Pull out the PostFilter to its own NonlinearBeamformer API

webrtc/modules/audio_processing/beamformer/nonlinear_beamformer.h

Index: webrtc/modules/audio_processing/beamformer/nonlinear_beamformer.h
diff --git a/webrtc/modules/audio_processing/beamformer/nonlinear_beamformer.h b/webrtc/modules/audio_processing/beamformer/nonlinear_beamformer.h
index b8953b0a4f8d2d4c1e0083b8de7cba0c37e80dc5..063107df356175e0c4925885cbf3593925793c66 100644
--- a/webrtc/modules/audio_processing/beamformer/nonlinear_beamformer.h
+++ b/webrtc/modules/audio_processing/beamformer/nonlinear_beamformer.h
@@ -21,20 +21,37 @@
 
 #include "webrtc/common_audio/lapped_transform.h"
 #include "webrtc/common_audio/channel_buffer.h"
-#include "webrtc/modules/audio_processing/beamformer/beamformer.h"
+#include "webrtc/modules/audio_processing/beamformer/array_util.h"
 #include "webrtc/modules/audio_processing/beamformer/complex_matrix.h"
 
 namespace webrtc {
 
+class PostFilterTransform : public LappedTransform::Callback {
+ public:
+  PostFilterTransform(size_t chunk_length, float* window, size_t fft_size);
+
+  void ProcessChunk(float* const* data, float* final_mask);
+
+ protected:
+  void ProcessAudioBlock(const complex<float>* const* input,
+                         size_t num_input_channels,
+                         size_t num_freq_bins,
+                         size_t num_output_channels,
+                         complex<float>* const* output) override;
+
+ private:
+  LappedTransform transform_;
+  const size_t num_freq_bins_;
+  float* final_mask_;
+};
+
 // Enhances sound sources coming directly in front of a uniform linear array
 // and suppresses sound sources coming from all other directions. Operates on
 // multichannel signals and produces single-channel output.
 //
 // The implemented nonlinear postfilter algorithm taken from "A Robust Nonlinear
 // Beamforming Postprocessor" by Bastiaan Kleijn.
-class NonlinearBeamformer
-  : public Beamformer<float>,
-    public LappedTransform::Callback {
+class NonlinearBeamformer : public LappedTransform::Callback {
  public:
   static const float kHalfBeamWidthRadians;
 
@@ -45,24 +62,28 @@ class NonlinearBeamformer
 
   // Sample rate corresponds to the lower band.
   // Needs to be called before the NonlinearBeamformer can be used.
-  void Initialize(int chunk_size_ms, int sample_rate_hz) override;
+  void Initialize(int chunk_size_ms, int sample_rate_hz);
 
-  // Process one time-domain chunk of audio. The audio is expected to be split
+  // Analyzes one time-domain chunk of audio. The audio is expected to be split
   // into frequency bands inside the ChannelBuffer. The number of frames and
-  // channels must correspond to the constructor parameters. The same
-  // ChannelBuffer can be passed in as |input| and |output|.
-  void ProcessChunk(const ChannelBuffer<float>& input,
-                    ChannelBuffer<float>* output) override;
+  // channels must correspond to the constructor parameters.
+  void AnalyzeChunk(const ChannelBuffer<float>& data);
+  // Applies the postfilter mask to one chunk of audio. The audio is expected to
+  // be split into frequency bands inside the ChannelBuffer. The number of
+  // frames must correspond to the constructor parameters and the number of
+  // channels is expected to be 1, since that is the output number of channels
+  // of ProcessChunk().
+  void PostFilter(ChannelBuffer<float>* data);
 
-  void AimAt(const SphericalPointf& target_direction) override;
+  void AimAt(const SphericalPointf& target_direction);
 
-  bool IsInBeam(const SphericalPointf& spherical_point) override;
+  bool IsInBeam(const SphericalPointf& spherical_point);
 
   // After processing each block |is_target_present_| is set to true if the
   // target signal es present and to false otherwise. This methods can be called
   // to know if the data is target signal or interference and process it
   // accordingly.
-  bool is_target_present() override { return is_target_present_; }
+  bool is_target_present() { return is_target_present_; }
 
  protected:
   // Process one frequency-domain block of audio. This is where the fun
@@ -116,8 +137,8 @@ class NonlinearBeamformer
   // Compute the means needed for the above frequency correction.
   float MaskRangeMean(size_t start_bin, size_t end_bin);
 
-  // Applies both sets of masks to |input| and store in |output|.
-  void ApplyMasks(const complex_f* const* input, complex_f* const* output);
+  // Applies post-filter mask to |input| and store in |output|.
+  void ApplyPostFilter(const complex_f* input, complex_f* output);
 
   void EstimateTargetPresence();
 
@@ -126,9 +147,12 @@ class NonlinearBeamformer
 
   // Deals with the fft transform and blocking.
   size_t chunk_length_;
-  std::unique_ptr<LappedTransform> lapped_transform_;
+  std::unique_ptr<LappedTransform> process_transform_;
+  std::unique_ptr<PostFilterTransform> postfilter_transform_;
   float window_[kFftSize];
 
+  std::unique_ptr<ChannelBuffer<float>> dummy_out_;

[peah-webrtc, 2016/06/08 12:04:55]

If possible, it would be nice to avoid this one. Please see my other comments
about that.

[aluebs-webrtc, 2016/06/09 02:11:46]

Removed.

+
   // Parameters exposed to the user.
   const size_t num_input_channels_;
   int sample_rate_hz_;
@@ -161,7 +185,6 @@ class NonlinearBeamformer
 
   // Array of length |kNumFreqBins|, Matrix of size |1| x |num_channels_|.
   ComplexMatrixF delay_sum_masks_[kNumFreqBins];
-  ComplexMatrixF normalized_delay_sum_masks_[kNumFreqBins];
 
   // Arrays of length |kNumFreqBins|, Matrix of size |num_input_channels_| x
   // |num_input_channels_|.
@@ -186,6 +209,7 @@ class NonlinearBeamformer
 
   // For processing the high-frequency input signal.
   float high_pass_postfilter_mask_;
+  float old_high_pass_mask_;
 
   // True when the target signal is present.
   bool is_target_present_;