Pull the Voice Activity Detector out from the AGC

webrtc/modules/audio_processing/vad/voice_activity_detector.h

Index: webrtc/modules/audio_processing/vad/voice_activity_detector.h
diff --git a/webrtc/modules/audio_processing/vad/voice_activity_detector.h b/webrtc/modules/audio_processing/vad/voice_activity_detector.h
new file mode 100644
index 0000000000000000000000000000000000000000..8c3325606c47ffb91c3f79a6443b0b7419593f31
--- /dev/null
+++ b/webrtc/modules/audio_processing/vad/voice_activity_detector.h
@@ -0,0 +1,64 @@
+/*

[Andrew MacDonald, 2015/06/15 04:32:25]

After all reviewers have LGTM'd, please run git cl format on your totally new
files.

[aluebs-webrtc, 2015/06/16 01:17:52]

Acknowledged.

+ *  Copyright (c) 2015 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.
+ */
+
+#ifndef WEBRTC_MODULES_AUDIO_PROCESSING_VAD_VOICE_ACTIVITY_DETECTOR_H_
+#define WEBRTC_MODULES_AUDIO_PROCESSING_VAD_VOICE_ACTIVITY_DETECTOR_H_
+
+#include <vector>
+
+#include "webrtc/base/scoped_ptr.h"
+#include "webrtc/common_audio/resampler/include/resampler.h"
+#include "webrtc/modules/audio_processing/vad/vad_audio_proc.h"
+#include "webrtc/modules/audio_processing/vad/common.h"
+#include "webrtc/modules/audio_processing/vad/pitch_based_vad.h"
+#include "webrtc/modules/audio_processing/vad/standalone_vad.h"
+
+namespace webrtc {
+
+// A Voice Activity Detector (VAD) that combines the voice probability from the
+// StandaloneVad and PitchBasedVad to get a more robust estimation.
+class VoiceActivityDetector {
+ public:
+  VoiceActivityDetector();
+
+  // Processes each capture audio chunk and estimates the voice probability. The
+  // maximum supported sample rate is 32kHz.
+  void ProcessCaptureAudio(const int16_t* audio, int length);

[Andrew MacDonald, 2015/06/15 04:32:25]

Perhaps ProcessChunk instead? It doesn't have to be captured audio.

[aluebs-webrtc, 2015/06/16 01:17:53]

Good point! Done.

+
+  // Because ISAC has a different chunk length, it returns a zero size vector

[Andrew MacDonald, 2015/06/15 04:32:25]

This is an implementation detail; you could put it in the implementation of
ProcessCaptureAudio. Here, I think you should discuss what this method returns
exactly (and for RMS below).

[aluebs-webrtc, 2015/06/16 01:17:52]

Done.

+  // every time there is no new data and then return a few chunk's data at once.

[bloch, 2015/06/12 23:34:38]

I'm thinking about what we discussed; I think I (personally) prefer this method.
It's closer to honestly reporting chunk-by-chunk data. However, will the
returned vector have all identical values? If so, it might make sense to just
return one. If not, it'll likely be too late to update the probability of a past
chunk--assuming you're using software that works in real-time.

[aluebs-webrtc, 2015/06/16 01:17:52]

It will not have identical values. As you point out, in real time, you can't
update the probability of a past chunk. But it is used by the AGC for updating a
histograms with all the values.

+  std::vector<double> chunkwise_voice_probabilities() const {
+    return chunkwise_voice_probabilities_;

[Andrew MacDonald, 2015/06/15 04:32:26]

Since this is called on every chunk, you don't want to return by value here.
Return a "const std::vector<double>&" instead.

[aluebs-webrtc, 2015/06/16 01:17:52]

Done.

+  }
+  std::vector<double> chunkwise_rms() const { return chunkwise_rms_; }

[Andrew MacDonald, 2015/06/15 04:32:25]

As above, return a const reference.

[aluebs-webrtc, 2015/06/16 01:17:52]

Done.

+
+  // Returns the last voice probability, regardless of the internal
+  // implementation, although it has a few chunks of delay.
+  float last_voice_probability() const { return last_voice_probability_; }

[Andrew MacDonald, 2015/06/15 04:32:25]

Should definitely convert things to float later given this inconsistency :)

[aluebs-webrtc, 2015/06/16 01:17:53]

Fixed this inconsistency in this CL, since it makes no sense, but will convert
everything to float on another CL.

+
+ private:
+  std::vector<double> chunkwise_voice_probabilities_;
+  std::vector<double> chunkwise_rms_;
+
+  double last_voice_probability_;

[bloch, 2015/06/12 23:34:38]

I know you're just refactoring old code, but is there a reason we're returning
doubles instead of floats? There's no way we have enough significant digits from
the calculation to need double the precision of just a floating-point number,
right?

[Andrew MacDonald, 2015/06/15 04:32:26]

Agreed, I'm sure these could be floats. However it involves touching the moved
code as well. Alex, do you want to do that in a follow-up?

[aluebs-webrtc, 2015/06/16 01:17:52]

Yes, I can do this in a follow up CL. But I changed to float here, to avoid a
cast.

+
+  Resampler resampler_;
+  VadAudioProc audio_processing_;
+
+  rtc::scoped_ptr<StandaloneVad> standalone_vad_;
+  PitchBasedVad pitch_based_vad_;
+
+  int16_t resampled_[kLength10Ms];
+  AudioFeatures features_;
+};
+
+}  // namespace webrtc
+
+#endif  // WEBRTC_MODULES_AUDIO_PROCESSING_VAD_VOICE_ACTIVITY_DETECTOR_H_