| Index: webrtc/modules/audio_coding/neteq/neteq_impl.cc
|
| diff --git a/webrtc/modules/audio_coding/neteq/neteq_impl.cc b/webrtc/modules/audio_coding/neteq/neteq_impl.cc
|
| index bbd3220c00b20f33523b68d31508bcdfc03f6d51..5fa2cbd5867afe317a15cb639dc9cf012f9d8785 100644
|
| --- a/webrtc/modules/audio_coding/neteq/neteq_impl.cc
|
| +++ b/webrtc/modules/audio_coding/neteq/neteq_impl.cc
|
| @@ -124,11 +124,6 @@ int NetEqImpl::InsertPacket(const WebRtcRTPHeader& rtp_header,
|
| rtc::ArrayView<const uint8_t> payload,
|
| uint32_t receive_timestamp) {
|
| CriticalSectionScoped lock(crit_sect_.get());
|
| - LOG(LS_VERBOSE) << "InsertPacket: ts=" << rtp_header.header.timestamp
|
| - << ", sn=" << rtp_header.header.sequenceNumber
|
| - << ", pt=" << static_cast<int>(rtp_header.header.payloadType)
|
| - << ", ssrc=" << rtp_header.header.ssrc
|
| - << ", len=" << payload.size();
|
| int error =
|
| InsertPacketInternal(rtp_header, payload, receive_timestamp, false);
|
| if (error != 0) {
|
| @@ -141,12 +136,6 @@ int NetEqImpl::InsertPacket(const WebRtcRTPHeader& rtp_header,
|
| int NetEqImpl::InsertSyncPacket(const WebRtcRTPHeader& rtp_header,
|
| uint32_t receive_timestamp) {
|
| CriticalSectionScoped lock(crit_sect_.get());
|
| - LOG(LS_VERBOSE) << "InsertPacket-Sync: ts="
|
| - << rtp_header.header.timestamp <<
|
| - ", sn=" << rtp_header.header.sequenceNumber <<
|
| - ", pt=" << static_cast<int>(rtp_header.header.payloadType) <<
|
| - ", ssrc=" << rtp_header.header.ssrc;
|
| -
|
| const uint8_t kSyncPayload[] = { 's', 'y', 'n', 'c' };
|
| int error =
|
| InsertPacketInternal(rtp_header, kSyncPayload, receive_timestamp, true);
|
| @@ -162,11 +151,8 @@ int NetEqImpl::GetAudio(size_t max_length, int16_t* output_audio,
|
| size_t* samples_per_channel, int* num_channels,
|
| NetEqOutputType* type) {
|
| CriticalSectionScoped lock(crit_sect_.get());
|
| - LOG(LS_VERBOSE) << "GetAudio";
|
| int error = GetAudioInternal(max_length, output_audio, samples_per_channel,
|
| num_channels);
|
| - LOG(LS_VERBOSE) << "Produced " << *samples_per_channel <<
|
| - " samples/channel for " << *num_channels << " channel(s)";
|
| if (error != 0) {
|
| error_code_ = error;
|
| return kFail;
|
| @@ -751,8 +737,6 @@ int NetEqImpl::GetAudioInternal(size_t max_length,
|
| last_mode_ = kModeError;
|
| return return_value;
|
| }
|
| - LOG(LS_VERBOSE) << "GetDecision returned operation=" << operation <<
|
| - " and " << packet_list.size() << " packet(s)";
|
|
|
| AudioDecoder::SpeechType speech_type;
|
| int length = 0;
|
| @@ -867,9 +851,6 @@ int NetEqImpl::GetAudioInternal(size_t max_length,
|
| sync_buffer_->GetNextAudioInterleaved(num_output_samples_per_channel,
|
| output);
|
| *num_channels = static_cast<int>(sync_buffer_->Channels());
|
| - LOG(LS_VERBOSE) << "Sync buffer (" << *num_channels << " channel(s)):" <<
|
| - " insert " << algorithm_buffer_->Size() << " samples, extract " <<
|
| - samples_from_sync << " samples";
|
| if (sync_buffer_->FutureLength() < expand_->overlap_length()) {
|
| // The sync buffer should always contain |overlap_length| samples, but now
|
| // too many samples have been extracted. Reinstall the |overlap_length|
|
| @@ -1327,7 +1308,6 @@ int NetEqImpl::DecodeCng(AudioDecoder* decoder, int* decoded_length,
|
| &decoded_buffer_[*decoded_length], speech_type);
|
| if (length > 0) {
|
| *decoded_length += length;
|
| - LOG(LS_VERBOSE) << "Decoded " << length << " CNG samples";
|
| } else {
|
| // Error.
|
| LOG(LS_WARNING) << "Failed to decode CNG";
|
| @@ -1367,34 +1347,17 @@ int NetEqImpl::DecodeLoop(PacketList* packet_list, const Operations& operation,
|
| int decode_length;
|
| if (packet->sync_packet) {
|
| // Decode to silence with the same frame size as the last decode.
|
| - LOG(LS_VERBOSE) << "Decoding sync-packet: " <<
|
| - " ts=" << packet->header.timestamp <<
|
| - ", sn=" << packet->header.sequenceNumber <<
|
| - ", pt=" << static_cast<int>(packet->header.payloadType) <<
|
| - ", ssrc=" << packet->header.ssrc <<
|
| - ", len=" << packet->payload_length;
|
| memset(&decoded_buffer_[*decoded_length], 0,
|
| decoder_frame_length_ * decoder->Channels() *
|
| sizeof(decoded_buffer_[0]));
|
| decode_length = rtc::checked_cast<int>(decoder_frame_length_);
|
| } else if (!packet->primary) {
|
| // This is a redundant payload; call the special decoder method.
|
| - LOG(LS_VERBOSE) << "Decoding packet (redundant):" <<
|
| - " ts=" << packet->header.timestamp <<
|
| - ", sn=" << packet->header.sequenceNumber <<
|
| - ", pt=" << static_cast<int>(packet->header.payloadType) <<
|
| - ", ssrc=" << packet->header.ssrc <<
|
| - ", len=" << packet->payload_length;
|
| decode_length = decoder->DecodeRedundant(
|
| packet->payload, packet->payload_length, fs_hz_,
|
| (decoded_buffer_length_ - *decoded_length) * sizeof(int16_t),
|
| &decoded_buffer_[*decoded_length], speech_type);
|
| } else {
|
| - LOG(LS_VERBOSE) << "Decoding packet: ts=" << packet->header.timestamp <<
|
| - ", sn=" << packet->header.sequenceNumber <<
|
| - ", pt=" << static_cast<int>(packet->header.payloadType) <<
|
| - ", ssrc=" << packet->header.ssrc <<
|
| - ", len=" << packet->payload_length;
|
| decode_length =
|
| decoder->Decode(
|
| packet->payload, packet->payload_length, fs_hz_,
|
| @@ -1410,9 +1373,6 @@ int NetEqImpl::DecodeLoop(PacketList* packet_list, const Operations& operation,
|
| // Update |decoder_frame_length_| with number of samples per channel.
|
| decoder_frame_length_ =
|
| static_cast<size_t>(decode_length) / decoder->Channels();
|
| - LOG(LS_VERBOSE) << "Decoded " << decode_length << " samples ("
|
| - << decoder->Channels() << " channel(s) -> "
|
| - << decoder_frame_length_ << " samples per channel)";
|
| } else if (decode_length < 0) {
|
| // Error.
|
| LOG(LS_WARNING) << "Decode " << decode_length << " " << payload_length;
|
|
|