368 lines
12 KiB
C++
368 lines
12 KiB
C++
|
/*
|
||
|
* Copyright (c) 2011 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/decoding_state.h"
|
||
|
|
||
|
#include "common_video/h264/h264_common.h"
|
||
|
#include "modules/include/module_common_types_public.h"
|
||
|
#include "modules/video_coding/frame_buffer.h"
|
||
|
#include "modules/video_coding/jitter_buffer_common.h"
|
||
|
#include "modules/video_coding/packet.h"
|
||
|
#include "rtc_base/logging.h"
|
||
|
|
||
|
namespace webrtc {
|
||
|
|
||
|
VCMDecodingState::VCMDecodingState()
|
||
|
: sequence_num_(0),
|
||
|
time_stamp_(0),
|
||
|
picture_id_(kNoPictureId),
|
||
|
temporal_id_(kNoTemporalIdx),
|
||
|
tl0_pic_id_(kNoTl0PicIdx),
|
||
|
full_sync_(true),
|
||
|
in_initial_state_(true) {
|
||
|
memset(frame_decoded_, 0, sizeof(frame_decoded_));
|
||
|
}
|
||
|
|
||
|
VCMDecodingState::~VCMDecodingState() {}
|
||
|
|
||
|
void VCMDecodingState::Reset() {
|
||
|
// TODO(mikhal): Verify - not always would want to reset the sync
|
||
|
sequence_num_ = 0;
|
||
|
time_stamp_ = 0;
|
||
|
picture_id_ = kNoPictureId;
|
||
|
temporal_id_ = kNoTemporalIdx;
|
||
|
tl0_pic_id_ = kNoTl0PicIdx;
|
||
|
full_sync_ = true;
|
||
|
in_initial_state_ = true;
|
||
|
memset(frame_decoded_, 0, sizeof(frame_decoded_));
|
||
|
received_sps_.clear();
|
||
|
received_pps_.clear();
|
||
|
}
|
||
|
|
||
|
uint32_t VCMDecodingState::time_stamp() const {
|
||
|
return time_stamp_;
|
||
|
}
|
||
|
|
||
|
uint16_t VCMDecodingState::sequence_num() const {
|
||
|
return sequence_num_;
|
||
|
}
|
||
|
|
||
|
bool VCMDecodingState::IsOldFrame(const VCMFrameBuffer* frame) const {
|
||
|
assert(frame != NULL);
|
||
|
if (in_initial_state_)
|
||
|
return false;
|
||
|
return !IsNewerTimestamp(frame->Timestamp(), time_stamp_);
|
||
|
}
|
||
|
|
||
|
bool VCMDecodingState::IsOldPacket(const VCMPacket* packet) const {
|
||
|
assert(packet != NULL);
|
||
|
if (in_initial_state_)
|
||
|
return false;
|
||
|
return !IsNewerTimestamp(packet->timestamp, time_stamp_);
|
||
|
}
|
||
|
|
||
|
void VCMDecodingState::SetState(const VCMFrameBuffer* frame) {
|
||
|
assert(frame != NULL && frame->GetHighSeqNum() >= 0);
|
||
|
if (!UsingFlexibleMode(frame))
|
||
|
UpdateSyncState(frame);
|
||
|
sequence_num_ = static_cast<uint16_t>(frame->GetHighSeqNum());
|
||
|
time_stamp_ = frame->Timestamp();
|
||
|
picture_id_ = frame->PictureId();
|
||
|
temporal_id_ = frame->TemporalId();
|
||
|
tl0_pic_id_ = frame->Tl0PicId();
|
||
|
|
||
|
for (const NaluInfo& nalu : frame->GetNaluInfos()) {
|
||
|
if (nalu.type == H264::NaluType::kPps) {
|
||
|
if (nalu.pps_id < 0) {
|
||
|
RTC_LOG(LS_WARNING) << "Received pps without pps id.";
|
||
|
} else if (nalu.sps_id < 0) {
|
||
|
RTC_LOG(LS_WARNING) << "Received pps without sps id.";
|
||
|
} else {
|
||
|
received_pps_[nalu.pps_id] = nalu.sps_id;
|
||
|
}
|
||
|
} else if (nalu.type == H264::NaluType::kSps) {
|
||
|
if (nalu.sps_id < 0) {
|
||
|
RTC_LOG(LS_WARNING) << "Received sps without sps id.";
|
||
|
} else {
|
||
|
received_sps_.insert(nalu.sps_id);
|
||
|
}
|
||
|
}
|
||
|
}
|
||
|
|
||
|
if (UsingFlexibleMode(frame)) {
|
||
|
uint16_t frame_index = picture_id_ % kFrameDecodedLength;
|
||
|
if (in_initial_state_) {
|
||
|
frame_decoded_cleared_to_ = frame_index;
|
||
|
} else if (frame->FrameType() == VideoFrameType::kVideoFrameKey) {
|
||
|
memset(frame_decoded_, 0, sizeof(frame_decoded_));
|
||
|
frame_decoded_cleared_to_ = frame_index;
|
||
|
} else {
|
||
|
if (AheadOfFramesDecodedClearedTo(frame_index)) {
|
||
|
while (frame_decoded_cleared_to_ != frame_index) {
|
||
|
frame_decoded_cleared_to_ =
|
||
|
(frame_decoded_cleared_to_ + 1) % kFrameDecodedLength;
|
||
|
frame_decoded_[frame_decoded_cleared_to_] = false;
|
||
|
}
|
||
|
}
|
||
|
}
|
||
|
frame_decoded_[frame_index] = true;
|
||
|
}
|
||
|
|
||
|
in_initial_state_ = false;
|
||
|
}
|
||
|
|
||
|
void VCMDecodingState::CopyFrom(const VCMDecodingState& state) {
|
||
|
sequence_num_ = state.sequence_num_;
|
||
|
time_stamp_ = state.time_stamp_;
|
||
|
picture_id_ = state.picture_id_;
|
||
|
temporal_id_ = state.temporal_id_;
|
||
|
tl0_pic_id_ = state.tl0_pic_id_;
|
||
|
full_sync_ = state.full_sync_;
|
||
|
in_initial_state_ = state.in_initial_state_;
|
||
|
frame_decoded_cleared_to_ = state.frame_decoded_cleared_to_;
|
||
|
memcpy(frame_decoded_, state.frame_decoded_, sizeof(frame_decoded_));
|
||
|
received_sps_ = state.received_sps_;
|
||
|
received_pps_ = state.received_pps_;
|
||
|
}
|
||
|
|
||
|
bool VCMDecodingState::UpdateEmptyFrame(const VCMFrameBuffer* frame) {
|
||
|
bool empty_packet = frame->GetHighSeqNum() == frame->GetLowSeqNum();
|
||
|
if (in_initial_state_ && empty_packet) {
|
||
|
// Drop empty packets as long as we are in the initial state.
|
||
|
return true;
|
||
|
}
|
||
|
if ((empty_packet && ContinuousSeqNum(frame->GetHighSeqNum())) ||
|
||
|
ContinuousFrame(frame)) {
|
||
|
// Continuous empty packets or continuous frames can be dropped if we
|
||
|
// advance the sequence number.
|
||
|
sequence_num_ = frame->GetHighSeqNum();
|
||
|
time_stamp_ = frame->Timestamp();
|
||
|
return true;
|
||
|
}
|
||
|
return false;
|
||
|
}
|
||
|
|
||
|
void VCMDecodingState::UpdateOldPacket(const VCMPacket* packet) {
|
||
|
assert(packet != NULL);
|
||
|
if (packet->timestamp == time_stamp_) {
|
||
|
// Late packet belonging to the last decoded frame - make sure we update the
|
||
|
// last decoded sequence number.
|
||
|
sequence_num_ = LatestSequenceNumber(packet->seqNum, sequence_num_);
|
||
|
}
|
||
|
}
|
||
|
|
||
|
void VCMDecodingState::SetSeqNum(uint16_t new_seq_num) {
|
||
|
sequence_num_ = new_seq_num;
|
||
|
}
|
||
|
|
||
|
bool VCMDecodingState::in_initial_state() const {
|
||
|
return in_initial_state_;
|
||
|
}
|
||
|
|
||
|
bool VCMDecodingState::full_sync() const {
|
||
|
return full_sync_;
|
||
|
}
|
||
|
|
||
|
void VCMDecodingState::UpdateSyncState(const VCMFrameBuffer* frame) {
|
||
|
if (in_initial_state_)
|
||
|
return;
|
||
|
if (frame->TemporalId() == kNoTemporalIdx ||
|
||
|
frame->Tl0PicId() == kNoTl0PicIdx) {
|
||
|
full_sync_ = true;
|
||
|
} else if (frame->FrameType() == VideoFrameType::kVideoFrameKey ||
|
||
|
frame->LayerSync()) {
|
||
|
full_sync_ = true;
|
||
|
} else if (full_sync_) {
|
||
|
// Verify that we are still in sync.
|
||
|
// Sync will be broken if continuity is true for layers but not for the
|
||
|
// other methods (PictureId and SeqNum).
|
||
|
if (UsingPictureId(frame)) {
|
||
|
// First check for a valid tl0PicId.
|
||
|
if (frame->Tl0PicId() - tl0_pic_id_ > 1) {
|
||
|
full_sync_ = false;
|
||
|
} else {
|
||
|
full_sync_ = ContinuousPictureId(frame->PictureId());
|
||
|
}
|
||
|
} else {
|
||
|
full_sync_ =
|
||
|
ContinuousSeqNum(static_cast<uint16_t>(frame->GetLowSeqNum()));
|
||
|
}
|
||
|
}
|
||
|
}
|
||
|
|
||
|
bool VCMDecodingState::ContinuousFrame(const VCMFrameBuffer* frame) const {
|
||
|
// Check continuity based on the following hierarchy:
|
||
|
// - Temporal layers (stop here if out of sync).
|
||
|
// - Picture Id when available.
|
||
|
// - Sequence numbers.
|
||
|
// Return true when in initial state.
|
||
|
// Note that when a method is not applicable it will return false.
|
||
|
assert(frame != NULL);
|
||
|
// A key frame is always considered continuous as it doesn't refer to any
|
||
|
// frames and therefore won't introduce any errors even if prior frames are
|
||
|
// missing.
|
||
|
if (frame->FrameType() == VideoFrameType::kVideoFrameKey &&
|
||
|
HaveSpsAndPps(frame->GetNaluInfos())) {
|
||
|
return true;
|
||
|
}
|
||
|
// When in the initial state we always require a key frame to start decoding.
|
||
|
if (in_initial_state_)
|
||
|
return false;
|
||
|
if (ContinuousLayer(frame->TemporalId(), frame->Tl0PicId()))
|
||
|
return true;
|
||
|
// tl0picId is either not used, or should remain unchanged.
|
||
|
if (frame->Tl0PicId() != tl0_pic_id_)
|
||
|
return false;
|
||
|
// Base layers are not continuous or temporal layers are inactive.
|
||
|
// In the presence of temporal layers, check for Picture ID/sequence number
|
||
|
// continuity if sync can be restored by this frame.
|
||
|
if (!full_sync_ && !frame->LayerSync())
|
||
|
return false;
|
||
|
if (UsingPictureId(frame)) {
|
||
|
if (UsingFlexibleMode(frame)) {
|
||
|
return ContinuousFrameRefs(frame);
|
||
|
} else {
|
||
|
return ContinuousPictureId(frame->PictureId());
|
||
|
}
|
||
|
} else {
|
||
|
return ContinuousSeqNum(static_cast<uint16_t>(frame->GetLowSeqNum())) &&
|
||
|
HaveSpsAndPps(frame->GetNaluInfos());
|
||
|
}
|
||
|
}
|
||
|
|
||
|
bool VCMDecodingState::ContinuousPictureId(int picture_id) const {
|
||
|
int next_picture_id = picture_id_ + 1;
|
||
|
if (picture_id < picture_id_) {
|
||
|
// Wrap
|
||
|
if (picture_id_ >= 0x80) {
|
||
|
// 15 bits used for picture id
|
||
|
return ((next_picture_id & 0x7FFF) == picture_id);
|
||
|
} else {
|
||
|
// 7 bits used for picture id
|
||
|
return ((next_picture_id & 0x7F) == picture_id);
|
||
|
}
|
||
|
}
|
||
|
// No wrap
|
||
|
return (next_picture_id == picture_id);
|
||
|
}
|
||
|
|
||
|
bool VCMDecodingState::ContinuousSeqNum(uint16_t seq_num) const {
|
||
|
return seq_num == static_cast<uint16_t>(sequence_num_ + 1);
|
||
|
}
|
||
|
|
||
|
bool VCMDecodingState::ContinuousLayer(int temporal_id, int tl0_pic_id) const {
|
||
|
// First, check if applicable.
|
||
|
if (temporal_id == kNoTemporalIdx || tl0_pic_id == kNoTl0PicIdx)
|
||
|
return false;
|
||
|
// If this is the first frame to use temporal layers, make sure we start
|
||
|
// from base.
|
||
|
else if (tl0_pic_id_ == kNoTl0PicIdx && temporal_id_ == kNoTemporalIdx &&
|
||
|
temporal_id == 0)
|
||
|
return true;
|
||
|
|
||
|
// Current implementation: Look for base layer continuity.
|
||
|
if (temporal_id != 0)
|
||
|
return false;
|
||
|
return (static_cast<uint8_t>(tl0_pic_id_ + 1) == tl0_pic_id);
|
||
|
}
|
||
|
|
||
|
bool VCMDecodingState::ContinuousFrameRefs(const VCMFrameBuffer* frame) const {
|
||
|
uint8_t num_refs = frame->CodecSpecific()->codecSpecific.VP9.num_ref_pics;
|
||
|
for (uint8_t r = 0; r < num_refs; ++r) {
|
||
|
uint16_t frame_ref = frame->PictureId() -
|
||
|
frame->CodecSpecific()->codecSpecific.VP9.p_diff[r];
|
||
|
uint16_t frame_index = frame_ref % kFrameDecodedLength;
|
||
|
if (AheadOfFramesDecodedClearedTo(frame_index) ||
|
||
|
!frame_decoded_[frame_index]) {
|
||
|
return false;
|
||
|
}
|
||
|
}
|
||
|
return true;
|
||
|
}
|
||
|
|
||
|
bool VCMDecodingState::UsingPictureId(const VCMFrameBuffer* frame) const {
|
||
|
return (frame->PictureId() != kNoPictureId && picture_id_ != kNoPictureId);
|
||
|
}
|
||
|
|
||
|
bool VCMDecodingState::UsingFlexibleMode(const VCMFrameBuffer* frame) const {
|
||
|
bool is_flexible_mode =
|
||
|
frame->CodecSpecific()->codecType == kVideoCodecVP9 &&
|
||
|
frame->CodecSpecific()->codecSpecific.VP9.flexible_mode;
|
||
|
if (is_flexible_mode && frame->PictureId() == kNoPictureId) {
|
||
|
RTC_LOG(LS_WARNING) << "Frame is marked as using flexible mode but no"
|
||
|
"picture id is set.";
|
||
|
return false;
|
||
|
}
|
||
|
return is_flexible_mode;
|
||
|
}
|
||
|
|
||
|
// TODO(philipel): change how check work, this check practially
|
||
|
// limits the max p_diff to 64.
|
||
|
bool VCMDecodingState::AheadOfFramesDecodedClearedTo(uint16_t index) const {
|
||
|
// No way of knowing for sure if we are actually ahead of
|
||
|
// frame_decoded_cleared_to_. We just make the assumption
|
||
|
// that we are not trying to reference back to a very old
|
||
|
// index, but instead are referencing a newer index.
|
||
|
uint16_t diff =
|
||
|
index > frame_decoded_cleared_to_
|
||
|
? kFrameDecodedLength - (index - frame_decoded_cleared_to_)
|
||
|
: frame_decoded_cleared_to_ - index;
|
||
|
return diff > kFrameDecodedLength / 2;
|
||
|
}
|
||
|
|
||
|
bool VCMDecodingState::HaveSpsAndPps(const std::vector<NaluInfo>& nalus) const {
|
||
|
std::set<int> new_sps;
|
||
|
std::map<int, int> new_pps;
|
||
|
for (const NaluInfo& nalu : nalus) {
|
||
|
// Check if this nalu actually contains sps/pps information or dependencies.
|
||
|
if (nalu.sps_id == -1 && nalu.pps_id == -1)
|
||
|
continue;
|
||
|
switch (nalu.type) {
|
||
|
case H264::NaluType::kPps:
|
||
|
if (nalu.pps_id < 0) {
|
||
|
RTC_LOG(LS_WARNING) << "Received pps without pps id.";
|
||
|
} else if (nalu.sps_id < 0) {
|
||
|
RTC_LOG(LS_WARNING) << "Received pps without sps id.";
|
||
|
} else {
|
||
|
new_pps[nalu.pps_id] = nalu.sps_id;
|
||
|
}
|
||
|
break;
|
||
|
case H264::NaluType::kSps:
|
||
|
if (nalu.sps_id < 0) {
|
||
|
RTC_LOG(LS_WARNING) << "Received sps without sps id.";
|
||
|
} else {
|
||
|
new_sps.insert(nalu.sps_id);
|
||
|
}
|
||
|
break;
|
||
|
default: {
|
||
|
int needed_sps = -1;
|
||
|
auto pps_it = new_pps.find(nalu.pps_id);
|
||
|
if (pps_it != new_pps.end()) {
|
||
|
needed_sps = pps_it->second;
|
||
|
} else {
|
||
|
auto pps_it2 = received_pps_.find(nalu.pps_id);
|
||
|
if (pps_it2 == received_pps_.end()) {
|
||
|
return false;
|
||
|
}
|
||
|
needed_sps = pps_it2->second;
|
||
|
}
|
||
|
if (new_sps.find(needed_sps) == new_sps.end() &&
|
||
|
received_sps_.find(needed_sps) == received_sps_.end()) {
|
||
|
return false;
|
||
|
}
|
||
|
break;
|
||
|
}
|
||
|
}
|
||
|
}
|
||
|
return true;
|
||
|
}
|
||
|
|
||
|
} // namespace webrtc
|