H.264 video codec support using OpenH264/FFmpeg

webrtc/modules/video_coding/codecs/h264/h264_encoder_impl.cc

Index: webrtc/modules/video_coding/codecs/h264/h264_encoder_impl.cc
diff --git a/webrtc/modules/video_coding/codecs/h264/h264_encoder_impl.cc b/webrtc/modules/video_coding/codecs/h264/h264_encoder_impl.cc
new file mode 100644
index 0000000000000000000000000000000000000000..9881783f00d233f0cc51d51c14c2823f3d6f0249
--- /dev/null
+++ b/webrtc/modules/video_coding/codecs/h264/h264_encoder_impl.cc
@@ -0,0 +1,380 @@
+/*
+ *  Copyright (c) 2015 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/h264/h264_encoder_impl.h"
+
+#include <limits>
+
+#include "third_party/openh264/src/codec/api/svc/codec_api.h"
+#include "third_party/openh264/src/codec/api/svc/codec_app_def.h"
+#include "third_party/openh264/src/codec/api/svc/codec_def.h"
+
+#include "webrtc/base/checks.h"
+#include "webrtc/base/logging.h"
+#include "webrtc/common_video/libyuv/include/webrtc_libyuv.h"
+
+namespace webrtc {
+
+namespace {
+const bool kOpenH264EncoderDetailedLogging = false;
+}  // namespace
+
+static FrameType EVideoFrameType_to_FrameType(EVideoFrameType type) {
+  switch (type) {
+    case videoFrameTypeInvalid:
+      return kEmptyFrame;
+    case videoFrameTypeIDR:
+      return kVideoFrameKey;
+    case videoFrameTypeSkip:
+    case videoFrameTypeI:
+    case videoFrameTypeP:
+    case videoFrameTypeIPMixed:
+      return kVideoFrameDelta;
+    default:
+      LOG(LS_WARNING) << "Unknown EVideoFrameType: " << type;
+      return kVideoFrameDelta;
+  }
+}
+
+// Helper method used by H264EncoderImpl::Encode.
+// Copies the encoded bytes from |info| to |encoded_image| and updates the
+// fragmentation information of |frag_header|. The |encoded_image->_buffer| may
+// be deleted and reallocated if a bigger buffer is required.
+//
+// After OpenH264 encoding, the encoded bytes are stored in |info| spread out
+// over a number of layers and "NAL units". Each NAL unit is a fragment starting
+// with the four-byte start code {0,0,0,1}. All of this data (including the
+// start codes) is copied to the |encoded_image->_buffer| and the |frag_header|
+// is updated to point to each fragment, with offsets and lengths set as to
+// exclude the start codes.
+static void RtpFragmentize(EncodedImage* encoded_image,
+                           rtc::scoped_ptr<uint8_t[]>* encoded_image_buffer,
+                           const VideoFrame& frame,
+                           SFrameBSInfo* info,
+                           RTPFragmentationHeader* frag_header) {
+  // Calculate minimum buffer size required to hold encoded data.
+  size_t required_size = 0;
+  size_t fragments_count = 0;
+  for (int layer = 0; layer < info->iLayerNum; ++layer) {
+    const SLayerBSInfo& layerInfo = info->sLayerInfo[layer];
+    for (int nal = 0; nal < layerInfo.iNalCount; ++nal) {
+      RTC_CHECK_GE(layerInfo.pNalLengthInByte[nal], 0);
+      // Ensure |required_size| will not overflow.
+      RTC_CHECK_LE(static_cast<size_t>(layerInfo.pNalLengthInByte[nal]),
+                   std::numeric_limits<size_t>::max() - required_size);
+      required_size += layerInfo.pNalLengthInByte[nal];
+      ++fragments_count;
+    }
+  }
+  if (encoded_image->_size < required_size) {
+    // Increase buffer size. Allocate enough to hold an unencoded image, this
+    // should be more than enough to hold any encoded data of future frames of
+    // the same size (avoiding possible future reallocation due to variations in
+    // required size).
+    encoded_image->_size = CalcBufferSize(kI420, frame.width(), frame.height());
+    if (encoded_image->_size < required_size) {
+      // Encoded data > unencoded data, wtf? Allocate required bytes.

[mflodman, 2016/01/12 10:31:25]

Seems like this really shouldn't happen, is it worth having this logic and will
we in that case ever catch this log message? Should we have some kind of check
instead?

[hta-webrtc, 2016/01/12 10:53:34]

Video encoding is complex enough that we can't guarantee it won't.
I suspect that the only way to get here is specially crafted images that are
designed to be uncompressible (possibly part of an attack), but there's no way
to guarantee it, and no way to guarantee it won't happen by accident either.

For instance, white noise images are, by definition almost, uncompressible.

[stefan-webrtc, 2016/01/12 11:18:33]

Agree, even though very unlikely, it's better not to crash if this happens.

[mflodman, 2016/01/12 11:25:34]

Fair enough, but then I'd prefer to remove the 'wtf' and maybe even the log if
it's not useful for anything. WDYT?

[hbos, 2016/01/12 13:56:27]

I was thinking the same. (Imagine if screensharing random pixels made the other
end hit a CHECK, nice way to troll a public hangout?) Removed the "wtf" :)

With the logging I'm thinking its unlikely enough that it wouldn't happen under
normal circumstances and could be interesting to know if it happened. But not
feeling strongly about this, if you want me to remove the log I will.

[mflodman, 2016/01/12 13:58:12]

Acknowledged.

+      LOG(LS_WARNING) << "Encoding produced more bytes than the original image "
+                      << "data! Original bytes: " << encoded_image->_size
+                      << ", encoded bytes: " << required_size << ".";
+      encoded_image->_size = required_size;
+    }
+    encoded_image->_buffer = new uint8_t[encoded_image->_size];
+    encoded_image_buffer->reset(encoded_image->_buffer);
+  }
+
+  // Iterate layers and NAL units, note each NAL unit as a fragment and copy
+  // the data to |encoded_image->_buffer|.
+  const uint8_t start_code[4] = {0, 0, 0, 1};
+  frag_header->VerifyAndAllocateFragmentationHeader(fragments_count);
+  size_t frag = 0;
+  encoded_image->_length = 0;
+  for (int layer = 0; layer < info->iLayerNum; ++layer) {
+    const SLayerBSInfo& layerInfo = info->sLayerInfo[layer];
+    // Iterate NAL units making up this layer, noting fragments.
+    size_t layer_len = 0;
+    for (int nal = 0; nal < layerInfo.iNalCount; ++nal, ++frag) {
+      RTC_DCHECK_EQ(layerInfo.pBsBuf[layer_len+0], start_code[0]);
+      RTC_DCHECK_EQ(layerInfo.pBsBuf[layer_len+1], start_code[1]);
+      RTC_DCHECK_EQ(layerInfo.pBsBuf[layer_len+2], start_code[2]);
+      RTC_DCHECK_EQ(layerInfo.pBsBuf[layer_len+3], start_code[3]);
+      frag_header->fragmentationOffset[frag] =
+          encoded_image->_length + layer_len + sizeof(start_code);
+      frag_header->fragmentationLength[frag] =
+          layerInfo.pNalLengthInByte[nal] - sizeof(start_code);
+      layer_len += layerInfo.pNalLengthInByte[nal];
+    }
+    // Copy the entire layer's data (including start codes).
+    memcpy(encoded_image->_buffer + encoded_image->_length,
+           layerInfo.pBsBuf,
+           layer_len);
+    encoded_image->_length += layer_len;
+  }
+}
+
+H264EncoderImpl::H264EncoderImpl()
+    : openh264_encoder_(nullptr),
+      encoded_image_callback_(nullptr) {
+}
+
+H264EncoderImpl::~H264EncoderImpl() {
+  Release();
+}
+
+int32_t H264EncoderImpl::InitEncode(const VideoCodec* codec_settings,
+                                    int32_t /*number_of_cores*/,
+                                    size_t /*max_payload_size*/) {
+  if (!codec_settings ||
+      codec_settings->codecType != kVideoCodecH264) {
+    return WEBRTC_VIDEO_CODEC_ERR_PARAMETER;
+  }
+  if (codec_settings->maxFramerate == 0)
+    return WEBRTC_VIDEO_CODEC_ERR_PARAMETER;
+  if (codec_settings->width < 1 || codec_settings->height < 1)
+    return WEBRTC_VIDEO_CODEC_ERR_PARAMETER;
+
+  int32_t release_ret = Release();
+  if (release_ret != WEBRTC_VIDEO_CODEC_OK)
+    return release_ret;
+  RTC_DCHECK(!openh264_encoder_);
+
+  // Create encoder.
+  if (WelsCreateSVCEncoder(&openh264_encoder_) != 0) {
+    // Failed to create encoder.
+    LOG(LS_ERROR) << "Failed to create OpenH264 encoder";
+    RTC_DCHECK(!openh264_encoder_);
+    return WEBRTC_VIDEO_CODEC_ERROR;
+  }
+  RTC_DCHECK(openh264_encoder_);
+  if (kOpenH264EncoderDetailedLogging) {
+    int trace_level = WELS_LOG_DETAIL;
+    openh264_encoder_->SetOption(ENCODER_OPTION_TRACE_LEVEL,
+                                 &trace_level);
+  }
+  // else WELS_LOG_DEFAULT is used by default.
+
+  codec_settings_ = *codec_settings;
+  if (codec_settings_.targetBitrate == 0)
+    codec_settings_.targetBitrate = codec_settings_.startBitrate;
+
+  // Initialization parameters.
+  // There are two ways to initialize. There is SEncParamBase (cleared with
+  // memset(&p, 0, sizeof(SEncParamBase)) used in Initialize, and SEncParamExt
+  // which is a superset of SEncParamBase (cleared with GetDefaultParams) used
+  // in InitializeExt.
+  SEncParamExt init_params;
+  openh264_encoder_->GetDefaultParams(&init_params);
+  if (codec_settings_.mode == kRealtimeVideo) {
+    init_params.iUsageType = CAMERA_VIDEO_REAL_TIME;
+  } else if (codec_settings_.mode == kScreensharing) {
+    init_params.iUsageType = SCREEN_CONTENT_REAL_TIME;
+  } else {
+    return WEBRTC_VIDEO_CODEC_ERR_PARAMETER;
+  }
+  init_params.iPicWidth = codec_settings_.width;
+  init_params.iPicHeight = codec_settings_.height;
+  // |init_params| uses bit/s, |codec_settings_| uses kbit/s.
+  init_params.iTargetBitrate = codec_settings_.targetBitrate * 1000;
+  init_params.iMaxBitrate = codec_settings_.maxBitrate * 1000;
+  // Rate Control mode
+  init_params.iRCMode = RC_BITRATE_MODE;
+  init_params.fMaxFrameRate = static_cast<float>(codec_settings_.maxFramerate);
+
+  // The following parameters are extension parameters (they're in SEncParamExt,
+  // not in SEncParamBase).
+  init_params.bEnableFrameSkip =
+      codec_settings_.codecSpecific.H264.frameDroppingOn;
+  // |uiIntraPeriod|    - multiple of GOP size
+  // |keyFrameInterval| - number of frames
+  init_params.uiIntraPeriod =
+      codec_settings_.codecSpecific.H264.keyFrameInterval;
+  init_params.uiMaxNalSize = 0;
+  // Threading model: use auto.
+  //  0: auto (dynamic imp. internal encoder)
+  //  1: single thread (default value)
+  // >1: number of threads
+  init_params.iMultipleThreadIdc = 0;

[mflodman, 2016/01/12 10:31:25]

Do we know what this leads to in the Chrome sandbox, where I assume (maybe
incorrectly) the encoder can't read the number of cores?

[hta-webrtc, 2016/01/12 10:53:34]

I think Tommi added the ability to read number of cores from within the sandbox
recently.

[mflodman, 2016/01/12 11:25:34]

I know we do get the correct number through the input parameter to this method,
but I didn't know it was solved for the generic case. Great if that is the case,
but if it isn't I'd prefer to use 'number_of_cores' and explicitly set the
number of threads.

Great if that is the case, if not I think it's better to explicitly set the
number of threads.

[hbos, 2016/01/12 13:56:27]

Ok. I'll use the NumberOfThreads function copied from vp9_impl.cc (which I
previously used on the decode side but removed).

+  // The base spatial layer 0 is the only one we use.
+  init_params.sSpatialLayers[0].iVideoWidth        = init_params.iPicWidth;
+  init_params.sSpatialLayers[0].iVideoHeight       = init_params.iPicHeight;
+  init_params.sSpatialLayers[0].fFrameRate         = init_params.fMaxFrameRate;
+  init_params.sSpatialLayers[0].iSpatialBitrate    = init_params.iTargetBitrate;
+  init_params.sSpatialLayers[0].iMaxSpatialBitrate = init_params.iMaxBitrate;
+  // Slice num according to number of threads.
+  init_params.sSpatialLayers[0].sSliceCfg.uiSliceMode = SM_AUTO_SLICE;
+
+  // Initialize.
+  if (openh264_encoder_->InitializeExt(&init_params) != 0) {
+    LOG(LS_ERROR) << "Failed to initialize OpenH264 encoder";
+    Release();
+    return WEBRTC_VIDEO_CODEC_ERROR;
+  }
+  int video_format = EVideoFormatType::videoFormatI420;
+  openh264_encoder_->SetOption(ENCODER_OPTION_DATAFORMAT,
+                               &video_format);
+
+  // Initialize encoded image. Default buffer size: size of unencoded data.
+  encoded_image_._size = CalcBufferSize(
+      kI420, codec_settings_.width, codec_settings_.height);
+  encoded_image_._buffer = new uint8_t[encoded_image_._size];
+  encoded_image_buffer_.reset(encoded_image_._buffer);
+  encoded_image_._completeFrame = true;
+  encoded_image_._encodedWidth = 0;
+  encoded_image_._encodedHeight = 0;
+  encoded_image_._length = 0;
+  return WEBRTC_VIDEO_CODEC_OK;
+}
+
+int32_t H264EncoderImpl::Release() {
+  if (openh264_encoder_) {
+    int uninit_ret = openh264_encoder_->Uninitialize();
+    if (uninit_ret != 0) {
+      LOG(LS_WARNING) << "OpenH264 encoder's Uninitialize() returned "
+                      << "unsuccessful: " << uninit_ret;
+    }
+    WelsDestroySVCEncoder(openh264_encoder_);
+    openh264_encoder_ = nullptr;
+  }
+  if (encoded_image_._buffer != nullptr) {
+    encoded_image_._buffer = nullptr;
+    encoded_image_buffer_.reset();
+  }
+  return WEBRTC_VIDEO_CODEC_OK;
+}
+
+int32_t H264EncoderImpl::RegisterEncodeCompleteCallback(
+    EncodedImageCallback* callback) {
+  encoded_image_callback_ = callback;
+  return WEBRTC_VIDEO_CODEC_OK;
+}
+
+int32_t H264EncoderImpl::SetRates(uint32_t bitrate, uint32_t framerate) {
+  if (bitrate <= 0 || framerate <= 0) {
+    return WEBRTC_VIDEO_CODEC_ERR_PARAMETER;
+  }
+  codec_settings_.targetBitrate = bitrate;
+  codec_settings_.maxFramerate = framerate;
+
+  SBitrateInfo target_bitrate;
+  memset(&target_bitrate, 0, sizeof(SBitrateInfo));
+  target_bitrate.iLayer = SPATIAL_LAYER_ALL,
+  target_bitrate.iBitrate = codec_settings_.targetBitrate * 1000;
+  openh264_encoder_->SetOption(ENCODER_OPTION_BITRATE,
+                               &target_bitrate);
+  float max_framerate = static_cast<float>(codec_settings_.maxFramerate);
+  openh264_encoder_->SetOption(ENCODER_OPTION_FRAME_RATE,
+                               &max_framerate);
+  return WEBRTC_VIDEO_CODEC_OK;
+}
+
+int32_t H264EncoderImpl::Encode(
+    const VideoFrame& frame, const CodecSpecificInfo* codec_specific_info,
+    const std::vector<FrameType>* frame_types) {
+  if (!IsInitialized())
+    return WEBRTC_VIDEO_CODEC_UNINITIALIZED;
+  if (frame.IsZeroSize())
+    return WEBRTC_VIDEO_CODEC_ERR_PARAMETER;
+  if (!encoded_image_callback_) {
+    LOG(LS_WARNING) << "InitEncode() has been called, but a callback function "
+                    << "has not been set with RegisterEncodeCompleteCallback()";
+    return WEBRTC_VIDEO_CODEC_UNINITIALIZED;
+  }
+  if (frame.width()  != codec_settings_.width ||
+      frame.height() != codec_settings_.height) {
+    LOG(LS_WARNING) << "Encoder initialized for " << codec_settings_.width
+                    << "x" << codec_settings_.height << " but trying to encode "
+                    << frame.width() << "x" << frame.height() << " frame.";
+    return WEBRTC_VIDEO_CODEC_ERR_SIZE;
+  }
+
+  bool force_key_frame = false;
+  if (frame_types != nullptr) {
+    // We only support a single stream.
+    RTC_DCHECK_EQ(frame_types->size(), static_cast<size_t>(1));
+    // Skip frame?
+    if ((*frame_types)[0] == kEmptyFrame) {
+      return WEBRTC_VIDEO_CODEC_OK;
+    }
+    // Force key frame?
+    force_key_frame = (*frame_types)[0] == kVideoFrameKey;
+  }
+  if (force_key_frame) {
+    // Only need to call ForceIntraFrame when true. API doc says
+    // ForceIntraFrame(false) does nothing but really if you call it for every
+    // frame it introduces massive delays and lag in the video stream.
+    openh264_encoder_->ForceIntraFrame(true);
+  }
+
+  // EncodeFrame input.
+  SSourcePicture picture;
+  memset(&picture, 0, sizeof(SSourcePicture));
+  picture.iPicWidth = frame.width();
+  picture.iPicHeight = frame.height();
+  picture.iColorFormat = EVideoFormatType::videoFormatI420;
+  picture.uiTimeStamp = frame.ntp_time_ms();
+  picture.iStride[0] = frame.stride(kYPlane);
+  picture.iStride[1] = frame.stride(kUPlane);
+  picture.iStride[2] = frame.stride(kVPlane);
+  picture.pData[0] = const_cast<uint8_t*>(frame.buffer(kYPlane));
+  picture.pData[1] = const_cast<uint8_t*>(frame.buffer(kUPlane));
+  picture.pData[2] = const_cast<uint8_t*>(frame.buffer(kVPlane));
+
+  // EncodeFrame output.
+  SFrameBSInfo info;
+  memset(&info, 0, sizeof(SFrameBSInfo));
+
+  // Encode!
+  int enc_ret = openh264_encoder_->EncodeFrame(&picture, &info);
+  if (enc_ret != 0) {
+    LOG(LS_ERROR) << "OpenH264 frame encoding failed, EncodeFrame returned "
+                  << enc_ret << ".";
+    return WEBRTC_VIDEO_CODEC_ERROR;
+  }
+
+  encoded_image_._encodedWidth = frame.width();
+  encoded_image_._encodedHeight = frame.height();
+  encoded_image_._timeStamp = frame.timestamp();
+  encoded_image_.ntp_time_ms_ = frame.ntp_time_ms();
+  encoded_image_.capture_time_ms_ = frame.render_time_ms();
+  encoded_image_._frameType = EVideoFrameType_to_FrameType(info.eFrameType);
+
+  // Split encoded image up into fragments. This also updates |encoded_image_|.
+  RTPFragmentationHeader frag_header;
+  RtpFragmentize(&encoded_image_, &encoded_image_buffer_, frame, &info,
+                 &frag_header);
+
+  // Encoder can skip frames to save bandwidth in which case
+  // |encoded_image_._length| == 0.
+  if (encoded_image_._length > 0) {
+    // Deliver encoded image.
+    encoded_image_callback_->Encoded(encoded_image_, codec_specific_info,
+                                     &frag_header);
+  }
+  return WEBRTC_VIDEO_CODEC_OK;
+}
+
+bool H264EncoderImpl::IsInitialized() const {
+  return openh264_encoder_ != nullptr;
+}
+
+int32_t H264EncoderImpl::SetChannelParameters(
+    uint32_t packet_loss, int64_t rtt) {
+  return WEBRTC_VIDEO_CODEC_OK;
+}
+
+int32_t H264EncoderImpl::SetPeriodicKeyFrames(bool enable) {
+  return WEBRTC_VIDEO_CODEC_OK;
+}
+
+void H264EncoderImpl::OnDroppedFrame() {
+}
+
+}  // namespace webrtc