[EXPERIMENTAL] Generic stereo codec with index header sending merged frames

webrtc/modules/video_coding/codecs/stereo/stereo_encoder_adapter.cc

Index: webrtc/modules/video_coding/codecs/stereo/stereo_encoder_adapter.cc
diff --git a/webrtc/modules/video_coding/codecs/stereo/stereo_encoder_adapter.cc b/webrtc/modules/video_coding/codecs/stereo/stereo_encoder_adapter.cc
new file mode 100644
index 0000000000000000000000000000000000000000..934aed4aed8a1f97250935f9e912b7918d4c4aa5
--- /dev/null
+++ b/webrtc/modules/video_coding/codecs/stereo/stereo_encoder_adapter.cc
@@ -0,0 +1,237 @@
+/*
+ *  Copyright (c) 2017 The WebRTC project authors. All Rights Reserved.
+ *
+ *  Use of this source code is governed by a BSD-style license
+ *  that can be found in the LICENSE file in the root of the source
+ *  tree. An additional intellectual property rights grant can be found
+ *  in the file PATENTS.  All contributing project authors may
+ *  be found in the AUTHORS file in the root of the source tree.
+ */
+
+#include "webrtc/modules/video_coding/codecs/stereo/include/stereo_encoder_adapter.h"
+
+#include "webrtc/base/keep_ref_until_done.h"
+#include "webrtc/base/logging.h"
+#include "webrtc/common_video/include/video_frame.h"
+#include "webrtc/common_video/include/video_frame_buffer.h"
+#include "webrtc/common_video/libyuv/include/webrtc_libyuv.h"
+#include "webrtc/modules/include/module_common_types.h"
+
+namespace webrtc {
+
+class StereoEncoderAdapter::AdapterEncodedImageCallback
+    : public webrtc::EncodedImageCallback {
+ public:
+  AdapterEncodedImageCallback(webrtc::StereoEncoderAdapter* adapter,
+                              StereoCodecStream stream_idx)
+      : adapter_(adapter), stream_idx_(stream_idx) {}
+
+  EncodedImageCallback::Result OnEncodedImage(
+      const EncodedImage& encoded_image,
+      const CodecSpecificInfo* codec_specific_info,
+      const RTPFragmentationHeader* fragmentation) override {
+    if (!adapter_)
+      return webrtc::EncodedImageCallback::Result::OK;
+    return adapter_->OnEncodedImage(stream_idx_, encoded_image,
+                                    codec_specific_info, fragmentation);
+  }
+
+ private:
+  StereoEncoderAdapter* adapter_;
+  const StereoCodecStream stream_idx_;
+};
+
+struct StereoEncoderAdapter::EncodedImageData {
+  explicit EncodedImageData(StereoCodecStream stream_idx)
+      : stream_idx_(stream_idx) {
+    RTC_DCHECK_EQ(kAXXStream, stream_idx);
+    encodedImage_._length = 0;
+  }
+  EncodedImageData(StereoCodecStream stream_idx,
+                   const EncodedImage& encodedImage,
+                   const CodecSpecificInfo* codecSpecificInfo,
+                   const RTPFragmentationHeader* fragmentation)
+      : stream_idx_(stream_idx),
+        encodedImage_(encodedImage),
+        codecSpecificInfo_(*codecSpecificInfo) {
+    fragmentation_.CopyFrom(*fragmentation);
+  }
+  const StereoCodecStream stream_idx_;
+  EncodedImage encodedImage_;
+  const CodecSpecificInfo codecSpecificInfo_;
+  RTPFragmentationHeader fragmentation_;
+
+ private:
+  RTC_DISALLOW_IMPLICIT_CONSTRUCTORS(EncodedImageData);
+};
+
+StereoEncoderAdapter::StereoEncoderAdapter(VideoEncoderFactory* factory)
+    : factory_(factory), encoded_complete_callback_(nullptr) {}
+
+StereoEncoderAdapter::~StereoEncoderAdapter() {
+  Release();
+}
+
+int StereoEncoderAdapter::InitEncode(const VideoCodec* inst,
+                                    int number_of_cores,
+                                    size_t max_payload_size) {
+  const size_t buffer_size =
+      CalcBufferSize(VideoType::kI420, inst->width, inst->height);
+  stereo_dummy_planes_.resize(buffer_size);
+  // It is more expensive to encode 0x00, so use 0x80 instead.
+  std::fill(stereo_dummy_planes_.begin(), stereo_dummy_planes_.end(), 0x80);
+
+  for (size_t i = 0; i < kStereoCodecStreams; ++i) {
+    VideoEncoder* encoder = factory_->Create();
+    const int rv = encoder->InitEncode(inst, number_of_cores, max_payload_size);
+    if (rv)
+      return rv;
+    encoders_.push_back(encoder);
+    adapter_callbacks_.emplace_back(new AdapterEncodedImageCallback(
+        this, static_cast<StereoCodecStream>(i)));
+    encoder->RegisterEncodeCompleteCallback(adapter_callbacks_.back().get());
+  }
+  return WEBRTC_VIDEO_CODEC_OK;
+}
+
+int StereoEncoderAdapter::Encode(const VideoFrame& input_image,
+                                 const CodecSpecificInfo* codec_specific_info,
+                                 const std::vector<FrameType>* frame_types) {
+  if (!encoded_complete_callback_) {
+    return WEBRTC_VIDEO_CODEC_UNINITIALIZED;
+  }
+
+  // Encode AXX
+  rtc::scoped_refptr<I420BufferInterface> yuva_buffer =
+      input_image.video_frame_buffer()->ToI420();
+  if (yuva_buffer->HasAlpha()) {
+    rtc::scoped_refptr<WrappedI420Buffer> alpha_buffer(
+        new rtc::RefCountedObject<webrtc::WrappedI420Buffer>(
+            input_image.width(), input_image.height(), yuva_buffer->DataA(),
+            yuva_buffer->StrideA(), stereo_dummy_planes_.data(),
+            yuva_buffer->StrideU(), stereo_dummy_planes_.data(),
+            yuva_buffer->StrideV(),
+            rtc::KeepRefUntilDone(input_image.video_frame_buffer())));
+    VideoFrame alpha_image(alpha_buffer, input_image.timestamp(),
+                           input_image.render_time_ms(),
+                           input_image.rotation());
+    encoders_[kAXXStream]->Encode(alpha_image, codec_specific_info,
+                                       frame_types);
+  } else {
+    RTC_DCHECK(encoded_data_.find(input_image.timestamp()) ==
+               encoded_data_.end());
+    encoded_data_.emplace(std::piecewise_construct,
+                          std::forward_as_tuple(input_image.timestamp()),
+                          std::forward_as_tuple(kAXXStream));
+  }
+
+  // Encode YUV
+  int rv = encoders_[kYUVStream]->Encode(input_image, codec_specific_info,
+                                         frame_types);
+  // RTC_DCHECK(rv);
+  return rv;
+}
+
+int StereoEncoderAdapter::RegisterEncodeCompleteCallback(
+    EncodedImageCallback* callback) {
+  encoded_complete_callback_ = callback;
+  return WEBRTC_VIDEO_CODEC_OK;
+}
+
+int StereoEncoderAdapter::SetChannelParameters(uint32_t packet_loss,
+                                              int64_t rtt) {
+  for (auto encoder : encoders_) {
+    const int rv = encoder->SetChannelParameters(packet_loss, rtt);
+    if (rv)
+      return rv;
+  }
+  return WEBRTC_VIDEO_CODEC_OK;
+}
+
+int StereoEncoderAdapter::SetRateAllocation(const BitrateAllocation& bitrate,
+                                           uint32_t new_framerate) {
+  for (auto encoder : encoders_) {
+    const int rv = encoder->SetRateAllocation(bitrate, new_framerate);
+    if (rv)
+      return rv;
+  }
+  return WEBRTC_VIDEO_CODEC_OK;
+}
+
+int StereoEncoderAdapter::Release() {
+  for (auto encoder : encoders_) {
+    const int rv = encoder->Release();
+    if (rv)
+      return rv;
+    factory_->Destroy(encoder);
+  }
+  encoders_.clear();
+  adapter_callbacks_.clear();
+  return WEBRTC_VIDEO_CODEC_OK;
+}
+
+EncodedImageCallback::Result StereoEncoderAdapter::OnEncodedImage(
+    StereoCodecStream stream_idx,
+    const EncodedImage& encodedImage,
+    const CodecSpecificInfo* codecSpecificInfo,
+    const RTPFragmentationHeader* fragmentation) {
+  const auto& other_encoded_data_it =
+      encoded_data_.find(encodedImage._timeStamp);
+  if (other_encoded_data_it != encoded_data_.end()) {
+    const auto& other_image_data = other_encoded_data_it->second;
+    EncodedImageCallback::Result res = EncodedImageCallback::Result::OK;
+    if (stream_idx == kYUVStream) {
+      RTC_DCHECK_EQ(kAXXStream, other_image_data.stream_idx_);
+      res = SendEncodedImages(encodedImage, codecSpecificInfo, fragmentation,
+                               other_image_data.encodedImage_,
+                               &other_image_data.codecSpecificInfo_,
+                               &other_image_data.fragmentation_);
+    } else {
+      RTC_DCHECK_EQ(kYUVStream, other_image_data.stream_idx_);
+      RTC_DCHECK_EQ(kAXXStream, stream_idx);
+      res = SendEncodedImages(other_image_data.encodedImage_,
+                               &other_image_data.codecSpecificInfo_,
+                               &other_image_data.fragmentation_, encodedImage,
+                               codecSpecificInfo, fragmentation);
+    }
+    encoded_data_.erase(encoded_data_.begin(), other_encoded_data_it);
+    return res;
+  }
+  RTC_DCHECK(encoded_data_.find(encodedImage._timeStamp) ==
+             encoded_data_.end());
+  encoded_data_.emplace(
+      std::piecewise_construct, std::forward_as_tuple(encodedImage._timeStamp),
+      std::forward_as_tuple(stream_idx, encodedImage, codecSpecificInfo,
+                            fragmentation));
+  return webrtc::EncodedImageCallback::Result::OK;
+}
+
+EncodedImageCallback::Result StereoEncoderAdapter::SendEncodedImages(
+    const EncodedImage& encoded_image,
+    const CodecSpecificInfo* codec_specific_info,
+    const RTPFragmentationHeader* fragmentation,
+    const EncodedImage& stereo_encoded_image,
+    const CodecSpecificInfo* stereo_codec_specific_info,
+    const RTPFragmentationHeader* stereo_fragmentation) {
+  const bool has_alpha = stereo_encoded_image._length != 0;
+
+  CodecSpecificInfo* yuv_codec =
+      const_cast<CodecSpecificInfo*>(codec_specific_info);
+  yuv_codec->codecType = kVideoCodecStereo;
+  yuv_codec->codec_name = "stereo-vp9";
+  yuv_codec->stereoInfo.stereoCodecType = kVideoCodecVP9;
+  if (!has_alpha) {
+    yuv_codec->stereoInfo.num_frames = 0;
+    return encoded_complete_callback_->OnEncodedImage(encoded_image, yuv_codec,
+                                                      fragmentation);
+  }
+
+  yuv_codec->stereoInfo.num_frames = 1;
+  yuv_codec->stereoInfo.encoded_images[0] = &encoded_image;
+  yuv_codec->stereoInfo.codec_specific_infos[0] = stereo_codec_specific_info;
+  yuv_codec->stereoInfo.fragmentations[0] = stereo_fragmentation;
+  return encoded_complete_callback_->OnEncodedImage(encoded_image, yuv_codec,
+                                                    fragmentation);
+}
+
+}  // namespace webrtc