1 /*
2  *  Copyright (c) 2014 The WebRTC project authors. All Rights Reserved.
3  *
4  *  Use of this source code is governed by a BSD-style license
5  *  that can be found in the LICENSE file in the root of the source
6  *  tree. An additional intellectual property rights grant can be found
7  *  in the file PATENTS.  All contributing project authors may
8  *  be found in the AUTHORS file in the root of the source tree.
9  */
10 
11 #ifndef WEBRTC_MODULES_AUDIO_CODING_CODECS_AUDIO_ENCODER_H_
12 #define WEBRTC_MODULES_AUDIO_CODING_CODECS_AUDIO_ENCODER_H_
13 
14 #include <algorithm>
15 #include <vector>
16 
17 #include "webrtc/base/array_view.h"
18 #include "webrtc/typedefs.h"
19 
20 namespace webrtc {
21 
22 // This is the interface class for encoders in AudioCoding module. Each codec
23 // type must have an implementation of this class.
24 class AudioEncoder {
25  public:
26   struct EncodedInfoLeaf {
27     size_t encoded_bytes = 0;
28     uint32_t encoded_timestamp = 0;
29     int payload_type = 0;
30     bool send_even_if_empty = false;
31     bool speech = true;
32   };
33 
34   // This is the main struct for auxiliary encoding information. Each encoded
35   // packet should be accompanied by one EncodedInfo struct, containing the
36   // total number of |encoded_bytes|, the |encoded_timestamp| and the
37   // |payload_type|. If the packet contains redundant encodings, the |redundant|
38   // vector will be populated with EncodedInfoLeaf structs. Each struct in the
39   // vector represents one encoding; the order of structs in the vector is the
40   // same as the order in which the actual payloads are written to the byte
41   // stream. When EncoderInfoLeaf structs are present in the vector, the main
42   // struct's |encoded_bytes| will be the sum of all the |encoded_bytes| in the
43   // vector.
44   struct EncodedInfo : public EncodedInfoLeaf {
45     EncodedInfo();
46     ~EncodedInfo();
47 
48     std::vector<EncodedInfoLeaf> redundant;
49   };
50 
51   virtual ~AudioEncoder() = default;
52 
53   // Returns the maximum number of bytes that can be produced by the encoder
54   // at each Encode() call. The caller can use the return value to determine
55   // the size of the buffer that needs to be allocated. This value is allowed
56   // to depend on encoder parameters like bitrate, frame size etc., so if
57   // any of these change, the caller of Encode() is responsible for checking
58   // that the buffer is large enough by calling MaxEncodedBytes() again.
59   virtual size_t MaxEncodedBytes() const = 0;
60 
61   // Returns the input sample rate in Hz and the number of input channels.
62   // These are constants set at instantiation time.
63   virtual int SampleRateHz() const = 0;
64   virtual size_t NumChannels() const = 0;
65 
66   // Returns the rate at which the RTP timestamps are updated. The default
67   // implementation returns SampleRateHz().
68   virtual int RtpTimestampRateHz() const;
69 
70   // Returns the number of 10 ms frames the encoder will put in the next
71   // packet. This value may only change when Encode() outputs a packet; i.e.,
72   // the encoder may vary the number of 10 ms frames from packet to packet, but
73   // it must decide the length of the next packet no later than when outputting
74   // the preceding packet.
75   virtual size_t Num10MsFramesInNextPacket() const = 0;
76 
77   // Returns the maximum value that can be returned by
78   // Num10MsFramesInNextPacket().
79   virtual size_t Max10MsFramesInAPacket() const = 0;
80 
81   // Returns the current target bitrate in bits/s. The value -1 means that the
82   // codec adapts the target automatically, and a current target cannot be
83   // provided.
84   virtual int GetTargetBitrate() const = 0;
85 
86   // Accepts one 10 ms block of input audio (i.e., SampleRateHz() / 100 *
87   // NumChannels() samples). Multi-channel audio must be sample-interleaved.
88   // The encoder produces zero or more bytes of output in |encoded| and
89   // returns additional encoding information.
90   // The caller is responsible for making sure that |max_encoded_bytes| is
91   // not smaller than the number of bytes actually produced by the encoder.
92   // Encode() checks some preconditions, calls EncodeInternal() which does the
93   // actual work, and then checks some postconditions.
94   EncodedInfo Encode(uint32_t rtp_timestamp,
95                      rtc::ArrayView<const int16_t> audio,
96                      size_t max_encoded_bytes,
97                      uint8_t* encoded);
98 
99   virtual EncodedInfo EncodeInternal(uint32_t rtp_timestamp,
100                                      rtc::ArrayView<const int16_t> audio,
101                                      size_t max_encoded_bytes,
102                                      uint8_t* encoded) = 0;
103 
104   // Resets the encoder to its starting state, discarding any input that has
105   // been fed to the encoder but not yet emitted in a packet.
106   virtual void Reset() = 0;
107 
108   // Enables or disables codec-internal FEC (forward error correction). Returns
109   // true if the codec was able to comply. The default implementation returns
110   // true when asked to disable FEC and false when asked to enable it (meaning
111   // that FEC isn't supported).
112   virtual bool SetFec(bool enable);
113 
114   // Enables or disables codec-internal VAD/DTX. Returns true if the codec was
115   // able to comply. The default implementation returns true when asked to
116   // disable DTX and false when asked to enable it (meaning that DTX isn't
117   // supported).
118   virtual bool SetDtx(bool enable);
119 
120   // Sets the application mode. Returns true if the codec was able to comply.
121   // The default implementation just returns false.
122   enum class Application { kSpeech, kAudio };
123   virtual bool SetApplication(Application application);
124 
125   // Tells the encoder about the highest sample rate the decoder is expected to
126   // use when decoding the bitstream. The encoder would typically use this
127   // information to adjust the quality of the encoding. The default
128   // implementation does nothing.
129   virtual void SetMaxPlaybackRate(int frequency_hz);
130 
131   // Tells the encoder what the projected packet loss rate is. The rate is in
132   // the range [0.0, 1.0]. The encoder would typically use this information to
133   // adjust channel coding efforts, such as FEC. The default implementation
134   // does nothing.
135   virtual void SetProjectedPacketLossRate(double fraction);
136 
137   // Tells the encoder what average bitrate we'd like it to produce. The
138   // encoder is free to adjust or disregard the given bitrate (the default
139   // implementation does the latter).
140   virtual void SetTargetBitrate(int target_bps);
141 };
142 }  // namespace webrtc
143 #endif  // WEBRTC_MODULES_AUDIO_CODING_CODECS_AUDIO_ENCODER_H_
144