w4kek/WSJT-X
1
0
Fork 0
mirror of https://github.com/saitohirga/WSJT-X.git synced 2024-11-25 05:38:46 -05:00
Code Issues Packages Projects Releases Wiki Activity
446c255eb5
WSJT-X/Modulator/Modulator.cpp
Bill Somerville 542ffe8311
Improve audio device handling and error recovery 
where possible  audio devices that  disappear are not  forgotten until
the user selects another device, this should allow temporarily missing
devices or forgetting  to switch on devices before  starting WSJT-X to
be  handled more  cleanly. If  all else  fails, visiting  the Settings
dialog and  clicking OK should  get things  going again. Note  that we
still  do not  have  a  reliable way  of  detecting  failed audio  out
devices, in that  case selecting another device and  then returning to
the original should work.

Enumerating  audio devices  is expensive  and on  Linux may  take many
seconds per  device. To avoid  lengthy blocking behaviour until  it is
absolutely necessary,  audio devices are  not enumerated until  one of
the "Settings->Audio" device drop-down lists is opened. Elsewhere when
devices  must be  discovered  the  enumeration stops  as  soon as  the
configured device is  discovered. A status bar message  is posted when
audio devices are being enumerated as a reminder that the UI may block
while this is happening.

The message box warning about  unaccounted-for input audio samples now
only triggers when  >5 seconds of audio appears to  be missing or over
provided. Hopefully this will make the warning less annoying for those
that are  using audio sources  with high and/or variable  latencies. A
status  bar message  is still  posted for  any amount  of audio  input
samples  unaccounted for  >1/5 second,  this message  appearing a  lot
should be considered as notification that  there is a problem with the
audio sub-system, system load is  too high, or time synchronization is
stepping the PC clock rather  than adjusting the frequency to maintain
monotonic clock ticks.
2020-09-20 18:20:16 +01:00

382 lines
12 KiB
C++
Raw Blame History

#include "Modulator.hpp"
#include <limits>
#include <qmath.h>
#include <QDateTime>
#if QT_VERSION >= QT_VERSION_CHECK(5, 15, 0)
#include <QRandomGenerator>
#endif
#include <QDebug>
#include "widgets/mainwindow.h" // TODO: G4WJS - break this dependency
#include "Audio/soundout.h"
#include "commons.h"
#include "moc_Modulator.cpp"
extern float gran(); // Noise generator (for tests only)
#define RAMP_INCREMENT 64 // MUST be an integral factor of 2^16
#if defined (WSJT_SOFT_KEYING)
# define SOFT_KEYING WSJT_SOFT_KEYING
#else
# define SOFT_KEYING 1
#endif
double constexpr Modulator::m_twoPi;
// float wpm=20.0;
// unsigned m_nspd=1.2*48000.0/wpm;
// m_nspd=3072; //18.75 WPM
Modulator::Modulator (unsigned frameRate, double periodLengthInSeconds,
QObject * parent)
: AudioDevice {parent}
, m_quickClose {false}
, m_phi {0.0}
, m_toneSpacing {0.0}
, m_fSpread {0.0}
, m_period {periodLengthInSeconds}
, m_frameRate {frameRate}
, m_state {Idle}
, m_tuning {false}
, m_cwLevel {false}
, m_j0 {-1}
, m_toneFrequency0 {1500.0}
{
}
void Modulator::start (QString mode, unsigned symbolsLength, double framesPerSymbol,
double frequency, double toneSpacing,
SoundOutput * stream, Channel channel,
bool synchronize, bool fastMode, double dBSNR, double TRperiod)
{
// qDebug () << "mode:" << mode << "symbolsLength:" << symbolsLength << "framesPerSymbol:" << framesPerSymbol << "frequency:" << frequency << "toneSpacing:" << toneSpacing << "channel:" << channel << "synchronize:" << synchronize << "fastMode:" << fastMode << "dBSNR:" << dBSNR << "TRperiod:" << TRperiod;
Q_ASSERT (stream);
// Time according to this computer which becomes our base time
qint64 ms0 = QDateTime::currentMSecsSinceEpoch() % 86400000;
unsigned mstr = ms0 % int(1000.0*m_period); // ms into the nominal Tx start time
if(m_state != Idle) stop();
m_quickClose = false;
m_symbolsLength = symbolsLength;
m_isym0 = std::numeric_limits<unsigned>::max (); // big number
m_frequency0 = 0.;
m_phi = 0.;
m_addNoise = dBSNR < 0.;
m_nsps = framesPerSymbol;
m_frequency = frequency;
m_amp = std::numeric_limits<qint16>::max ();
m_toneSpacing = toneSpacing;
m_bFastMode=fastMode;
m_TRperiod=TRperiod;
unsigned delay_ms=1000;
if(mode=="FT8" or (mode=="FST4" and m_nsps==720)) delay_ms=500; //FT8, FST4-15
if(mode=="FT4") delay_ms=300; //FT4
// noise generator parameters
if (m_addNoise) {
m_snr = qPow (10.0, 0.05 * (dBSNR - 6.0));
m_fac = 3000.0;
if (m_snr > 1.0) m_fac = 3000.0 / m_snr;
}
m_silentFrames = 0;
m_ic=0;
if (!m_tuning && !m_bFastMode)
{
// calculate number of silent frames to send, so that audio will
// start at the nominal time "delay_ms" into the Tx sequence.
if (synchronize)
{
if(delay_ms > mstr) m_silentFrames = (delay_ms - mstr) * m_frameRate / 1000;
}
// adjust for late starts
if(!m_silentFrames && mstr >= delay_ms)
{
m_ic = (mstr - delay_ms) * m_frameRate / 1000;
}
}
initialize (QIODevice::ReadOnly, channel);
Q_EMIT stateChanged ((m_state = (synchronize && m_silentFrames) ?
Synchronizing : Active));
// qDebug() << "delay_ms:" << delay_ms << "mstr:" << mstr << "m_silentFrames:" << m_silentFrames << "m_ic:" << m_ic << "m_state:" << m_state;
m_stream = stream;
if (m_stream)
{
m_stream->restart (this);
}
else
{
qDebug () << "Modulator::start: no audio output stream assigned";
}
}
void Modulator::tune (bool newState)
{
m_tuning = newState;
if (!m_tuning) stop (true);
}
void Modulator::stop (bool quick)
{
m_quickClose = quick;
close ();
}
void Modulator::close ()
{
if (m_stream)
{
if (m_quickClose)
{
m_stream->reset ();
}
else
{
m_stream->stop ();
}
}
if (m_state != Idle)
{
Q_EMIT stateChanged ((m_state = Idle));
}
AudioDevice::close ();
}
qint64 Modulator::readData (char * data, qint64 maxSize)
{
double toneFrequency=1500.0;
if(m_nsps==6) {
toneFrequency=1000.0;
m_frequency=1000.0;
m_frequency0=1000.0;
}
if(maxSize==0) return 0;
Q_ASSERT (!(maxSize % qint64 (bytesPerFrame ()))); // no torn frames
Q_ASSERT (isOpen ());
qint64 numFrames (maxSize / bytesPerFrame ());
qint16 * samples (reinterpret_cast<qint16 *> (data));
qint16 * end (samples + numFrames * (bytesPerFrame () / sizeof (qint16)));
qint64 framesGenerated (0);
// if(m_ic==0) qDebug() << "aa" << 0.001*(QDateTime::currentMSecsSinceEpoch() % qint64(1000*m_TRperiod))
// << m_state << m_TRperiod << m_silentFrames << m_ic << foxcom_.wave[m_ic];
switch (m_state)
{
case Synchronizing:
{
if (m_silentFrames) { // send silence up to end of start delay
framesGenerated = qMin (m_silentFrames, numFrames);
do
{
samples = load (0, samples); // silence
} while (--m_silentFrames && samples != end);
if (!m_silentFrames)
{
Q_EMIT stateChanged ((m_state = Active));
}
}
m_cwLevel = false;
m_ramp = 0; // prepare for CW wave shaping
}
// fall through
case Active:
{
unsigned int isym=0;
if(!m_tuning) isym=m_ic/(4.0*m_nsps); // Actual fsample=48000
bool slowCwId=((isym >= m_symbolsLength) && (icw[0] > 0)) && (!m_bFastMode);
if(m_TRperiod==3.0) slowCwId=false;
bool fastCwId=false;
static bool bCwId=false;
qint64 ms = QDateTime::currentMSecsSinceEpoch();
float tsec=0.001*(ms % int(1000*m_TRperiod));
if(m_bFastMode and (icw[0]>0) and (tsec > (m_TRperiod-5.0))) fastCwId=true;
if(!m_bFastMode) m_nspd=2560; // 22.5 WPM
// qDebug() << "Mod A" << m_ic << isym << tsec;
if(slowCwId or fastCwId) { // Transmit CW ID?
m_dphi = m_twoPi*m_frequency/m_frameRate;
if(m_bFastMode and !bCwId) {
m_frequency=1500; // Set params for CW ID
m_dphi = m_twoPi*m_frequency/m_frameRate;
m_symbolsLength=126;
m_nsps=4096.0*12000.0/11025.0;
m_ic=2246949;
m_nspd=2560; // 22.5 WPM
if(icw[0]*m_nspd/48000.0 > 4.0) m_nspd=4.0*48000.0/icw[0]; //Faster CW for long calls
}
bCwId=true;
unsigned ic0 = m_symbolsLength * 4 * m_nsps;
unsigned j(0);
while (samples != end) {
j = (m_ic - ic0)/m_nspd + 1; // symbol of this sample
bool level {bool (icw[j])};
m_phi += m_dphi;
if (m_phi > m_twoPi) m_phi -= m_twoPi;
qint16 sample=0;
float amp=32767.0;
float x=0;
if(m_ramp!=0) {
x=qSin(float(m_phi));
if(SOFT_KEYING) {
amp=qAbs(qint32(m_ramp));
if(amp>32767.0) amp=32767.0;
}
sample=round(amp*x);
}
if(m_bFastMode) {
sample=0;
if(level) sample=32767.0*x;
}
if (int (j) <= icw[0] && j < NUM_CW_SYMBOLS) { // stop condition
samples = load (postProcessSample (sample), samples);
++framesGenerated;
++m_ic;
} else {
Q_EMIT stateChanged ((m_state = Idle));
return framesGenerated * bytesPerFrame ();
}
// adjust ramp
if ((m_ramp != 0 && m_ramp != std::numeric_limits<qint16>::min ()) || level != m_cwLevel) {
// either ramp has terminated at max/min or direction has changed
m_ramp += RAMP_INCREMENT; // ramp
}
m_cwLevel = level;
}
return framesGenerated * bytesPerFrame ();
} else {
bCwId=false;
} //End of code for CW ID
double const baud (12000.0 / m_nsps);
// fade out parameters (no fade out for tuning)
unsigned int i0,i1;
if(m_tuning) {
i1 = i0 = (m_bFastMode ? 999999 : 9999) * m_nsps;
} else {
i0=(m_symbolsLength - 0.017) * 4.0 * m_nsps;
i1= m_symbolsLength * 4.0 * m_nsps;
}
if(m_bFastMode and !m_tuning) {
i1=m_TRperiod*48000.0 - 24000.0;
i0=i1-816;
}
qint16 sample;
while (samples != end && m_ic <= i1) {
isym=0;
if(!m_tuning and m_TRperiod!=3.0) isym=m_ic/(4.0*m_nsps); //Actual fsample=48000
if(m_bFastMode) isym=isym%m_symbolsLength;
if (isym != m_isym0 || m_frequency != m_frequency0) {
if(itone[0]>=100) {
m_toneFrequency0=itone[0];
} else {
if(m_toneSpacing==0.0) {
m_toneFrequency0=m_frequency + itone[isym]*baud;
} else {
m_toneFrequency0=m_frequency + itone[isym]*m_toneSpacing;
}
}
m_dphi = m_twoPi * m_toneFrequency0 / m_frameRate;
m_isym0 = isym;
m_frequency0 = m_frequency; //???
}
int j=m_ic/480;
if(m_fSpread>0.0 and j!=m_j0) {
#if QT_VERSION >= QT_VERSION_CHECK(5, 15, 0)
float x1=QRandomGenerator::global ()->generateDouble ();
float x2=QRandomGenerator::global ()->generateDouble ();
#else
float x1=(float)qrand()/RAND_MAX;
float x2=(float)qrand()/RAND_MAX;
#endif
toneFrequency = m_toneFrequency0 + 0.5*m_fSpread*(x1+x2-1.0);
m_dphi = m_twoPi * toneFrequency / m_frameRate;
m_j0=j;
}
m_phi += m_dphi;
if (m_phi > m_twoPi) m_phi -= m_twoPi;
if (m_ic > i0) m_amp = 0.98 * m_amp;
if (m_ic > i1) m_amp = 0.0;
sample=qRound(m_amp*qSin(m_phi));
//Here's where we transmit from a precomputed wave[] array:
if(!m_tuning and (m_toneSpacing < 0)) {
m_amp=32767.0;
sample=qRound(m_amp*foxcom_.wave[m_ic]);
}
/*
if((m_ic<1000 or (4*m_symbolsLength*m_nsps - m_ic) < 1000) and (m_ic%10)==0) {
qDebug() << "cc" << QDateTime::currentDateTimeUtc().toString("hh:mm:ss.zzz") << m_ic << sample;
}
*/
samples = load(postProcessSample(sample), samples);
++framesGenerated;
++m_ic;
}
// qDebug() << "dd" << QDateTime::currentDateTimeUtc().toString("hh:mm:ss.zzz")
// << m_ic << i1 << foxcom_.wave[m_ic] << framesGenerated;
if (m_amp == 0.0) { // TODO G4WJS: compare double with zero might not be wise
if (icw[0] == 0) {
// no CW ID to send
Q_EMIT stateChanged ((m_state = Idle));
return framesGenerated * bytesPerFrame ();
}
m_phi = 0.0;
}
m_frequency0 = m_frequency;
// done for this chunk - continue on next call
// qDebug() << "Mod B" << m_ic << i1 << 0.001*(QDateTime::currentMSecsSinceEpoch() % (1000*m_TRperiod));
while (samples != end) // pad block with silence
{
samples = load (0, samples);
++framesGenerated;
}
return framesGenerated * bytesPerFrame ();
}
// fall through
case Idle:
break;
}
Q_ASSERT (Idle == m_state);
return 0;
}
qint16 Modulator::postProcessSample (qint16 sample) const
{
if (m_addNoise) { // Test frame, we'll add noise
qint32 s = m_fac * (gran () + sample * m_snr / 32768.0);
if (s > std::numeric_limits<qint16>::max ()) {
s = std::numeric_limits<qint16>::max ();
}
if (s < std::numeric_limits<qint16>::min ()) {
s = std::numeric_limits<qint16>::min ();
}
sample = s;
}
return sample;
}
Reference in New Issue View Git Blame Copy Permalink