AudioEncoder subclass for G722
BUG=3926 R=henrik.lundin@webrtc.org, kjellander@webrtc.org Review URL: https://webrtc-codereview.appspot.com/30259004 git-svn-id: http://webrtc.googlecode.com/svn/trunk@7779 4adac7df-926f-26a2-2b94-8c16560cd09d
This commit is contained in:
@@ -172,6 +172,8 @@ config("g722_config") {
|
|||||||
|
|
||||||
source_set("g722") {
|
source_set("g722") {
|
||||||
sources = [
|
sources = [
|
||||||
|
"codecs/g722/audio_encoder_g722.cc",
|
||||||
|
"codecs/g722/include/audio_encoder_g722.h",
|
||||||
"codecs/g722/include/g722_interface.h",
|
"codecs/g722/include/g722_interface.h",
|
||||||
"codecs/g722/g722_interface.c",
|
"codecs/g722/g722_interface.c",
|
||||||
"codecs/g722/g722_encode.c",
|
"codecs/g722/g722_encode.c",
|
||||||
|
|||||||
119
webrtc/modules/audio_coding/codecs/g722/audio_encoder_g722.cc
Normal file
119
webrtc/modules/audio_coding/codecs/g722/audio_encoder_g722.cc
Normal file
@@ -0,0 +1,119 @@
|
|||||||
|
/*
|
||||||
|
* Copyright (c) 2014 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.
|
||||||
|
*/
|
||||||
|
|
||||||
|
#include "webrtc/modules/audio_coding/codecs/g722/include/audio_encoder_g722.h"
|
||||||
|
|
||||||
|
#include <limits>
|
||||||
|
#include "webrtc/base/checks.h"
|
||||||
|
#include "webrtc/modules/audio_coding/codecs/g722/include/g722_interface.h"
|
||||||
|
|
||||||
|
namespace webrtc {
|
||||||
|
|
||||||
|
namespace {
|
||||||
|
|
||||||
|
const int kSampleRateHz = 16000;
|
||||||
|
|
||||||
|
} // namespace
|
||||||
|
|
||||||
|
AudioEncoderG722::EncoderState::EncoderState() {
|
||||||
|
CHECK_EQ(0, WebRtcG722_CreateEncoder(&encoder));
|
||||||
|
CHECK_EQ(0, WebRtcG722_EncoderInit(encoder));
|
||||||
|
}
|
||||||
|
|
||||||
|
AudioEncoderG722::EncoderState::~EncoderState() {
|
||||||
|
CHECK_EQ(0, WebRtcG722_FreeEncoder(encoder));
|
||||||
|
}
|
||||||
|
|
||||||
|
AudioEncoderG722::AudioEncoderG722(const Config& config)
|
||||||
|
: num_channels_(config.num_channels),
|
||||||
|
num_10ms_frames_per_packet_(config.frame_size_ms / 10),
|
||||||
|
num_10ms_frames_buffered_(0),
|
||||||
|
first_timestamp_in_buffer_(0),
|
||||||
|
encoders_(new EncoderState[num_channels_]),
|
||||||
|
interleave_buffer_(new uint8_t[2 * num_channels_]) {
|
||||||
|
CHECK_EQ(config.frame_size_ms % 10, 0)
|
||||||
|
<< "Frame size must be an integer multiple of 10 ms.";
|
||||||
|
const int samples_per_channel =
|
||||||
|
kSampleRateHz / 100 * num_10ms_frames_per_packet_;
|
||||||
|
for (int i = 0; i < num_channels_; ++i) {
|
||||||
|
encoders_[i].speech_buffer.reset(new int16_t[samples_per_channel]);
|
||||||
|
encoders_[i].encoded_buffer.reset(new uint8_t[samples_per_channel / 2]);
|
||||||
|
}
|
||||||
|
}
|
||||||
|
|
||||||
|
AudioEncoderG722::~AudioEncoderG722() {}
|
||||||
|
|
||||||
|
int AudioEncoderG722::sample_rate_hz() const {
|
||||||
|
return kSampleRateHz;
|
||||||
|
}
|
||||||
|
int AudioEncoderG722::num_channels() const {
|
||||||
|
return num_channels_;
|
||||||
|
}
|
||||||
|
int AudioEncoderG722::Num10MsFramesInNextPacket() const {
|
||||||
|
return num_10ms_frames_per_packet_;
|
||||||
|
}
|
||||||
|
|
||||||
|
bool AudioEncoderG722::Encode(uint32_t timestamp,
|
||||||
|
const int16_t* audio,
|
||||||
|
size_t max_encoded_bytes,
|
||||||
|
uint8_t* encoded,
|
||||||
|
size_t* encoded_bytes,
|
||||||
|
EncodedInfo* info) {
|
||||||
|
const int samples_per_channel =
|
||||||
|
kSampleRateHz / 100 * num_10ms_frames_per_packet_;
|
||||||
|
CHECK_GE(max_encoded_bytes,
|
||||||
|
static_cast<size_t>(samples_per_channel) / 2 * num_channels_);
|
||||||
|
|
||||||
|
if (num_10ms_frames_buffered_ == 0)
|
||||||
|
first_timestamp_in_buffer_ = timestamp;
|
||||||
|
|
||||||
|
// Deinterleave samples and save them in each channel's buffer.
|
||||||
|
const int start = kSampleRateHz / 100 * num_10ms_frames_buffered_;
|
||||||
|
for (int i = 0; i < kSampleRateHz / 100; ++i)
|
||||||
|
for (int j = 0; j < num_channels_; ++j)
|
||||||
|
encoders_[j].speech_buffer[start + i] = audio[i * num_channels_ + j];
|
||||||
|
|
||||||
|
// If we don't yet have enough samples for a packet, we're done for now.
|
||||||
|
if (++num_10ms_frames_buffered_ < num_10ms_frames_per_packet_) {
|
||||||
|
*encoded_bytes = 0;
|
||||||
|
return true;
|
||||||
|
}
|
||||||
|
|
||||||
|
// Encode each channel separately.
|
||||||
|
CHECK_EQ(num_10ms_frames_buffered_, num_10ms_frames_per_packet_);
|
||||||
|
num_10ms_frames_buffered_ = 0;
|
||||||
|
for (int i = 0; i < num_channels_; ++i) {
|
||||||
|
const int encoded = WebRtcG722_Encode(
|
||||||
|
encoders_[i].encoder, encoders_[i].speech_buffer.get(),
|
||||||
|
samples_per_channel, encoders_[i].encoded_buffer.get());
|
||||||
|
if (encoded < 0)
|
||||||
|
return false;
|
||||||
|
CHECK_EQ(encoded, samples_per_channel / 2);
|
||||||
|
}
|
||||||
|
|
||||||
|
// Interleave the encoded bytes of the different channels. Each separate
|
||||||
|
// channel and the interleaved stream encodes two samples per byte, most
|
||||||
|
// significant half first.
|
||||||
|
for (int i = 0; i < samples_per_channel / 2; ++i) {
|
||||||
|
for (int j = 0; j < num_channels_; ++j) {
|
||||||
|
uint8_t two_samples = encoders_[j].encoded_buffer[i];
|
||||||
|
interleave_buffer_[j] = two_samples >> 4;
|
||||||
|
interleave_buffer_[num_channels_ + j] = two_samples & 0xf;
|
||||||
|
}
|
||||||
|
for (int j = 0; j < num_channels_; ++j)
|
||||||
|
encoded[i * num_channels_ + j] =
|
||||||
|
interleave_buffer_[2 * j] << 4 | interleave_buffer_[2 * j + 1];
|
||||||
|
}
|
||||||
|
*encoded_bytes = samples_per_channel / 2 * num_channels_;
|
||||||
|
info->encoded_timestamp = first_timestamp_in_buffer_;
|
||||||
|
return true;
|
||||||
|
}
|
||||||
|
|
||||||
|
} // namespace webrtc
|
||||||
@@ -21,6 +21,8 @@
|
|||||||
],
|
],
|
||||||
},
|
},
|
||||||
'sources': [
|
'sources': [
|
||||||
|
'audio_encoder_g722.cc',
|
||||||
|
'include/audio_encoder_g722.h',
|
||||||
'include/g722_interface.h',
|
'include/g722_interface.h',
|
||||||
'g722_interface.c',
|
'g722_interface.c',
|
||||||
'g722_encode.c',
|
'g722_encode.c',
|
||||||
|
|||||||
@@ -46,9 +46,9 @@ int16_t WebRtcG722_FreeEncoder(G722EncInst *G722enc_inst)
|
|||||||
}
|
}
|
||||||
|
|
||||||
int16_t WebRtcG722_Encode(G722EncInst *G722enc_inst,
|
int16_t WebRtcG722_Encode(G722EncInst *G722enc_inst,
|
||||||
int16_t *speechIn,
|
const int16_t* speechIn,
|
||||||
int16_t len,
|
int16_t len,
|
||||||
int16_t *encoded)
|
uint8_t* encoded)
|
||||||
{
|
{
|
||||||
unsigned char *codechar = (unsigned char*) encoded;
|
unsigned char *codechar = (unsigned char*) encoded;
|
||||||
// Encode the input speech vector
|
// Encode the input speech vector
|
||||||
|
|||||||
@@ -0,0 +1,64 @@
|
|||||||
|
/*
|
||||||
|
* Copyright (c) 2014 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_CODING_CODECS_G722_INCLUDE_AUDIO_ENCODER_G722_H_
|
||||||
|
#define WEBRTC_MODULES_AUDIO_CODING_CODECS_G722_INCLUDE_AUDIO_ENCODER_G722_H_
|
||||||
|
|
||||||
|
#include "webrtc/modules/audio_coding/codecs/audio_encoder.h"
|
||||||
|
#include "webrtc/modules/audio_coding/codecs/g722/include/g722_interface.h"
|
||||||
|
#include "webrtc/system_wrappers/interface/scoped_ptr.h"
|
||||||
|
|
||||||
|
namespace webrtc {
|
||||||
|
|
||||||
|
class AudioEncoderG722 : public AudioEncoder {
|
||||||
|
public:
|
||||||
|
struct Config {
|
||||||
|
Config() : payload_type(9), frame_size_ms(20), num_channels(1) {}
|
||||||
|
|
||||||
|
int payload_type;
|
||||||
|
int frame_size_ms;
|
||||||
|
int num_channels;
|
||||||
|
};
|
||||||
|
|
||||||
|
explicit AudioEncoderG722(const Config& config);
|
||||||
|
virtual ~AudioEncoderG722();
|
||||||
|
|
||||||
|
virtual int sample_rate_hz() const OVERRIDE;
|
||||||
|
virtual int num_channels() const OVERRIDE;
|
||||||
|
virtual int Num10MsFramesInNextPacket() const OVERRIDE;
|
||||||
|
|
||||||
|
protected:
|
||||||
|
virtual bool Encode(uint32_t timestamp,
|
||||||
|
const int16_t* audio,
|
||||||
|
size_t max_encoded_bytes,
|
||||||
|
uint8_t* encoded,
|
||||||
|
size_t* encoded_bytes,
|
||||||
|
EncodedInfo* info) OVERRIDE;
|
||||||
|
|
||||||
|
private:
|
||||||
|
// The encoder state for one channel.
|
||||||
|
struct EncoderState {
|
||||||
|
G722EncInst* encoder;
|
||||||
|
scoped_ptr<int16_t[]> speech_buffer; // Queued up for encoding.
|
||||||
|
scoped_ptr<uint8_t[]> encoded_buffer; // Already encoded.
|
||||||
|
EncoderState();
|
||||||
|
~EncoderState();
|
||||||
|
};
|
||||||
|
|
||||||
|
const int num_channels_;
|
||||||
|
const int num_10ms_frames_per_packet_;
|
||||||
|
int num_10ms_frames_buffered_;
|
||||||
|
uint32_t first_timestamp_in_buffer_;
|
||||||
|
const scoped_ptr<EncoderState[]> encoders_;
|
||||||
|
const scoped_ptr<uint8_t[]> interleave_buffer_;
|
||||||
|
};
|
||||||
|
|
||||||
|
} // namespace webrtc
|
||||||
|
#endif // WEBRTC_MODULES_AUDIO_CODING_CODECS_G722_INCLUDE_AUDIO_ENCODER_G722_H_
|
||||||
@@ -96,9 +96,9 @@ int16_t WebRtcG722_FreeEncoder(G722EncInst *G722enc_inst);
|
|||||||
*/
|
*/
|
||||||
|
|
||||||
int16_t WebRtcG722_Encode(G722EncInst* G722enc_inst,
|
int16_t WebRtcG722_Encode(G722EncInst* G722enc_inst,
|
||||||
int16_t *speechIn,
|
const int16_t* speechIn,
|
||||||
int16_t len,
|
int16_t len,
|
||||||
int16_t *encoded);
|
uint8_t* encoded);
|
||||||
|
|
||||||
|
|
||||||
/****************************************************************************
|
/****************************************************************************
|
||||||
|
|||||||
@@ -62,7 +62,7 @@ int main(int argc, char* argv[])
|
|||||||
int16_t stream_len = 0;
|
int16_t stream_len = 0;
|
||||||
int16_t shortdata[960];
|
int16_t shortdata[960];
|
||||||
int16_t decoded[960];
|
int16_t decoded[960];
|
||||||
int16_t streamdata[80*3];
|
uint8_t streamdata[80 * 6];
|
||||||
int16_t speechType[1];
|
int16_t speechType[1];
|
||||||
|
|
||||||
/* handling wrong input arguments in the command line */
|
/* handling wrong input arguments in the command line */
|
||||||
@@ -124,7 +124,9 @@ int main(int argc, char* argv[])
|
|||||||
|
|
||||||
/* G.722 encoding + decoding */
|
/* G.722 encoding + decoding */
|
||||||
stream_len = WebRtcG722_Encode((G722EncInst *)G722enc_inst, shortdata, framelength, streamdata);
|
stream_len = WebRtcG722_Encode((G722EncInst *)G722enc_inst, shortdata, framelength, streamdata);
|
||||||
err = WebRtcG722_Decode((G722DecInst *)G722dec_inst, streamdata, stream_len, decoded, speechType);
|
err = WebRtcG722_Decode(G722dec_inst,
|
||||||
|
reinterpret_cast<int16_t*>(streamdata),
|
||||||
|
stream_len, decoded, speechType);
|
||||||
|
|
||||||
/* Stop clock after call to encoder and decoder */
|
/* Stop clock after call to encoder and decoder */
|
||||||
runtime += (double)((clock()/(double)CLOCKS_PER_SEC_G722)-starttime);
|
runtime += (double)((clock()/(double)CLOCKS_PER_SEC_G722)-starttime);
|
||||||
|
|||||||
@@ -115,11 +115,11 @@ int16_t ACMG722::InternalEncode(uint8_t* bitstream,
|
|||||||
}
|
}
|
||||||
len_in_bytes = WebRtcG722_Encode(
|
len_in_bytes = WebRtcG722_Encode(
|
||||||
encoder_inst_ptr_, left_channel, frame_len_smpl_,
|
encoder_inst_ptr_, left_channel, frame_len_smpl_,
|
||||||
reinterpret_cast<int16_t*>(out_left));
|
out_left);
|
||||||
len_in_bytes += WebRtcG722_Encode(encoder_inst_ptr_right_,
|
len_in_bytes += WebRtcG722_Encode(encoder_inst_ptr_right_,
|
||||||
right_channel,
|
right_channel,
|
||||||
frame_len_smpl_,
|
frame_len_smpl_,
|
||||||
reinterpret_cast<int16_t*>(out_right));
|
out_right);
|
||||||
*bitstream_len_byte = len_in_bytes;
|
*bitstream_len_byte = len_in_bytes;
|
||||||
|
|
||||||
// Interleave the 4 bits per sample from left and right channel
|
// Interleave the 4 bits per sample from left and right channel
|
||||||
@@ -130,7 +130,7 @@ int16_t ACMG722::InternalEncode(uint8_t* bitstream,
|
|||||||
} else {
|
} else {
|
||||||
*bitstream_len_byte = WebRtcG722_Encode(
|
*bitstream_len_byte = WebRtcG722_Encode(
|
||||||
encoder_inst_ptr_, &in_audio_[in_audio_ix_read_], frame_len_smpl_,
|
encoder_inst_ptr_, &in_audio_[in_audio_ix_read_], frame_len_smpl_,
|
||||||
reinterpret_cast<int16_t*>(bitstream));
|
bitstream);
|
||||||
}
|
}
|
||||||
|
|
||||||
// increment the read index this tell the caller how far
|
// increment the read index this tell the caller how far
|
||||||
|
|||||||
@@ -22,7 +22,7 @@
|
|||||||
#endif
|
#endif
|
||||||
#include "webrtc/modules/audio_coding/codecs/g711/include/g711_interface.h"
|
#include "webrtc/modules/audio_coding/codecs/g711/include/g711_interface.h"
|
||||||
#include "webrtc/modules/audio_coding/codecs/g711/include/audio_encoder_pcm.h"
|
#include "webrtc/modules/audio_coding/codecs/g711/include/audio_encoder_pcm.h"
|
||||||
#include "webrtc/modules/audio_coding/codecs/g722/include/g722_interface.h"
|
#include "webrtc/modules/audio_coding/codecs/g722/include/audio_encoder_g722.h"
|
||||||
#include "webrtc/modules/audio_coding/codecs/ilbc/interface/ilbc.h"
|
#include "webrtc/modules/audio_coding/codecs/ilbc/interface/ilbc.h"
|
||||||
#include "webrtc/modules/audio_coding/codecs/isac/fix/interface/isacfix.h"
|
#include "webrtc/modules/audio_coding/codecs/isac/fix/interface/isacfix.h"
|
||||||
#include "webrtc/modules/audio_coding/codecs/isac/main/interface/isac.h"
|
#include "webrtc/modules/audio_coding/codecs/isac/main/interface/isac.h"
|
||||||
@@ -483,67 +483,26 @@ class AudioDecoderG722Test : public AudioDecoderTest {
|
|||||||
data_length_ = 10 * frame_size_;
|
data_length_ = 10 * frame_size_;
|
||||||
decoder_ = new AudioDecoderG722;
|
decoder_ = new AudioDecoderG722;
|
||||||
assert(decoder_);
|
assert(decoder_);
|
||||||
WebRtcG722_CreateEncoder(&encoder_);
|
AudioEncoderG722::Config config;
|
||||||
|
config.frame_size_ms = 10;
|
||||||
|
config.num_channels = 1;
|
||||||
|
audio_encoder_.reset(new AudioEncoderG722(config));
|
||||||
}
|
}
|
||||||
|
|
||||||
~AudioDecoderG722Test() {
|
|
||||||
WebRtcG722_FreeEncoder(encoder_);
|
|
||||||
}
|
|
||||||
|
|
||||||
virtual void InitEncoder() {
|
|
||||||
ASSERT_EQ(0, WebRtcG722_EncoderInit(encoder_));
|
|
||||||
}
|
|
||||||
|
|
||||||
virtual int EncodeFrame(const int16_t* input, size_t input_len_samples,
|
|
||||||
uint8_t* output) {
|
|
||||||
int enc_len_bytes =
|
|
||||||
WebRtcG722_Encode(encoder_, const_cast<int16_t*>(input),
|
|
||||||
static_cast<int>(input_len_samples),
|
|
||||||
reinterpret_cast<int16_t*>(output));
|
|
||||||
EXPECT_EQ(80, enc_len_bytes);
|
|
||||||
return enc_len_bytes;
|
|
||||||
}
|
|
||||||
|
|
||||||
G722EncInst* encoder_;
|
|
||||||
};
|
};
|
||||||
|
|
||||||
class AudioDecoderG722StereoTest : public AudioDecoderG722Test {
|
class AudioDecoderG722StereoTest : public AudioDecoderTest {
|
||||||
protected:
|
protected:
|
||||||
AudioDecoderG722StereoTest() : AudioDecoderG722Test() {
|
AudioDecoderG722StereoTest() : AudioDecoderTest() {
|
||||||
channels_ = 2;
|
channels_ = 2;
|
||||||
// Delete the |decoder_| that was created by AudioDecoderG722Test and
|
codec_input_rate_hz_ = 16000;
|
||||||
// create an AudioDecoderG722Stereo object instead.
|
frame_size_ = 160;
|
||||||
delete decoder_;
|
data_length_ = 10 * frame_size_;
|
||||||
decoder_ = new AudioDecoderG722Stereo;
|
decoder_ = new AudioDecoderG722Stereo;
|
||||||
assert(decoder_);
|
assert(decoder_);
|
||||||
}
|
AudioEncoderG722::Config config;
|
||||||
|
config.frame_size_ms = 10;
|
||||||
virtual int EncodeFrame(const int16_t* input, size_t input_len_samples,
|
config.num_channels = 2;
|
||||||
uint8_t* output) {
|
audio_encoder_.reset(new AudioEncoderG722(config));
|
||||||
uint8_t* temp_output = new uint8_t[data_length_ * 2];
|
|
||||||
// Encode a mono payload using the base test class.
|
|
||||||
int mono_enc_len_bytes =
|
|
||||||
AudioDecoderG722Test::EncodeFrame(input, input_len_samples,
|
|
||||||
temp_output);
|
|
||||||
// The bit-stream consists of 4-bit samples:
|
|
||||||
// +--------+--------+--------+
|
|
||||||
// | s0 s1 | s2 s3 | s4 s5 |
|
|
||||||
// +--------+--------+--------+
|
|
||||||
//
|
|
||||||
// Duplicate them to the |output| such that the stereo stream becomes:
|
|
||||||
// +--------+--------+--------+
|
|
||||||
// | s0 s0 | s1 s1 | s2 s2 |
|
|
||||||
// +--------+--------+--------+
|
|
||||||
EXPECT_LE(mono_enc_len_bytes * 2, static_cast<int>(data_length_ * 2));
|
|
||||||
uint8_t* output_ptr = output;
|
|
||||||
for (int i = 0; i < mono_enc_len_bytes; ++i) {
|
|
||||||
*output_ptr = (temp_output[i] & 0xF0) + (temp_output[i] >> 4);
|
|
||||||
++output_ptr;
|
|
||||||
*output_ptr = (temp_output[i] << 4) + (temp_output[i] & 0x0F);
|
|
||||||
++output_ptr;
|
|
||||||
}
|
|
||||||
delete [] temp_output;
|
|
||||||
return mono_enc_len_bytes * 2;
|
|
||||||
}
|
}
|
||||||
};
|
};
|
||||||
|
|
||||||
|
|||||||
@@ -1615,7 +1615,7 @@ int NetEQTest_encode(int coder, int16_t *indata, int frameLen, unsigned char * e
|
|||||||
#endif
|
#endif
|
||||||
#ifdef CODEC_G722
|
#ifdef CODEC_G722
|
||||||
else if (coder==webrtc::kDecoderG722) { /*g722 */
|
else if (coder==webrtc::kDecoderG722) { /*g722 */
|
||||||
cdlen=WebRtcG722_Encode(g722EncState[k], indata, frameLen, (int16_t*)encoded);
|
cdlen=WebRtcG722_Encode(g722EncState[k], indata, frameLen, encoded);
|
||||||
assert(cdlen == frameLen>>1);
|
assert(cdlen == frameLen>>1);
|
||||||
}
|
}
|
||||||
#endif
|
#endif
|
||||||
|
|||||||
Reference in New Issue
Block a user