// 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_decoder_speex.h" #include #include "base/basictypes.h" #include "base/memory/scoped_ptr.h" #include "base/logging.h" #include "base/stl_util.h" #include "remoting/proto/audio.pb.h" #include "third_party/speex/include/speex/speex_callbacks.h" #include "third_party/speex/include/speex/speex_stereo.h" namespace remoting { namespace { // Hosts will never generate more than 100 frames in a single packet. const int kMaxFramesPerPacket = 100; } // namespace AudioDecoderSpeex::AudioDecoderSpeex() { // Create and initialize the Speex structures. speex_bits_.reset(new SpeexBits()); speex_bits_init(speex_bits_.get()); speex_state_ = speex_decoder_init(&speex_wb_mode); // Create and initialize the Speex stereo state. speex_stereo_state_ = speex_stereo_state_init(); // Create and initialize the stereo callback. speex_callback_.reset(new SpeexCallback()); speex_callback_->callback_id = SPEEX_INBAND_STEREO; speex_callback_->func = speex_std_stereo_request_handler; speex_callback_->data = speex_stereo_state_; int result; // Turn on perceptual enhancer, which will make the audio sound better, // at the price of further distorting the decoded samples. int enhancer = 1; result = speex_decoder_ctl(speex_state_, SPEEX_SET_ENH, &enhancer); CHECK_EQ(result, 0); // Get the frame size, so that we know the size of output when we decode // frame by frame. result = speex_decoder_ctl(speex_state_, SPEEX_GET_FRAME_SIZE, &speex_frame_size_); CHECK_EQ(result, 0); // Set the stereo callback, so that the Speex decoder can get the intensity // stereo information. result = speex_decoder_ctl(speex_state_, SPEEX_SET_HANDLER, speex_callback_.get()); CHECK_EQ(result, 0); } AudioDecoderSpeex::~AudioDecoderSpeex() { speex_stereo_state_destroy(speex_stereo_state_); speex_decoder_destroy(speex_state_); speex_bits_destroy(speex_bits_.get()); } scoped_ptr AudioDecoderSpeex::Decode( scoped_ptr packet) { if ((packet->encoding() != AudioPacket::ENCODING_SPEEX) || (packet->bytes_per_sample() != AudioPacket::BYTES_PER_SAMPLE_2) || (packet->sampling_rate() == AudioPacket::SAMPLING_RATE_INVALID) || (packet->channels() != AudioPacket::CHANNELS_STEREO)) { LOG(WARNING) << "Received an unsupported packet."; return scoped_ptr(NULL); } if (packet->data_size() > kMaxFramesPerPacket) { LOG(WARNING) << "Received an packet with too many frames."; return scoped_ptr(); } // Create a new packet of decoded data. scoped_ptr decoded_packet(new AudioPacket()); decoded_packet->set_encoding(AudioPacket::ENCODING_RAW); decoded_packet->set_sampling_rate(packet->sampling_rate()); decoded_packet->set_bytes_per_sample(packet->bytes_per_sample()); decoded_packet->set_channels(packet->channels()); std::string* decoded_data = decoded_packet->add_data(); decoded_data->resize(packet->data_size() * speex_frame_size_ * packet->bytes_per_sample() * packet->channels()); int16* samples = reinterpret_cast(string_as_array(decoded_data)); for (int i = 0; i < packet->data_size(); ++i) { // Read the bytes into the bits structure. speex_bits_read_from(speex_bits_.get(), string_as_array(packet->mutable_data(i)), packet->data(i).size()); // Decode the frame and store it in the buffer. int status = speex_decode_int(speex_state_, speex_bits_.get(), samples); if (status < 0) { LOG(ERROR) << "Error in decoding Speex data."; return scoped_ptr(NULL); } // Transform mono to stereo. speex_decode_stereo_int(samples, speex_frame_size_, speex_stereo_state_); samples += (speex_frame_size_ * packet->channels()); } return decoded_packet.Pass(); } } // namespace remoting