3 * Copyright 2004 Google Inc.
5 * Redistribution and use in source and binary forms, with or without
6 * modification, are permitted provided that the following conditions are met:
8 * 1. Redistributions of source code must retain the above copyright notice,
9 * this list of conditions and the following disclaimer.
10 * 2. Redistributions in binary form must reproduce the above copyright notice,
11 * this list of conditions and the following disclaimer in the documentation
12 * and/or other materials provided with the distribution.
13 * 3. The name of the author may not be used to endorse or promote products
14 * derived from this software without specific prior written permission.
16 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR IMPLIED
17 * WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
18 * MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO
19 * EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
20 * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
21 * PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS;
22 * OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY,
23 * WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR
24 * OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF
25 * ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
28 #ifndef TALK_MEDIA_BASE_MEDIACHANNEL_H_
29 #define TALK_MEDIA_BASE_MEDIACHANNEL_H_
34 #include "talk/base/basictypes.h"
35 #include "talk/base/buffer.h"
36 #include "talk/base/dscp.h"
37 #include "talk/base/logging.h"
38 #include "talk/base/sigslot.h"
39 #include "talk/base/socket.h"
40 #include "talk/base/window.h"
41 #include "talk/media/base/codec.h"
42 #include "talk/media/base/constants.h"
43 #include "talk/media/base/streamparams.h"
44 // TODO(juberti): re-evaluate this include
45 #include "talk/session/media/audiomonitor.h"
62 const int kMinRtpHeaderExtensionId = 1;
63 const int kMaxRtpHeaderExtensionId = 255;
64 const int kScreencastDefaultFps = 5;
65 const int kHighStartBitrate = 1500;
67 // Used in AudioOptions and VideoOptions to signify "unset" values.
71 Settable() : set_(false), val_() {}
72 explicit Settable(T val) : set_(true), val_(val) {}
78 bool Get(T* out) const {
83 T GetWithDefaultIfUnset(const T& default_value) const {
84 return set_ ? val_ : default_value;
87 virtual void Set(T val) {
97 void SetFrom(const Settable<T>& o) {
98 // Set this value based on the value of o, iff o is set. If this value is
99 // set and o is unset, the current value will be unchanged.
106 std::string ToString() const {
107 return set_ ? talk_base::ToString(val_) : "";
110 bool operator==(const Settable<T>& o) const {
111 // Equal if both are unset with any value or both set with the same value.
112 return (set_ == o.set_) && (!set_ || (val_ == o.val_));
115 bool operator!=(const Settable<T>& o) const {
116 return !operator==(o);
120 void InitializeValue(const T &val) {
129 class SettablePercent : public Settable<float> {
131 virtual void Set(float val) {
138 Settable<float>::Set(val);
143 static std::string ToStringIfSet(const char* key, const Settable<T>& val) {
148 str += val.ToString();
154 // Options that can be applied to a VoiceMediaChannel or a VoiceMediaEngine.
155 // Used to be flags, but that makes it hard to selectively apply options.
156 // We are moving all of the setting of options to structs like this,
157 // but some things currently still use flags.
158 struct AudioOptions {
159 void SetAll(const AudioOptions& change) {
160 echo_cancellation.SetFrom(change.echo_cancellation);
161 auto_gain_control.SetFrom(change.auto_gain_control);
162 rx_auto_gain_control.SetFrom(change.rx_auto_gain_control);
163 noise_suppression.SetFrom(change.noise_suppression);
164 highpass_filter.SetFrom(change.highpass_filter);
165 stereo_swapping.SetFrom(change.stereo_swapping);
166 typing_detection.SetFrom(change.typing_detection);
167 aecm_generate_comfort_noise.SetFrom(change.aecm_generate_comfort_noise);
168 conference_mode.SetFrom(change.conference_mode);
169 adjust_agc_delta.SetFrom(change.adjust_agc_delta);
170 experimental_agc.SetFrom(change.experimental_agc);
171 experimental_aec.SetFrom(change.experimental_aec);
172 experimental_ns.SetFrom(change.experimental_ns);
173 aec_dump.SetFrom(change.aec_dump);
174 tx_agc_target_dbov.SetFrom(change.tx_agc_target_dbov);
175 tx_agc_digital_compression_gain.SetFrom(
176 change.tx_agc_digital_compression_gain);
177 tx_agc_limiter.SetFrom(change.tx_agc_limiter);
178 rx_agc_target_dbov.SetFrom(change.rx_agc_target_dbov);
179 rx_agc_digital_compression_gain.SetFrom(
180 change.rx_agc_digital_compression_gain);
181 rx_agc_limiter.SetFrom(change.rx_agc_limiter);
182 recording_sample_rate.SetFrom(change.recording_sample_rate);
183 playout_sample_rate.SetFrom(change.playout_sample_rate);
184 dscp.SetFrom(change.dscp);
185 opus_fec.SetFrom(change.opus_fec);
188 bool operator==(const AudioOptions& o) const {
189 return echo_cancellation == o.echo_cancellation &&
190 auto_gain_control == o.auto_gain_control &&
191 rx_auto_gain_control == o.rx_auto_gain_control &&
192 noise_suppression == o.noise_suppression &&
193 highpass_filter == o.highpass_filter &&
194 stereo_swapping == o.stereo_swapping &&
195 typing_detection == o.typing_detection &&
196 aecm_generate_comfort_noise == o.aecm_generate_comfort_noise &&
197 conference_mode == o.conference_mode &&
198 experimental_agc == o.experimental_agc &&
199 experimental_aec == o.experimental_aec &&
200 experimental_ns == o.experimental_ns &&
201 adjust_agc_delta == o.adjust_agc_delta &&
202 aec_dump == o.aec_dump &&
203 tx_agc_target_dbov == o.tx_agc_target_dbov &&
204 tx_agc_digital_compression_gain == o.tx_agc_digital_compression_gain &&
205 tx_agc_limiter == o.tx_agc_limiter &&
206 rx_agc_target_dbov == o.rx_agc_target_dbov &&
207 rx_agc_digital_compression_gain == o.rx_agc_digital_compression_gain &&
208 rx_agc_limiter == o.rx_agc_limiter &&
209 recording_sample_rate == o.recording_sample_rate &&
210 playout_sample_rate == o.playout_sample_rate &&
212 opus_fec == o.opus_fec;
215 std::string ToString() const {
216 std::ostringstream ost;
217 ost << "AudioOptions {";
218 ost << ToStringIfSet("aec", echo_cancellation);
219 ost << ToStringIfSet("agc", auto_gain_control);
220 ost << ToStringIfSet("rx_agc", rx_auto_gain_control);
221 ost << ToStringIfSet("ns", noise_suppression);
222 ost << ToStringIfSet("hf", highpass_filter);
223 ost << ToStringIfSet("swap", stereo_swapping);
224 ost << ToStringIfSet("typing", typing_detection);
225 ost << ToStringIfSet("comfort_noise", aecm_generate_comfort_noise);
226 ost << ToStringIfSet("conference", conference_mode);
227 ost << ToStringIfSet("agc_delta", adjust_agc_delta);
228 ost << ToStringIfSet("experimental_agc", experimental_agc);
229 ost << ToStringIfSet("experimental_aec", experimental_aec);
230 ost << ToStringIfSet("experimental_ns", experimental_ns);
231 ost << ToStringIfSet("aec_dump", aec_dump);
232 ost << ToStringIfSet("tx_agc_target_dbov", tx_agc_target_dbov);
233 ost << ToStringIfSet("tx_agc_digital_compression_gain",
234 tx_agc_digital_compression_gain);
235 ost << ToStringIfSet("tx_agc_limiter", tx_agc_limiter);
236 ost << ToStringIfSet("rx_agc_target_dbov", rx_agc_target_dbov);
237 ost << ToStringIfSet("rx_agc_digital_compression_gain",
238 rx_agc_digital_compression_gain);
239 ost << ToStringIfSet("rx_agc_limiter", rx_agc_limiter);
240 ost << ToStringIfSet("recording_sample_rate", recording_sample_rate);
241 ost << ToStringIfSet("playout_sample_rate", playout_sample_rate);
242 ost << ToStringIfSet("dscp", dscp);
243 ost << ToStringIfSet("opus_fec", opus_fec);
248 // Audio processing that attempts to filter away the output signal from
249 // later inbound pickup.
250 Settable<bool> echo_cancellation;
251 // Audio processing to adjust the sensitivity of the local mic dynamically.
252 Settable<bool> auto_gain_control;
253 // Audio processing to apply gain to the remote audio.
254 Settable<bool> rx_auto_gain_control;
255 // Audio processing to filter out background noise.
256 Settable<bool> noise_suppression;
257 // Audio processing to remove background noise of lower frequencies.
258 Settable<bool> highpass_filter;
259 // Audio processing to swap the left and right channels.
260 Settable<bool> stereo_swapping;
261 // Audio processing to detect typing.
262 Settable<bool> typing_detection;
263 Settable<bool> aecm_generate_comfort_noise;
264 Settable<bool> conference_mode;
265 Settable<int> adjust_agc_delta;
266 Settable<bool> experimental_agc;
267 Settable<bool> experimental_aec;
268 Settable<bool> experimental_ns;
269 Settable<bool> aec_dump;
270 // Note that tx_agc_* only applies to non-experimental AGC.
271 Settable<uint16> tx_agc_target_dbov;
272 Settable<uint16> tx_agc_digital_compression_gain;
273 Settable<bool> tx_agc_limiter;
274 Settable<uint16> rx_agc_target_dbov;
275 Settable<uint16> rx_agc_digital_compression_gain;
276 Settable<bool> rx_agc_limiter;
277 Settable<uint32> recording_sample_rate;
278 Settable<uint32> playout_sample_rate;
279 // Set DSCP value for packet sent from audio channel.
282 Settable<bool> opus_fec;
285 // Options that can be applied to a VideoMediaChannel or a VideoMediaEngine.
286 // Used to be flags, but that makes it hard to selectively apply options.
287 // We are moving all of the setting of options to structs like this,
288 // but some things currently still use flags.
289 struct VideoOptions {
290 enum HighestBitrate {
297 process_adaptation_threshhold.Set(kProcessCpuThreshold);
298 system_low_adaptation_threshhold.Set(kLowSystemCpuThreshold);
299 system_high_adaptation_threshhold.Set(kHighSystemCpuThreshold);
300 unsignalled_recv_stream_limit.Set(kNumDefaultUnsignalledVideoRecvStreams);
303 void SetAll(const VideoOptions& change) {
304 adapt_input_to_encoder.SetFrom(change.adapt_input_to_encoder);
305 adapt_input_to_cpu_usage.SetFrom(change.adapt_input_to_cpu_usage);
306 adapt_cpu_with_smoothing.SetFrom(change.adapt_cpu_with_smoothing);
307 adapt_view_switch.SetFrom(change.adapt_view_switch);
308 video_adapt_third.SetFrom(change.video_adapt_third);
309 video_noise_reduction.SetFrom(change.video_noise_reduction);
310 video_one_layer_screencast.SetFrom(change.video_one_layer_screencast);
311 video_high_bitrate.SetFrom(change.video_high_bitrate);
312 video_start_bitrate.SetFrom(change.video_start_bitrate);
313 video_temporal_layer_screencast.SetFrom(
314 change.video_temporal_layer_screencast);
315 video_temporal_layer_realtime.SetFrom(
316 change.video_temporal_layer_realtime);
317 video_leaky_bucket.SetFrom(change.video_leaky_bucket);
318 video_highest_bitrate.SetFrom(change.video_highest_bitrate);
319 cpu_overuse_detection.SetFrom(change.cpu_overuse_detection);
320 cpu_underuse_threshold.SetFrom(change.cpu_underuse_threshold);
321 cpu_overuse_threshold.SetFrom(change.cpu_overuse_threshold);
322 cpu_underuse_encode_rsd_threshold.SetFrom(
323 change.cpu_underuse_encode_rsd_threshold);
324 cpu_overuse_encode_rsd_threshold.SetFrom(
325 change.cpu_overuse_encode_rsd_threshold);
326 cpu_overuse_encode_usage.SetFrom(change.cpu_overuse_encode_usage);
327 conference_mode.SetFrom(change.conference_mode);
328 process_adaptation_threshhold.SetFrom(change.process_adaptation_threshhold);
329 system_low_adaptation_threshhold.SetFrom(
330 change.system_low_adaptation_threshhold);
331 system_high_adaptation_threshhold.SetFrom(
332 change.system_high_adaptation_threshhold);
333 buffered_mode_latency.SetFrom(change.buffered_mode_latency);
334 lower_min_bitrate.SetFrom(change.lower_min_bitrate);
335 dscp.SetFrom(change.dscp);
336 suspend_below_min_bitrate.SetFrom(change.suspend_below_min_bitrate);
337 unsignalled_recv_stream_limit.SetFrom(change.unsignalled_recv_stream_limit);
338 use_simulcast_adapter.SetFrom(change.use_simulcast_adapter);
339 skip_encoding_unused_streams.SetFrom(change.skip_encoding_unused_streams);
340 screencast_min_bitrate.SetFrom(change.screencast_min_bitrate);
341 use_improved_wifi_bandwidth_estimator.SetFrom(
342 change.use_improved_wifi_bandwidth_estimator);
343 use_payload_padding.SetFrom(change.use_payload_padding);
346 bool operator==(const VideoOptions& o) const {
347 return adapt_input_to_encoder == o.adapt_input_to_encoder &&
348 adapt_input_to_cpu_usage == o.adapt_input_to_cpu_usage &&
349 adapt_cpu_with_smoothing == o.adapt_cpu_with_smoothing &&
350 adapt_view_switch == o.adapt_view_switch &&
351 video_adapt_third == o.video_adapt_third &&
352 video_noise_reduction == o.video_noise_reduction &&
353 video_one_layer_screencast == o.video_one_layer_screencast &&
354 video_high_bitrate == o.video_high_bitrate &&
355 video_start_bitrate == o.video_start_bitrate &&
356 video_temporal_layer_screencast == o.video_temporal_layer_screencast &&
357 video_temporal_layer_realtime == o.video_temporal_layer_realtime &&
358 video_leaky_bucket == o.video_leaky_bucket &&
359 video_highest_bitrate == o.video_highest_bitrate &&
360 cpu_overuse_detection == o.cpu_overuse_detection &&
361 cpu_underuse_threshold == o.cpu_underuse_threshold &&
362 cpu_overuse_threshold == o.cpu_overuse_threshold &&
363 cpu_underuse_encode_rsd_threshold ==
364 o.cpu_underuse_encode_rsd_threshold &&
365 cpu_overuse_encode_rsd_threshold ==
366 o.cpu_overuse_encode_rsd_threshold &&
367 cpu_overuse_encode_usage == o.cpu_overuse_encode_usage &&
368 conference_mode == o.conference_mode &&
369 process_adaptation_threshhold == o.process_adaptation_threshhold &&
370 system_low_adaptation_threshhold ==
371 o.system_low_adaptation_threshhold &&
372 system_high_adaptation_threshhold ==
373 o.system_high_adaptation_threshhold &&
374 buffered_mode_latency == o.buffered_mode_latency &&
375 lower_min_bitrate == o.lower_min_bitrate &&
377 suspend_below_min_bitrate == o.suspend_below_min_bitrate &&
378 unsignalled_recv_stream_limit == o.unsignalled_recv_stream_limit &&
379 use_simulcast_adapter == o.use_simulcast_adapter &&
380 skip_encoding_unused_streams == o.skip_encoding_unused_streams &&
381 screencast_min_bitrate == o.screencast_min_bitrate &&
382 use_improved_wifi_bandwidth_estimator ==
383 o.use_improved_wifi_bandwidth_estimator &&
384 use_payload_padding == o.use_payload_padding;
387 std::string ToString() const {
388 std::ostringstream ost;
389 ost << "VideoOptions {";
390 ost << ToStringIfSet("encoder adaption", adapt_input_to_encoder);
391 ost << ToStringIfSet("cpu adaption", adapt_input_to_cpu_usage);
392 ost << ToStringIfSet("cpu adaptation smoothing", adapt_cpu_with_smoothing);
393 ost << ToStringIfSet("adapt view switch", adapt_view_switch);
394 ost << ToStringIfSet("video adapt third", video_adapt_third);
395 ost << ToStringIfSet("noise reduction", video_noise_reduction);
396 ost << ToStringIfSet("1 layer screencast", video_one_layer_screencast);
397 ost << ToStringIfSet("high bitrate", video_high_bitrate);
398 ost << ToStringIfSet("start bitrate", video_start_bitrate);
399 ost << ToStringIfSet("video temporal layer screencast",
400 video_temporal_layer_screencast);
401 ost << ToStringIfSet("video temporal layer realtime",
402 video_temporal_layer_realtime);
403 ost << ToStringIfSet("leaky bucket", video_leaky_bucket);
404 ost << ToStringIfSet("highest video bitrate", video_highest_bitrate);
405 ost << ToStringIfSet("cpu overuse detection", cpu_overuse_detection);
406 ost << ToStringIfSet("cpu underuse threshold", cpu_underuse_threshold);
407 ost << ToStringIfSet("cpu overuse threshold", cpu_overuse_threshold);
408 ost << ToStringIfSet("cpu underuse encode rsd threshold",
409 cpu_underuse_encode_rsd_threshold);
410 ost << ToStringIfSet("cpu overuse encode rsd threshold",
411 cpu_overuse_encode_rsd_threshold);
412 ost << ToStringIfSet("cpu overuse encode usage",
413 cpu_overuse_encode_usage);
414 ost << ToStringIfSet("conference mode", conference_mode);
415 ost << ToStringIfSet("process", process_adaptation_threshhold);
416 ost << ToStringIfSet("low", system_low_adaptation_threshhold);
417 ost << ToStringIfSet("high", system_high_adaptation_threshhold);
418 ost << ToStringIfSet("buffered mode latency", buffered_mode_latency);
419 ost << ToStringIfSet("lower min bitrate", lower_min_bitrate);
420 ost << ToStringIfSet("dscp", dscp);
421 ost << ToStringIfSet("suspend below min bitrate",
422 suspend_below_min_bitrate);
423 ost << ToStringIfSet("num channels for early receive",
424 unsignalled_recv_stream_limit);
425 ost << ToStringIfSet("use simulcast adapter", use_simulcast_adapter);
426 ost << ToStringIfSet("skip encoding unused streams",
427 skip_encoding_unused_streams);
428 ost << ToStringIfSet("screencast min bitrate", screencast_min_bitrate);
429 ost << ToStringIfSet("improved wifi bwe",
430 use_improved_wifi_bandwidth_estimator);
431 ost << ToStringIfSet("payload padding", use_payload_padding);
436 // Encoder adaption, which is the gd callback in LMI, and TBA in WebRTC.
437 Settable<bool> adapt_input_to_encoder;
438 // Enable CPU adaptation?
439 Settable<bool> adapt_input_to_cpu_usage;
440 // Enable CPU adaptation smoothing?
441 Settable<bool> adapt_cpu_with_smoothing;
442 // Enable Adapt View Switch?
443 Settable<bool> adapt_view_switch;
444 // Enable video adapt third?
445 Settable<bool> video_adapt_third;
447 Settable<bool> video_noise_reduction;
448 // Experimental: Enable one layer screencast?
449 Settable<bool> video_one_layer_screencast;
450 // Experimental: Enable WebRtc higher bitrate?
451 Settable<bool> video_high_bitrate;
452 // Experimental: Enable WebRtc higher start bitrate?
453 Settable<int> video_start_bitrate;
454 // Experimental: Enable WebRTC layered screencast.
455 Settable<bool> video_temporal_layer_screencast;
456 // Experimental: Enable WebRTC temporal layer strategy for realtime video.
457 Settable<bool> video_temporal_layer_realtime;
458 // Enable WebRTC leaky bucket when sending media packets.
459 Settable<bool> video_leaky_bucket;
460 // Set highest bitrate mode for video.
461 Settable<HighestBitrate> video_highest_bitrate;
462 // Enable WebRTC Cpu Overuse Detection, which is a new version of the CPU
463 // adaptation algorithm. So this option will override the
464 // |adapt_input_to_cpu_usage|.
465 Settable<bool> cpu_overuse_detection;
466 // Low threshold (t1) for cpu overuse adaptation. (Adapt up)
467 // Metric: encode usage (m1). m1 < t1 => underuse.
468 Settable<int> cpu_underuse_threshold;
469 // High threshold (t1) for cpu overuse adaptation. (Adapt down)
470 // Metric: encode usage (m1). m1 > t1 => overuse.
471 Settable<int> cpu_overuse_threshold;
472 // Low threshold (t2) for cpu overuse adaptation. (Adapt up)
473 // Metric: relative standard deviation of encode time (m2).
474 // Optional threshold. If set, (m1 < t1 && m2 < t2) => underuse.
475 // Note: t2 will have no effect if t1 is not set.
476 Settable<int> cpu_underuse_encode_rsd_threshold;
477 // High threshold (t2) for cpu overuse adaptation. (Adapt down)
478 // Metric: relative standard deviation of encode time (m2).
479 // Optional threshold. If set, (m1 > t1 || m2 > t2) => overuse.
480 // Note: t2 will have no effect if t1 is not set.
481 Settable<int> cpu_overuse_encode_rsd_threshold;
482 // Use encode usage for cpu detection.
483 Settable<bool> cpu_overuse_encode_usage;
484 // Use conference mode?
485 Settable<bool> conference_mode;
486 // Threshhold for process cpu adaptation. (Process limit)
487 SettablePercent process_adaptation_threshhold;
488 // Low threshhold for cpu adaptation. (Adapt up)
489 SettablePercent system_low_adaptation_threshhold;
490 // High threshhold for cpu adaptation. (Adapt down)
491 SettablePercent system_high_adaptation_threshhold;
492 // Specify buffered mode latency in milliseconds.
493 Settable<int> buffered_mode_latency;
494 // Make minimum configured send bitrate even lower than usual, at 30kbit.
495 Settable<bool> lower_min_bitrate;
496 // Set DSCP value for packet sent from video channel.
498 // Enable WebRTC suspension of video. No video frames will be sent when the
499 // bitrate is below the configured minimum bitrate.
500 Settable<bool> suspend_below_min_bitrate;
501 // Limit on the number of early receive channels that can be created.
502 Settable<int> unsignalled_recv_stream_limit;
503 // Enable use of simulcast adapter.
504 Settable<bool> use_simulcast_adapter;
505 // Enables the encoder to skip encoding stream not actually sent due to too
506 // low available bit rate.
507 Settable<bool> skip_encoding_unused_streams;
508 // Force screencast to use a minimum bitrate
509 Settable<int> screencast_min_bitrate;
510 // Enable improved bandwidth estiamtor on wifi.
511 Settable<bool> use_improved_wifi_bandwidth_estimator;
512 // Enable payload padding.
513 Settable<bool> use_payload_padding;
516 // A class for playing out soundclips.
517 class SoundclipMedia {
519 enum SoundclipFlags {
523 virtual ~SoundclipMedia() {}
525 // Plays a sound out to the speakers with the given audio stream. The stream
526 // must be 16-bit little-endian 16 kHz PCM. If a stream is already playing
527 // on this SoundclipMedia, it is stopped. If clip is NULL, nothing is played.
528 // Returns whether it was successful.
529 virtual bool PlaySound(const char *clip, int len, int flags) = 0;
532 struct RtpHeaderExtension {
533 RtpHeaderExtension() : id(0) {}
534 RtpHeaderExtension(const std::string& u, int i) : uri(u), id(i) {}
537 // TODO(juberti): SendRecv direction;
539 bool operator==(const RtpHeaderExtension& ext) const {
540 // id is a reserved word in objective-c. Therefore the id attribute has to
541 // be a fully qualified name in order to compile on IOS.
542 return this->id == ext.id &&
547 // Returns the named header extension if found among all extensions, NULL
549 inline const RtpHeaderExtension* FindHeaderExtension(
550 const std::vector<RtpHeaderExtension>& extensions,
551 const std::string& name) {
552 for (std::vector<RtpHeaderExtension>::const_iterator it = extensions.begin();
553 it != extensions.end(); ++it) {
560 enum MediaChannelOptions {
561 // Tune the stream for conference mode.
562 OPT_CONFERENCE = 0x0001
565 enum VoiceMediaChannelOptions {
566 // Tune the audio stream for vcs with different target levels.
567 OPT_AGC_MINUS_10DB = 0x80000000
570 // DTMF flags to control if a DTMF tone should be played and/or sent.
576 class MediaChannel : public sigslot::has_slots<> {
578 class NetworkInterface {
580 enum SocketType { ST_RTP, ST_RTCP };
581 virtual bool SendPacket(
582 talk_base::Buffer* packet,
583 talk_base::DiffServCodePoint dscp = talk_base::DSCP_NO_CHANGE) = 0;
584 virtual bool SendRtcp(
585 talk_base::Buffer* packet,
586 talk_base::DiffServCodePoint dscp = talk_base::DSCP_NO_CHANGE) = 0;
587 virtual int SetOption(SocketType type, talk_base::Socket::Option opt,
589 virtual ~NetworkInterface() {}
592 MediaChannel() : network_interface_(NULL) {}
593 virtual ~MediaChannel() {}
595 // Sets the abstract interface class for sending RTP/RTCP data.
596 virtual void SetInterface(NetworkInterface *iface) {
597 talk_base::CritScope cs(&network_interface_crit_);
598 network_interface_ = iface;
601 // Called when a RTP packet is received.
602 virtual void OnPacketReceived(talk_base::Buffer* packet,
603 const talk_base::PacketTime& packet_time) = 0;
604 // Called when a RTCP packet is received.
605 virtual void OnRtcpReceived(talk_base::Buffer* packet,
606 const talk_base::PacketTime& packet_time) = 0;
607 // Called when the socket's ability to send has changed.
608 virtual void OnReadyToSend(bool ready) = 0;
609 // Creates a new outgoing media stream with SSRCs and CNAME as described
611 virtual bool AddSendStream(const StreamParams& sp) = 0;
612 // Removes an outgoing media stream.
613 // ssrc must be the first SSRC of the media stream if the stream uses
615 virtual bool RemoveSendStream(uint32 ssrc) = 0;
616 // Creates a new incoming media stream with SSRCs and CNAME as described
618 virtual bool AddRecvStream(const StreamParams& sp) = 0;
619 // Removes an incoming media stream.
620 // ssrc must be the first SSRC of the media stream if the stream uses
622 virtual bool RemoveRecvStream(uint32 ssrc) = 0;
624 // Mutes the channel.
625 virtual bool MuteStream(uint32 ssrc, bool on) = 0;
627 // Sets the RTP extension headers and IDs to use when sending RTP.
628 virtual bool SetRecvRtpHeaderExtensions(
629 const std::vector<RtpHeaderExtension>& extensions) = 0;
630 virtual bool SetSendRtpHeaderExtensions(
631 const std::vector<RtpHeaderExtension>& extensions) = 0;
632 // Returns the absoulte sendtime extension id value from media channel.
633 virtual int GetRtpSendTimeExtnId() const {
636 // Sets the initial bandwidth to use when sending starts.
637 virtual bool SetStartSendBandwidth(int bps) = 0;
638 // Sets the maximum allowed bandwidth to use when sending data.
639 virtual bool SetMaxSendBandwidth(int bps) = 0;
641 // Base method to send packet using NetworkInterface.
642 bool SendPacket(talk_base::Buffer* packet) {
643 return DoSendPacket(packet, false);
646 bool SendRtcp(talk_base::Buffer* packet) {
647 return DoSendPacket(packet, true);
650 int SetOption(NetworkInterface::SocketType type,
651 talk_base::Socket::Option opt,
653 talk_base::CritScope cs(&network_interface_crit_);
654 if (!network_interface_)
657 return network_interface_->SetOption(type, opt, option);
661 // This method sets DSCP |value| on both RTP and RTCP channels.
662 int SetDscp(talk_base::DiffServCodePoint value) {
664 ret = SetOption(NetworkInterface::ST_RTP,
665 talk_base::Socket::OPT_DSCP,
668 ret = SetOption(NetworkInterface::ST_RTCP,
669 talk_base::Socket::OPT_DSCP,
676 bool DoSendPacket(talk_base::Buffer* packet, bool rtcp) {
677 talk_base::CritScope cs(&network_interface_crit_);
678 if (!network_interface_)
681 return (!rtcp) ? network_interface_->SendPacket(packet) :
682 network_interface_->SendRtcp(packet);
685 // |network_interface_| can be accessed from the worker_thread and
686 // from any MediaEngine threads. This critical section is to protect accessing
687 // of network_interface_ object.
688 talk_base::CriticalSection network_interface_crit_;
689 NetworkInterface* network_interface_;
698 // The stats information is structured as follows:
699 // Media are represented by either MediaSenderInfo or MediaReceiverInfo.
700 // Media contains a vector of SSRC infos that are exclusively used by this
701 // media. (SSRCs shared between media streams can't be represented.)
703 // Information about an SSRC.
704 // This data may be locally recorded, or received in an RTCP SR or RR.
705 struct SsrcSenderInfo {
711 double timestamp; // NTP timestamp, represented as seconds since epoch.
714 struct SsrcReceiverInfo {
723 struct MediaSenderInfo {
731 void add_ssrc(const SsrcSenderInfo& stat) {
732 local_stats.push_back(stat);
734 // Temporary utility function for call sites that only provide SSRC.
735 // As more info is added into SsrcSenderInfo, this function should go away.
736 void add_ssrc(uint32 ssrc) {
741 // Utility accessor for clients that are only interested in ssrc numbers.
742 std::vector<uint32> ssrcs() const {
743 std::vector<uint32> retval;
744 for (std::vector<SsrcSenderInfo>::const_iterator it = local_stats.begin();
745 it != local_stats.end(); ++it) {
746 retval.push_back(it->ssrc);
750 // Utility accessor for clients that make the assumption only one ssrc
752 // This will eventually go away.
753 uint32 ssrc() const {
754 if (local_stats.size() > 0) {
755 return local_stats[0].ssrc;
765 std::string codec_name;
766 std::vector<SsrcSenderInfo> local_stats;
767 std::vector<SsrcReceiverInfo> remote_stats;
771 struct VariableInfo {
784 struct MediaReceiverInfo {
791 void add_ssrc(const SsrcReceiverInfo& stat) {
792 local_stats.push_back(stat);
794 // Temporary utility function for call sites that only provide SSRC.
795 // As more info is added into SsrcSenderInfo, this function should go away.
796 void add_ssrc(uint32 ssrc) {
797 SsrcReceiverInfo stat;
801 std::vector<uint32> ssrcs() const {
802 std::vector<uint32> retval;
803 for (std::vector<SsrcReceiverInfo>::const_iterator it = local_stats.begin();
804 it != local_stats.end(); ++it) {
805 retval.push_back(it->ssrc);
809 // Utility accessor for clients that make the assumption only one ssrc
811 // This will eventually go away.
812 uint32 ssrc() const {
813 if (local_stats.size() > 0) {
814 return local_stats[0].ssrc;
824 std::string codec_name;
825 std::vector<SsrcReceiverInfo> local_stats;
826 std::vector<SsrcSenderInfo> remote_stats;
829 struct VoiceSenderInfo : public MediaSenderInfo {
834 aec_quality_min(0.0),
835 echo_delay_median_ms(0),
836 echo_delay_std_ms(0),
838 echo_return_loss_enhancement(0),
839 typing_noise_detected(false) {
845 float aec_quality_min;
846 int echo_delay_median_ms;
847 int echo_delay_std_ms;
848 int echo_return_loss;
849 int echo_return_loss_enhancement;
850 bool typing_noise_detected;
853 struct VoiceReceiverInfo : public MediaReceiverInfo {
858 jitter_buffer_preferred_ms(0),
859 delay_estimate_ms(0),
862 decoding_calls_to_silence_generator(0),
863 decoding_calls_to_neteq(0),
868 capture_start_ntp_time_ms(-1) {
873 int jitter_buffer_ms;
874 int jitter_buffer_preferred_ms;
875 int delay_estimate_ms;
877 // fraction of synthesized speech inserted through pre-emptive expansion
879 int decoding_calls_to_silence_generator;
880 int decoding_calls_to_neteq;
884 int decoding_plc_cng;
885 // Estimated capture start time in NTP time in ms.
886 int64 capture_start_ntp_time_ms;
889 struct VideoSenderInfo : public MediaSenderInfo {
895 input_frame_width(0),
896 input_frame_height(0),
898 send_frame_height(0),
902 preferred_bitrate(0),
904 capture_jitter_ms(0),
906 encode_usage_percent(0),
908 capture_queue_delay_ms_per_s(0) {
911 std::vector<SsrcGroup> ssrc_groups;
916 int input_frame_width;
917 int input_frame_height;
918 int send_frame_width;
919 int send_frame_height;
923 int preferred_bitrate;
925 int capture_jitter_ms;
927 int encode_usage_percent;
929 int capture_queue_delay_ms_per_s;
930 VariableInfo<int> adapt_frame_drops;
931 VariableInfo<int> effects_frame_drops;
932 VariableInfo<double> capturer_frame_time;
935 struct VideoReceiverInfo : public MediaReceiverInfo {
937 : packets_concealed(0),
944 framerate_decoded(0),
946 framerate_render_input(0),
947 framerate_render_output(0),
951 min_playout_delay_ms(0),
955 capture_start_ntp_time_ms(-1) {
958 std::vector<SsrcGroup> ssrc_groups;
959 int packets_concealed;
966 int framerate_decoded;
967 int framerate_output;
968 // Framerate as sent to the renderer.
969 int framerate_render_input;
970 // Framerate that the renderer reports.
971 int framerate_render_output;
973 // All stats below are gathered per-VideoReceiver, but some will be correlated
974 // across MediaStreamTracks. NOTE(hta): when sinking stats into per-SSRC
975 // structures, reflect this in the new layout.
977 // Current frame decode latency.
979 // Maximum observed frame decode latency.
981 // Jitter (network-related) latency.
982 int jitter_buffer_ms;
983 // Requested minimum playout latency.
984 int min_playout_delay_ms;
985 // Requested latency to account for rendering delay.
987 // Target overall delay: network+decode+render, accounting for
988 // min_playout_delay_ms.
990 // Current overall delay, possibly ramping towards target_delay_ms.
991 int current_delay_ms;
993 // Estimated capture start time in NTP time in ms.
994 int64 capture_start_ntp_time_ms;
997 struct DataSenderInfo : public MediaSenderInfo {
1005 struct DataReceiverInfo : public MediaReceiverInfo {
1013 struct BandwidthEstimationInfo {
1014 BandwidthEstimationInfo()
1015 : available_send_bandwidth(0),
1016 available_recv_bandwidth(0),
1017 target_enc_bitrate(0),
1018 actual_enc_bitrate(0),
1019 retransmit_bitrate(0),
1020 transmit_bitrate(0),
1022 total_received_propagation_delta_ms(0) {
1025 int available_send_bandwidth;
1026 int available_recv_bandwidth;
1027 int target_enc_bitrate;
1028 int actual_enc_bitrate;
1029 int retransmit_bitrate;
1030 int transmit_bitrate;
1032 // The following stats are only valid when
1033 // StatsOptions::include_received_propagation_stats is true.
1034 int total_received_propagation_delta_ms;
1035 std::vector<int> recent_received_propagation_delta_ms;
1036 std::vector<int64> recent_received_packet_group_arrival_time_ms;
1039 struct VoiceMediaInfo {
1044 std::vector<VoiceSenderInfo> senders;
1045 std::vector<VoiceReceiverInfo> receivers;
1048 struct VideoMediaInfo {
1052 bw_estimations.clear();
1054 std::vector<VideoSenderInfo> senders;
1055 std::vector<VideoReceiverInfo> receivers;
1056 std::vector<BandwidthEstimationInfo> bw_estimations;
1059 struct DataMediaInfo {
1064 std::vector<DataSenderInfo> senders;
1065 std::vector<DataReceiverInfo> receivers;
1068 struct StatsOptions {
1069 StatsOptions() : include_received_propagation_stats(false) {}
1071 bool include_received_propagation_stats;
1074 class VoiceMediaChannel : public MediaChannel {
1077 ERROR_NONE = 0, // No error.
1078 ERROR_OTHER, // Other errors.
1079 ERROR_REC_DEVICE_OPEN_FAILED = 100, // Could not open mic.
1080 ERROR_REC_DEVICE_MUTED, // Mic was muted by OS.
1081 ERROR_REC_DEVICE_SILENT, // No background noise picked up.
1082 ERROR_REC_DEVICE_SATURATION, // Mic input is clipping.
1083 ERROR_REC_DEVICE_REMOVED, // Mic was removed while active.
1084 ERROR_REC_RUNTIME_ERROR, // Processing is encountering errors.
1085 ERROR_REC_SRTP_ERROR, // Generic SRTP failure.
1086 ERROR_REC_SRTP_AUTH_FAILED, // Failed to authenticate packets.
1087 ERROR_REC_TYPING_NOISE_DETECTED, // Typing noise is detected.
1088 ERROR_PLAY_DEVICE_OPEN_FAILED = 200, // Could not open playout.
1089 ERROR_PLAY_DEVICE_MUTED, // Playout muted by OS.
1090 ERROR_PLAY_DEVICE_REMOVED, // Playout removed while active.
1091 ERROR_PLAY_RUNTIME_ERROR, // Errors in voice processing.
1092 ERROR_PLAY_SRTP_ERROR, // Generic SRTP failure.
1093 ERROR_PLAY_SRTP_AUTH_FAILED, // Failed to authenticate packets.
1094 ERROR_PLAY_SRTP_REPLAY, // Packet replay detected.
1097 VoiceMediaChannel() {}
1098 virtual ~VoiceMediaChannel() {}
1099 // Sets the codecs/payload types to be used for incoming media.
1100 virtual bool SetRecvCodecs(const std::vector<AudioCodec>& codecs) = 0;
1101 // Sets the codecs/payload types to be used for outgoing media.
1102 virtual bool SetSendCodecs(const std::vector<AudioCodec>& codecs) = 0;
1103 // Starts or stops playout of received audio.
1104 virtual bool SetPlayout(bool playout) = 0;
1105 // Starts or stops sending (and potentially capture) of local audio.
1106 virtual bool SetSend(SendFlags flag) = 0;
1107 // Sets the renderer object to be used for the specified remote audio stream.
1108 virtual bool SetRemoteRenderer(uint32 ssrc, AudioRenderer* renderer) = 0;
1109 // Sets the renderer object to be used for the specified local audio stream.
1110 virtual bool SetLocalRenderer(uint32 ssrc, AudioRenderer* renderer) = 0;
1111 // Gets current energy levels for all incoming streams.
1112 virtual bool GetActiveStreams(AudioInfo::StreamList* actives) = 0;
1113 // Get the current energy level of the stream sent to the speaker.
1114 virtual int GetOutputLevel() = 0;
1115 // Get the time in milliseconds since last recorded keystroke, or negative.
1116 virtual int GetTimeSinceLastTyping() = 0;
1117 // Temporarily exposed field for tuning typing detect options.
1118 virtual void SetTypingDetectionParameters(int time_window,
1119 int cost_per_typing, int reporting_threshold, int penalty_decay,
1120 int type_event_delay) = 0;
1121 // Set left and right scale for speaker output volume of the specified ssrc.
1122 virtual bool SetOutputScaling(uint32 ssrc, double left, double right) = 0;
1123 // Get left and right scale for speaker output volume of the specified ssrc.
1124 virtual bool GetOutputScaling(uint32 ssrc, double* left, double* right) = 0;
1125 // Specifies a ringback tone to be played during call setup.
1126 virtual bool SetRingbackTone(const char *buf, int len) = 0;
1127 // Plays or stops the aforementioned ringback tone
1128 virtual bool PlayRingbackTone(uint32 ssrc, bool play, bool loop) = 0;
1129 // Returns if the telephone-event has been negotiated.
1130 virtual bool CanInsertDtmf() { return false; }
1131 // Send and/or play a DTMF |event| according to the |flags|.
1132 // The DTMF out-of-band signal will be used on sending.
1133 // The |ssrc| should be either 0 or a valid send stream ssrc.
1134 // The valid value for the |event| are 0 to 15 which corresponding to
1135 // DTMF event 0-9, *, #, A-D.
1136 virtual bool InsertDtmf(uint32 ssrc, int event, int duration, int flags) = 0;
1137 // Gets quality stats for the channel.
1138 virtual bool GetStats(VoiceMediaInfo* info) = 0;
1139 // Gets last reported error for this media channel.
1140 virtual void GetLastMediaError(uint32* ssrc,
1141 VoiceMediaChannel::Error* error) {
1142 ASSERT(error != NULL);
1143 *error = ERROR_NONE;
1145 // Sets the media options to use.
1146 virtual bool SetOptions(const AudioOptions& options) = 0;
1147 virtual bool GetOptions(AudioOptions* options) const = 0;
1149 // Signal errors from MediaChannel. Arguments are:
1150 // ssrc(uint32), and error(VoiceMediaChannel::Error).
1151 sigslot::signal2<uint32, VoiceMediaChannel::Error> SignalMediaError;
1154 class VideoMediaChannel : public MediaChannel {
1157 ERROR_NONE = 0, // No error.
1158 ERROR_OTHER, // Other errors.
1159 ERROR_REC_DEVICE_OPEN_FAILED = 100, // Could not open camera.
1160 ERROR_REC_DEVICE_NO_DEVICE, // No camera.
1161 ERROR_REC_DEVICE_IN_USE, // Device is in already use.
1162 ERROR_REC_DEVICE_REMOVED, // Device is removed.
1163 ERROR_REC_SRTP_ERROR, // Generic sender SRTP failure.
1164 ERROR_REC_SRTP_AUTH_FAILED, // Failed to authenticate packets.
1165 ERROR_REC_CPU_MAX_CANT_DOWNGRADE, // Can't downgrade capture anymore.
1166 ERROR_PLAY_SRTP_ERROR = 200, // Generic receiver SRTP failure.
1167 ERROR_PLAY_SRTP_AUTH_FAILED, // Failed to authenticate packets.
1168 ERROR_PLAY_SRTP_REPLAY, // Packet replay detected.
1171 VideoMediaChannel() : renderer_(NULL) {}
1172 virtual ~VideoMediaChannel() {}
1173 // Sets the codecs/payload types to be used for incoming media.
1174 virtual bool SetRecvCodecs(const std::vector<VideoCodec>& codecs) = 0;
1175 // Sets the codecs/payload types to be used for outgoing media.
1176 virtual bool SetSendCodecs(const std::vector<VideoCodec>& codecs) = 0;
1177 // Gets the currently set codecs/payload types to be used for outgoing media.
1178 virtual bool GetSendCodec(VideoCodec* send_codec) = 0;
1179 // Sets the format of a specified outgoing stream.
1180 virtual bool SetSendStreamFormat(uint32 ssrc, const VideoFormat& format) = 0;
1181 // Starts or stops playout of received video.
1182 virtual bool SetRender(bool render) = 0;
1183 // Starts or stops transmission (and potentially capture) of local video.
1184 virtual bool SetSend(bool send) = 0;
1185 // Sets the renderer object to be used for the specified stream.
1186 // If SSRC is 0, the renderer is used for the 'default' stream.
1187 virtual bool SetRenderer(uint32 ssrc, VideoRenderer* renderer) = 0;
1188 // If |ssrc| is 0, replace the default capturer (engine capturer) with
1189 // |capturer|. If |ssrc| is non zero create a new stream with |ssrc| as SSRC.
1190 virtual bool SetCapturer(uint32 ssrc, VideoCapturer* capturer) = 0;
1191 // Gets quality stats for the channel.
1192 virtual bool GetStats(const StatsOptions& options, VideoMediaInfo* info) = 0;
1193 // This is needed for MediaMonitor to use the same template for voice, video
1194 // and data MediaChannels.
1195 bool GetStats(VideoMediaInfo* info) {
1196 return GetStats(StatsOptions(), info);
1199 // Send an intra frame to the receivers.
1200 virtual bool SendIntraFrame() = 0;
1201 // Reuqest each of the remote senders to send an intra frame.
1202 virtual bool RequestIntraFrame() = 0;
1203 // Sets the media options to use.
1204 virtual bool SetOptions(const VideoOptions& options) = 0;
1205 virtual bool GetOptions(VideoOptions* options) const = 0;
1206 virtual void UpdateAspectRatio(int ratio_w, int ratio_h) = 0;
1208 // Signal errors from MediaChannel. Arguments are:
1209 // ssrc(uint32), and error(VideoMediaChannel::Error).
1210 sigslot::signal2<uint32, Error> SignalMediaError;
1213 VideoRenderer *renderer_;
1216 enum DataMessageType {
1217 // Chrome-Internal use only. See SctpDataMediaChannel for the actual PPID
1225 // Info about data received in DataMediaChannel. For use in
1226 // DataMediaChannel::SignalDataReceived and in all of the signals that
1227 // signal fires, on up the chain.
1228 struct ReceiveDataParams {
1229 // The in-packet stream indentifier.
1230 // For SCTP, this is really SID, not SSRC.
1232 // The type of message (binary, text, or control).
1233 DataMessageType type;
1234 // A per-stream value incremented per packet in the stream.
1236 // A per-stream value monotonically increasing with time.
1239 ReceiveDataParams() :
1247 struct SendDataParams {
1248 // The in-packet stream indentifier.
1249 // For SCTP, this is really SID, not SSRC.
1251 // The type of message (binary, text, or control).
1252 DataMessageType type;
1254 // For SCTP, whether to send messages flagged as ordered or not.
1255 // If false, messages can be received out of order.
1257 // For SCTP, whether the messages are sent reliably or not.
1258 // If false, messages may be lost.
1260 // For SCTP, if reliable == false, provide partial reliability by
1261 // resending up to this many times. Either count or millis
1262 // is supported, not both at the same time.
1264 // For SCTP, if reliable == false, provide partial reliability by
1265 // resending for up to this many milliseconds. Either count or millis
1266 // is supported, not both at the same time.
1272 // TODO(pthatcher): Make these true by default?
1280 enum SendDataResult { SDR_SUCCESS, SDR_ERROR, SDR_BLOCK };
1282 class DataMediaChannel : public MediaChannel {
1285 ERROR_NONE = 0, // No error.
1286 ERROR_OTHER, // Other errors.
1287 ERROR_SEND_SRTP_ERROR = 200, // Generic SRTP failure.
1288 ERROR_SEND_SRTP_AUTH_FAILED, // Failed to authenticate packets.
1289 ERROR_RECV_SRTP_ERROR, // Generic SRTP failure.
1290 ERROR_RECV_SRTP_AUTH_FAILED, // Failed to authenticate packets.
1291 ERROR_RECV_SRTP_REPLAY, // Packet replay detected.
1294 virtual ~DataMediaChannel() {}
1296 virtual bool SetSendCodecs(const std::vector<DataCodec>& codecs) = 0;
1297 virtual bool SetRecvCodecs(const std::vector<DataCodec>& codecs) = 0;
1299 virtual bool MuteStream(uint32 ssrc, bool on) { return false; }
1300 // TODO(pthatcher): Implement this.
1301 virtual bool GetStats(DataMediaInfo* info) { return true; }
1303 virtual bool SetSend(bool send) = 0;
1304 virtual bool SetReceive(bool receive) = 0;
1306 virtual bool SendData(
1307 const SendDataParams& params,
1308 const talk_base::Buffer& payload,
1309 SendDataResult* result = NULL) = 0;
1310 // Signals when data is received (params, data, len)
1311 sigslot::signal3<const ReceiveDataParams&,
1313 size_t> SignalDataReceived;
1314 // Signal errors from MediaChannel. Arguments are:
1315 // ssrc(uint32), and error(DataMediaChannel::Error).
1316 sigslot::signal2<uint32, DataMediaChannel::Error> SignalMediaError;
1317 // Signal when the media channel is ready to send the stream. Arguments are:
1319 sigslot::signal1<bool> SignalReadyToSend;
1320 // Signal for notifying that the remote side has closed the DataChannel.
1321 sigslot::signal1<uint32> SignalStreamClosedRemotely;
1324 } // namespace cricket
1326 #endif // TALK_MEDIA_BASE_MEDIACHANNEL_H_