1
0
mirror of https://github.com/f4exb/sdrangel.git synced 2024-11-26 09:48:45 -05:00

IQ correction with phase imbalance: floating point implementation

This commit is contained in:
f4exb 2018-02-04 10:49:13 +01:00
parent 58f0145705
commit b9b2c41ba2
5 changed files with 220 additions and 39 deletions

View File

@ -23,6 +23,7 @@
#include <stdio.h>
#include <QDebug>
#include "dsp/dspcommands.h"
#include "util/fixed.h"
#include "samplesinkfifo.h"
#include "threadedbasebandsamplesink.h"
@ -182,14 +183,41 @@ void DSPDeviceSourceEngine::iqCorrections(SampleVector::iterator begin, SampleVe
if (imbalanceCorrection)
{
int32_t xi = it->m_real - (int32_t) m_iBeta;
int32_t xq = it->m_imag - (int32_t) m_qBeta;
m_avgII(xi*xi);
m_avgIQ(xi*xq);
// TODO
// DC correction and conversion
float xi = (it->m_real - (int32_t) m_iBeta) / SDR_RX_SCALEF;
float xq = (it->m_imag - (int32_t) m_qBeta) / SDR_RX_SCALEF;
// phase imbalance
m_avgII(xi*xi); // <I", I">
m_avgIQ(xi*xq); // <I", Q">
if (m_avgII.asDouble() != 0) {
m_avgPhi(m_avgIQ.asDouble()/m_avgII.asDouble());
}
float yi = xi - m_avgPhi.asDouble()*xq;
float yq = xq;
// amplitude I/Q imbalance
m_avgII2(yi*yi); // <I, I>
m_avgQQ2(yq*yq); // <Q, Q>
if (m_avgQQ2.asDouble() != 0) {
m_avgAmp(sqrt(m_avgII2.asDouble() / m_avgQQ2.asDouble()));
}
// final correction
float zi = yi;
float zq = m_avgAmp.asDouble() * yq;
// convert and store
it->m_real = zi * SDR_RX_SCALEF;
it->m_imag = zq * SDR_RX_SCALEF;
}
else
{
// DC correction only
it->m_real -= (int32_t) m_iBeta;
it->m_imag -= (int32_t) m_qBeta;
}
@ -198,11 +226,6 @@ void DSPDeviceSourceEngine::iqCorrections(SampleVector::iterator begin, SampleVe
void DSPDeviceSourceEngine::dcOffset(SampleVector::iterator begin, SampleVector::iterator end)
{
// double count;
// int io = 0;
// int qo = 0;
// Sample corr((FixReal)m_iOffset, (FixReal)m_qOffset);
// sum and correct in one pass
for(SampleVector::iterator it = begin; it < end; it++)
{
@ -210,15 +233,7 @@ void DSPDeviceSourceEngine::dcOffset(SampleVector::iterator begin, SampleVector:
m_qBeta(it->imag());
it->m_real -= (int32_t) m_iBeta;
it->m_imag -= (int32_t) m_qBeta;
// io += it->real();
// qo += it->imag();
// *it -= corr;
}
// // moving average
// count = end - begin;
// m_iOffset = (15.0 * m_iOffset + (double)io / count) / 16.0;
// m_qOffset = (15.0 * m_qOffset + (double)qo / count) / 16.0;
}
void DSPDeviceSourceEngine::imbalance(SampleVector::iterator begin, SampleVector::iterator end)
@ -289,15 +304,20 @@ void DSPDeviceSourceEngine::work()
if (part1begin != part1end)
{
// correct stuff
if (m_dcOffsetCorrection)
{
dcOffset(part1begin, part1end);
}
if (m_dcOffsetCorrection)
{
iqCorrections(part1begin, part1end, m_iqImbalanceCorrection);
}
if (m_iqImbalanceCorrection)
{
imbalance(part1begin, part1end);
}
// if (m_dcOffsetCorrection)
// {
// dcOffset(part1begin, part1end);
// }
//
// if (m_iqImbalanceCorrection)
// {
// imbalance(part1begin, part1end);
// }
// feed data to direct sinks
for (BasebandSampleSinks::const_iterator it = m_basebandSampleSinks.begin(); it != m_basebandSampleSinks.end(); ++it)
@ -316,15 +336,20 @@ void DSPDeviceSourceEngine::work()
if(part2begin != part2end)
{
// correct stuff
if (m_dcOffsetCorrection)
{
dcOffset(part2begin, part2end);
}
if (m_dcOffsetCorrection)
{
iqCorrections(part2begin, part2end, m_iqImbalanceCorrection);
}
if (m_iqImbalanceCorrection)
{
imbalance(part2begin, part2end);
}
// if (m_dcOffsetCorrection)
// {
// dcOffset(part2begin, part2end);
// }
//
// if (m_iqImbalanceCorrection)
// {
// imbalance(part2begin, part2end);
// }
// feed data to direct sinks
for (BasebandSampleSinks::const_iterator it = m_basebandSampleSinks.begin(); it != m_basebandSampleSinks.end(); it++)

View File

@ -104,8 +104,14 @@ private:
double m_iOffset, m_qOffset;
MovingAverageUtil<int32_t, int64_t, 1024> m_iBeta;
MovingAverageUtil<int32_t, int64_t, 1024> m_qBeta;
MovingAverageUtil<int64_t, int64_t, 1024> m_avgII;
MovingAverageUtil<int64_t, int64_t, 1024> m_avgIQ;
MovingAverageUtil<float, double, 128> m_avgII;
MovingAverageUtil<float, double, 128> m_avgIQ;
MovingAverageUtil<float, double, 128> m_avgII2;
MovingAverageUtil<float, double, 128> m_avgQQ2;
MovingAverageUtil<double, double, 128> m_avgPhi;
MovingAverageUtil<double, double, 128> m_avgAmp;
// MovingAverageUtil<int64_t, int64_t, 1024> m_avgII;
// MovingAverageUtil<int64_t, int64_t, 1024> m_avgIQ;
qint32 m_iRange;
qint32 m_qRange;
qint32 m_imbalance;

View File

@ -12,6 +12,8 @@
// Modified as fully templatized class with variable size and type internal //
// representation //
// //
// sqrt requires even IntBits //
// //
// This program is free software; you can redistribute it and/or modify //
// it under the terms of the GNU General Public License as published by //
// the Free Software Foundation as version 3 of the License, or //
@ -479,9 +481,9 @@ Fixed<IntType, IntBits>& Fixed<IntType, IntBits>::operator/=(Fixed<IntType, IntB
template<typename IntType, uint32_t IntBits>
Fixed<IntType, IntBits> Fixed<IntType, IntBits>::sqrt() const
{
unsigned const max_shift = 62;
unsigned int const max_shift = 62;
uint64_t a_squared = 1LL << max_shift;
unsigned b_shift = (max_shift + FixedTraits<IntBits>::fixed_resolution_shift) / 2;
unsigned int b_shift = (max_shift + FixedTraits<IntBits>::fixed_resolution_shift) / 2;
uint64_t a = 1LL << b_shift;
uint64_t x = m_nVal;

View File

@ -282,6 +282,138 @@ const int64_t FixedTraits<16>::arctantab[32] = {
0LL
};
// 1.0 = 2^23 internal representation
const int64_t FixedTraits<23>::log_two_power_n_reversed[40] = {
232581599LL,
226767059LL,
220952519LL,
215137979LL,
209323439LL,
203508899LL,
197694359LL,
191879819LL,
186065279LL,
180250740LL,
174436200LL,
168621660LL,
162807120LL,
156992580LL,
151178040LL,
145363500LL,
139548960LL,
133734420LL,
127919880LL,
122105340LL,
116290800LL,
110476260LL,
104661720LL,
98847180LL,
93032640LL,
87218100LL,
81403560LL,
75589020LL,
69774480LL,
63959940LL,
58145400LL,
52330860LL,
46516320LL,
40701780LL,
34887240LL,
29072700LL,
23258160LL,
17443620LL,
11629080LL,
5814540LL
};
const int64_t FixedTraits<23>::log_one_plus_two_power_minus_n[23] = {
3401288LL,
1871864LL,
988036LL,
508556LL,
258131LL,
130059LL,
65281LL,
32704LL,
16368LL,
8188LL,
4095LL,
2048LL,
1024LL,
512LL,
256LL,
128LL,
64LL,
32LL,
16LL,
8LL,
4LL,
2LL,
1LL
};
const int64_t FixedTraits<23>::log_one_over_one_minus_two_power_minus_n[23] = {
5814540LL,
2413252LL,
1120143LL,
541388LL,
266327LL,
132107LL,
65793LL,
32832LL,
16400LL,
8196LL,
4097LL,
2048LL,
1024LL,
512LL,
256LL,
128LL,
64LL,
32LL,
16LL,
8LL,
4LL,
2LL,
1LL
};
const int64_t FixedTraits<23>::arctantab[32] = {
9287436LL,
6588397LL,
3889358LL,
2055029LL,
1043165LL,
523606LL,
262058LL,
131061LL,
65534LL,
32767LL,
16383LL,
8192LL,
4096LL,
2048LL,
1024LL,
512LL,
256LL,
128LL,
64LL,
32LL,
16LL,
8LL,
4LL,
2LL,
1LL,
0LL,
0LL,
0LL,
0LL,
0LL,
0LL,
0LL
};
// 1.0 = 2^24 internal representation
const int64_t FixedTraits<24>::log_two_power_n_reversed[39] = {

View File

@ -57,6 +57,22 @@ struct FixedTraits<16>
static const int64_t arctantab[32];
};
template<>
struct FixedTraits<23>
{
static const uint32_t fixed_resolution_shift = 23;
static const int64_t fixed_resolution = 1LL << fixed_resolution_shift;
static const int32_t max_power = 63 - fixed_resolution_shift;
static const int64_t internal_pi = 26353589;
static const int64_t internal_two_pi = 52707179;
static const int64_t internal_half_pi = 13176795;
static const int64_t internal_quarter_pi = 6588397;
static const int64_t log_two_power_n_reversed[40]; // 40 = 63 - 23
static const int64_t log_one_plus_two_power_minus_n[23];
static const int64_t log_one_over_one_minus_two_power_minus_n[23];
static const int64_t arctantab[32];
};
template<>
struct FixedTraits<24>
{
@ -67,7 +83,7 @@ struct FixedTraits<24>
static const int64_t internal_two_pi = 105414357;
static const int64_t internal_half_pi = 26353589;
static const int64_t internal_quarter_pi = 13176795;
static const int64_t log_two_power_n_reversed[39]; // 39 = 63 - 16
static const int64_t log_two_power_n_reversed[39]; // 39 = 63 - 24
static const int64_t log_one_plus_two_power_minus_n[24];
static const int64_t log_one_over_one_minus_two_power_minus_n[24];
static const int64_t arctantab[32];