778 lines
32 KiB
C++
778 lines
32 KiB
C++
//=========================================================================
|
|
// Name: TxRxThread.h
|
|
// Purpose: Implements the main processing thread for audio I/O.
|
|
//
|
|
// Authors: Mooneer Salem
|
|
// License:
|
|
//
|
|
// All rights reserved.
|
|
//
|
|
// This program is free software; you can redistribute it and/or modify
|
|
// it under the terms of the GNU General Public License version 2.1,
|
|
// as published by the Free Software Foundation. This program is
|
|
// distributed in the hope that it will be useful, but WITHOUT ANY
|
|
// WARRANTY; without even the implied warranty of MERCHANTABILITY or
|
|
// FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public
|
|
// License for more details.
|
|
//
|
|
// You should have received a copy of the GNU General Public License
|
|
// along with this program; if not, see <http://www.gnu.org/licenses/>.
|
|
//
|
|
//=========================================================================
|
|
|
|
#include <chrono>
|
|
using namespace std::chrono_literals;
|
|
|
|
// This forces us to use freedv-gui's version rather than another one.
|
|
// TBD -- may not be needed once we fully switch over to the audio pipeline.
|
|
#include "../defines.h"
|
|
|
|
#include "TxRxThread.h"
|
|
#include "paCallbackData.h"
|
|
|
|
#include "PlaybackStep.h"
|
|
#include "EitherOrStep.h"
|
|
#include "SpeexStep.h"
|
|
#include "EqualizerStep.h"
|
|
#include "ResamplePlotStep.h"
|
|
#include "ResampleStep.h"
|
|
#include "TapStep.h"
|
|
#include "LevelAdjustStep.h"
|
|
#include "FreeDVTransmitStep.h"
|
|
#include "RecordStep.h"
|
|
#include "ToneInterfererStep.h"
|
|
#include "ComputeRfSpectrumStep.h"
|
|
#include "FreeDVReceiveStep.h"
|
|
#include "ExclusiveAccessStep.h"
|
|
#include "MuteStep.h"
|
|
#include "LinkStep.h"
|
|
|
|
#include "util/logging/ulog.h"
|
|
|
|
#include <wx/stopwatch.h>
|
|
|
|
// External globals
|
|
// TBD -- work on fully removing the need for these.
|
|
extern paCallBackData* g_rxUserdata;
|
|
extern int g_analog;
|
|
extern int g_nSoundCards;
|
|
extern bool g_half_duplex;
|
|
extern int g_tx;
|
|
extern int g_dump_fifo_state;
|
|
extern bool endingTx;
|
|
extern bool g_playFileToMicIn;
|
|
extern int g_sfTxFs;
|
|
extern bool g_loopPlayFileToMicIn;
|
|
extern float g_TxFreqOffsetHz;
|
|
extern struct FIFO* g_plotSpeechInFifo;
|
|
extern struct FIFO* g_plotDemodInFifo;
|
|
extern struct FIFO* g_plotSpeechOutFifo;
|
|
extern int g_mode;
|
|
extern bool g_recFileFromModulator;
|
|
extern int g_txLevel;
|
|
extern int g_dump_timing;
|
|
extern bool g_queueResync;
|
|
extern int g_resyncs;
|
|
extern bool g_recFileFromRadio;
|
|
extern unsigned int g_recFromRadioSamples;
|
|
extern bool g_playFileFromRadio;
|
|
extern int g_sfFs;
|
|
extern bool g_loopPlayFileFromRadio;
|
|
extern int g_SquelchActive;
|
|
extern float g_SquelchLevel;
|
|
extern float g_tone_phase;
|
|
extern float g_avmag[MODEM_STATS_NSPEC];
|
|
extern int g_State;
|
|
extern int g_channel_noise;
|
|
extern float g_RxFreqOffsetHz;
|
|
extern float g_sig_pwr_av;
|
|
extern bool g_voice_keyer_tx;
|
|
extern bool g_eoo_enqueued;
|
|
|
|
#include <speex/speex_preprocess.h>
|
|
|
|
#include "../freedv_interface.h"
|
|
extern FreeDVInterface freedvInterface;
|
|
|
|
#include <wx/wx.h>
|
|
#include "../main.h"
|
|
extern wxMutex txModeChangeMutex;
|
|
extern wxMutex g_mutexProtectingCallbackData;
|
|
extern wxWindow* g_parent;
|
|
|
|
#include <sndfile.h>
|
|
extern SNDFILE* g_sfPlayFile;
|
|
extern SNDFILE* g_sfRecFileFromModulator;
|
|
extern SNDFILE* g_sfRecFile;
|
|
extern SNDFILE* g_sfRecMicFile;
|
|
extern SNDFILE* g_sfPlayFileFromRadio;
|
|
|
|
extern bool g_recFileFromMic;
|
|
extern bool g_recVoiceKeyerFile;
|
|
|
|
// TBD -- shouldn't be needed once we've fully converted over
|
|
extern int resample(SRC_STATE *src,
|
|
short output_short[],
|
|
short input_short[],
|
|
int output_sample_rate,
|
|
int input_sample_rate,
|
|
int length_output_short, // maximum output array length in samples
|
|
int length_input_short
|
|
);
|
|
#include "sox_biquad.h"
|
|
|
|
void TxRxThread::initializePipeline_()
|
|
{
|
|
// Function definitions shared across both pipelines.
|
|
auto callbackLockFn = []() {
|
|
g_mutexProtectingCallbackData.Lock();
|
|
};
|
|
|
|
auto callbackUnlockFn = []() {
|
|
g_mutexProtectingCallbackData.Unlock();
|
|
};
|
|
|
|
if (m_tx)
|
|
{
|
|
pipeline_ = std::shared_ptr<AudioPipeline>(new AudioPipeline(inputSampleRate_, outputSampleRate_));
|
|
|
|
// Record from mic step (optional)
|
|
auto recordMicStep = new RecordStep(
|
|
inputSampleRate_,
|
|
[]() { return g_sfRecMicFile; },
|
|
[](int numSamples) {
|
|
// Recording stops when the user explicitly tells us to,
|
|
// no action required here.
|
|
}
|
|
);
|
|
auto recordMicPipeline = new AudioPipeline(inputSampleRate_, inputSampleRate_);
|
|
recordMicPipeline->appendPipelineStep(std::shared_ptr<IPipelineStep>(recordMicStep));
|
|
|
|
auto recordMicTap = new TapStep(inputSampleRate_, recordMicPipeline);
|
|
auto bypassRecordMic = new AudioPipeline(inputSampleRate_, inputSampleRate_);
|
|
|
|
auto eitherOrRecordMic = new EitherOrStep(
|
|
[]() { return (g_recVoiceKeyerFile || g_recFileFromMic) && (g_sfRecMicFile != NULL); },
|
|
std::shared_ptr<IPipelineStep>(recordMicTap),
|
|
std::shared_ptr<IPipelineStep>(bypassRecordMic)
|
|
);
|
|
auto recordMicLockStep = new ExclusiveAccessStep(eitherOrRecordMic, callbackLockFn, callbackUnlockFn);
|
|
pipeline_->appendPipelineStep(std::shared_ptr<IPipelineStep>(recordMicLockStep));
|
|
|
|
// Mic In playback step (optional)
|
|
auto eitherOrBypassPlay = new AudioPipeline(inputSampleRate_, inputSampleRate_);
|
|
auto eitherOrPlayMicIn = new AudioPipeline(inputSampleRate_, inputSampleRate_);
|
|
auto playMicIn = new PlaybackStep(
|
|
inputSampleRate_,
|
|
[]() { return g_sfTxFs; },
|
|
[]() { return g_sfPlayFile; },
|
|
[]() {
|
|
if (g_loopPlayFileToMicIn)
|
|
sf_seek(g_sfPlayFile, 0, SEEK_SET);
|
|
else {
|
|
log_info("playFileFromRadio finished, issuing event!");
|
|
g_parent->CallAfter(&MainFrame::StopPlayFileToMicIn);
|
|
}
|
|
}
|
|
);
|
|
eitherOrPlayMicIn->appendPipelineStep(std::shared_ptr<IPipelineStep>(playMicIn));
|
|
|
|
auto eitherOrPlayStep = new EitherOrStep(
|
|
[]() { return g_playFileToMicIn && (g_sfPlayFile != NULL); },
|
|
std::shared_ptr<IPipelineStep>(eitherOrPlayMicIn),
|
|
std::shared_ptr<IPipelineStep>(eitherOrBypassPlay));
|
|
auto playMicLockStep = new ExclusiveAccessStep(eitherOrPlayStep, callbackLockFn, callbackUnlockFn);
|
|
pipeline_->appendPipelineStep(std::shared_ptr<IPipelineStep>(playMicLockStep));
|
|
|
|
// Speex step (optional)
|
|
auto eitherOrProcessSpeex = new AudioPipeline(inputSampleRate_, inputSampleRate_);
|
|
auto eitherOrBypassSpeex = new AudioPipeline(inputSampleRate_, inputSampleRate_);
|
|
|
|
auto speexStep = new SpeexStep(inputSampleRate_);
|
|
eitherOrProcessSpeex->appendPipelineStep(std::shared_ptr<IPipelineStep>(speexStep));
|
|
|
|
auto eitherOrSpeexStep = new EitherOrStep(
|
|
[]() { return wxGetApp().appConfiguration.filterConfiguration.speexppEnable; },
|
|
std::shared_ptr<IPipelineStep>(eitherOrProcessSpeex),
|
|
std::shared_ptr<IPipelineStep>(eitherOrBypassSpeex));
|
|
auto speexLockStep = new ExclusiveAccessStep(eitherOrSpeexStep, callbackLockFn, callbackUnlockFn);
|
|
pipeline_->appendPipelineStep(std::shared_ptr<IPipelineStep>(speexLockStep));
|
|
|
|
// Equalizer step (optional based on filter state)
|
|
auto equalizerStep = new EqualizerStep(
|
|
inputSampleRate_,
|
|
&g_rxUserdata->micInEQEnable,
|
|
&g_rxUserdata->sbqMicInBass,
|
|
&g_rxUserdata->sbqMicInMid,
|
|
&g_rxUserdata->sbqMicInTreble,
|
|
&g_rxUserdata->sbqMicInVol);
|
|
auto equalizerLockStep = new ExclusiveAccessStep(equalizerStep, callbackLockFn, callbackUnlockFn);
|
|
pipeline_->appendPipelineStep(std::shared_ptr<IPipelineStep>(equalizerLockStep));
|
|
|
|
// Take TX audio post-equalizer and send it to RX for possible monitoring use.
|
|
if (equalizedMicAudioLink_ != nullptr)
|
|
{
|
|
auto micAudioPipeline = new AudioPipeline(inputSampleRate_, equalizedMicAudioLink_->getSampleRate());
|
|
micAudioPipeline->appendPipelineStep(equalizedMicAudioLink_->getInputPipelineStep());
|
|
|
|
auto micAudioTap = std::make_shared<TapStep>(inputSampleRate_, micAudioPipeline);
|
|
pipeline_->appendPipelineStep(micAudioTap);
|
|
}
|
|
|
|
// Resample for plot step
|
|
auto resampleForPlotStep = new ResampleForPlotStep(g_plotSpeechInFifo);
|
|
auto resampleForPlotPipeline = new AudioPipeline(inputSampleRate_, resampleForPlotStep->getOutputSampleRate());
|
|
resampleForPlotPipeline->appendPipelineStep(std::shared_ptr<IPipelineStep>(resampleForPlotStep));
|
|
|
|
auto resampleForPlotTap = new TapStep(inputSampleRate_, resampleForPlotPipeline);
|
|
pipeline_->appendPipelineStep(std::shared_ptr<IPipelineStep>(resampleForPlotTap));
|
|
|
|
// FreeDV TX step (analog leg)
|
|
auto doubleLevelStep = new LevelAdjustStep(inputSampleRate_, []() { return 2.0; });
|
|
auto analogTxPipeline = new AudioPipeline(inputSampleRate_, outputSampleRate_);
|
|
analogTxPipeline->appendPipelineStep(std::shared_ptr<IPipelineStep>(doubleLevelStep));
|
|
|
|
auto digitalTxStep = freedvInterface.createTransmitPipeline(inputSampleRate_, outputSampleRate_, []() { return g_TxFreqOffsetHz; });
|
|
auto digitalTxPipeline = new AudioPipeline(inputSampleRate_, outputSampleRate_);
|
|
digitalTxPipeline->appendPipelineStep(std::shared_ptr<IPipelineStep>(digitalTxStep));
|
|
|
|
auto eitherOrDigitalAnalog = new EitherOrStep(
|
|
[]() { return g_analog; },
|
|
std::shared_ptr<IPipelineStep>(analogTxPipeline),
|
|
std::shared_ptr<IPipelineStep>(digitalTxPipeline));
|
|
pipeline_->appendPipelineStep(std::shared_ptr<IPipelineStep>(eitherOrDigitalAnalog));
|
|
|
|
// Record modulated output (optional)
|
|
auto recordModulatedStep = new RecordStep(
|
|
outputSampleRate_,
|
|
[]() { return g_sfRecFileFromModulator; },
|
|
[](int numSamples) {
|
|
// empty
|
|
});
|
|
auto recordModulatedPipeline = new AudioPipeline(outputSampleRate_, recordModulatedStep->getOutputSampleRate());
|
|
recordModulatedPipeline->appendPipelineStep(std::shared_ptr<IPipelineStep>(recordModulatedStep));
|
|
|
|
auto recordModulatedTap = new TapStep(outputSampleRate_, recordModulatedPipeline);
|
|
auto recordModulatedTapPipeline = new AudioPipeline(outputSampleRate_, outputSampleRate_);
|
|
recordModulatedTapPipeline->appendPipelineStep(std::shared_ptr<IPipelineStep>(recordModulatedTap));
|
|
|
|
auto bypassRecordModulated = new AudioPipeline(outputSampleRate_, outputSampleRate_);
|
|
|
|
auto eitherOrRecordModulated = new EitherOrStep(
|
|
[]() { return g_recFileFromModulator && (g_sfRecFileFromModulator != NULL); },
|
|
std::shared_ptr<IPipelineStep>(recordModulatedTapPipeline),
|
|
std::shared_ptr<IPipelineStep>(bypassRecordModulated));
|
|
auto recordModulatedLockStep = new ExclusiveAccessStep(eitherOrRecordModulated, callbackLockFn, callbackUnlockFn);
|
|
pipeline_->appendPipelineStep(std::shared_ptr<IPipelineStep>(recordModulatedLockStep));
|
|
|
|
// TX attenuation step
|
|
auto txAttenuationStep = new LevelAdjustStep(outputSampleRate_, []() {
|
|
double dbLoss = g_txLevel / 10.0;
|
|
double scaleFactor = exp(dbLoss/20.0 * log(10.0));
|
|
return scaleFactor;
|
|
});
|
|
pipeline_->appendPipelineStep(std::shared_ptr<IPipelineStep>(txAttenuationStep));
|
|
}
|
|
else
|
|
{
|
|
pipeline_ = std::shared_ptr<AudioPipeline>(new AudioPipeline(inputSampleRate_, outputSampleRate_));
|
|
|
|
// Record from radio step (optional)
|
|
auto recordRadioStep = new RecordStep(
|
|
inputSampleRate_,
|
|
[]() { return g_sfRecFile; },
|
|
[](int numSamples) {
|
|
g_recFromRadioSamples -= numSamples;
|
|
if (g_recFromRadioSamples <= 0)
|
|
{
|
|
// call stop record menu item, should be thread safe
|
|
g_parent->CallAfter(&MainFrame::StopRecFileFromRadio);
|
|
}
|
|
}
|
|
);
|
|
auto recordRadioPipeline = new AudioPipeline(inputSampleRate_, inputSampleRate_);
|
|
recordRadioPipeline->appendPipelineStep(std::shared_ptr<IPipelineStep>(recordRadioStep));
|
|
|
|
auto recordRadioTap = new TapStep(inputSampleRate_, recordRadioPipeline);
|
|
auto bypassRecordRadio = new AudioPipeline(inputSampleRate_, inputSampleRate_);
|
|
|
|
auto eitherOrRecordRadio = new EitherOrStep(
|
|
[]() { return g_recFileFromRadio && (g_sfRecFile != NULL); },
|
|
std::shared_ptr<IPipelineStep>(recordRadioTap),
|
|
std::shared_ptr<IPipelineStep>(bypassRecordRadio)
|
|
);
|
|
auto recordRadioLockStep = new ExclusiveAccessStep(eitherOrRecordRadio, callbackLockFn, callbackUnlockFn);
|
|
pipeline_->appendPipelineStep(std::shared_ptr<IPipelineStep>(recordRadioLockStep));
|
|
|
|
// Play from radio step (optional)
|
|
auto eitherOrBypassPlayRadio = new AudioPipeline(inputSampleRate_, inputSampleRate_);
|
|
auto eitherOrPlayRadio = new AudioPipeline(inputSampleRate_, inputSampleRate_);
|
|
auto playRadio = new PlaybackStep(
|
|
inputSampleRate_,
|
|
[]() { return g_sfFs; },
|
|
[]() { return g_sfPlayFileFromRadio; },
|
|
[]() {
|
|
if (g_loopPlayFileFromRadio)
|
|
sf_seek(g_sfPlayFileFromRadio, 0, SEEK_SET);
|
|
else {
|
|
log_info("playFileFromRadio finished, issuing event!");
|
|
g_parent->CallAfter(&MainFrame::StopPlaybackFileFromRadio);
|
|
}
|
|
}
|
|
);
|
|
eitherOrPlayRadio->appendPipelineStep(std::shared_ptr<IPipelineStep>(playRadio));
|
|
|
|
auto eitherOrPlayRadioStep = new EitherOrStep(
|
|
[]() {
|
|
g_mutexProtectingCallbackData.Lock();
|
|
auto result = g_playFileFromRadio && (g_sfPlayFileFromRadio != NULL);
|
|
g_mutexProtectingCallbackData.Unlock();
|
|
return result;
|
|
},
|
|
std::shared_ptr<IPipelineStep>(eitherOrPlayRadio),
|
|
std::shared_ptr<IPipelineStep>(eitherOrBypassPlayRadio));
|
|
auto playRadioLockStep = new ExclusiveAccessStep(eitherOrPlayRadioStep, callbackLockFn, callbackUnlockFn);
|
|
pipeline_->appendPipelineStep(std::shared_ptr<IPipelineStep>(playRadioLockStep));
|
|
|
|
// Resample for plot step (demod in)
|
|
auto resampleForPlotStep = new ResampleForPlotStep(g_plotDemodInFifo);
|
|
auto resampleForPlotPipeline = new AudioPipeline(inputSampleRate_, resampleForPlotStep->getOutputSampleRate());
|
|
resampleForPlotPipeline->appendPipelineStep(std::shared_ptr<IPipelineStep>(resampleForPlotStep));
|
|
|
|
auto resampleForPlotTap = new TapStep(inputSampleRate_, resampleForPlotPipeline);
|
|
pipeline_->appendPipelineStep(std::shared_ptr<IPipelineStep>(resampleForPlotTap));
|
|
|
|
// Tone interferer step (optional)
|
|
auto bypassToneInterferer = new AudioPipeline(inputSampleRate_, inputSampleRate_);
|
|
auto toneInterfererStep = new ToneInterfererStep(
|
|
inputSampleRate_,
|
|
[]() { return wxGetApp().m_tone_freq_hz; },
|
|
[]() { return wxGetApp().m_tone_amplitude; },
|
|
[]() { return &g_tone_phase; }
|
|
);
|
|
auto eitherOrToneInterferer = new EitherOrStep(
|
|
[]() { return wxGetApp().m_tone; },
|
|
std::shared_ptr<IPipelineStep>(toneInterfererStep),
|
|
std::shared_ptr<IPipelineStep>(bypassToneInterferer)
|
|
);
|
|
pipeline_->appendPipelineStep(std::shared_ptr<IPipelineStep>(eitherOrToneInterferer));
|
|
|
|
// RF spectrum computation step
|
|
auto computeRfSpectrumStep = new ComputeRfSpectrumStep(
|
|
[]() { return freedvInterface.getCurrentRxModemStats(); },
|
|
[]() { return &g_avmag[0]; }
|
|
);
|
|
auto computeRfSpectrumPipeline = new AudioPipeline(
|
|
inputSampleRate_, computeRfSpectrumStep->getOutputSampleRate());
|
|
computeRfSpectrumPipeline->appendPipelineStep(std::shared_ptr<IPipelineStep>(computeRfSpectrumStep));
|
|
|
|
auto computeRfSpectrumTap = new TapStep(inputSampleRate_, computeRfSpectrumPipeline);
|
|
pipeline_->appendPipelineStep(std::shared_ptr<IPipelineStep>(computeRfSpectrumTap));
|
|
|
|
// RX demodulation step
|
|
auto bypassRfDemodulationPipeline = new AudioPipeline(inputSampleRate_, outputSampleRate_);
|
|
auto rfDemodulationPipeline = new AudioPipeline(inputSampleRate_, outputSampleRate_);
|
|
auto rfDemodulationStep = freedvInterface.createReceivePipeline(
|
|
inputSampleRate_, outputSampleRate_,
|
|
[]() { return &g_State; },
|
|
[]() { return g_channel_noise; },
|
|
[]() { return wxGetApp().appConfiguration.noiseSNR; },
|
|
[]() { return g_RxFreqOffsetHz; },
|
|
[]() { return &g_sig_pwr_av; }
|
|
);
|
|
rfDemodulationPipeline->appendPipelineStep(std::shared_ptr<IPipelineStep>(rfDemodulationStep));
|
|
|
|
// Replace received audio with microphone audio if we're monitoring TX/voice keyer recording.
|
|
if (equalizedMicAudioLink_ != nullptr)
|
|
{
|
|
auto bypassMonitorAudio = new AudioPipeline(inputSampleRate_, outputSampleRate_);
|
|
auto mutePipeline = new AudioPipeline(inputSampleRate_, outputSampleRate_);
|
|
|
|
auto monitorPipeline = new AudioPipeline(inputSampleRate_, outputSampleRate_);
|
|
monitorPipeline->appendPipelineStep(equalizedMicAudioLink_->getOutputPipelineStep());
|
|
|
|
auto monitorLevelStep = std::make_shared<LevelAdjustStep>(outputSampleRate_, [&]() {
|
|
double volInDb = 0;
|
|
if (g_voice_keyer_tx && wxGetApp().appConfiguration.monitorVoiceKeyerAudio)
|
|
{
|
|
volInDb = wxGetApp().appConfiguration.monitorVoiceKeyerAudioVol;
|
|
}
|
|
else
|
|
{
|
|
volInDb = wxGetApp().appConfiguration.monitorTxAudioVol;
|
|
}
|
|
|
|
return std::exp(volInDb/20.0f * std::log(10.0f));
|
|
});
|
|
monitorPipeline->appendPipelineStep(monitorLevelStep);
|
|
|
|
auto muteStep = new MuteStep(outputSampleRate_);
|
|
|
|
mutePipeline->appendPipelineStep(std::shared_ptr<IPipelineStep>(muteStep));
|
|
|
|
auto eitherOrMuteStep = new EitherOrStep(
|
|
[]() { return g_recVoiceKeyerFile; },
|
|
std::shared_ptr<IPipelineStep>(mutePipeline),
|
|
std::shared_ptr<IPipelineStep>(bypassMonitorAudio)
|
|
);
|
|
|
|
auto eitherOrMicMonitorStep = new EitherOrStep(
|
|
[]() { return
|
|
(g_voice_keyer_tx && wxGetApp().appConfiguration.monitorVoiceKeyerAudio) ||
|
|
(g_tx && wxGetApp().appConfiguration.monitorTxAudio); },
|
|
std::shared_ptr<IPipelineStep>(monitorPipeline),
|
|
std::shared_ptr<IPipelineStep>(eitherOrMuteStep)
|
|
);
|
|
bypassRfDemodulationPipeline->appendPipelineStep(std::shared_ptr<IPipelineStep>(eitherOrMicMonitorStep));
|
|
}
|
|
|
|
auto eitherOrRfDemodulationStep = new EitherOrStep(
|
|
[this]() { return g_analog ||
|
|
(equalizedMicAudioLink_ != nullptr && (
|
|
(g_recVoiceKeyerFile) ||
|
|
(g_voice_keyer_tx && wxGetApp().appConfiguration.monitorVoiceKeyerAudio) ||
|
|
(g_tx && wxGetApp().appConfiguration.monitorTxAudio)
|
|
)); },
|
|
std::shared_ptr<IPipelineStep>(bypassRfDemodulationPipeline),
|
|
std::shared_ptr<IPipelineStep>(rfDemodulationPipeline)
|
|
);
|
|
|
|
pipeline_->appendPipelineStep(std::shared_ptr<IPipelineStep>(eitherOrRfDemodulationStep));
|
|
|
|
// Equalizer step (optional based on filter state)
|
|
auto equalizerStep = new EqualizerStep(
|
|
outputSampleRate_,
|
|
&g_rxUserdata->spkOutEQEnable,
|
|
&g_rxUserdata->sbqSpkOutBass,
|
|
&g_rxUserdata->sbqSpkOutMid,
|
|
&g_rxUserdata->sbqSpkOutTreble,
|
|
&g_rxUserdata->sbqSpkOutVol);
|
|
auto equalizerLockStep = new ExclusiveAccessStep(equalizerStep, callbackLockFn, callbackUnlockFn);
|
|
pipeline_->appendPipelineStep(std::shared_ptr<IPipelineStep>(equalizerLockStep));
|
|
|
|
// Resample for plot step (speech out)
|
|
auto resampleForPlotOutStep = new ResampleForPlotStep(g_plotSpeechOutFifo);
|
|
auto resampleForPlotOutPipeline = new AudioPipeline(outputSampleRate_, resampleForPlotOutStep->getOutputSampleRate());
|
|
resampleForPlotOutPipeline->appendPipelineStep(std::shared_ptr<IPipelineStep>(resampleForPlotOutStep));
|
|
|
|
auto resampleForPlotOutTap = new TapStep(outputSampleRate_, resampleForPlotOutPipeline);
|
|
pipeline_->appendPipelineStep(std::shared_ptr<IPipelineStep>(resampleForPlotOutTap));
|
|
|
|
// Clear anything in the FIFO before resuming decode.
|
|
clearFifos_();
|
|
}
|
|
}
|
|
|
|
void* TxRxThread::Entry()
|
|
{
|
|
initializePipeline_();
|
|
|
|
while (m_run)
|
|
{
|
|
#if defined(__linux__)
|
|
const char* threadName = nullptr;
|
|
if (m_tx) threadName = "FreeDV txThread";
|
|
else threadName = "FreeDV rxThread";
|
|
pthread_setname_np(pthread_self(), threadName);
|
|
#endif // defined(__linux__)
|
|
|
|
#if 0
|
|
{
|
|
std::unique_lock<std::mutex> lk(m_processingMutex);
|
|
if (m_processingCondVar.wait_for(lk, std::chrono::milliseconds(100)) == std::cv_status::timeout)
|
|
{
|
|
log_warn("txRxThread: timeout while waiting for CV, tx = %d", m_tx);
|
|
}
|
|
}
|
|
#endif
|
|
|
|
auto currentTime = std::chrono::steady_clock::now();
|
|
|
|
if (!m_run) break;
|
|
if (m_tx) txProcessing_();
|
|
else rxProcessing_();
|
|
|
|
std::this_thread::sleep_for(10ms);
|
|
}
|
|
|
|
// Force pipeline to delete itself when we're done with the thread.
|
|
pipeline_ = nullptr;
|
|
|
|
return NULL;
|
|
}
|
|
|
|
void TxRxThread::OnExit()
|
|
{
|
|
// No actions required for exit.
|
|
}
|
|
|
|
void TxRxThread::terminateThread()
|
|
{
|
|
m_run = 0;
|
|
notify();
|
|
}
|
|
|
|
void TxRxThread::notify()
|
|
{
|
|
#if 0
|
|
std::unique_lock<std::mutex> lk(m_processingMutex);
|
|
m_processingCondVar.notify_all();
|
|
#endif
|
|
}
|
|
|
|
void TxRxThread::clearFifos_()
|
|
{
|
|
paCallBackData *cbData = g_rxUserdata;
|
|
|
|
if (equalizedMicAudioLink_ != nullptr && !g_tx)
|
|
{
|
|
equalizedMicAudioLink_->clearFifo();
|
|
}
|
|
|
|
if (m_tx)
|
|
{
|
|
auto used = codec2_fifo_used(cbData->outfifo1);
|
|
if (used > 0)
|
|
{
|
|
short* temp = new short[used];
|
|
assert(temp != nullptr);
|
|
codec2_fifo_read(cbData->outfifo1, temp, used);
|
|
delete[] temp;
|
|
}
|
|
|
|
used = codec2_fifo_used(cbData->infifo2);
|
|
if (used > 0)
|
|
{
|
|
short* temp = new short[used];
|
|
assert(temp != nullptr);
|
|
codec2_fifo_read(cbData->infifo2, temp, used);
|
|
delete[] temp;
|
|
}
|
|
}
|
|
else
|
|
{
|
|
auto used = codec2_fifo_used(cbData->infifo1);
|
|
if (used > 0)
|
|
{
|
|
short* temp = new short[used];
|
|
assert(temp != nullptr);
|
|
codec2_fifo_read(cbData->infifo1, temp, used);
|
|
delete[] temp;
|
|
}
|
|
|
|
auto outFifo = (g_nSoundCards == 1) ? cbData->outfifo1 : cbData->outfifo2;
|
|
used = codec2_fifo_used(outFifo);
|
|
if (used > 0)
|
|
{
|
|
short* temp = new short[used];
|
|
assert(temp != nullptr);
|
|
codec2_fifo_read(outFifo, temp, used);
|
|
delete[] temp;
|
|
}
|
|
}
|
|
}
|
|
|
|
//---------------------------------------------------------------------------------------------
|
|
// Main real time processing for tx and rx of FreeDV signals, run in its own threads
|
|
//---------------------------------------------------------------------------------------------
|
|
|
|
void TxRxThread::txProcessing_()
|
|
{
|
|
wxStopWatch sw;
|
|
paCallBackData *cbData = g_rxUserdata;
|
|
|
|
// Buffers re-used by tx and rx processing. We take samples from
|
|
// the sound card, and resample them for the freedv modem input
|
|
// sample rate. Typically the sound card is running at 48 or 44.1
|
|
// kHz, and the modem at 8kHz
|
|
|
|
//
|
|
// TX side processing --------------------------------------------
|
|
//
|
|
|
|
if (((g_nSoundCards == 2) && ((g_half_duplex && g_tx) || !g_half_duplex || g_voice_keyer_tx || g_recVoiceKeyerFile || g_recFileFromMic))) {
|
|
// Lock the mode mutex so that TX state doesn't change on us during processing.
|
|
txModeChangeMutex.Lock();
|
|
|
|
if (pipeline_ == nullptr)
|
|
{
|
|
initializePipeline_();
|
|
}
|
|
|
|
// This while loop locks the modulator to the sample rate of
|
|
// the input sound card. We want to make sure that modulator samples
|
|
// are uninterrupted by differences in sample rate between
|
|
// this sound card and the output sound card.
|
|
|
|
// Run code inside this while loop as soon as we have enough
|
|
// room for one frame of modem samples. Aim is to keep
|
|
// outfifo1 nice and full so we don't have any gaps in tx
|
|
// signal.
|
|
|
|
unsigned int nsam_one_modem_frame = freedvInterface.getTxNNomModemSamples() * ((float)outputSampleRate_ / (float)freedvInterface.getTxModemSampleRate());
|
|
|
|
if (g_dump_fifo_state) {
|
|
// If this drops to zero we have a problem as we will run out of output samples
|
|
// to send to the sound driver
|
|
log_debug("outfifo1 used: %6d free: %6d nsam_one_modem_frame: %d",
|
|
codec2_fifo_used(cbData->outfifo1), codec2_fifo_free(cbData->outfifo1), nsam_one_modem_frame);
|
|
}
|
|
|
|
int nsam_in_48 = freedvInterface.getTxNumSpeechSamples() * ((float)inputSampleRate_ / (float)freedvInterface.getTxSpeechSampleRate());
|
|
assert(nsam_in_48 > 0);
|
|
|
|
short insound_card[nsam_in_48];
|
|
int nout;
|
|
|
|
|
|
while((unsigned)codec2_fifo_free(cbData->outfifo1) >= nsam_one_modem_frame) {
|
|
// OK to generate a frame of modem output samples we need
|
|
// an input frame of speech samples from the microphone.
|
|
|
|
// infifo2 is written to by another sound card so it may
|
|
// over or underflow, but we don't really care. It will
|
|
// just result in a short interruption in audio being fed
|
|
// to codec2_enc, possibly making a click every now and
|
|
// again in the decoded audio at the other end.
|
|
|
|
// zero speech input just in case infifo2 underflows
|
|
memset(insound_card, 0, nsam_in_48*sizeof(short));
|
|
|
|
// There may be recorded audio left to encode while ending TX. To handle this,
|
|
// we keep reading from the FIFO until we have less than nsam_in_48 samples available.
|
|
int nread = codec2_fifo_read(cbData->infifo2, insound_card, nsam_in_48);
|
|
if (nread != 0 && endingTx)
|
|
{
|
|
if (freedvInterface.getCurrentMode() >= FREEDV_MODE_RADE)
|
|
{
|
|
if (!hasEooBeenSent_)
|
|
{
|
|
log_info("Triggering sending of EOO");
|
|
|
|
// Special case for handling RADE EOT
|
|
freedvInterface.restartTxVocoder();
|
|
hasEooBeenSent_ = true;
|
|
}
|
|
|
|
short* inputSamples = new short[1];
|
|
auto inputSamplesPtr = std::shared_ptr<short>(inputSamples, std::default_delete<short[]>());
|
|
auto outputSamples = pipeline_->execute(inputSamplesPtr, 0, &nout);
|
|
if (nout > 0 && outputSamples.get() != nullptr)
|
|
{
|
|
log_debug("Injecting %d samples of resampled EOO into TX stream", nout);
|
|
if (codec2_fifo_write(cbData->outfifo1, outputSamples.get(), nout) != 0)
|
|
{
|
|
log_warn("Could not inject resampled EOO samples (space remaining in FIFO = %d)", cbData->outfifo1);
|
|
}
|
|
}
|
|
else
|
|
{
|
|
g_eoo_enqueued = true;
|
|
}
|
|
}
|
|
break;
|
|
}
|
|
else
|
|
{
|
|
g_eoo_enqueued = false;
|
|
hasEooBeenSent_ = false;
|
|
}
|
|
|
|
short* inputSamples = new short[nsam_in_48];
|
|
memcpy(inputSamples, insound_card, nsam_in_48 * sizeof(short));
|
|
|
|
auto inputSamplesPtr = std::shared_ptr<short>(inputSamples, std::default_delete<short[]>());
|
|
auto outputSamples = pipeline_->execute(inputSamplesPtr, nsam_in_48, &nout);
|
|
|
|
if (g_dump_fifo_state) {
|
|
log_info(" nout: %d", nout);
|
|
}
|
|
|
|
if (outputSamples.get() != nullptr)
|
|
{
|
|
codec2_fifo_write(cbData->outfifo1, outputSamples.get(), nout);
|
|
}
|
|
}
|
|
|
|
txModeChangeMutex.Unlock();
|
|
}
|
|
else
|
|
{
|
|
// Deallocates TX pipeline when not in use. This is needed to reset the state of
|
|
// certain TX pipeline steps (such as Speex).
|
|
pipeline_ = nullptr;
|
|
|
|
// Wipe anything added in the FIFO to prevent pops on next TX.
|
|
clearFifos_();
|
|
}
|
|
|
|
if (g_dump_timing) {
|
|
log_info("%4ld", sw.Time());
|
|
}
|
|
}
|
|
|
|
void TxRxThread::rxProcessing_()
|
|
{
|
|
wxStopWatch sw;
|
|
paCallBackData *cbData = g_rxUserdata;
|
|
|
|
// Buffers re-used by tx and rx processing. We take samples from
|
|
// the sound card, and resample them for the freedv modem input
|
|
// sample rate. Typically the sound card is running at 48 or 44.1
|
|
// kHz, and the modem at 8kHz.
|
|
|
|
//
|
|
// RX side processing --------------------------------------------
|
|
//
|
|
|
|
if (g_queueResync)
|
|
{
|
|
log_debug("Unsyncing per user request.");
|
|
g_queueResync = false;
|
|
freedvInterface.setSync(FREEDV_SYNC_UNSYNC);
|
|
g_resyncs++;
|
|
}
|
|
|
|
// Attempt to read one processing frame (about 20ms) of receive samples, we
|
|
// keep this frame duration constant across modes and sound card sample rates
|
|
int nsam = (int)(inputSampleRate_ * FRAME_DURATION);
|
|
assert(nsam > 0);
|
|
|
|
short insound_card[nsam];
|
|
int nout;
|
|
|
|
|
|
bool processInputFifo =
|
|
(g_voice_keyer_tx && wxGetApp().appConfiguration.monitorVoiceKeyerAudio) ||
|
|
(g_tx && wxGetApp().appConfiguration.monitorTxAudio) ||
|
|
(!g_voice_keyer_tx && ((g_half_duplex && !g_tx) || !g_half_duplex));
|
|
if (!processInputFifo)
|
|
{
|
|
clearFifos_();
|
|
}
|
|
|
|
// while we have enough input samples available ...
|
|
while (codec2_fifo_read(cbData->infifo1, insound_card, nsam) == 0 && processInputFifo) {
|
|
|
|
// send latest squelch level to FreeDV API, as it handles squelch internally
|
|
freedvInterface.setSquelch(g_SquelchActive, g_SquelchLevel);
|
|
|
|
short* inputSamples = new short[nsam];
|
|
memcpy(inputSamples, insound_card, nsam * sizeof(short));
|
|
|
|
auto inputSamplesPtr = std::shared_ptr<short>(inputSamples, std::default_delete<short[]>());
|
|
auto outputSamples = pipeline_->execute(inputSamplesPtr, nsam, &nout);
|
|
auto outFifo = (g_nSoundCards == 1) ? cbData->outfifo1 : cbData->outfifo2;
|
|
|
|
if (nout > 0 && outputSamples.get() != nullptr)
|
|
{
|
|
codec2_fifo_write(outFifo, outputSamples.get(), nout);
|
|
}
|
|
|
|
processInputFifo =
|
|
(g_voice_keyer_tx && wxGetApp().appConfiguration.monitorVoiceKeyerAudio) ||
|
|
(g_tx && wxGetApp().appConfiguration.monitorTxAudio) ||
|
|
(!g_voice_keyer_tx && ((g_half_duplex && !g_tx) || !g_half_duplex));
|
|
}
|
|
}
|