Pull out the PostFilter to its own NonlinearBeamformer API

webrtc/modules/audio_processing/audio_processing_impl.cc

 /*
  *  Copyright (c) 2012 The WebRTC project authors. All Rights Reserved.
  *
  *  Use of this source code is governed by a BSD-style license
  *  that can be found in the LICENSE file in the root of the source
  *  tree. An additional intellectual property rights grant can be found
  *  in the file PATENTS.  All contributing project authors may
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
@@ skipping 109 matching lines @@
   std::unique_ptr<VoiceDetectionImpl> voice_detection;
   std::unique_ptr<GainControlForExperimentalAgc>
       gain_control_for_experimental_agc;
 
   // Accessed internally from both render and capture.
   std::unique_ptr<TransientSuppressor> transient_suppressor;
   std::unique_ptr<IntelligibilityEnhancer> intelligibility_enhancer;
 };
 
 struct AudioProcessingImpl::ApmPrivateSubmodules {
-  explicit ApmPrivateSubmodules(Beamformer<float>* beamformer)
+  explicit ApmPrivateSubmodules(NonlinearBeamformer* beamformer)
       : beamformer(beamformer) {}
   // Accessed internally from capture or during initialization
-  std::unique_ptr<Beamformer<float>> beamformer;
+  std::unique_ptr<NonlinearBeamformer> beamformer;
   std::unique_ptr<AgcManagerDirect> agc_manager;
 };
 
 AudioProcessing* AudioProcessing::Create() {
   Config config;
   return Create(config, nullptr);
 }
 
 AudioProcessing* AudioProcessing::Create(const Config& config) {
   return Create(config, nullptr);
 }
 
 AudioProcessing* AudioProcessing::Create(const Config& config,
-                                         Beamformer<float>* beamformer) {
+                                         NonlinearBeamformer* beamformer) {
   AudioProcessingImpl* apm = new AudioProcessingImpl(config, beamformer);
   if (apm->Initialize() != kNoError) {
     delete apm;
     apm = nullptr;
   }
 
   return apm;
 }
 
 AudioProcessingImpl::AudioProcessingImpl(const Config& config)
     : AudioProcessingImpl(config, nullptr) {}
 
 AudioProcessingImpl::AudioProcessingImpl(const Config& config,
-                                         Beamformer<float>* beamformer)
+                                         NonlinearBeamformer* beamformer)
     : public_submodules_(new ApmPublicSubmodules()),
       private_submodules_(new ApmPrivateSubmodules(beamformer)),
       constants_(config.Get<ExperimentalAgc>().startup_min_volume,
 #if defined(WEBRTC_ANDROID) || defined(WEBRTC_IOS)
                  false),
 #else
                  config.Get<ExperimentalAgc>().enabled),
 #endif
 #if defined(WEBRTC_ANDROID) || defined(WEBRTC_IOS)
       capture_(false,
@@ skipping 507 matching lines @@
     private_submodules_->agc_manager->AnalyzePreProcess(
         ca->channels()[0], ca->num_channels(),
         capture_nonlocked_.fwd_proc_format.num_frames());
   }
 
   if (fwd_analysis_needed()) {
     ca->SplitIntoFrequencyBands();
   }
 
   if (capture_nonlocked_.beamformer_enabled) {
-    private_submodules_->beamformer->ProcessChunk(*ca->split_data_f(),
-                                                  ca->split_data_f());
+    private_submodules_->beamformer->AnalyzeChunk(*ca->split_data_f());
+    // Discards all channels by the leftmost one.
     ca->set_num_channels(1);

[peah-webrtc, 2016/06/23 22:02:40]

The handling of the number of channels is a bit messy.

As it is now, all the pipeline supports multichannel signals (maybe not in
practice, but at least in theory).

Would it not make sense to here set the number of channels equal to the desired
number of output channels? I.e.:
formats_.api_format.output_stream().num_channels()?

As it is now, that format is overridden when the beamformer is activated which
is not that nice. It may have been necessary to do it like that before this CL,
but now the pipeline supports multichannel processing which would motivate
changing this.

WDYT?

[aluebs-webrtc, 2016/06/24 03:00:54]

That makes sense, but it should be a different CL. It has nothing to do with
pulling out the PostFilter from the Beamformer.

   }
 
   public_submodules_->high_pass_filter->ProcessCaptureAudio(ca);
   RETURN_ON_ERR(public_submodules_->gain_control->AnalyzeCaptureAudio(ca));
   public_submodules_->noise_suppression->AnalyzeCaptureAudio(ca);
 
   // Ensure that the stream delay was set before the call to the
   // AEC ProcessCaptureAudio function.
   if (public_submodules_->echo_cancellation->is_enabled() &&
       !was_stream_delay_set()) {
@@ skipping 20 matching lines @@
   // Ensure that the stream delay was set before the call to the
   // AECM ProcessCaptureAudio function.
   if (public_submodules_->echo_control_mobile->is_enabled() &&
       !was_stream_delay_set()) {
     return AudioProcessing::kStreamParameterNotSetError;
   }
 
   RETURN_ON_ERR(public_submodules_->echo_control_mobile->ProcessCaptureAudio(
       ca, stream_delay_ms()));
 
+  if (capture_nonlocked_.beamformer_enabled) {
+    private_submodules_->beamformer->PostFilter(ca->split_data_f());
+  }
+
   public_submodules_->voice_detection->ProcessCaptureAudio(ca);
 
   if (constants_.use_experimental_agc &&
       public_submodules_->gain_control->is_enabled() &&
       (!capture_nonlocked_.beamformer_enabled ||
        private_submodules_->beamformer->is_target_present())) {
     private_submodules_->agc_manager->Process(
         ca->split_bands_const(0)[kBand0To8kHz], ca->num_frames_per_band(),
         capture_nonlocked_.split_rate);
   }
@@ skipping 465 matching lines @@
         capture_nonlocked_.fwd_proc_format.sample_rate_hz(),
         capture_nonlocked_.split_rate,
         num_proc_channels());
   }
 }
 
 void AudioProcessingImpl::InitializeBeamformer() {
   if (capture_nonlocked_.beamformer_enabled) {
     if (!private_submodules_->beamformer) {
       private_submodules_->beamformer.reset(new NonlinearBeamformer(
-          capture_.array_geometry, capture_.target_direction));
+          capture_.array_geometry, 1u, capture_.target_direction));
     }
     private_submodules_->beamformer->Initialize(kChunkSizeMs,
                                                 capture_nonlocked_.split_rate);
   }
 }
 
 void AudioProcessingImpl::InitializeIntelligibility() {
   if (capture_nonlocked_.intelligibility_enabled) {
     public_submodules_->intelligibility_enhancer.reset(
         new IntelligibilityEnhancer(capture_nonlocked_.split_rate,
@@ skipping 242 matching lines @@
   debug_dump_.capture.event_msg->mutable_config()->CopyFrom(config);
 
   RETURN_ON_ERR(WriteMessageToDebugFile(debug_dump_.debug_file.get(),
                                         &debug_dump_.num_bytes_left_for_log_,
                                         &crit_debug_, &debug_dump_.capture));
   return kNoError;
 }
 #endif  // WEBRTC_AUDIOPROC_DEBUG_DUMP
 
 }  // namespace webrtc