Add a deinterleaved float interface to AudioProcessing.

This is mainly to support the native audio format in Chrome. Although
this implementation just moves the float->int conversion under the hood,
we will transition AudioProcessing towards supporting this format
throughout.

- Add a test which verifies we get identical output with the float and
int interfaces.
- The float and int wrappers are tasked with conversion to the
AudioBuffer format. A new shared Process/Analyze method does most of
the work.
- Add a new field to the debug.proto to hold deinterleaved data.
- Add helpers to audio_utils.cc, and start using numeric_limits.
- Note that there was no performance difference between numeric_limits
and a literal value when measured on Linux using gcc or clang.

BUG=2894
R=aluebs@webrtc.org, bjornv@webrtc.org, henrikg@webrtc.org, tommi@webrtc.org, turaj@webrtc.org, xians@webrtc.org

Review URL: https://webrtc-codereview.appspot.com/9179004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@5641 4adac7df-926f-26a2-2b94-8c16560cd09d
diff --git a/webrtc/modules/audio_processing/include/audio_processing.h b/webrtc/modules/audio_processing/include/audio_processing.h
index f41c886..394bef8 100644
--- a/webrtc/modules/audio_processing/include/audio_processing.h
+++ b/webrtc/modules/audio_processing/include/audio_processing.h
@@ -135,6 +135,16 @@
 //
 class AudioProcessing {
  public:
+  enum ChannelLayout {
+    kMono,
+    // Left, right.
+    kStereo,
+    // Mono, keyboard mic.
+    kMonoAndKeyboard,
+    // Left, right, keyboard mic.
+    kStereoAndKeyboard
+  };
+
   // Creates an APM instance. Use one instance for every primary audio stream
   // requiring processing. On the client-side, this would typically be one
   // instance for the near-end stream, and additional instances for each far-end
@@ -205,6 +215,17 @@
   // method, it will trigger an initialization.
   virtual int ProcessStream(AudioFrame* frame) = 0;
 
+  // Accepts deinterleaved float audio with the range [-1, 1]. Each element
+  // of |data| points to a channel buffer, arranged according to
+  // |input_layout|. At output, the channels will be arranged according to
+  // |output_layout|.
+  // TODO(ajm): Output layout conversion does not yet work.
+  virtual int ProcessStream(float* const* data,
+                            int samples_per_channel,
+                            int sample_rate_hz,
+                            ChannelLayout input_layout,
+                            ChannelLayout output_layout) = 0;
+
   // Analyzes a 10 ms |frame| of the reverse direction audio stream. The frame
   // will not be modified. On the client-side, this is the far-end (or to be
   // rendered) audio.
@@ -222,6 +243,13 @@
   // TODO(ajm): add const to input; requires an implementation fix.
   virtual int AnalyzeReverseStream(AudioFrame* frame) = 0;
 
+  // Accepts deinterleaved float audio with the range [-1, 1]. Each element
+  // of |data| points to a channel buffer, arranged according to |layout|.
+  virtual int AnalyzeReverseStream(const float* const* data,
+                                   int samples_per_channel,
+                                   int sample_rate_hz,
+                                   ChannelLayout layout) = 0;
+
   // This must be called if and only if echo processing is enabled.
   //
   // Sets the |delay| in ms between AnalyzeReverseStream() receiving a far-end