaries() const {
+  return color_space_info().ToGfxColorSpace().GetPrimariesString();
+}
+
+std::string VideoDecoderConfig::GetHumanReadableColorSpaceTransfer() const {
+  return color_space_info().ToGfxColorSpace().GetTransferString();
+}
+
+std::string VideoDecoderConfig::GetHumanReadableColorSpaceMatrix() const {
+  return color_space_info().ToGfxColorSpace().GetMatrixString();
+}
+
+std::string VideoDecoderConfig::GetHumanReadableColorSpaceRange() const {
+  return color_space_info().ToGfxColorSpace().GetRangeString();
+}
+
+float VideoDecoderConfig::GetHumanReadableCodedSizeWidth() const {
+  return coded_size().width();
+}
+
+float VideoDecoderConfig::GetHumanReadableCodedSizeHeight() const {
+  return coded_size().height();
+}
+
+float VideoDecoderConfig::GetHumanReadableNaturalSizeWidth() const {
+  return natural_size().width();
+}
+
+float VideoDecoderConfig::GetHumanReadableNaturalSizeHeight() const {
+  return natural_size().height();
+}
+
+float VideoDecoderConfig::GetHumanReadableVisibleRectX() const {
+  return visible_rect().x();
+}
+
+float VideoDecoderConfig::GetHumanReadableVisibleRectY() const {
+  return visible_rect().y();
+}
+
+float VideoDecoderConfig::GetHumanReadableVisibleRectWidth() const {
+  return visible_rect().width();
+}
+
+float VideoDecoderConfig::GetHumanReadableVisibleRectHeight() const {
+  return visible_rect().height();
+}
+
+bool VideoDecoderConfig::GetHumanReadableHasAlpha() const {
+  return alpha_mode() == AlphaMode::kHasAlpha ? true : false;
+}
+
+bool VideoDecoderConfig::GetHumanReadableHasExtraData() const {
+  return extra_data().empty() ? false : true;
+}
+
+bool VideoDecoderConfig::GetHumanReadableTransformationFlipped() const {
+  return video_transformation().mirrored;
+}
+
+std::string VideoDecoderConfig::GetHumanReadableTransformationRotation() const {
+  return VideoRotationToString(video_transformation().rotation);
+}
+
+std::string VideoDecoderConfig::GetHumanReadableEncryption() const {
+  std::ostringstream s;
+  s << encryption_scheme();
+  return s.str();
+}
+
+std::string VideoDecoderConfig::GetHumanReadableHDRMetadataLuminanceRange() const {
+  if (hdr_metadata().has_value()) {
+    std::ostringstream s;
+    s << hdr_metadata()->color_volume_metadata.luminance_min << "-"
+      << hdr_metadata()->color_volume_metadata.luminance_max;
+    return s.str();
+  } else {
+    return "";
+  }
+}
+
+std::string VideoDecoderConfig::GetHumanReadableHDRMetadataPrimaries() const {
+  if (hdr_metadata().has_value()) {
+    std::ostringstream s;
+    s << std::setprecision(4)
+      << "r("
+      << hdr_metadata()->color_volume_metadata.primary_r.x()
+      << ","
+      << hdr_metadata()->color_volume_metadata.primary_r.y()
+      << ") g("
+      << hdr_metadata()->color_volume_metadata.primary_g.x()
+      << ","
+      << hdr_metadata()->color_volume_metadata.primary_g.y()
+      << ") b("
+      << hdr_metadata()->color_volume_metadata.primary_b.x()
+      << ","
+      << hdr_metadata()->color_volume_metadata.primary_b.y()
+      << ") wp("
+      << hdr_metadata()->color_volume_metadata.white_point.x()
+      << ","
+      << hdr_metadata()->color_volume_metadata.white_point.y()
+      << ")";
+    return s.str();
+  } else {
+    return "";
+  }
+}
+
+unsigned VideoDecoderConfig::GetHumanReadableHDRMetadataMaxContentLightLevel() const {
+  if (hdr_metadata().has_value()) {
+    return hdr_metadata()->max_content_light_level;
+  } else {
+    return 0;
+  }
+}
+
+unsigned VideoDecoderConfig::GetHumanReadableHDRMetadataMaxFrameAverageLightLevel() const {
+  if (hdr_metadata().has_value()) {
+    return hdr_metadata()->max_frame_average_light_level;
+  } else {
+    return 0;
+  }
+}
+
+void VideoDecoderConfig::SetTotalStreamBitrate(int total_stream_bitrate) {
+  total_stream_bitrate_ = total_stream_bitrate;
+}
+
+void VideoDecoderConfig::SetVideoStreamBitrate(int video_stream_bitrate) {
+  video_stream_bitrate_ = video_stream_bitrate;
+}
+
 void VideoDecoderConfig::SetExtraData(const std::vector<uint8_t>& extra_data) {
   extra_data_ = extra_data;
 }
diff --git a/media/base/video_decoder_config.h b/media/base/video_decoder_config.h
index 1eb9bf07c46b7..59f0b1de48e87 100644
--- a/media/base/video_decoder_config.h
+++ b/media/base/video_decoder_config.h
@@ -74,7 +74,32 @@ class MEDIA_EXPORT VideoDecoderConfig {
   // Returns a human-readable string describing |*this|.
   std::string AsHumanReadableString() const;

+  int GetHumanReadableTotalBitrate() const;
+  int GetHumanReadableVideoBitrate() const;
   std::string GetHumanReadableCodecName() const;
+  std::string GetHumanReadableProfile() const;
+  std::string GetHumanReadableLevel() const;
+  std::string GetHumanReadableColorSpacePrimaries() const;
+  std::string GetHumanReadableColorSpaceTransfer() const;
+  std::string GetHumanReadableColorSpaceMatrix() const;
+  std::string GetHumanReadableColorSpaceRange() const;
+  float GetHumanReadableCodedSizeWidth() const;
+  float GetHumanReadableCodedSizeHeight() const;
+  float GetHumanReadableNaturalSizeWidth() const;
+  float GetHumanReadableNaturalSizeHeight() const;
+  float GetHumanReadableVisibleRectX() const;
+  float GetHumanReadableVisibleRectY() const;
+  float GetHumanReadableVisibleRectWidth() const;
+  float GetHumanReadableVisibleRectHeight() const;
+  bool GetHumanReadableHasAlpha() const;
+  bool GetHumanReadableHasExtraData() const;
+  bool GetHumanReadableTransformationFlipped() const;
+  std::string GetHumanReadableTransformationRotation() const;
+  std::string GetHumanReadableEncryption() const;
+  std::string GetHumanReadableHDRMetadataLuminanceRange() const;
+  std::string GetHumanReadableHDRMetadataPrimaries() const;
+  unsigned GetHumanReadableHDRMetadataMaxContentLightLevel() const;
+  unsigned GetHumanReadableHDRMetadataMaxFrameAverageLightLevel() const;

   VideoCodec codec() const { return codec_; }
   VideoCodecProfile profile() const { return profile_; }
@@ -155,6 +180,12 @@ class MEDIA_EXPORT VideoDecoderConfig {
   // useful for decryptors that decrypts an encrypted stream to a clear stream.
   void SetIsEncrypted(bool is_encrypted);

+  // Sets/Gets the audio + video / video bitrate
+  int totalStreamBitrate() const { return total_stream_bitrate_; }
+  void SetTotalStreamBitrate(int total_stream_bitrate);
+  int videoStreamBitrate() const { return video_stream_bitrate_; }
+  void SetVideoStreamBitrate(int video_stream_bitrate);
+
   // Sets whether this config is for WebRTC or not.
   void set_is_rtc(bool is_rtc) { is_rtc_ = is_rtc; }
   bool is_rtc() const { return is_rtc_; }
@@ -186,6 +217,9 @@ class MEDIA_EXPORT VideoDecoderConfig {
   absl::optional<gfx::HDRMetadata> hdr_metadata_;
   bool is_rtc_ = false;

+  int total_stream_bitrate_ = 0;
+  int video_stream_bitrate_ = 0;
+
   // Not using DISALLOW_COPY_AND_ASSIGN here intentionally to allow the compiler
   // generated copy constructor and assignment operator. Since the extra data is
   // typically small, the performance impact is minimal.

扩展 AudioDecoderConfig

我们需要将AudioDecoderConfig的属性提取成我们需要的类型,且能支持我们保存比特率,因此这里需要的额外扩展一下 DecoderConfig:

diff --git a/media/base/audio_decoder_config.cc b/media/base/audio_decoder_config.cc
index 1fcbb4119aa35..7c5d9056a5b6d 100644
--- a/media/base/audio_decoder_config.cc
+++ b/media/base/audio_decoder_config.cc
@@ -5,6 +5,7 @@
 #include "media/base/audio_decoder_config.h"

 #include "base/logging.h"
+#include "base/strings/string_number_conversions.h"
 #include "media/base/limits.h"
 #include "media/base/media_util.h"

@@ -99,9 +100,91 @@ std::string AudioDecoderConfig::AsHumanReadableString() const {
     << (should_discard_decoder_delay() ? "true" : "false")
     << ", target_output_channel_layout: "
     << ChannelLayoutToString(target_output_channel_layout());
+
+  if (totalStreamBitrate() > 0) {
+    s << ", total_bitrate: " << base::NumberToString(totalStreamBitrate())
+      << ", audio_bitrate: " << base::NumberToString(audioStreamBitrate());
+  }
+
+  return s.str();
+}
+
+int AudioDecoderConfig::GetHumanReadableTotalBitrate() const {
+  return totalStreamBitrate();
+}
+
+int AudioDecoderConfig::GetHumanReadableAudioBitrate() const {
+  return audioStreamBitrate();
+}
+
+std::string AudioDecoderConfig::GetHumanReadableCodecName() const {
+  return GetCodecName(codec());
+}
+
+std::string AudioDecoderConfig::GetHumanReadableProfile() const {
+  return GetProfileName(profile());
+}
+
+int AudioDecoderConfig::GetHumanReadableBytesPerChannel() const {
+  return bytes_per_channel();
+}
+
+std::string AudioDecoderConfig::GetHumanReadableChannelLayout() const {
+  return ChannelLayoutToString(channel_layout());
+}
+
+int AudioDecoderConfig::GetHumanReadableChannels() const {
+  return channels();
+}
+
+int AudioDecoderConfig::GetHumanReadableSamplesPerSecond() const {
+  return samples_per_second();
+}
+
+std::string AudioDecoderConfig::GetHumanReadableSampleFormat() const {
+  return SampleFormatToString(sample_format());
+}
+
+int AudioDecoderConfig::GetHumanReadableBytesPerFrame() const {
+  return bytes_per_frame();
+}
+
+std::string AudioDecoderConfig::GetHumanReadableSeekPreroll() const {
+  std::ostringstream s;
+  s << seek_preroll().InMicroseconds() << "us";
+  return s.str();
+}
+
+int AudioDecoderConfig::GetHumanReadableCodecDelay() const {
+  return codec_delay();
+}
+
+std::string AudioDecoderConfig::GetHumanReadableEncryption() const {
+  std::ostringstream s;
+  s << encryption_scheme();
   return s.str();
 }

+std::string AudioDecoderConfig::GetHumanReadableTargetOutputChannelLayout() const {
+  return ChannelLayoutToString(target_output_channel_layout());
+}
+
+bool AudioDecoderConfig::GetHumanReadableHasExtraData() const {
+  return extra_data().empty() ? false : true;
+}
+
+bool AudioDecoderConfig::GetHumanReadableShouldDiscardDecoderDelay() const {
+  return should_discard_decoder_delay() ? true : false;
+}
+
+void AudioDecoderConfig::SetTotalStreamBitrate(int total_stream_bitrate) {
+  total_stream_bitrate_ = total_stream_bitrate;
+}
+
+void AudioDecoderConfig::SetAudioStreamBitrate(int audio_stream_bitrate) {
+  audio_stream_bitrate_ = audio_stream_bitrate;
+}
+
 void AudioDecoderConfig::SetChannelsForDiscrete(int channels) {
   DCHECK(channel_layout_ == CHANNEL_LAYOUT_DISCRETE ||
          channels == ChannelLayoutToChannelCount(channel_layout_));
diff --git a/media/base/audio_decoder_config.h b/media/base/audio_decoder_config.h
index 1ec4ed073b995..c3f3a984a7735 100644
--- a/media/base/audio_decoder_config.h
+++ b/media/base/audio_decoder_config.h
@@ -62,6 +62,30 @@ class MEDIA_EXPORT AudioDecoderConfig {
   // Returns a human-readable string describing |*this|.
   std::string AsHumanReadableString() const;

+  // Returns a human-readable string for media decoder config
+  int GetHumanReadableTotalBitrate() const;
+  int GetHumanReadableAudioBitrate() const;
+  std::string GetHumanReadableCodecName() const;
+  std::string GetHumanReadableProfile() const;
+  int GetHumanReadableBytesPerChannel() const;
+  std::string GetHumanReadableChannelLayout() const;
+  int GetHumanReadableChannels() const;
+  int GetHumanReadableSamplesPerSecond() const;
+  std::string GetHumanReadableSampleFormat() const;
+  int GetHumanReadableBytesPerFrame() const;
+  std::string GetHumanReadableSeekPreroll() const;
+  int GetHumanReadableCodecDelay() const;
+  std::string GetHumanReadableEncryption() const;
+  std::string GetHumanReadableTargetOutputChannelLayout() const;
+  bool GetHumanReadableShouldDiscardDecoderDelay() const;
+  bool GetHumanReadableHasExtraData() const;
+
+  // Sets/Gets the video + audio bitrate and audio bitrate
+  int totalStreamBitrate() const { return total_stream_bitrate_; }
+  void SetTotalStreamBitrate(int total_stream_bitrate);
+  int audioStreamBitrate() const { return audio_stream_bitrate_; }
+  void SetAudioStreamBitrate(int audio_stream_bitrate);
+
   // Sets the number of channels if |channel_layout_| is CHANNEL_LAYOUT_DISCRETE
   void SetChannelsForDiscrete(int channels);

@@ -161,6 +185,10 @@ class MEDIA_EXPORT AudioDecoderConfig {
   // be manually set in `SetChannelsForDiscrete()`;
   int channels_ = 0;

+  // The bitrate of audio + video / audio
+  int total_stream_bitrate_ = 0;
+  int audio_stream_bitrate_ = 0;
+
   // Not using DISALLOW_COPY_AND_ASSIGN here intentionally to allow the compiler
   // generated copy constructor and assignment operator. Since the extra data is
   // typically small, the performance impact is minimal.

扩展 ColorSpace

我们需要将ColorSpace的属性提取成我们需要的类型,因此这里需要的额外扩展一下:

diff --git a/ui/gfx/color_space.cc b/ui/gfx/color_space.cc
index 02a4feec4481c..9d693abbf82fe 100644
--- a/ui/gfx/color_space.cc
+++ b/ui/gfx/color_space.cc
@@ -567,6 +567,151 @@ std::string ColorSpace::ToString() const {
   return ss.str();
 }

+std::string ColorSpace::GetPrimariesString() const {
+  std::stringstream ss;
+  ss << std::fixed << std::setprecision(4);
+  switch (primaries_) {
+    PRINT_ENUM_CASE(PrimaryID, INVALID)
+    PRINT_ENUM_CASE(PrimaryID, BT709)
+    PRINT_ENUM_CASE(PrimaryID, BT470M)
+    PRINT_ENUM_CASE(PrimaryID, BT470BG)
+    PRINT_ENUM_CASE(PrimaryID, SMPTE170M)
+    PRINT_ENUM_CASE(PrimaryID, SMPTE240M)
+    PRINT_ENUM_CASE(PrimaryID, FILM)
+    PRINT_ENUM_CASE(PrimaryID, BT2020)
+    PRINT_ENUM_CASE(PrimaryID, SMPTEST428_1)
+    PRINT_ENUM_CASE(PrimaryID, SMPTEST431_2)
+    PRINT_ENUM_CASE(PrimaryID, SMPTEST432_1)
+    PRINT_ENUM_CASE(PrimaryID, XYZ_D50)
+    PRINT_ENUM_CASE(PrimaryID, ADOBE_RGB)
+    PRINT_ENUM_CASE(PrimaryID, APPLE_GENERIC_RGB)
+    PRINT_ENUM_CASE(PrimaryID, WIDE_GAMUT_COLOR_SPIN)
+    case PrimaryID::CUSTOM:
+      // |custom_primary_matrix_| is in row-major order.
+      const float sum_R = custom_primary_matrix_[0] +
+                          custom_primary_matrix_[3] + custom_primary_matrix_[6];
+      const float sum_G = custom_primary_matrix_[1] +
+                          custom_primary_matrix_[4] + custom_primary_matrix_[7];
+      const float sum_B = custom_primary_matrix_[2] +
+                          custom_primary_matrix_[5] + custom_primary_matrix_[8];
+      if (IsAlmostZero(sum_R) || IsAlmostZero(sum_G) || IsAlmostZero(sum_B))
+        break;
+
+      ss << "[[" << (custom_primary_matrix_[0] / sum_R)
+        << ", " << (custom_primary_matrix_[3] / sum_R) << "], "
+        << " [" << (custom_primary_matrix_[1] / sum_G) << ", "
+        << (custom_primary_matrix_[4] / sum_G) << "], "
+        << " [" << (custom_primary_matrix_[2] / sum_B) << ", "
+        << (custom_primary_matrix_[5] / sum_B) << "]]";
+      break;
+  }
+
+  return ss.str();
+}
+
+std::string ColorSpace::GetTransferString() const {
+  std::stringstream ss;
+  ss << std::fixed << std::setprecision(4);
+  switch (transfer_) {
+    PRINT_ENUM_CASE(TransferID, INVALID)
+    PRINT_ENUM_CASE(TransferID, BT709)
+    PRINT_ENUM_CASE(TransferID, BT709_APPLE)
+    PRINT_ENUM_CASE(TransferID, GAMMA18)
+    PRINT_ENUM_CASE(TransferID, GAMMA22)
+    PRINT_ENUM_CASE(TransferID, GAMMA24)
+    PRINT_ENUM_CASE(TransferID, GAMMA28)
+    PRINT_ENUM_CASE(TransferID, SMPTE170M)
+    PRINT_ENUM_CASE(TransferID, SMPTE240M)
+    PRINT_ENUM_CASE(TransferID, LINEAR)
+    PRINT_ENUM_CASE(TransferID, LOG)
+    PRINT_ENUM_CASE(TransferID, LOG_SQRT)
+    PRINT_ENUM_CASE(TransferID, IEC61966_2_4)
+    PRINT_ENUM_CASE(TransferID, BT1361_ECG)
+    PRINT_ENUM_CASE(TransferID, IEC61966_2_1)
+    PRINT_ENUM_CASE(TransferID, BT2020_10)
+    PRINT_ENUM_CASE(TransferID, BT2020_12)
+    PRINT_ENUM_CASE(TransferID, SMPTEST428_1)
+    PRINT_ENUM_CASE(TransferID, IEC61966_2_1_HDR)
+    PRINT_ENUM_CASE(TransferID, LINEAR_HDR)
+    case TransferID::ARIB_STD_B67:
+      ss << "HLG (SDR white point ";
+      if (transfer_params_[0] == 0.f)
+        ss << "default " << kDefaultSDRWhiteLevel;
+      else
+        ss << transfer_params_[0];
+      ss << " nits)";
+      break;
+    case TransferID::SMPTEST2084:
+      ss << "PQ (SDR white point ";
+      if (transfer_params_[0] == 0.f)
+        ss << "default " << kDefaultSDRWhiteLevel;
+      else
+        ss << transfer_params_[0];
+      ss << " nits)";
+      break;
+    case TransferID::CUSTOM: {
+      skcms_TransferFunction fn;
+      GetTransferFunction(&fn);
+      ss << fn.c << "*x + " << fn.f << " if x < " << fn.d << " else (" << fn.a
+         << "*x + " << fn.b << ")**" << fn.g << " + " << fn.e;
+      break;
+    }
+    case TransferID::CUSTOM_HDR: {
+      skcms_TransferFunction fn;
+      GetTransferFunction(&fn);
+      if (fn.g == 1.0f && fn.a > 0.0f && fn.b == 0.0f && fn.c == 0.0f &&
+          fn.d == 0.0f && fn.e == 0.0f && fn.f == 0.0f) {
+        ss << "LINEAR_HDR (slope " << fn.a << ", SDR white point "
+           << kDefaultScrgbLinearSdrWhiteLevel / fn.a << " nits)";
+        break;
+      }
+      ss << fn.c << "*x + " << fn.f << " if |x| < " << fn.d << " else sign(x)*("
+         << fn.a << "*|x| + " << fn.b << ")**" << fn.g << " + " << fn.e;
+      break;
+    }
+    case TransferID::PIECEWISE_HDR: {
+      skcms_TransferFunction fn;
+      GetTransferFunction(&fn);
+      ss << "sRGB to 1 at " << transfer_params_[0] << ", linear to "
+         << transfer_params_[1] << " at 1";
+      break;
+    }
+  }
+  return ss.str();
+}
+
+std::string ColorSpace::GetMatrixString() const {
+  std::stringstream ss;
+  ss << std::fixed << std::setprecision(4);
+  switch (matrix_) {
+    PRINT_ENUM_CASE(MatrixID, INVALID)
+    PRINT_ENUM_CASE(MatrixID, RGB)
+    PRINT_ENUM_CASE(MatrixID, BT709)
+    PRINT_ENUM_CASE(MatrixID, FCC)
+    PRINT_ENUM_CASE(MatrixID, BT470BG)
+    PRINT_ENUM_CASE(MatrixID, SMPTE170M)
+    PRINT_ENUM_CASE(MatrixID, SMPTE240M)
+    PRINT_ENUM_CASE(MatrixID, YCOCG)
+    PRINT_ENUM_CASE(MatrixID, BT2020_NCL)
+    PRINT_ENUM_CASE(MatrixID, BT2020_CL)
+    PRINT_ENUM_CASE(MatrixID, YDZDX)
+    PRINT_ENUM_CASE(MatrixID, GBR)
+  }
+  return ss.str();
+}
+
+std::string ColorSpace::GetRangeString() const {
+  std::stringstream ss;
+  ss << std::fixed << std::setprecision(4);
+  switch (range_) {
+    PRINT_ENUM_CASE(RangeID, INVALID)
+    PRINT_ENUM_CASE(RangeID, LIMITED)
+    PRINT_ENUM_CASE(RangeID, FULL)
+    PRINT_ENUM_CASE(RangeID, DERIVED)
+  }
+  return ss.str();
+}
+
 #undef PRINT_ENUM_CASE

 ColorSpace ColorSpace::GetAsFullRangeRGB() const {
diff --git a/ui/gfx/color_space.h b/ui/gfx/color_space.h
index 4c3f068c86877..aa4482b3c258e 100644
--- a/ui/gfx/color_space.h
+++ b/ui/gfx/color_space.h
@@ -262,6 +262,10 @@ class COLOR_SPACE_EXPORT ColorSpace {
   bool operator<(const ColorSpace& other) const;
   size_t GetHash() const;
   std::string ToString() const;
+  std::string GetPrimariesString() const;
+  std::string GetTransferString() const;
+  std::string GetMatrixString() const;
+  std::string GetRangeString() const;

   bool IsWide() const;

定义 MediaDecoderConfig IDL

这里没什么内容,大部分时候就是都是字符串的搬运工,实现我们要的数据类型MediaDecoderConfig数据类型,即VideoDecoderConfig以及AudioDecoderConfig的合集。

+++ b/third_party/blink/renderer/core/html/media/media_decoder_config.cc
@@ -0,0 +1,173 @@
+
+#include "third_party/blink/renderer/core/html/media/media_decoder_config.h"
+
+// #include "third_party/blink/renderer/core/core_export.h"
+#include "third_party/blink/renderer/core/dom/document.h"
+#include "third_party/blink/renderer/core/frame/local_dom_window.h"
+#include "third_party/blink/renderer/core/timing/dom_window_performance.h"
+#include "third_party/blink/renderer/core/timing/window_performance.h"
+
+namespace blink {
+
+MediaDecoderConfig::MediaDecoderConfig() :
+      total_bitrate_(0),
+      video_bitrate_(0),
+      video_codec_(""),
+      video_profile_(""),
+      video_level_(""),
+      video_color_space_primaries_(""),
+      video_color_space_transfer_(""),
+      video_color_space_matrix_(""),
+      video_color_space_range_(""),
+      video_coded_size_width_(0),
+      video_coded_size_height_(0),
+      video_natural_size_width_(0),
+      video_natural_size_height_(0),
+      video_visible_rect_x_(0),
+      video_visible_rect_y_(0),
+      video_visible_rect_width_(0),
+      video_visible_rect_height_(0),
+      video_encryption_(""),
+      video_hdr_metadata_luminance_range_(""),
+      video_hdr_metadata_primaries_(""),
+      video_hdr_metadata_max_content_light_level_(0),
+      video_hdr_metadata_max_frame_average_light_level_(0),
+      video_transformation_rotation_(""),
+      video_transformation_flipped_(false),
+      video_has_alpha_(false),
+      video_has_extra_data_(false),
+      audio_bitrate_(0),
+      audio_codec_(""),
+      audio_profile_(""),
+      audio_bytes_per_channel_(0),
+      audio_channel_layout_(""),
+      audio_channels_(0),
+      audio_samples_per_second_(0),
+      audio_sample_format_(""),
+      audio_bytes_per_frame_(0),
+      audio_seek_preroll_(""),
+      audio_codec_delay_(0),
+      audio_encryption_(""),
+      audio_target_output_channel_layout_(""),
+      audio_should_discard_decoder_delay_(false),
+      audio_has_extra_data_(false) {}
+
+void MediaDecoderConfig::SetVideoConfig(int total_bitrate,
+                                        int video_bitrate,
+                                        AtomicString video_codec,
+                                        AtomicString video_profile,
+                                        AtomicString video_level,
+                                        AtomicString video_color_space_primaries,
+                                        AtomicString video_color_space_transfer,
+                                        AtomicString video_color_space_matrix,
+                                        AtomicString video_color_space_range,
+                                        float video_coded_size_width,
+                                        float video_coded_size_height,
+                                        float video_natural_size_width,
+                                        float video_natural_size_height,
+                                        float video_visible_rect_x,
+                                        float video_visible_rect_y,
+                                        float video_visible_rect_width,
+                                        float video_visible_rect_height,
+                                        AtomicString video_encryption,
+                                        AtomicString video_hdr_metadata_luminance_range,
+                                        AtomicString video_hdr_metadata_primaries,
+                                        unsigned video_hdr_metadata_max_content_light_level,
+                                        unsigned video_hdr_metadata_max_frame_average_light_level,
+                                        AtomicString video_transformation_rotation,
+                                        bool video_transformation_flipped,
+                                        bool video_has_alpha,
+                                        bool video_has_extra_data)
+  {
+      total_bitrate_ = total_bitrate;
+      video_bitrate_ = video_bitrate;
+      video_codec_ = video_codec;
+      video_profile_ = video_profile;
+      video_level_ = video_level;
+      video_color_space_primaries_ = video_color_space_primaries;
+      video_color_space_transfer_ = video_color_space_transfer;
+      video_color_space_matrix_ = video_color_space_matrix;
+      video_color_space_range_ = video_color_space_range;
+      video_coded_size_width_ = video_coded_size_width;
+      video_coded_size_height_ = video_coded_size_height;
+      video_natural_size_width_ = video_natural_size_width;
+      video_natural_size_height_ = video_natural_size_height;
+      video_visible_rect_x_ = video_visible_rect_x;
+      video_visible_rect_y_ = video_visible_rect_y;
+      video_visible_rect_width_ = video_visible_rect_width;
+      video_visible_rect_height_ = video_visible_rect_height;
+      video_encryption_ = video_encryption;
+      video_hdr_metadata_luminance_range_ = video_hdr_metadata_luminance_range;
+      video_hdr_metadata_primaries_ = video_hdr_metadata_primaries;
+      video_hdr_metadata_max_content_light_level_ = video_hdr_metadata_max_content_light_level;
+      video_hdr_metadata_max_frame_average_light_level_ = video_hdr_metadata_max_frame_average_light_level;
+      video_transformation_rotation_ = video_transformation_rotation;
+      video_transformation_flipped_ = video_transformation_flipped;
+      video_has_alpha_ = video_has_alpha;
+      video_has_extra_data_ = video_has_extra_data;
+  }
+
+void MediaDecoderConfig::SetAudioConfig(int total_bitrate,
+                                        int audio_bitrate,
+                                        AtomicString audio_codec,
+                                        AtomicString audio_profile,
+                                        int audio_bytes_per_channel,
+                                        AtomicString audio_channel_layout,
+                                        int audio_channels,
+                                        int audio_samples_per_second,
+                                        AtomicString audio_sample_format,
+                                        int audio_bytes_per_frame,
+                                        AtomicString audio_seek_preroll,
+                                        int audio_codec_delay,
+                                        AtomicString audio_encryption,
+                                        AtomicString audio_target_output_channel_layout,
+                                        bool audio_should_discard_decoder_delay,
+                                        bool audio_has_extra_data)
+  {
+      total_bitrate_ = total_bitrate;
+      audio_bitrate_ = audio_bitrate;
+      audio_codec_ = audio_codec;
+      audio_profile_ = audio_profile;
+      audio_bytes_per_channel_ = audio_bytes_per_channel;
+      audio_channel_layout_ = audio_channel_layout;
+      audio_channels_ = audio_channels;
+      audio_samples_per_second_ = audio_samples_per_second;
+      audio_sample_format_ = audio_sample_format;
+      audio_bytes_per_frame_ = audio_bytes_per_frame;
+      audio_seek_preroll_ = audio_seek_preroll;
+      audio_codec_delay_ = audio_codec_delay;
+      audio_encryption_ = audio_encryption;
+      audio_target_output_channel_layout_ = audio_target_output_channel_layout;
+      audio_should_discard_decoder_delay_ = audio_should_discard_decoder_delay;
+      audio_has_extra_data_ = audio_has_extra_data;
+  }
+}  // namespace blink
diff --git a/third_party/blink/renderer/core/html/media/media_decoder_config.h b/third_party/blink/renderer/core/html/media/media_decoder_config.h
new file mode 100644
index 0000000000000..942d209700909
--- /dev/null
+++ b/third_party/blink/renderer/core/html/media/media_decoder_config.h
@@ -0,0 +1,187 @@
+
+#ifndef THIRD_PARTY_BLINK_RENDERER_MODULES_MEDIA_DECODER_CONFIG_H_
+#define THIRD_PARTY_BLINK_RENDERER_MODULES_MEDIA_DECODER_CONFIG_H_
+
+#include "third_party/blink/renderer/platform/bindings/script_wrappable.h"
+#include "third_party/blink/renderer/platform/heap/handle.h"
+#include "third_party/blink/renderer/platform/wtf/text/wtf_string.h"
+
+namespace blink {
+
+class Document;
+
+class MediaDecoderConfig : public ScriptWrappable {
+  DEFINE_WRAPPERTYPEINFO();
+
+ public:
+  MediaDecoderConfig();
+  void SetVideoConfig(int total_bitrate,
+                      int video_bitrate,
+                      AtomicString video_codec,
+                      AtomicString video_profile,
+                      AtomicString video_level,
+                      AtomicString video_color_space_primaries,
+                      AtomicString video_color_space_transfer,
+                      AtomicString video_color_space_matrix,
+                      AtomicString video_color_space_range,
+                      float video_coded_size_width,
+                      float video_coded_size_height,
+                      float video_natural_size_width,
+                      float video_natural_size_height,
+                      float video_visible_rect_x,
+                      float video_visible_rect_y,
+                      float video_visible_rect_width,
+                      float video_visible_rect_height,
+                      AtomicString video_encryption,
+                      AtomicString video_hdr_metadata_luminance_range,
+                      AtomicString video_hdr_metadata_primaries,
+                      unsigned video_hdr_metadata_max_content_light_level,
+                      unsigned video_hdr_metadata_max_frame_average_light_level,
+                      AtomicString video_transformation_rotation,
+                      bool video_transformation_flipped,
+                      bool video_has_alpha,
+                      bool video_has_extra_data);
+
+  void SetAudioConfig(int total_bitrate,
+                      int audio_bitrate,
+                      AtomicString audio_codec,
+                      AtomicString audio_profile,
+                      int audio_bytes_per_channel,
+                      AtomicString audio_channel_layout,
+                      int audio_channels,
+                      int audio_samples_per_second,
+                      AtomicString audio_sample_format,
+                      int audio_bytes_per_frame,
+                      AtomicString audio_seek_preroll,
+                      int audio_codec_delay,
+                      AtomicString audio_encryption,
+                      AtomicString audio_target_output_channel_layout,
+                      bool audio_should_discard_decoder_delay,
+                      bool audio_has_extra_data);
+
+  int totalBitrate() const { return total_bitrate_; }
+  // video config
+  int videoBitrate() const { return video_bitrate_; }
+  AtomicString videoCodec() const { return video_codec_; }
+  AtomicString videoProfile() const { return video_profile_; }
+  AtomicString videoLevel() const { return video_level_; }
+  AtomicString videoColorSpacePrimaries() const { return video_color_space_primaries_; }
+  AtomicString videoColorSpaceTransfer() const { return video_color_space_transfer_; }
+  AtomicString videoColorSpaceMatrix() const { return video_color_space_matrix_; }
+  AtomicString videoColorSpaceRange() const { return video_color_space_range_; }
+  float videoCodedSizeWidth() const { return video_coded_size_width_; }
+  float videoCodedSizeHeight() const { return video_coded_size_height_; }
+  float videoNaturalSizeWidth() const { return video_natural_size_width_; }
+  float videoNaturalSizeHeight() const { return video_natural_size_height_; }
+  float videoVisibleRectX() const { return video_visible_rect_x_; }
+  float videoVisibleRectY() const { return video_visible_rect_y_; }
+  float videoVisibleRectWidth() const { return video_visible_rect_width_; }
+  float videoVisibleRectHeight() const { return video_visible_rect_height_; }
+  AtomicString videoEncryption() const { return video_encryption_; }
+  AtomicString videoHdrMetadataLuminanceRange() const { return video_hdr_metadata_luminance_range_; }
+  AtomicString videoHdrMetadataPrimaries() const { return video_hdr_metadata_primaries_; }
+  unsigned videoHdrMetadataMaxContentLightLevel() const { return video_hdr_metadata_max_content_light_level_; }
+  unsigned videoHdrMetadataMaxFrameAverageLightLevel() const { return video_hdr_metadata_max_frame_average_light_level_; }
+  AtomicString videoTransformationRotation() const { return video_transformation_rotation_; }
+  bool videoTransformationFlipped() const { return video_transformation_flipped_; }
+  bool videoHasAlpha() const { return video_has_alpha_; }
+  bool videoHasExtraData() const { return video_has_extra_data_; }
+
+
+  // audio config
+  int audioBitrate() const { return audio_bitrate_; }
+  AtomicString audioCodec() const { return audio_codec_; }
+  AtomicString audioProfile() const { return audio_profile_; }
+  int audioBytesPerChannel() const { return audio_bytes_per_channel_; }
+  AtomicString audioChannelLayout() const { return audio_channel_layout_; }
+  int audioChannels() const { return audio_channels_; }
+  int audioSamplesPerSecond() const { return audio_samples_per_second_; }
+  AtomicString audioSampleFormat() const { return audio_sample_format_; }
+  int audioBytesPerFrame() const { return audio_bytes_per_frame_; }
+  AtomicString audioSeekPreroll() const { return audio_seek_preroll_; }
+  int audioCodecDelay() const { return audio_codec_delay_; }
+  AtomicString audioEncryption() const { return audio_encryption_; }
+  AtomicString audioTargetOutputChannelLayout() const { return audio_target_output_channel_layout_; }
+  bool audioShouldDiscardDecoderDelay() const { return audio_should_discard_decoder_delay_; }
+  bool audioHasExtraData() const { return audio_has_extra_data_; }
+
+ private:
+  int total_bitrate_;
+
+  // video
+  int video_bitrate_;
+  AtomicString video_codec_;
+  AtomicString video_profile_;
+  AtomicString video_level_;
+  AtomicString video_color_space_primaries_;
+  AtomicString video_color_space_transfer_;
+  AtomicString video_color_space_matrix_;
+  AtomicString video_color_space_range_;
+  float video_coded_size_width_;
+  float video_coded_size_height_;
+  float video_natural_size_width_;
+  float video_natural_size_height_;
+  float video_visible_rect_x_;
+  float video_visible_rect_y_;
+  float video_visible_rect_width_;
+  float video_visible_rect_height_;
+  AtomicString video_encryption_;
+  AtomicString video_hdr_metadata_luminance_range_;
+  AtomicString video_hdr_metadata_primaries_;
+  unsigned video_hdr_metadata_max_content_light_level_;
+  unsigned video_hdr_metadata_max_frame_average_light_level_;
+  AtomicString video_transformation_rotation_;
+  bool video_transformation_flipped_;
+  bool video_has_alpha_;
+  bool video_has_extra_data_;
+
+  // audio
+  int audio_bitrate_;
+  AtomicString audio_codec_;
+  AtomicString audio_profile_;
+  int audio_bytes_per_channel_;
+  AtomicString audio_channel_layout_;
+  int audio_channels_;
+  int audio_samples_per_second_;
+  AtomicString audio_sample_format_;
+  int audio_bytes_per_frame_;
+  AtomicString audio_seek_preroll_;
+  int audio_codec_delay_;
+  AtomicString audio_encryption_;
+  AtomicString audio_target_output_channel_layout_;
+  bool audio_should_discard_decoder_delay_;
+  bool audio_has_extra_data_;
+};
+
+}  // namespace blink
+
+#endif  // THIRD_PARTY_BLINK_RENDERER_MODULES_MEDIA_DECODER_CONFIG_H_

在 blink 实现任何数据结构,都需要定义好 IDL,IDL 的数据结构即前端访问时实际获取到的数据类型,这里后面会单独写个文档给大家介绍 IDL 的一些定义属性及背后的含义。

+++ b/third_party/blink/renderer/core/html/media/media_decoder_config.idl
@@ -0,0 +1,46 @@
+[Exposed=Window]
+interface MediaDecoderConfig {
+    readonly attribute unsigned long totalBitrate;
+
+    readonly attribute unsigned long videoBitrate;
+    readonly attribute DOMString videoCodec;
+    readonly attribute DOMString videoProfile;
+    readonly attribute DOMString videoLevel;
+    readonly attribute DOMString videoColorSpacePrimaries;
+    readonly attribute DOMString videoColorSpaceTransfer;
+    readonly attribute DOMString videoColorSpaceMatrix;
+    readonly attribute DOMString videoColorSpaceRange;
+    readonly attribute float videoCodedSizeWidth;
+    readonly attribute float videoCodedSizeHeight;
+    readonly attribute float videoNaturalSizeWidth;
+    readonly attribute float videoNaturalSizeHeight;
+    readonly attribute float videoVisibleRectX;
+    readonly attribute float videoVisibleRectY;
+    readonly attribute float videoVisibleRectWidth;
+    readonly attribute float videoVisibleRectHeight;
+    readonly attribute DOMString videoEncryption;
+    readonly attribute DOMString videoHdrMetadataLuminanceRange;
+    readonly attribute DOMString videoHdrMetadataPrimaries;
+    readonly attribute unsigned long videoHdrMetadataMaxContentLightLevel;
+    readonly attribute unsigned long videoHdrMetadataMaxFrameAverageLightLevel;
+    readonly attribute DOMString videoTransformationRotation;
+    readonly attribute boolean videoTransformationFlipped;
+    readonly attribute boolean videoHasAlpha;
+    readonly attribute boolean videoHasExtraData;
+
+    readonly attribute unsigned long audioBitrate;
+    readonly attribute DOMString audioCodec;
+    readonly attribute DOMString audioProfile;
+    readonly attribute unsigned long audioBytesPerChannel;
+    readonly attribute DOMString audioChannelLayout;
+    readonly attribute unsigned long audioChannels;
+    readonly attribute unsigned long audioSamplesPerSecond;
+    readonly attribute DOMString audioSampleFormat;
+    readonly attribute unsigned long audioBytesPerFrame;
+    readonly attribute DOMString audioSeekPreroll;
+    readonly attribute unsigned long audioCodecDelay;
+    readonly attribute DOMString audioEncryption;
+    readonly attribute DOMString audioTargetOutputChannelLayout;
+    readonly attribute boolean audioShouldDiscardDecoderDelay;
+    readonly attribute boolean audioHasExtraData;
+};

接着将我们的新增文件引入:

diff --git a/third_party/blink/renderer/bindings/generated_in_core.gni b/third_party/blink/renderer/bindings/generated_in_core.gni
index dbd963f7118e5..d690cec5e216d 100644
--- a/third_party/blink/renderer/bindings/generated_in_core.gni
+++ b/third_party/blink/renderer/bindings/generated_in_core.gni
@@ -997,6 +997,8 @@ generated_interface_sources_in_core = [
   "$root_gen_dir/third_party/blink/renderer/bindings/core/v8/v8_location.h",
   "$root_gen_dir/third_party/blink/renderer/bindings/core/v8/v8_mathml_element.cc",
   "$root_gen_dir/third_party/blink/renderer/bindings/core/v8/v8_mathml_element.h",
+  "$root_gen_dir/third_party/blink/renderer/bindings/core/v8/v8_media_decoder_config.cc",
+  "$root_gen_dir/third_party/blink/renderer/bindings/core/v8/v8_media_decoder_config.h",
   "$root_gen_dir/third_party/blink/renderer/bindings/core/v8/v8_media_error.cc",
   "$root_gen_dir/third_party/blink/renderer/bindings/core/v8/v8_media_error.h",
   "$root_gen_dir/third_party/blink/renderer/bindings/core/v8/v8_media_list.cc",
diff --git a/third_party/blink/renderer/bindings/idl_in_core.gni b/third_party/blink/renderer/bindings/idl_in_core.gni
index bccaa0392bd7c..179f601947be6 100644
--- a/third_party/blink/renderer/bindings/idl_in_core.gni
+++ b/third_party/blink/renderer/bindings/idl_in_core.gni
@@ -435,6 +435,7 @@ static_idl_files_in_core = get_path_info(
           "//third_party/blink/renderer/core/html/media/html_audio_element.idl",
           "//third_party/blink/renderer/core/html/media/html_media_element.idl",
           "//third_party/blink/renderer/core/html/media/html_video_element.idl",
+          "//third_party/blink/renderer/core/html/media/media_decoder_config.idl",
           "//third_party/blink/renderer/core/html/media/media_error.idl",
           "//third_party/blink/renderer/core/html/portal/html_portal_element.idl",
           "//third_party/blink/renderer/core/html/portal/portal_activate_event.idl",

定义 mediaconfigchange 事件

为了支持 media source 变化时监听媒体信息的变化,我们需要新增一个mediaconfigchange事件,并支持onmediaconfigchange属性,用于绑定钩子函数。

// 创建一个video标签,或者audio标签
const media = document.createElement('video');
const cb = (e: Event) => console.log('媒体信息发生变化', media.getMediaConfig());

// 添加监听
media.addEventListener('mediaconfigchange', cb)

// 或者这样添加监听
media.onmediaconfigchange = cb;
diff --git a/third_party/blink/renderer/core/dom/global_event_handlers.h b/third_party/blink/renderer/core/dom/global_event_handlers.h
index 63d3e344b46a1..7c23f4fab9265 100644
--- a/third_party/blink/renderer/core/dom/global_event_handlers.h
+++ b/third_party/blink/renderer/core/dom/global_event_handlers.h
@@ -80,6 +80,7 @@ class GlobalEventHandlers {
   DEFINE_STATIC_ATTRIBUTE_EVENT_LISTENER(loadstart, kLoadstart)
   DEFINE_STATIC_ATTRIBUTE_EVENT_LISTENER(lostpointercapture,
                                          kLostpointercapture)
+  DEFINE_STATIC_ATTRIBUTE_EVENT_LISTENER(mediaconfigchange, kMediaconfigchange)
   DEFINE_STATIC_ATTRIBUTE_EVENT_LISTENER(mousedown, kMousedown)
   DEFINE_STATIC_ATTRIBUTE_EVENT_LISTENER(mouseenter, kMouseenter)
   DEFINE_STATIC_ATTRIBUTE_EVENT_LISTENER(mouseleave, kMouseleave)
diff --git a/third_party/blink/renderer/core/dom/global_event_handlers.idl b/third_party/blink/renderer/core/dom/global_event_handlers.idl
index c7657dde82327..3858e98d69a5c 100644
--- a/third_party/blink/renderer/core/dom/global_event_handlers.idl
+++ b/third_party/blink/renderer/core/dom/global_event_handlers.idl
@@ -66,6 +66,7 @@
     attribute EventHandler onloadeddata;
     attribute EventHandler onloadedmetadata;
     attribute EventHandler onloadstart;
+    attribute EventHandler onmediaconfigchange;
     attribute EventHandler onmousedown;
     [LegacyLenientThis] attribute EventHandler onmouseenter;
     [LegacyLenientThis] attribute EventHandler onmouseleave;
diff --git a/third_party/blink/renderer/core/events/event_type_names.json5 b/third_party/blink/renderer/core/events/event_type_names.json5
index fcb366c160ed4..210b2d051fef1 100644
--- a/third_party/blink/renderer/core/events/event_type_names.json5
+++ b/third_party/blink/renderer/core/events/event_type_names.json5
@@ -182,6 +182,7 @@
     "lostpointercapture",
     "managedconfigurationchange",
     "mark",
+    "mediaconfigchange",
     "message",
     "messageerror",
     "midimessage",
diff --git a/third_party/blink/renderer/core/html/html_element.cc b/third_party/blink/renderer/core/html/html_element.cc
index 46185c325204f..316458201959b 100644
--- a/third_party/blink/renderer/core/html/html_element.cc
+++ b/third_party/blink/renderer/core/html/html_element.cc
@@ -583,6 +583,8 @@ AttributeTriggers* HTMLElement::TriggersForAttributeName(
        event_type_names::kTouchstart, nullptr},
       {html_names::kOntransitionendAttr, kNoWebFeature,
        event_type_names::kWebkitTransitionEnd, nullptr},
+      {html_names::kOnmediaconfigchangeAttr, kNoWebFeature,
+       event_type_names::kMediaconfigchange, nullptr},
       {html_names::kOnvolumechangeAttr, kNoWebFeature,
        event_type_names::kVolumechange, nullptr},
       {html_names::kOnwaitingAttr, kNoWebFeature, event_type_names::kWaiting,
```C++

定义videoconfig与audioconfig属性

继续按照其他blink的属性实现方式,将我们需要的videoconfig和audioconfig以及onmediaconfigchange属性,添加到html\_attribute\_names.json5内。


diff --git a/third_party/blink/renderer/core/html/html_attribute_names.json5 b/third_party/blink/renderer/core/html/html_attribute_names.json5
index 04210d127dab1..6c4d0924101a9 100644
--- a/third_party/blink/renderer/core/html/html_attribute_names.json5
+++ b/third_party/blink/renderer/core/html/html_attribute_names.json5
@@ -29,6 +29,7 @@
     "attributionreportto",
     "attributionsourceeventid",
     "attributionsourcepriority",
+    "audioconfig",
     "autocapitalize",
     "autocomplete",
     "autocorrect",
@@ -205,6 +206,7 @@
     "onloadedmetadata",
     "onloadstart",
     "onlostpointercapture",
+    "onmediaconfigchange",
     "onmessage",
     "onmessageerror",
     "onmousedown",
@@ -338,6 +340,7 @@
     "vlink",
     "vspace",
     "virtualkeyboardpolicy",
+    "videoconfig",
     "webkitdirectory",
     "width",
     "wrap",

代码与使用示例

// 创建一个video标签,或者audio标签
const media = document.createElement('video');
const cb = (e: Event) => console.log('Media Config Changed:', video.getMediaConfig());

// 添加监听
media.addEventListener('mediaconfigchange', cb)

// 或者这样添加监听
media.onmediaconfigchange = cb;

// 设定src
// [CONSOLE]
// Media Config Changed:MediaDecoderConfig {totalBitrate: 11163352, videoBitrate: 11035352, videoCodec: 'mpeg4', videoProfile: 'unknown', videoLevel: 'not available', …}
media.src = 'https://a.b.c/xxx.avi';

// 重设src
// [CONSOLE]
// Media Config Changed:MediaDecoderConfig {totalBitrate: 13115756, videoBitrate: 12943517, videoCodec: 'hevc', videoProfile: 'hevc main', videoLevel: 'not available', …}
media.src = 'https://a.b.c/yyy.mp4';

可在 devtools,内定位到对应的 video 标签或者 audio 标签

修改 Source 示例

监听变化示例

总结

如上,为 Chromium 实现以上类似 MediaConfig 的私有 API,需要:

  1. 可以落地的场景,比如 Electron 或 CEF 客户端。
  2. 掌握 Chromium 的编译,调试,快速 Debug 的步骤。
  3. 明确想要是什么,在哪里实现。
  4. 反复不断的编译与尝试。

若都满足,就可以自己尝试实现一下啦。

Copyright© 2013-2020

All Rights Reserved 京ICP备2023019179号-8