Speex encoding/decoding.

remoting/codec/audio_encoder_speex.cc

+// Copyright (c) 2012 The Chromium Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style license that can be
+// found in the LICENSE file.
+
+#include "remoting/codec/audio_encoder_speex.h"
+
+#include <string>
+#include <sstream>
+
+#include "base/basictypes.h"
+#include "base/logging.h"
+#include "base/stl_util.h"
+#include "remoting/proto/audio.pb.h"
+#include "third_party/speex/speex.h"
+#include "third_party/speex/include/speex/speex_stereo.h"
+
+namespace {
+// A quality of 8 in wide band mode corresponds to 27,800 bits per second.
+const int kSpeexHighQuality = 8;
+const int kEncodedDataBufferSize = 0xFF;
+}  // namespace
+
+namespace remoting {
+
+AudioEncoderSpeex::AudioEncoderSpeex()
+    : leftover_frames_(0) {
+  // Create and initialize the Speex structures.
+  speex_bits_.reset(new SpeexBits());
+  speex_bits_init(speex_bits_.get());
+  speex_state_ = speex_encoder_init(&speex_wb_mode);
+
+  // Set the encoding quality.
+  int quality = kSpeexHighQuality;
+  speex_encoder_ctl(speex_state_, SPEEX_SET_QUALITY, &quality);
+
+  // Get the frame size and construct the input buffer accordingly.
+  int result = speex_encoder_ctl(speex_state_,
+                                 SPEEX_GET_FRAME_SIZE,
+                                 &speex_frame_size_);
+  leftover_buffer_.reset(
+      new int16[speex_frame_size_ * AudioPacket::CHANNELS_STEREO]);
+  if (result < 0) {
+    LOG(FATAL) << "Failed to get the frame size for the Speex encoder";
+    return;
+  }
+}
+
+AudioEncoderSpeex::~AudioEncoderSpeex() {
+  speex_encoder_destroy(speex_state_);
+  speex_bits_destroy(speex_bits_.get());
+}
+
+scoped_ptr<AudioPacket> AudioEncoderSpeex::Encode(
+    scoped_ptr<AudioPacket> packet) {
+  DCHECK_EQ(AudioPacket::ENCODING_RAW, packet->encoding());
+  DCHECK_EQ(1, packet->data_size());
+  DCHECK_EQ(AudioPacket::BYTES_PER_SAMPLE_2, packet->bytes_per_sample());
+  DCHECK_NE(AudioPacket::SAMPLING_RATE_INVALID, packet->sampling_rate());
+  DCHECK_EQ(AudioPacket::CHANNELS_STEREO, packet->channels());
+
+  int frames_left = packet->data(0).size() / packet->bytes_per_sample()
+      / packet->channels();

[Sergey Ulanov, 2012/08/21 17:32:23]

nit: / goes to the previous line.

[kxing, 2012/08/21 18:10:20]

Done.

+  const int16* next_sample =
+      reinterpret_cast<const int16*>(packet->data(0).data());
+
+  // Create a new packet of encoded data.
+  scoped_ptr<AudioPacket> encoded_packet(new AudioPacket());
+  encoded_packet->set_encoding(AudioPacket::ENCODING_SPEEX);
+  encoded_packet->set_sampling_rate(packet->sampling_rate());
+  encoded_packet->set_bytes_per_sample(packet->bytes_per_sample());
+  encoded_packet->set_channels(packet->channels());
+
+  while (leftover_frames_ + frames_left >= speex_frame_size_) {
+    int16* unencoded_buffer = NULL;
+    int frames_consumed = 0;
+
+    if (leftover_frames_ > 0) {
+      unencoded_buffer = leftover_buffer_.get();
+      frames_consumed = speex_frame_size_ - leftover_frames_;
+
+      memcpy(leftover_buffer_.get() + leftover_frames_ * packet->channels(),
+             next_sample,
+             frames_consumed * packet->bytes_per_sample() * packet->channels());
+
+      leftover_frames_ = 0;
+    } else {
+      unencoded_buffer = const_cast<int16*>(next_sample);
+      frames_consumed = speex_frame_size_;
+    }
+
+    // Transform stereo to mono.
+    speex_encode_stereo_int(unencoded_buffer,
+                            speex_frame_size_,
+                            speex_bits_.get());
+
+    // Encode the frame, treating all samples as integers.
+    speex_encode_int(speex_state_,
+                     unencoded_buffer,
+                     speex_bits_.get());
+
+    next_sample += frames_consumed * packet->channels();
+    frames_left -= frames_consumed;
+
+    std::string* new_data = encoded_packet->add_data();
+    new_data->resize(speex_bits_nbytes(speex_bits_.get()));
+
+    // Copy the encoded data from the bits structure into the buffer.
+    int bytes_written = speex_bits_write(speex_bits_.get(),
+                                         string_as_array(new_data),
+                                         new_data->size());
+
+    // Expect that the bytes are all written.
+    DCHECK_EQ(bytes_written, static_cast<int>(new_data->size()));
+
+    // Reset the bits structure for this frame.
+    speex_bits_reset(speex_bits_.get());
+  }
+
+  // Store the leftover samples.
+  if (frames_left > 0) {
+    CHECK_LE(leftover_frames_ + frames_left, speex_frame_size_);
+    memcpy(leftover_buffer_.get() + leftover_frames_ * packet->channels(),
+           next_sample,
+           frames_left * packet->bytes_per_sample() * packet->channels());
+    leftover_frames_ += frames_left;
+  }
+
+  // Return NULL if there's nothing in the packet.
+  if (encoded_packet->data_size() == 0)
+    return scoped_ptr<AudioPacket>();
+
+  return encoded_packet.Pass();
+}
+
+}  // namespace remoting