mirror of
https://github.com/mollyim/webrtc.git
synced 2025-05-12 21:30:45 +01:00

This is a reland of commit 7ae48c452a
with updated RtpVp9RefFinder
RtpVp9RefFinder relied on the fact that frames with (inter_pic_predicted=true && inter_layer_predicted=true) were marked as keyframes. Since this is not the case anymore, the related code paths in RtpVp9RefFinder have been deleted.
Calculation of gof_info_[] index for non-keyframes has been updated to account for that fact it is now possible to received multiple T0 frames belonging to the same temporal unit (we don't need to do "unwrapped_tl0 - 1" in this case).
Original change's description:
> Mark frames with inter_layer_predicted=true as delta frames
>
> As it is currently implemented, the VP9 depacketizer decides packet's frame type based on p_bit ("Inter-picture predicted layer frame"). p_bit is set to 0 for upper spatial layer frames of keyframe since they do not have temporal refs. This results in marking packets of upper spatial layer frames, and, eventually these frames, of SVC keyframes as "keyframe" while they are in fact delta frames.
>
> Normally spatial layer frames are merged into a superframe and the superframe is passed to decoder. But passing individual layers to a single decoder instance is a valid scenario too and is used in downstream projects. In this case, an upper layer frame marked as keyframe may cause decoder reset [2] and break decoding.
>
> This CL changes frame type decision logic in the VP9 depacketizer such that only packets with both P and D (inter-layer predicted) bits unset are considered as keyframe packets.
>
> When spatial layer frames are merged into a superframe in CombineAndDeleteFrames [1], frame type of the superframe is inferred from the lowest spatial layer frame.
>
> [1] https://source.chromium.org/chromium/chromium/src/+/main:third_party/webrtc/modules/video_coding/frame_helpers.cc;l=53
>
> [2] https://source.corp.google.com/piper///depot/google3/third_party/webrtc/files/stable/webrtc/modules/video_coding/codecs/vp9/libvpx_vp9_decoder.cc;l=209
>
> Bug: webrtc:15827
> Change-Id: Idc3445636f0eae0192dac998876fedec48628560
> Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/343342
> Reviewed-by: Danil Chapovalov <danilchap@webrtc.org>
> Commit-Queue: Sergey Silkin <ssilkin@webrtc.org>
> Cr-Commit-Position: refs/heads/main@{#41939}
Bug: webrtc:15827
Change-Id: Ic69b94989919cf6d353bceea85d0eba63bc500ee
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/344144
Reviewed-by: Philip Eliasson <philipel@webrtc.org>
Commit-Queue: Sergey Silkin <ssilkin@webrtc.org>
Reviewed-by: Danil Chapovalov <danilchap@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#41985}
363 lines
13 KiB
C++
363 lines
13 KiB
C++
/*
|
|
* Copyright (c) 2020 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 "modules/video_coding/rtp_vp9_ref_finder.h"
|
|
|
|
#include <algorithm>
|
|
#include <utility>
|
|
|
|
#include "rtc_base/logging.h"
|
|
|
|
namespace webrtc {
|
|
RtpFrameReferenceFinder::ReturnVector RtpVp9RefFinder::ManageFrame(
|
|
std::unique_ptr<RtpFrameObject> frame) {
|
|
const RTPVideoHeaderVP9& codec_header = absl::get<RTPVideoHeaderVP9>(
|
|
frame->GetRtpVideoHeader().video_type_header);
|
|
|
|
if (codec_header.temporal_idx != kNoTemporalIdx)
|
|
frame->SetTemporalIndex(codec_header.temporal_idx);
|
|
frame->SetSpatialIndex(codec_header.spatial_idx);
|
|
frame->SetId(codec_header.picture_id & (kFrameIdLength - 1));
|
|
|
|
FrameDecision decision;
|
|
if (codec_header.temporal_idx >= kMaxTemporalLayers ||
|
|
codec_header.spatial_idx >= kMaxSpatialLayers) {
|
|
decision = kDrop;
|
|
} else if (codec_header.flexible_mode) {
|
|
decision = ManageFrameFlexible(frame.get(), codec_header);
|
|
} else {
|
|
if (codec_header.tl0_pic_idx == kNoTl0PicIdx) {
|
|
RTC_LOG(LS_WARNING) << "TL0PICIDX is expected to be present in "
|
|
"non-flexible mode.";
|
|
decision = kDrop;
|
|
} else {
|
|
int64_t unwrapped_tl0 =
|
|
tl0_unwrapper_.Unwrap(codec_header.tl0_pic_idx & 0xFF);
|
|
decision = ManageFrameGof(frame.get(), codec_header, unwrapped_tl0);
|
|
|
|
if (decision == kStash) {
|
|
if (stashed_frames_.size() > kMaxStashedFrames) {
|
|
stashed_frames_.pop_back();
|
|
}
|
|
|
|
stashed_frames_.push_front(
|
|
{.unwrapped_tl0 = unwrapped_tl0, .frame = std::move(frame)});
|
|
}
|
|
}
|
|
}
|
|
|
|
RtpFrameReferenceFinder::ReturnVector res;
|
|
switch (decision) {
|
|
case kStash:
|
|
return res;
|
|
case kHandOff:
|
|
res.push_back(std::move(frame));
|
|
RetryStashedFrames(res);
|
|
return res;
|
|
case kDrop:
|
|
return res;
|
|
}
|
|
|
|
return res;
|
|
}
|
|
|
|
RtpVp9RefFinder::FrameDecision RtpVp9RefFinder::ManageFrameFlexible(
|
|
RtpFrameObject* frame,
|
|
const RTPVideoHeaderVP9& codec_header) {
|
|
if (codec_header.num_ref_pics > EncodedFrame::kMaxFrameReferences) {
|
|
return kDrop;
|
|
}
|
|
|
|
frame->num_references = codec_header.num_ref_pics;
|
|
for (size_t i = 0; i < frame->num_references; ++i) {
|
|
frame->references[i] =
|
|
Subtract<kFrameIdLength>(frame->Id(), codec_header.pid_diff[i]);
|
|
}
|
|
|
|
FlattenFrameIdAndRefs(frame, codec_header.inter_layer_predicted);
|
|
return kHandOff;
|
|
}
|
|
|
|
RtpVp9RefFinder::FrameDecision RtpVp9RefFinder::ManageFrameGof(
|
|
RtpFrameObject* frame,
|
|
const RTPVideoHeaderVP9& codec_header,
|
|
int64_t unwrapped_tl0) {
|
|
GofInfo* info;
|
|
if (codec_header.ss_data_available) {
|
|
if (codec_header.temporal_idx != 0) {
|
|
RTC_LOG(LS_WARNING) << "Received scalability structure on a non base "
|
|
"layer frame. Scalability structure ignored.";
|
|
} else {
|
|
if (codec_header.gof.num_frames_in_gof > kMaxVp9FramesInGof) {
|
|
return kDrop;
|
|
}
|
|
|
|
for (size_t i = 0; i < codec_header.gof.num_frames_in_gof; ++i) {
|
|
if (codec_header.gof.num_ref_pics[i] > kMaxVp9RefPics) {
|
|
return kDrop;
|
|
}
|
|
}
|
|
|
|
GofInfoVP9 gof = codec_header.gof;
|
|
if (gof.num_frames_in_gof == 0) {
|
|
RTC_LOG(LS_WARNING) << "Number of frames in GOF is zero. Assume "
|
|
"that stream has only one temporal layer.";
|
|
gof.SetGofInfoVP9(kTemporalStructureMode1);
|
|
}
|
|
|
|
current_ss_idx_ = Add<kMaxGofSaved>(current_ss_idx_, 1);
|
|
scalability_structures_[current_ss_idx_] = gof;
|
|
scalability_structures_[current_ss_idx_].pid_start = frame->Id();
|
|
gof_info_.emplace(
|
|
unwrapped_tl0,
|
|
GofInfo(&scalability_structures_[current_ss_idx_], frame->Id()));
|
|
}
|
|
|
|
const auto gof_info_it = gof_info_.find(unwrapped_tl0);
|
|
if (gof_info_it == gof_info_.end())
|
|
return kStash;
|
|
|
|
info = &gof_info_it->second;
|
|
|
|
if (frame->frame_type() == VideoFrameType::kVideoFrameKey) {
|
|
frame->num_references = 0;
|
|
FrameReceivedVp9(frame->Id(), info);
|
|
FlattenFrameIdAndRefs(frame, codec_header.inter_layer_predicted);
|
|
return kHandOff;
|
|
}
|
|
} else {
|
|
if (frame->frame_type() == VideoFrameType::kVideoFrameKey) {
|
|
RTC_LOG(LS_WARNING) << "Received keyframe without scalability structure";
|
|
return kDrop;
|
|
}
|
|
|
|
// tl0_idx is incremented on temporal_idx=0 frames of the lowest spatial
|
|
// layer (which spatial_idx is not necessarily zero). Upper spatial layer
|
|
// frames with inter-layer prediction use GOF info of their base spatial
|
|
// layer frames.
|
|
const bool use_prev_gof =
|
|
codec_header.temporal_idx == 0 && !codec_header.inter_layer_predicted;
|
|
auto gof_info_it =
|
|
gof_info_.find(use_prev_gof ? unwrapped_tl0 - 1 : unwrapped_tl0);
|
|
|
|
// Gof info for this frame is not available yet, stash this frame.
|
|
if (gof_info_it == gof_info_.end())
|
|
return kStash;
|
|
|
|
if (codec_header.temporal_idx == 0) {
|
|
gof_info_it = gof_info_
|
|
.emplace(unwrapped_tl0,
|
|
GofInfo(gof_info_it->second.gof, frame->Id()))
|
|
.first;
|
|
}
|
|
|
|
info = &gof_info_it->second;
|
|
}
|
|
|
|
// Clean up info for base layers that are too old.
|
|
int64_t old_tl0_pic_idx = unwrapped_tl0 - kMaxGofSaved;
|
|
auto clean_gof_info_to = gof_info_.lower_bound(old_tl0_pic_idx);
|
|
gof_info_.erase(gof_info_.begin(), clean_gof_info_to);
|
|
|
|
FrameReceivedVp9(frame->Id(), info);
|
|
|
|
// Make sure we don't miss any frame that could potentially have the
|
|
// up switch flag set.
|
|
if (MissingRequiredFrameVp9(frame->Id(), *info))
|
|
return kStash;
|
|
|
|
if (codec_header.temporal_up_switch)
|
|
up_switch_.emplace(frame->Id(), codec_header.temporal_idx);
|
|
|
|
// Clean out old info about up switch frames.
|
|
uint16_t old_picture_id = Subtract<kFrameIdLength>(frame->Id(), 50);
|
|
auto up_switch_erase_to = up_switch_.lower_bound(old_picture_id);
|
|
up_switch_.erase(up_switch_.begin(), up_switch_erase_to);
|
|
|
|
if (codec_header.inter_pic_predicted) {
|
|
size_t diff = ForwardDiff<uint16_t, kFrameIdLength>(info->gof->pid_start,
|
|
frame->Id());
|
|
size_t gof_idx = diff % info->gof->num_frames_in_gof;
|
|
|
|
if (info->gof->num_ref_pics[gof_idx] > EncodedFrame::kMaxFrameReferences) {
|
|
return kDrop;
|
|
}
|
|
|
|
// Populate references according to the scalability structure.
|
|
frame->num_references = info->gof->num_ref_pics[gof_idx];
|
|
for (size_t i = 0; i < frame->num_references; ++i) {
|
|
frame->references[i] = Subtract<kFrameIdLength>(
|
|
frame->Id(), info->gof->pid_diff[gof_idx][i]);
|
|
|
|
// If this is a reference to a frame earlier than the last up switch
|
|
// point, then ignore this reference.
|
|
if (UpSwitchInIntervalVp9(frame->Id(), codec_header.temporal_idx,
|
|
frame->references[i])) {
|
|
--frame->num_references;
|
|
}
|
|
}
|
|
} else {
|
|
frame->num_references = 0;
|
|
}
|
|
|
|
FlattenFrameIdAndRefs(frame, codec_header.inter_layer_predicted);
|
|
return kHandOff;
|
|
}
|
|
|
|
bool RtpVp9RefFinder::MissingRequiredFrameVp9(uint16_t picture_id,
|
|
const GofInfo& info) {
|
|
size_t diff =
|
|
ForwardDiff<uint16_t, kFrameIdLength>(info.gof->pid_start, picture_id);
|
|
size_t gof_idx = diff % info.gof->num_frames_in_gof;
|
|
size_t temporal_idx = info.gof->temporal_idx[gof_idx];
|
|
|
|
if (temporal_idx >= kMaxTemporalLayers) {
|
|
RTC_LOG(LS_WARNING) << "At most " << kMaxTemporalLayers
|
|
<< " temporal "
|
|
"layers are supported.";
|
|
return true;
|
|
}
|
|
|
|
// For every reference this frame has, check if there is a frame missing in
|
|
// the interval (`ref_pid`, `picture_id`) in any of the lower temporal
|
|
// layers. If so, we are missing a required frame.
|
|
uint8_t num_references = info.gof->num_ref_pics[gof_idx];
|
|
for (size_t i = 0; i < num_references; ++i) {
|
|
uint16_t ref_pid =
|
|
Subtract<kFrameIdLength>(picture_id, info.gof->pid_diff[gof_idx][i]);
|
|
for (size_t l = 0; l < temporal_idx; ++l) {
|
|
auto missing_frame_it = missing_frames_for_layer_[l].lower_bound(ref_pid);
|
|
if (missing_frame_it != missing_frames_for_layer_[l].end() &&
|
|
AheadOf<uint16_t, kFrameIdLength>(picture_id, *missing_frame_it)) {
|
|
return true;
|
|
}
|
|
}
|
|
}
|
|
return false;
|
|
}
|
|
|
|
void RtpVp9RefFinder::FrameReceivedVp9(uint16_t picture_id, GofInfo* info) {
|
|
int last_picture_id = info->last_picture_id;
|
|
size_t gof_size = std::min(info->gof->num_frames_in_gof, kMaxVp9FramesInGof);
|
|
|
|
// If there is a gap, find which temporal layer the missing frames
|
|
// belong to and add the frame as missing for that temporal layer.
|
|
// Otherwise, remove this frame from the set of missing frames.
|
|
if (AheadOf<uint16_t, kFrameIdLength>(picture_id, last_picture_id)) {
|
|
size_t diff = ForwardDiff<uint16_t, kFrameIdLength>(info->gof->pid_start,
|
|
last_picture_id);
|
|
size_t gof_idx = diff % gof_size;
|
|
|
|
last_picture_id = Add<kFrameIdLength>(last_picture_id, 1);
|
|
while (last_picture_id != picture_id) {
|
|
gof_idx = (gof_idx + 1) % gof_size;
|
|
RTC_CHECK(gof_idx < kMaxVp9FramesInGof);
|
|
|
|
size_t temporal_idx = info->gof->temporal_idx[gof_idx];
|
|
if (temporal_idx >= kMaxTemporalLayers) {
|
|
RTC_LOG(LS_WARNING) << "At most " << kMaxTemporalLayers
|
|
<< " temporal "
|
|
"layers are supported.";
|
|
return;
|
|
}
|
|
|
|
missing_frames_for_layer_[temporal_idx].insert(last_picture_id);
|
|
last_picture_id = Add<kFrameIdLength>(last_picture_id, 1);
|
|
}
|
|
|
|
info->last_picture_id = last_picture_id;
|
|
} else {
|
|
size_t diff =
|
|
ForwardDiff<uint16_t, kFrameIdLength>(info->gof->pid_start, picture_id);
|
|
size_t gof_idx = diff % gof_size;
|
|
RTC_CHECK(gof_idx < kMaxVp9FramesInGof);
|
|
|
|
size_t temporal_idx = info->gof->temporal_idx[gof_idx];
|
|
if (temporal_idx >= kMaxTemporalLayers) {
|
|
RTC_LOG(LS_WARNING) << "At most " << kMaxTemporalLayers
|
|
<< " temporal "
|
|
"layers are supported.";
|
|
return;
|
|
}
|
|
|
|
missing_frames_for_layer_[temporal_idx].erase(picture_id);
|
|
}
|
|
}
|
|
|
|
bool RtpVp9RefFinder::UpSwitchInIntervalVp9(uint16_t picture_id,
|
|
uint8_t temporal_idx,
|
|
uint16_t pid_ref) {
|
|
for (auto up_switch_it = up_switch_.upper_bound(pid_ref);
|
|
up_switch_it != up_switch_.end() &&
|
|
AheadOf<uint16_t, kFrameIdLength>(picture_id, up_switch_it->first);
|
|
++up_switch_it) {
|
|
if (up_switch_it->second < temporal_idx)
|
|
return true;
|
|
}
|
|
|
|
return false;
|
|
}
|
|
|
|
void RtpVp9RefFinder::RetryStashedFrames(
|
|
RtpFrameReferenceFinder::ReturnVector& res) {
|
|
bool complete_frame = false;
|
|
do {
|
|
complete_frame = false;
|
|
for (auto it = stashed_frames_.begin(); it != stashed_frames_.end();) {
|
|
const RTPVideoHeaderVP9& codec_header = absl::get<RTPVideoHeaderVP9>(
|
|
it->frame->GetRtpVideoHeader().video_type_header);
|
|
RTC_DCHECK(!codec_header.flexible_mode);
|
|
FrameDecision decision =
|
|
ManageFrameGof(it->frame.get(), codec_header, it->unwrapped_tl0);
|
|
|
|
switch (decision) {
|
|
case kStash:
|
|
++it;
|
|
break;
|
|
case kHandOff:
|
|
complete_frame = true;
|
|
res.push_back(std::move(it->frame));
|
|
[[fallthrough]];
|
|
case kDrop:
|
|
it = stashed_frames_.erase(it);
|
|
}
|
|
}
|
|
} while (complete_frame);
|
|
}
|
|
|
|
void RtpVp9RefFinder::FlattenFrameIdAndRefs(RtpFrameObject* frame,
|
|
bool inter_layer_predicted) {
|
|
for (size_t i = 0; i < frame->num_references; ++i) {
|
|
frame->references[i] =
|
|
unwrapper_.Unwrap(frame->references[i]) * kMaxSpatialLayers +
|
|
*frame->SpatialIndex();
|
|
}
|
|
frame->SetId(unwrapper_.Unwrap(frame->Id()) * kMaxSpatialLayers +
|
|
*frame->SpatialIndex());
|
|
|
|
if (inter_layer_predicted &&
|
|
frame->num_references + 1 <= EncodedFrame::kMaxFrameReferences) {
|
|
frame->references[frame->num_references] = frame->Id() - 1;
|
|
++frame->num_references;
|
|
}
|
|
}
|
|
|
|
void RtpVp9RefFinder::ClearTo(uint16_t seq_num) {
|
|
auto it = stashed_frames_.begin();
|
|
while (it != stashed_frames_.end()) {
|
|
if (AheadOf<uint16_t>(seq_num, it->frame->first_seq_num())) {
|
|
it = stashed_frames_.erase(it);
|
|
} else {
|
|
++it;
|
|
}
|
|
}
|
|
}
|
|
|
|
} // namespace webrtc
|