AEC3: Improve dominant nearend detection

This change makes the dominant nearend detection more accurate.
- The hangover is increased not leave nearend state between words.
- The SNR requirement is increased to not enter nearend state without
  speech activity.
- An early exit mechanism has been added to leave nearend state quickly
  when the echo is strong.

Bug: chromium:897701,webrtc:9897
Change-Id: I9e0f3e6ecb80eee1c0c917d4835f110555f74acf
Reviewed-on: https://webrtc-review.googlesource.com/c/107347
Reviewed-by: Per Åhgren <peah@webrtc.org>
Commit-Queue: Gustaf Ullberg <gustaf@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#25299}

api/audio/echo_canceller3_config.h

@@ -181,10 +181,11 @@ struct RTC_EXPORT EchoCanceller3Config {
                                    0.25f);
 
     struct DominantNearendDetection {
-      float enr_threshold = 6.f;
-      float snr_threshold = 6.f;
-      int hold_duration = 5;
-      int trigger_threshold = 15;
+      float enr_threshold = 4.f;
+      float enr_exit_threshold = .1f;
+      float snr_threshold = 30.f;
+      int hold_duration = 50;
+      int trigger_threshold = 12;
     } dominant_nearend_detection;
 
     struct HighBandsSuppression {

api/audio/echo_canceller3_config_json.cc

@@ -287,6 +287,8 @@ EchoCanceller3Config Aec3ConfigFromJsonString(absl::string_view json_string) {
                                     &subsection)) {
       ReadParam(subsection, "enr_threshold",
                 &cfg.suppressor.dominant_nearend_detection.enr_threshold);
+      ReadParam(subsection, "enr_exit_threshold",
+                &cfg.suppressor.dominant_nearend_detection.enr_exit_threshold);
       ReadParam(subsection, "snr_threshold",
                 &cfg.suppressor.dominant_nearend_detection.snr_threshold);
       ReadParam(subsection, "hold_duration",
@@ -520,6 +522,8 @@ std::string Aec3ConfigToJsonString(const EchoCanceller3Config& config) {
   ost << "\"dominant_nearend_detection\": {";
   ost << "\"enr_threshold\": "
       << config.suppressor.dominant_nearend_detection.enr_threshold << ",";
+  ost << "\"enr_exit_threshold\": "
+      << config.suppressor.dominant_nearend_detection.enr_exit_threshold << ",";
   ost << "\"snr_threshold\": "
       << config.suppressor.dominant_nearend_detection.snr_threshold << ",";
   ost << "\"hold_duration\": "

modules/audio_processing/aec3/suppression_gain.cc

@@ -398,6 +398,7 @@ bool SuppressionGain::LowNoiseRenderDetector::Detect(
 SuppressionGain::DominantNearendDetector::DominantNearendDetector(
     const EchoCanceller3Config::Suppressor::DominantNearendDetection config)
     : enr_threshold_(config.enr_threshold),
+      enr_exit_threshold_(config.enr_exit_threshold),
       snr_threshold_(config.snr_threshold),
       hold_duration_(config.hold_duration),
       trigger_threshold_(config.trigger_threshold) {}
@@ -428,6 +429,12 @@ void SuppressionGain::DominantNearendDetector::Update(
     trigger_counter_ = std::max(0, trigger_counter_ - 1);
   }
 
+  // Exit nearend-state early at strong echo.
+  if (ne_sum < enr_exit_threshold_ * echo_sum &&
+      echo_sum > snr_threshold_ * noise_sum) {
+    hold_counter_ = 0;
+  }
+
   // Remain in any nearend mode for a certain duration.
   hold_counter_ = std::max(0, hold_counter_ - 1);
   nearend_state_ = hold_counter_ > 0;

modules/audio_processing/aec3/suppression_gain.h

@@ -106,6 +106,7 @@ class SuppressionGain {
 
    private:
     const float enr_threshold_;
+    const float enr_exit_threshold_;
     const float snr_threshold_;
     const int hold_duration_;
     const int trigger_threshold_;