FFT256iq corrected windowing subscripts, added xAxis fcn, nAve

pull/6/merge
boblark 3 years ago
parent c621d6cb0d
commit 10b56112d1
  1. 78
      analyze_fft256_iq_F32.cpp
  2. 81
      analyze_fft256_iq_F32.h
  3. 57
      examples/TestFFT256iq/TestFFT256iq.ino

@ -39,15 +39,13 @@
#include "analyze_fft256_iq_F32.h"
// Move audio data from audio_block_f32_t to the interleaved FFT instance buffer.
static void copy_to_fft_buffer1(void *destination, const void *sourceI, const void *sourceQ) {
static void copy_to_fft_buffer0(void *destination, const void *sourceI, const void *sourceQ) {
const float *srcI = (const float *)sourceI;
const float *srcQ = (const float *)sourceQ;
float *dst = (float *)destination;
for (int i=0; i < AUDIO_BLOCK_SAMPLES; i++) {
*dst++ = *srcI++; // real sample, interleave
//*dst++ = 0.0f;
*dst++ = *srcQ++; // imag
//*dst++ = 0.0f;
}
}
@ -55,13 +53,14 @@ static void apply_window_to_fft_buffer1(void *fft_buffer, const void *window) {
float *buf = (float *)fft_buffer; // 0th entry is real (do window) 1th is imag
const float *win = (float *)window;
for (int i=0; i < 256; i++) {
buf[2*i] *= *win++; // real
buf[2*i] *= *win; // real
buf[2*i + 1] *= *win++; // imag
}
}
void AudioAnalyzeFFT256_IQ_F32::update(void) {
audio_block_f32_t *block_i,*block_q;
int ii;
block_i = receiveReadOnly_f32(0);
if (!block_i) return;
@ -78,40 +77,67 @@ void AudioAnalyzeFFT256_IQ_F32::update(void) {
prevblock_q = block_q;
return; // Nothing to release
}
// FFT is 256 and blocks are 128, so we need 2 blocks. We still do
// this every 128 samples to get 50% overlap on FFT data to roughly
// compensate for windowing.
// ( dest, i-source, q-source )
copy_to_fft_buffer1(fft_buffer, prevblock_i->data, prevblock_q->data);
copy_to_fft_buffer1(fft_buffer+256, block_i->data, block_q->data);
copy_to_fft_buffer0(fft_buffer, prevblock_i->data, prevblock_q->data);
copy_to_fft_buffer0(fft_buffer+256, block_i->data, block_q->data);
if (pWin)
apply_window_to_fft_buffer1(fft_buffer, window);
arm_cfft_radix4_f32(&fft_inst, fft_buffer); // Finally the FFT
#if defined(__IMXRT1062__)
// Teensyduino core for T4.x supports arm_cfft_f32
// arm_cfft_f32 (const arm_cfft_instance_f32 *S, float32_t *p1, uint8_t ifftFlag, uint8_t bitReverseFlag)
arm_cfft_f32(&Sfft, fft_buffer, 0, 1);
#else
// For T3.x go back to old (deprecated) style
arm_cfft_radix4_f32(&fft_inst, fft_buffer);
#endif
count++;
for (int i=0; i < 256; i++) {
float ss = fft_buffer[2*i]*fft_buffer[2*i] + fft_buffer[2*i+1]*fft_buffer[2*i+1];
if(count==1) // Starting new average
sumsq[i] = ss;
else if (count <= nAverage) // Adding on to average
sumsq[i] += ss;
for (int i = 0; i < 128; i++) {
// From complex FFT the "negative frequencies" are mirrors of the frequencies above fs/2. So, we get
// frequencies from 0 to fs by re-arranging the coefficients. These are powers (not Volts)
// See DD4WH SDR (Note - here and at "if(xAxis & xxxx)" below, we may have redundancy in index changing.
// Leave as is for now.)
float ss0 = fft_buffer[2 * i] * fft_buffer[2 * i] +
fft_buffer[2 * i + 1] * fft_buffer[2 * i + 1];
float ss1 = fft_buffer[2 * (i + 128)] * fft_buffer[2 * (i + 128)] +
fft_buffer[2 * (i + 128) + 1] * fft_buffer[2 * (i + 128) + 1];
if(count==1) { // Starting new average
sumsq[i+128] = ss0;
sumsq[i] = ss1;
}
else if (count <= nAverage) { // Adding on to average
sumsq[i+128] += ss0;
sumsq[i] += ss1;
}
}
if (count >= nAverage) { // Average is finished
count = 0;
float inAf = 1.0f/(float)nAverage;
for (int i=0; i < 256; i++) {
int ii = 255 - (i ^ 128);
if(outputType==FFT_RMS)
output[ii] = sqrtf(inAf*sumsq[ii]);
else if(outputType==FFT_POWER)
output[ii] = inAf*sumsq[ii];
else if(outputType==FFT_DBFS)
output[ii] = 10.0f*log10f(inAf*sumsq[ii])-42.1442f; // Scaled to FS sine wave
else
output[ii] = 0.0f;
}
count = 0;
float inAf = 1.0f/(float)nAverage;
for (int i=0; i < 256; i++) {
// xAxis, bit 0 left/right; bit 1 low to high
if(xAxis & 0X02)
ii = i;
else
ii = i^128;
if(xAxis & 0X01)
ii = (255 - ii);
if(outputType==FFT_RMS)
output[i] = sqrtf(inAf*sumsq[ii]);
else if(outputType==FFT_POWER)
output[i] = inAf*sumsq[ii];
else if(outputType==FFT_DBFS)
output[i] = 10.0f*log10f(inAf*sumsq[ii])-42.1442f; // Scaled to FS sine wave
else
output[i] = 0.0f;
}
}
outputflag = true;
release(prevblock_i); // Release the 2 blocks that were block_i
release(prevblock_q); // and block_q on last time through update()

@ -1,15 +1,27 @@
/* analyze_fft256_iq_F32.h
/* analyze_fft256_iq_F32.h Assembled by Bob Larkin 6 Mar 2021
*
* Converted to F32 floating point input and also extended
* for complex I and Q inputs
* * Adapted all I/O to be F32 floating point for OpenAudio_ArduinoLibrary
* * Future: Add outputs for I & Q FFT x2 for overlapped FFT
* Rev 6 Mar 2021 - Added setXAxis()
* Rev 7 Mar 2021 - Corrected bug in applying windowing
*
* Does Fast Fourier Transform of a 256 point complex (I-Q) input.
* Output is one of three measures of the power in each of the 256
* output bins, Power, RMS level or dB relative to a full scale
* sine wave. Windowing of the input data is provided for to reduce
* spreading of the power in the output bins. All inputs are Teensy
* floating point extension (_F32) and all outputs are floating point.
*
* Features include:
* * I and Q inputs are OpenAudio_Arduino Library F32 compatible.
* * FFT output for every 128 inputs to overlapped FFTs to
* compensate for windowing.
* * Windowing None, Hann, Kaiser and Blackman-Harris.
* * Multiple bin-sum output to simulate wider bins.
* * Power averaging of multiple FFT
* * Soon: F32 audio outputs for I & Q
*
* Conversion Copyright (c) 2021 Bob Larkin
* Same MIT license as PJRC:
*
*
* Audio Library for Teensy 3.X
* Copyright (c) 2014, Paul Stoffregen, paul@pjrc.com
*
@ -36,8 +48,8 @@
* THE SOFTWARE.
*/
/* Does complex input FFT of 1024 points. Output is not audio, and is magnitude
* only. Multiple output formats of RMS (same as I16 version, and default),
/* Does complex input FFT of 256 points. Multiple non-audio (via functions)
* output formats of RMS (same as I16 version, and default),
* Power or dBFS (full scale). Output can be bin by bin or a pointer to
* the output array is available. Several window functions are provided by
* in-class design, or a custom window can be provided from the INO.
@ -51,7 +63,17 @@
* float* getData(void)
* float* getWindow(void)
* void putWindow(float *pwin)
* void setNAverage(int NAve) // >=1
* void setOutputType(int _type)
* void setXAxis(uint8_t _xAxis) // 0, 1, 2, 3
*
* x-Axis direction and offset per setXAxis(xAxis) for sine to I
* and cosine to Q.
* If xAxis=0 f=fs/2 in middle, f=0 on right edge
* If xAxis=1 f=fs/2 in middle, f=0 on left edge
* If xAxis=2 f=fs/2 on left edge, f=0 in middle
* If xAxis=3 f=fs/2 on right edgr, f=0 in middle
* If there is 180 degree phase shift to I or Q these all get reversed.
*
* Timing, max is longest update() time:
* T3.6 Windowed, RMS out, - uSec max
@ -75,13 +97,13 @@
#ifndef analyze_fft256iq_h_
#define analyze_fft256iq_h_
//#include "AudioStream.h"
//#include "arm_math.h"
#include "Arduino.h"
#include "AudioStream_F32.h"
#include "arm_math.h"
#include "mathDSP_F32.h"
#if defined(__IMXRT1062__)
#include "arm_const_structs.h"
#endif
#define FFT_RMS 0
#define FFT_POWER 1
@ -97,10 +119,21 @@ class AudioAnalyzeFFT256_IQ_F32 : public AudioStream_F32 {
//GUI: inputs:2, outputs:4 //this line used for automatic generation of GUI node
//GUI: shortName:AnalyzeFFT256IQ
public:
AudioAnalyzeFFT256_IQ_F32() : AudioStream_F32(2, inputQueueArray) { // NEEDS SETTINGS etc <<<<<<<<
arm_cfft_radix4_init_f32(&fft_inst, 256, 0, 1);
AudioAnalyzeFFT256_IQ_F32() : AudioStream_F32(2, inputQueueArray) {
// __MK20DX128__ T_LC; __MKL26Z64__ T3.0; __MK20DX256__T3.1 and T3.2
// __MK64FX512__) T3.5; __MK66FX1M0__ T3.6; __IMXRT1062__ T4.0 and T4.1
#if defined(__IMXRT1062__)
// Teensy4 core library has the right files for new FFT
// arm CMSIS library has predefined structures of type arm_cfft_instance_f32
Sfft = arm_cfft_sR_f32_len256; // This is one of the structures
#else
arm_cfft_radix4_init_f32(&fft_inst, 256, 0, 1); // for T3.x
#endif
useHanningWindow();
}
// There is no varient for "settings," as blocks other than 128 are
// not supported and, nothing depends on sample rate so we don't need that.
bool available() {
if (outputflag == true) {
@ -181,7 +214,18 @@ public:
outputType = _type;
}
virtual void update(void);
// Output power (non-coherent) averaging
// i.e., the number of FFT powers averaged in the output
void setNAverage(int _nAverage) {
nAverage = _nAverage;
}
// xAxis, bit 0 left/right; bit 1 low to high; default 0X03
void setXAxis(uint8_t _xAxis) {
xAxis = _xAxis;
}
virtual void update(void);
private:
float output[256];
@ -193,10 +237,17 @@ private:
bool outputflag = false;
audio_block_f32_t *inputQueueArray[2];
audio_block_f32_t *prevblock_i,*prevblock_q;
#if defined(__IMXRT1062__)
// For T4.x
// const static arm_cfft_instance_f32 arm_cfft_sR_f32_len256;
arm_cfft_instance_f32 Sfft;
#else
arm_cfft_radix4_instance_f32 fft_inst;
#endif
int outputType = FFT_RMS; //Same type as I16 version init
int count = 0;
int nAverage = 1;
uint8_t xAxis = 3;
// The Hann window is a good all-around window
void useHanningWindow(void) {
@ -238,8 +289,6 @@ private:
kbes = 1.0f / mathEqualizer.i0f(beta); // An additional derived parameter used in loop
for (int n=0; n<128; n++) {
xn2 = 0.5f+(float32_t)n;
// 4/(1023^2)=0.00000382215877f
// xn2 = 0.00000382215877f*xn2*xn2;
// 4/(255^2)=0.000061514802f
xn2 = 0.000061514802f*xn2*xn2;
window[127 - n]=kbes*(mathEqualizer.i0f(beta*sqrtf(1.0-xn2)));

@ -10,12 +10,11 @@
#include <SerialFlash.h>
// GUItool: begin automatically generated code
AudioSynthSineCosine_F32 sine_cos1; //xy=76,532
AudioSynthSineCosine_F32 sine_cos1; //xy=76,532
AudioAnalyzeFFT256_IQ_F32 FFT256iq1; //xy=243,532
AudioOutputI2S_F32 audioOutI2S1; //xy=246,591
AudioConnection_F32 patchCord1(sine_cos1, 0, FFT256iq1, 0);
AudioConnection_F32 patchCord2(sine_cos1, 1, FFT256iq1, 1);
//AudioControlSGTL5000 sgtl5000_1;
AudioOutputI2S_F32 audioOutI2S1; //xy=246,591
AudioConnection_F32 patchCord1(sine_cos1, 0, FFT256iq1, 0);
AudioConnection_F32 patchCord2(sine_cos1, 1, FFT256iq1, 1);
// GUItool: end automatically generated code
void setup(void) {
@ -24,27 +23,55 @@ void setup(void) {
Serial.begin(9600);
delay(1000);
AudioMemory_F32(20);
Serial.println("FFT256IQ Test");
// sgtl5000_1.enable(); //start the audio board
// sgtl5000_1.inputSelect(AUDIO_INPUT_LINEIN); // or AUDIO_INPUT_MIC
Serial.println("FFT256IQ Test v2");
sine_cos1.amplitude(0.5); // Initialize Waveform Generator
sine_cos1.amplitude(1.0); // Initialize Waveform Generator
// bin spacing = 44117.648/256 = 172.335 172.3 * 4 = 689.335 Hz (T3.6)
// Half bin higher is 775.3 for testing windows
//sine_cos1.frequency(689.34f);
sine_cos1.frequency(1723.35f);
// Pick T3.6 bin center
//sine_cos1.frequency(689.33);
// or pick T4.x bin center
//sine_cos1.frequency(689.0625f);
// or pick any old frequency
sine_cos1.frequency(7100.0);
// elect the output format
FFT256iq1.setOutputType(FFT_DBFS);
// Select the wndow function
//FFT256iq1.windowFunction(AudioWindowNone);
//FFT256iq1.windowFunction(AudioWindowHanning256);
//FFT256iq1.windowFunction(AudioWindowKaiser256, 55.0f);
FFT256iq1.windowFunction(AudioWindowBlackmanHarris256);
// Uncomment to Serial print window function
//float* pw = FFT256iq1.getWindow(); // Print window
//for (int i=0; i<512; i++) Serial.println(pw[i], 4);
// xAxis, bit 0 left/right; bit 1 low to high; default 0X03
FFT256iq1.setXAxis(0X03);
FFT256iq1.windowFunction(AudioWindowBlackmanHarris256);
//float* pw = FFT256iq1.getWindow(); // Print window
//for (int i=0; i<256; i++) Serial.println(pw[i], 4);
delay(1000);
if( FFT256iq1.available() )
pPwr = FFT256iq1.getData();
// Do power averaging (outputs appear less often, as well)
FFT256iq1.setNAverage(5); // nAverage >= 1
for(int i=0; i<256; i++)
Serial.println(*(pPwr + i), 8 );
delay(1000);
if( FFT256iq1.available() ) {
pPwr = FFT256iq1.getData();
for(int i=0; i<256; i++) {
Serial.print(i);
Serial.print(",");
Serial.println(*(pPwr + i), 8 );
}
Serial.print("\n\n");
}
}
void loop(void) {

Loading…
Cancel
Save