Expose new audio stats on the API Several new audio stats were recently standardized and implemented in WebRTC in https://webrtc-review.googlesource.com/c/src/+/133887. This CL adds these to the GetStats API. Bug: webrtc:10442, webrtc:10443, webrtc:10444 Change-Id: I0e898ac14777e82b1a9099b5e0a5584eb9cb5934 Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/134213 Reviewed-by: Karl Wiberg <kwiberg@webrtc.org> Reviewed-by: Henrik Boström <hbos@webrtc.org> Commit-Queue: Ivo Creusen <ivoc@webrtc.org> Cr-Commit-Position: refs/heads/master@{#27839}
diff --git a/api/stats/rtcstats_objects.h b/api/stats/rtcstats_objects.h index 96c7a03..d614a86 100644 --- a/api/stats/rtcstats_objects.h +++ b/api/stats/rtcstats_objects.h
@@ -315,7 +315,10 @@ RTCStatsMember<uint64_t> total_samples_received; RTCStatsMember<double> total_samples_duration; RTCStatsMember<uint64_t> concealed_samples; + RTCStatsMember<uint64_t> silent_concealed_samples; RTCStatsMember<uint64_t> concealment_events; + RTCStatsMember<uint64_t> inserted_samples_for_deceleration; + RTCStatsMember<uint64_t> removed_samples_for_acceleration; // Non-standard audio-only member // TODO(kuddai): Add description to standard. crbug.com/webrtc/10042 RTCNonStandardStatsMember<uint64_t> jitter_buffer_flushes; @@ -399,6 +402,8 @@ ~RTCInboundRTPStreamStats() override; RTCStatsMember<uint32_t> packets_received; + RTCStatsMember<uint64_t> fec_packets_received; + RTCStatsMember<uint64_t> fec_packets_discarded; RTCStatsMember<uint64_t> bytes_received; RTCStatsMember<int32_t> packets_lost; // Signed per RFC 3550 RTCStatsMember<double> last_packet_received_timestamp;
diff --git a/audio/audio_receive_stream.cc b/audio/audio_receive_stream.cc index 677ee20..b98c213 100644 --- a/audio/audio_receive_stream.cc +++ b/audio/audio_receive_stream.cc
@@ -206,15 +206,20 @@ // Get jitter buffer and total delay (alg + jitter + playout) stats. auto ns = channel_receive_->GetNetworkStatistics(); + stats.fec_packets_received = ns.fecPacketsReceived; + stats.fec_packets_discarded = ns.fecPacketsDiscarded; stats.jitter_buffer_ms = ns.currentBufferSize; stats.jitter_buffer_preferred_ms = ns.preferredBufferSize; stats.total_samples_received = ns.totalSamplesReceived; stats.concealed_samples = ns.concealedSamples; + stats.silent_concealed_samples = ns.silentConcealedSamples; stats.concealment_events = ns.concealmentEvents; stats.jitter_buffer_delay_seconds = static_cast<double>(ns.jitterBufferDelayMs) / static_cast<double>(rtc::kNumMillisecsPerSec); stats.jitter_buffer_emitted_count = ns.jitterBufferEmittedCount; + stats.inserted_samples_for_deceleration = ns.insertedSamplesForDeceleration; + stats.removed_samples_for_acceleration = ns.removedSamplesForAcceleration; stats.expand_rate = Q14ToFloat(ns.currentExpandRate); stats.speech_expand_rate = Q14ToFloat(ns.currentSpeechExpandRate); stats.secondary_decoded_rate = Q14ToFloat(ns.currentSecondaryDecodedRate);
diff --git a/audio/audio_receive_stream_unittest.cc b/audio/audio_receive_stream_unittest.cc index 127f2ff..303e0e8 100644 --- a/audio/audio_receive_stream_unittest.cc +++ b/audio/audio_receive_stream_unittest.cc
@@ -67,8 +67,9 @@ 123, {"codec_name_recv", 96000, 0}}; const NetworkStatistics kNetworkStats = { - 123, 456, false, 789012, 3456, 123, 456, 789, 0, {}, 789, - 12, 345, 678, 901, 0, -1, -1, -1, -1, -1, 0}; + 123, 456, false, 789012, 3456, 123, 456, 789, 543, 432, + 321, 123, 101, 0, {}, 789, 12, 345, 678, 901, + 0, -1, -1, -1, -1, -1, 0, 0, 0, 0}; const AudioDecodingCallStats kAudioDecodeStats = MakeAudioDecodeStatsForTest(); struct ConfigHelper {
diff --git a/call/audio_receive_stream.h b/call/audio_receive_stream.h index 9091afd..35c6ef7 100644 --- a/call/audio_receive_stream.h +++ b/call/audio_receive_stream.h
@@ -38,6 +38,8 @@ uint32_t remote_ssrc = 0; int64_t bytes_rcvd = 0; uint32_t packets_rcvd = 0; + uint64_t fec_packets_received = 0; + uint64_t fec_packets_discarded = 0; uint32_t packets_lost = 0; float fraction_lost = 0.0f; std::string codec_name; @@ -54,9 +56,12 @@ uint64_t total_samples_received = 0; double total_output_duration = 0.0; uint64_t concealed_samples = 0; + uint64_t silent_concealed_samples = 0; uint64_t concealment_events = 0; double jitter_buffer_delay_seconds = 0.0; uint64_t jitter_buffer_emitted_count = 0; + uint64_t inserted_samples_for_deceleration = 0; + uint64_t removed_samples_for_acceleration = 0; // Stats below DO NOT correspond directly to anything in the WebRTC stats float expand_rate = 0.0f; float speech_expand_rate = 0.0f;
diff --git a/media/base/media_channel.h b/media/base/media_channel.h index 69570e7..15721c5 100644 --- a/media/base/media_channel.h +++ b/media/base/media_channel.h
@@ -479,9 +479,14 @@ uint64_t total_samples_received = 0; double total_output_duration = 0.0; uint64_t concealed_samples = 0; + uint64_t silent_concealed_samples = 0; uint64_t concealment_events = 0; double jitter_buffer_delay_seconds = 0.0; uint64_t jitter_buffer_emitted_count = 0; + uint64_t inserted_samples_for_deceleration = 0; + uint64_t removed_samples_for_acceleration = 0; + uint64_t fec_packets_received = 0; + uint64_t fec_packets_discarded = 0; // Stats below DO NOT correspond directly to anything in the WebRTC stats // fraction of synthesized audio inserted through expansion. float expand_rate = 0.0f;
diff --git a/media/engine/webrtc_voice_engine.cc b/media/engine/webrtc_voice_engine.cc index 3b85064..a3b375a 100644 --- a/media/engine/webrtc_voice_engine.cc +++ b/media/engine/webrtc_voice_engine.cc
@@ -2226,6 +2226,8 @@ rinfo.add_ssrc(stats.remote_ssrc); rinfo.bytes_rcvd = stats.bytes_rcvd; rinfo.packets_rcvd = stats.packets_rcvd; + rinfo.fec_packets_received = stats.fec_packets_received; + rinfo.fec_packets_discarded = stats.fec_packets_discarded; rinfo.packets_lost = stats.packets_lost; rinfo.fraction_lost = stats.fraction_lost; rinfo.codec_name = stats.codec_name; @@ -2240,9 +2242,14 @@ rinfo.total_samples_received = stats.total_samples_received; rinfo.total_output_duration = stats.total_output_duration; rinfo.concealed_samples = stats.concealed_samples; + rinfo.silent_concealed_samples = stats.silent_concealed_samples; rinfo.concealment_events = stats.concealment_events; rinfo.jitter_buffer_delay_seconds = stats.jitter_buffer_delay_seconds; rinfo.jitter_buffer_emitted_count = stats.jitter_buffer_emitted_count; + rinfo.inserted_samples_for_deceleration = + stats.inserted_samples_for_deceleration; + rinfo.removed_samples_for_acceleration = + stats.removed_samples_for_acceleration; rinfo.expand_rate = stats.expand_rate; rinfo.speech_expand_rate = stats.speech_expand_rate; rinfo.secondary_decoded_rate = stats.secondary_decoded_rate;
diff --git a/modules/audio_coding/acm2/acm_receiver.cc b/modules/audio_coding/acm2/acm_receiver.cc index c10a71c..3bce0c4 100644 --- a/modules/audio_coding/acm2/acm_receiver.cc +++ b/modules/audio_coding/acm2/acm_receiver.cc
@@ -251,6 +251,8 @@ NetEqLifetimeStatistics neteq_lifetime_stat = neteq_->GetLifetimeStatistics(); acm_stat->totalSamplesReceived = neteq_lifetime_stat.total_samples_received; acm_stat->concealedSamples = neteq_lifetime_stat.concealed_samples; + acm_stat->silentConcealedSamples = + neteq_lifetime_stat.silent_concealed_samples; acm_stat->concealmentEvents = neteq_lifetime_stat.concealment_events; acm_stat->jitterBufferDelayMs = neteq_lifetime_stat.jitter_buffer_delay_ms; acm_stat->jitterBufferEmittedCount = @@ -262,6 +264,12 @@ acm_stat->interruptionCount = neteq_lifetime_stat.interruption_count; acm_stat->totalInterruptionDurationMs = neteq_lifetime_stat.total_interruption_duration_ms; + acm_stat->insertedSamplesForDeceleration = + neteq_lifetime_stat.inserted_samples_for_deceleration; + acm_stat->removedSamplesForAcceleration = + neteq_lifetime_stat.removed_samples_for_acceleration; + acm_stat->fecPacketsReceived = neteq_lifetime_stat.fec_packets_received; + acm_stat->fecPacketsDiscarded = neteq_lifetime_stat.fec_packets_discarded; NetEqOperationsAndState neteq_operations_and_state = neteq_->GetOperationsAndState();
diff --git a/modules/audio_coding/include/audio_coding_module_typedefs.h b/modules/audio_coding/include/audio_coding_module_typedefs.h index 621c478..d256fd1 100644 --- a/modules/audio_coding/include/audio_coding_module_typedefs.h +++ b/modules/audio_coding/include/audio_coding_module_typedefs.h
@@ -84,9 +84,14 @@ // https://w3c.github.io/webrtc-stats/#dom-rtcmediastreamtrackstats uint64_t totalSamplesReceived; uint64_t concealedSamples; + uint64_t silentConcealedSamples; uint64_t concealmentEvents; uint64_t jitterBufferDelayMs; uint64_t jitterBufferEmittedCount; + uint64_t insertedSamplesForDeceleration; + uint64_t removedSamplesForAcceleration; + uint64_t fecPacketsReceived; + uint64_t fecPacketsDiscarded; // Stats below DO NOT correspond directly to anything in the WebRTC stats // Loss rate (network + late); fraction between 0 and 1, scaled to Q14. uint16_t currentPacketLossRate;
diff --git a/pc/rtc_stats_collector.cc b/pc/rtc_stats_collector.cc index 8849d86..0d1fbba 100644 --- a/pc/rtc_stats_collector.cc +++ b/pc/rtc_stats_collector.cc
@@ -250,6 +250,10 @@ *voice_receiver_info.last_packet_received_timestamp_ms) / rtc::kNumMillisecsPerSec; } + inbound_audio->fec_packets_received = + voice_receiver_info.fec_packets_received; + inbound_audio->fec_packets_discarded = + voice_receiver_info.fec_packets_discarded; } void SetInboundRTPStreamStatsFromVideoReceiverInfo( @@ -475,6 +479,10 @@ voice_receiver_info.jitter_buffer_delay_seconds; audio_track_stats->jitter_buffer_emitted_count = voice_receiver_info.jitter_buffer_emitted_count; + audio_track_stats->inserted_samples_for_deceleration = + voice_receiver_info.inserted_samples_for_deceleration; + audio_track_stats->removed_samples_for_acceleration = + voice_receiver_info.removed_samples_for_acceleration; audio_track_stats->total_audio_energy = voice_receiver_info.total_output_energy; audio_track_stats->total_samples_received = @@ -482,6 +490,8 @@ audio_track_stats->total_samples_duration = voice_receiver_info.total_output_duration; audio_track_stats->concealed_samples = voice_receiver_info.concealed_samples; + audio_track_stats->silent_concealed_samples = + voice_receiver_info.silent_concealed_samples; audio_track_stats->concealment_events = voice_receiver_info.concealment_events; audio_track_stats->jitter_buffer_flushes = @@ -921,7 +931,7 @@ void RTCStatsCollector::ProducePartialResultsOnNetworkThread( int64_t timestamp_us) { RTC_DCHECK(network_thread_->IsCurrent()); - // Touching |network_report_| on this thread is safe by this method because + // Touching |network_report_| on this thread is safe by this method because // |network_report_event_| is reset before this method is invoked. network_report_ = RTCStatsReport::Create(timestamp_us);
diff --git a/pc/rtc_stats_collector_unittest.cc b/pc/rtc_stats_collector_unittest.cc index 0539379..a6531d2 100644 --- a/pc/rtc_stats_collector_unittest.cc +++ b/pc/rtc_stats_collector_unittest.cc
@@ -1424,6 +1424,9 @@ voice_receiver_info.total_output_duration = 0.25; voice_receiver_info.concealed_samples = 123; voice_receiver_info.concealment_events = 12; + voice_receiver_info.inserted_samples_for_deceleration = 987; + voice_receiver_info.removed_samples_for_acceleration = 876; + voice_receiver_info.silent_concealed_samples = 765; voice_receiver_info.jitter_buffer_delay_seconds = 3456; voice_receiver_info.jitter_buffer_emitted_count = 13; voice_receiver_info.jitter_buffer_flushes = 7; @@ -1463,6 +1466,9 @@ expected_remote_audio_track.total_samples_duration = 0.25; expected_remote_audio_track.concealed_samples = 123; expected_remote_audio_track.concealment_events = 12; + expected_remote_audio_track.inserted_samples_for_deceleration = 987; + expected_remote_audio_track.removed_samples_for_acceleration = 876; + expected_remote_audio_track.silent_concealed_samples = 765; expected_remote_audio_track.jitter_buffer_delay = 3456; expected_remote_audio_track.jitter_buffer_emitted_count = 13; expected_remote_audio_track.jitter_buffer_flushes = 7; @@ -1625,6 +1631,8 @@ voice_media_info.receivers[0].local_stats[0].ssrc = 1; voice_media_info.receivers[0].packets_lost = -1; // Signed per RFC3550 voice_media_info.receivers[0].packets_rcvd = 2; + voice_media_info.receivers[0].fec_packets_discarded = 5566; + voice_media_info.receivers[0].fec_packets_received = 6677; voice_media_info.receivers[0].bytes_rcvd = 3; voice_media_info.receivers[0].codec_payload_type = 42; voice_media_info.receivers[0].jitter_ms = 4500; @@ -1660,6 +1668,8 @@ expected_audio.transport_id = "RTCTransport_TransportName_1"; expected_audio.codec_id = "RTCCodec_AudioMid_Inbound_42"; expected_audio.packets_received = 2; + expected_audio.fec_packets_discarded = 5566; + expected_audio.fec_packets_received = 6677; expected_audio.bytes_received = 3; expected_audio.packets_lost = -1; // |expected_audio.last_packet_received_timestamp| should be undefined.
diff --git a/pc/rtc_stats_integrationtest.cc b/pc/rtc_stats_integrationtest.cc index d576c60..a751599 100644 --- a/pc/rtc_stats_integrationtest.cc +++ b/pc/rtc_stats_integrationtest.cc
@@ -649,6 +649,12 @@ verifier.TestMemberIsNonNegative<uint64_t>( media_stream_track.concealment_events); verifier.TestMemberIsNonNegative<uint64_t>( + media_stream_track.inserted_samples_for_deceleration); + verifier.TestMemberIsNonNegative<uint64_t>( + media_stream_track.removed_samples_for_acceleration); + verifier.TestMemberIsNonNegative<uint64_t>( + media_stream_track.silent_concealed_samples); + verifier.TestMemberIsNonNegative<uint64_t>( media_stream_track.jitter_buffer_flushes); verifier.TestMemberIsNonNegative<uint64_t>( media_stream_track.delayed_packet_outage_samples); @@ -722,6 +728,13 @@ verifier.TestMemberIsUndefined(inbound_stream.qp_sum); } verifier.TestMemberIsNonNegative<uint32_t>(inbound_stream.packets_received); + if (inbound_stream.media_type.is_defined() && + *inbound_stream.media_type == "audio") { + verifier.TestMemberIsNonNegative<uint64_t>( + inbound_stream.fec_packets_received); + verifier.TestMemberIsNonNegative<uint64_t>( + inbound_stream.fec_packets_discarded); + } verifier.TestMemberIsNonNegative<uint64_t>(inbound_stream.bytes_received); // packets_lost is defined as signed, but this should never happen in // this test. See RFC 3550.
diff --git a/stats/rtcstats_objects.cc b/stats/rtcstats_objects.cc index 6c914e7..2aa2cd2 100644 --- a/stats/rtcstats_objects.cc +++ b/stats/rtcstats_objects.cc
@@ -433,7 +433,10 @@ total_samples_received("totalSamplesReceived"), total_samples_duration("totalSamplesDuration"), concealed_samples("concealedSamples"), + silent_concealed_samples("silentConcealedSamples"), concealment_events("concealmentEvents"), + inserted_samples_for_deceleration("insertedSamplesForDeceleration"), + removed_samples_for_acceleration("removedSamplesForAcceleration"), jitter_buffer_flushes( "jitterBufferFlushes", {NonStandardGroupId::kRtcAudioJitterBufferMaxPackets}), @@ -484,7 +487,11 @@ total_samples_received(other.total_samples_received), total_samples_duration(other.total_samples_duration), concealed_samples(other.concealed_samples), + silent_concealed_samples(other.silent_concealed_samples), concealment_events(other.concealment_events), + inserted_samples_for_deceleration( + other.inserted_samples_for_deceleration), + removed_samples_for_acceleration(other.removed_samples_for_acceleration), jitter_buffer_flushes(other.jitter_buffer_flushes), delayed_packet_outage_samples(other.delayed_packet_outage_samples), relative_packet_arrival_delay(other.relative_packet_arrival_delay), @@ -610,6 +617,8 @@ int64_t timestamp_us) : RTCRTPStreamStats(std::move(id), timestamp_us), packets_received("packetsReceived"), + fec_packets_received("fecPacketsReceived"), + fec_packets_discarded("fecPacketsDiscarded"), bytes_received("bytesReceived"), packets_lost("packetsLost"), last_packet_received_timestamp("lastPacketReceivedTimestamp"), @@ -633,6 +642,8 @@ const RTCInboundRTPStreamStats& other) : RTCRTPStreamStats(other), packets_received(other.packets_received), + fec_packets_received(other.fec_packets_received), + fec_packets_discarded(other.fec_packets_discarded), bytes_received(other.bytes_received), packets_lost(other.packets_lost), last_packet_received_timestamp(other.last_packet_received_timestamp),