Removed the issue with the leading semicolon in the audio processing module experiment description…

webrtc/modules/audio_processing/echo_cancellation_impl.cc

Index: webrtc/modules/audio_processing/echo_cancellation_impl.cc
diff --git a/webrtc/modules/audio_processing/echo_cancellation_impl.cc b/webrtc/modules/audio_processing/echo_cancellation_impl.cc
index c18ed94bd2c1b97bb4157e610c647149bab41560..5a7aef6fdf970e2b8cae8e779df6298b9552efa6 100644
--- a/webrtc/modules/audio_processing/echo_cancellation_impl.cc
+++ b/webrtc/modules/audio_processing/echo_cancellation_impl.cc
@@ -415,9 +415,9 @@ bool EchoCancellationImpl::is_aec3_enabled() const {
 
 std::string EchoCancellationImpl::GetExperimentsDescription() {
   rtc::CritScope cs(crit_capture_);
-  std::string description = (aec3_enabled_ ? "AEC3" : "");
+  std::string description = (aec3_enabled_ ? "AEC3;" : "");
   if (refined_adaptive_filter_enabled_) {
-    description += ";RefinedAdaptiveFilter";
+    description += "RefinedAdaptiveFilter;";
   }
   return description;
 }