First try on a fake analog volume control interface for use in audioproc_f.

webrtc/modules/audio_processing/test/analog_volume_mapper.h

+/*
+ *  Copyright (c) 2017 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_TEST_ANALOG_VOLUME_MAPPER_H_
+#define WEBRTC_MODULES_AUDIO_PROCESSING_TEST_ANALOG_VOLUME_MAPPER_H_
+
+#include "webrtc/base/checks.h"
+
+namespace webrtc {
+
+// Class for simulating an analog gain controller controlled by
+// webrtc::GainControl. The class wraps a mapping from the current
+// level to a floating point scaling factor abstracting
+// non-linearities in the gain curve of real analog microphones. The
+// intended mode of operation is to use get_scaling_factor() to apply
+// a gain factor to a signal.
+class AnalogLevelMapper {

[AleBzk, 2017/04/21 10:35:49]

I thought about the class name and personally I'm ok with it, but I'd like to
hear Per's opinion.

[peah-webrtc, 2017/04/21 12:04:39]

I definitely think that this CL should include the code hooking this up into
audioproc_f. It is hard to properly review this code without seeing the vision
for how it is to be used.

[peah-webrtc, 2017/04/21 12:04:39]

An alternative to an interface would be to have a class that implements a number
of mappings.

The actual mapping is typically very short, sometimes even a one-liner, and
having a separate class for each mapping causes a lot of boilerplate code to be
re-created just to add a new mapping.

My suggestion is to wait with creating an interface until you have two
implementations of the interface. That is, go with a proper class for the first
implementation, and then change to having an interface when you add the second
implementation. Then you will know what overlap there is on the code.

[aleloi, 2017/04/24 11:34:45]

Can I make a dependent CL for that after
https://codereview.webrtc.org/2834643002/ has landed?

[aleloi, 2017/04/24 11:34:45]

That makes sense. I've changed into const enum to choose the mapping. The enum
is set in the c-tor.

+ public:
+  virtual ~AnalogLevelMapper() = default;
+
+  // As the analog level limits in webrtc::GainControl, |level| must
+  // be within [0, 65535].

[peah-webrtc, 2017/04/21 12:04:39]

Is the maximum level really that high, I thought it was lower, maybe 128 or 256.

[aleloi, 2017/04/24 11:34:45]

In the documentation to set_analog_level_limits it says [0, 65535]. In practice,
0 to 255 seems to be used: k{Max, Min}VolumeLevel in voice_engine_defines.h is 0
and 255. In unit tests the level is set to larger values.

+  void set_analog_level(int level) {
+    RTC_DCHECK_LE(level_, 65535);
+    RTC_DCHECK_LE(0, level_);
+    level_ = level;
+  }
+
+  virtual int analog_level() const { return level_; }
+
+  virtual float get_scaling_factor() const = 0;
+
+ private:
+  int level_ = 0;
+};
+}  // namespace webrtc
+
+#endif  // WEBRTC_MODULES_AUDIO_PROCESSING_TEST_ANALOG_VOLUME_MAPPER_H_