webrtc/talk/media/base/hybridvideoengine.h

286 lines
10 KiB
C++

/*
* libjingle
* Copyright 2004 Google Inc.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions are met:
*
* 1. Redistributions of source code must retain the above copyright notice,
* this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright notice,
* this list of conditions and the following disclaimer in the documentation
* and/or other materials provided with the distribution.
* 3. The name of the author may not be used to endorse or promote products
* derived from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR IMPLIED
* WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
* MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO
* EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
* SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
* PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS;
* OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY,
* WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR
* OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF
* ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
#ifndef TALK_MEDIA_BASE_HYBRIDVIDEOENGINE_H_
#define TALK_MEDIA_BASE_HYBRIDVIDEOENGINE_H_
#include <string>
#include <vector>
#include "talk/base/logging.h"
#include "talk/base/sigslotrepeater.h"
#include "talk/media/base/codec.h"
#include "talk/media/base/mediachannel.h"
#include "talk/media/base/videocapturer.h"
#include "talk/media/base/videocommon.h"
namespace cricket {
struct Device;
struct VideoFormat;
class HybridVideoEngineInterface;
class VideoCapturer;
class VideoFrame;
class VideoRenderer;
// HybridVideoMediaChannels work with a HybridVideoEngine to combine
// two unrelated VideoMediaChannel implementations into a single class.
class HybridVideoMediaChannel : public VideoMediaChannel {
public:
HybridVideoMediaChannel(HybridVideoEngineInterface* engine,
VideoMediaChannel* channel1,
VideoMediaChannel* channel2);
virtual ~HybridVideoMediaChannel();
// VideoMediaChannel methods
virtual void SetInterface(NetworkInterface* iface);
virtual bool SetOptions(const VideoOptions& options);
virtual bool GetOptions(VideoOptions* options) const;
virtual bool AddSendStream(const StreamParams& sp);
virtual bool RemoveSendStream(uint32 ssrc);
virtual bool SetRenderer(uint32 ssrc, VideoRenderer* renderer);
virtual bool SetRender(bool render);
virtual bool MuteStream(uint32 ssrc, bool muted);
virtual bool SetRecvCodecs(const std::vector<VideoCodec>& codecs);
virtual bool SetRecvRtpHeaderExtensions(
const std::vector<RtpHeaderExtension>& extensions);
virtual bool SetSendCodecs(const std::vector<VideoCodec>& codecs);
virtual bool GetSendCodec(VideoCodec* codec);
virtual bool SetSendStreamFormat(uint32 ssrc, const VideoFormat& format);
virtual bool SetSendRtpHeaderExtensions(
const std::vector<RtpHeaderExtension>& extensions);
virtual bool SetSendBandwidth(bool autobw, int bps);
virtual bool SetSend(bool send);
virtual bool AddRecvStream(const StreamParams& sp);
virtual bool RemoveRecvStream(uint32 ssrc);
virtual bool SetCapturer(uint32 ssrc, VideoCapturer* capturer);
virtual bool SendIntraFrame();
virtual bool RequestIntraFrame();
virtual bool GetStats(VideoMediaInfo* info);
virtual void OnPacketReceived(talk_base::Buffer* packet,
const talk_base::PacketTime& packet_time);
virtual void OnRtcpReceived(talk_base::Buffer* packet,
const talk_base::PacketTime& packet_time);
virtual void OnReadyToSend(bool ready);
virtual void UpdateAspectRatio(int ratio_w, int ratio_h);
void OnLocalFrame(VideoCapturer*, const VideoFrame*);
void OnLocalFrameFormat(VideoCapturer*, const VideoFormat*);
bool sending() const { return sending_; }
private:
bool SelectActiveChannel(const std::vector<VideoCodec>& codecs);
void SplitCodecs(const std::vector<VideoCodec>& codecs,
std::vector<VideoCodec>* codecs1,
std::vector<VideoCodec>* codecs2);
void OnMediaError(uint32 ssrc, Error error);
HybridVideoEngineInterface* engine_;
talk_base::scoped_ptr<VideoMediaChannel> channel1_;
talk_base::scoped_ptr<VideoMediaChannel> channel2_;
VideoMediaChannel* active_channel_;
bool sending_;
};
// Interface class for HybridVideoChannels to talk to the engine.
class HybridVideoEngineInterface {
public:
virtual ~HybridVideoEngineInterface() {}
virtual bool HasCodec1(const VideoCodec& codec) = 0;
virtual bool HasCodec2(const VideoCodec& codec) = 0;
virtual void OnSendChange1(VideoMediaChannel* channel1, bool send) = 0;
virtual void OnSendChange2(VideoMediaChannel* channel1, bool send) = 0;
virtual void OnNewSendResolution(int width, int height) = 0;
};
// The HybridVideoEngine class combines two unrelated VideoEngine impls
// into a single class. It creates HybridVideoMediaChannels that also contain
// a VideoMediaChannel implementation from each engine. Policy is then used
// during call setup to determine which VideoMediaChannel should be used.
// Currently, this policy is based on what codec the remote side wants to use.
template<class VIDEO1, class VIDEO2>
class HybridVideoEngine : public HybridVideoEngineInterface {
public:
HybridVideoEngine() {
// Unify the codec lists.
codecs_ = video1_.codecs();
codecs_.insert(codecs_.end(), video2_.codecs().begin(),
video2_.codecs().end());
rtp_header_extensions_ = video1_.rtp_header_extensions();
rtp_header_extensions_.insert(rtp_header_extensions_.end(),
video2_.rtp_header_extensions().begin(),
video2_.rtp_header_extensions().end());
SignalCaptureStateChange.repeat(video2_.SignalCaptureStateChange);
}
bool Init(talk_base::Thread* worker_thread) {
if (!video1_.Init(worker_thread)) {
LOG(LS_ERROR) << "Failed to init VideoEngine1";
return false;
}
if (!video2_.Init(worker_thread)) {
LOG(LS_ERROR) << "Failed to init VideoEngine2";
video1_.Terminate();
return false;
}
return true;
}
void Terminate() {
video1_.Terminate();
video2_.Terminate();
}
int GetCapabilities() {
return (video1_.GetCapabilities() | video2_.GetCapabilities());
}
HybridVideoMediaChannel* CreateChannel(VoiceMediaChannel* channel) {
talk_base::scoped_ptr<VideoMediaChannel> channel1(
video1_.CreateChannel(channel));
if (!channel1) {
LOG(LS_ERROR) << "Failed to create VideoMediaChannel1";
return NULL;
}
talk_base::scoped_ptr<VideoMediaChannel> channel2(
video2_.CreateChannel(channel));
if (!channel2) {
LOG(LS_ERROR) << "Failed to create VideoMediaChannel2";
return NULL;
}
return new HybridVideoMediaChannel(this,
channel1.release(), channel2.release());
}
bool SetOptions(const VideoOptions& options) {
return video1_.SetOptions(options) && video2_.SetOptions(options);
}
bool SetDefaultEncoderConfig(const VideoEncoderConfig& config) {
VideoEncoderConfig conf = config;
if (video1_.codecs().size() > 0) {
conf.max_codec.name = video1_.codecs()[0].name;
if (!video1_.SetDefaultEncoderConfig(conf)) {
LOG(LS_ERROR) << "Failed to SetDefaultEncoderConfig for video1";
return false;
}
}
if (video2_.codecs().size() > 0) {
conf.max_codec.name = video2_.codecs()[0].name;
if (!video2_.SetDefaultEncoderConfig(conf)) {
LOG(LS_ERROR) << "Failed to SetDefaultEncoderConfig for video2";
return false;
}
}
return true;
}
VideoEncoderConfig GetDefaultEncoderConfig() const {
// This looks pretty strange, but, in practice, it'll do sane things if
// GetDefaultEncoderConfig is only called after SetDefaultEncoderConfig,
// since both engines should be essentially equivalent at that point. If it
// hasn't been called, though, we'll use the first meaningful encoder
// config, or the config from the second video engine if neither are
// meaningful.
VideoEncoderConfig config = video1_.GetDefaultEncoderConfig();
if (config.max_codec.width != 0) {
return config;
} else {
return video2_.GetDefaultEncoderConfig();
}
}
const std::vector<VideoCodec>& codecs() const {
return codecs_;
}
const std::vector<RtpHeaderExtension>& rtp_header_extensions() const {
return rtp_header_extensions_;
}
void SetLogging(int min_sev, const char* filter) {
video1_.SetLogging(min_sev, filter);
video2_.SetLogging(min_sev, filter);
}
VideoFormat GetStartCaptureFormat() const {
return video2_.GetStartCaptureFormat();
}
// TODO(juberti): Remove these functions after we do the capturer refactoring.
// For now they are set to always use the second engine for capturing, which
// is convenient given our intended use case.
bool SetCaptureDevice(const Device* device) {
return video2_.SetCaptureDevice(device);
}
VideoCapturer* GetVideoCapturer() const {
return video2_.GetVideoCapturer();
}
bool SetLocalRenderer(VideoRenderer* renderer) {
return video2_.SetLocalRenderer(renderer);
}
sigslot::repeater2<VideoCapturer*, CaptureState> SignalCaptureStateChange;
virtual bool HasCodec1(const VideoCodec& codec) {
return HasCodec(video1_, codec);
}
virtual bool HasCodec2(const VideoCodec& codec) {
return HasCodec(video2_, codec);
}
template<typename VIDEO>
bool HasCodec(const VIDEO& engine, const VideoCodec& codec) const {
for (std::vector<VideoCodec>::const_iterator i = engine.codecs().begin();
i != engine.codecs().end();
++i) {
if (i->Matches(codec)) {
return true;
}
}
return false;
}
virtual void OnSendChange1(VideoMediaChannel* channel1, bool send) {
}
virtual void OnSendChange2(VideoMediaChannel* channel2, bool send) {
}
virtual void OnNewSendResolution(int width, int height) {
}
protected:
VIDEO1 video1_;
VIDEO2 video2_;
std::vector<VideoCodec> codecs_;
std::vector<RtpHeaderExtension> rtp_header_extensions_;
};
} // namespace cricket
#endif // TALK_MEDIA_BASE_HYBRIDVIDEOENGINE_H_