diff --git a/gui/demos/1_fir_twotone.cpp b/gui/demos/1_fir_twotone.cpp new file mode 100644 index 0000000..2f33b45 --- /dev/null +++ b/gui/demos/1_fir_twotone.cpp @@ -0,0 +1,474 @@ +// Digilent Waveforms +// Custom signal: +// 4096 samples, X from 0 to 4095 +// sin(PI/8000*X*770)+sin(PI/8000*X*1336) + +#include +using float32_t = float; + +typedef struct +{ + uint16_t numTaps; /**< number of filter coefficients in the filter. */ + float32_t *pState; /**< points to the state variable array. The array is of length numTaps+blockSize-1. */ + float32_t *pCoeffs; /**< points to the coefficient array. The array is of length numTaps. */ +} arm_fir_instance_f32; + +static void arm_fir_f32(const arm_fir_instance_f32 * S, float32_t * pSrc, float32_t * pDst, uint32_t blockSize); + +adcsample_t *process_data(adcsample_t *samples, unsigned int size) +{ + // 1. Define our array sizes (Be sure to set Run > Set buffer size... to below value!) + constexpr unsigned int buffer_size = 500; + constexpr unsigned int filter_size = 100; + + // 2. Define our filter and the working arrays + static float filter[filter_size] = { +0.0004387887025183802,0.000832200098857514,0.0015307184662295217,0.0025082006379108486,0.0037734534845577567,0.005298545881839843,0.007011536790726042,0.008794237700486741,0.0104866537299476,0.011898670086639574,0.01282927474471195,0.013091387203266143,0.012539923791204348,0.011099382508792464,0.008786721087738127,0.005724922305741446,0.002144437816471804,-0.0016305355316754216,-0.005210998401135279,-0.00818522521443848,-0.0101701012867913,-0.010865871085906236,-0.01010594970823667,-0.007895785082463622,-0.00443176063435828,-0.00009684904191641966,0.004570418146137933,0.008932018058761999,0.012326603606274784,0.014157852725048914,0.013983789156387687,0.011595484863822138,0.007073052408314296,0.0008096544104778903,-0.006503129471685042,-0.013929880283520692,-0.020382390416343286,-0.024738365522063946,-0.02597487538349127,-0.02330063345756659,-0.016270714607748243,-0.004868040016583383,0.010460249064783475,0.028818398768187533,0.04893049916926163,0.06925561893929223,0.0881379889511425,0.10397551230005074,0.11538733811709222,0.12136089030896859,0.12136089030896859,0.11538733811709222,0.10397551230005074,0.0881379889511425,0.06925561893929223,0.04893049916926163,0.028818398768187533,0.010460249064783475,-0.004868040016583383,-0.016270714607748243,-0.02330063345756659,-0.02597487538349127,-0.024738365522063946,-0.020382390416343286,-0.013929880283520692,-0.006503129471685042,0.0008096544104778903,0.007073052408314296,0.011595484863822138,0.013983789156387687,0.014157852725048914,0.012326603606274784,0.008932018058761999,0.004570418146137933,-0.00009684904191641966,-0.00443176063435828,-0.007895785082463622,-0.01010594970823667,-0.010865871085906236,-0.0101701012867913,-0.00818522521443848,-0.005210998401135279,-0.0016305355316754216,0.002144437816471804,0.005724922305741446,0.008786721087738127,0.011099382508792464,0.012539923791204348,0.013091387203266143,0.01282927474471195,0.011898670086639574,0.0104866537299476,0.008794237700486741,0.007011536790726042,0.005298545881839843,0.0037734534845577567,0.0025082006379108486,0.0015307184662295217,0.000832200098857514,0.0004387887025183802 + }; + static float input[buffer_size]; + static float output[buffer_size]; + static float working[buffer_size + filter_size]; + + // 3. Scale 0-4095 interger sample values to +/- 1.0 floats + for (unsigned int i = 0; i < size; i++) + input[i] = (samples[i] - 2048) / 2048.f; + + // 4. Compute the FIR + arm_fir_instance_f32 fir { filter_size, working, filter }; + arm_fir_f32(&fir, input, output, size); + + // 5. Convert float results back to 0-4095 range for output + for (unsigned int i = 0; i < size; i++) + samples[i] = output[i] * 2048.f + 2048; + + return samples; +} + +// Below taken from the CMSIS DSP Library (find it on GitHub) +void arm_fir_f32( + const arm_fir_instance_f32 * S, + float32_t * pSrc, + float32_t * pDst, + uint32_t blockSize) +{ + float32_t *pState = S->pState; /* State pointer */ + float32_t *pCoeffs = S->pCoeffs; /* Coefficient pointer */ + float32_t *pStateCurnt; /* Points to the current sample of the state */ + float32_t *px, *pb; /* Temporary pointers for state and coefficient buffers */ + float32_t acc0, acc1, acc2, acc3, acc4, acc5, acc6, acc7; /* Accumulators */ + float32_t x0, x1, x2, x3, x4, x5, x6, x7, c0; /* Temporary variables to hold state and coefficient values */ + uint32_t numTaps = S->numTaps; /* Number of filter coefficients in the filter */ + uint32_t i, tapCnt, blkCnt; /* Loop counters */ + float32_t p0,p1,p2,p3,p4,p5,p6,p7; /* Temporary product values */ + + /* S->pState points to state array which contains previous frame (numTaps - 1) samples */ + /* pStateCurnt points to the location where the new input data should be written */ + pStateCurnt = &(S->pState[(numTaps - 1u)]); + + /* Apply loop unrolling and compute 8 output values simultaneously. + * The variables acc0 ... acc7 hold output values that are being computed: + * + * acc0 = b[numTaps-1] * x[n-numTaps-1] + b[numTaps-2] * x[n-numTaps-2] + b[numTaps-3] * x[n-numTaps-3] +...+ b[0] * x[0] + * acc1 = b[numTaps-1] * x[n-numTaps] + b[numTaps-2] * x[n-numTaps-1] + b[numTaps-3] * x[n-numTaps-2] +...+ b[0] * x[1] + * acc2 = b[numTaps-1] * x[n-numTaps+1] + b[numTaps-2] * x[n-numTaps] + b[numTaps-3] * x[n-numTaps-1] +...+ b[0] * x[2] + * acc3 = b[numTaps-1] * x[n-numTaps+2] + b[numTaps-2] * x[n-numTaps+1] + b[numTaps-3] * x[n-numTaps] +...+ b[0] * x[3] + */ + blkCnt = blockSize >> 3; + + /* First part of the processing with loop unrolling. Compute 8 outputs at a time. + ** a second loop below computes the remaining 1 to 7 samples. */ + while(blkCnt > 0u) + { + /* Copy four new input samples into the state buffer */ + *pStateCurnt++ = *pSrc++; + *pStateCurnt++ = *pSrc++; + *pStateCurnt++ = *pSrc++; + *pStateCurnt++ = *pSrc++; + + /* Set all accumulators to zero */ + acc0 = 0.0f; + acc1 = 0.0f; + acc2 = 0.0f; + acc3 = 0.0f; + acc4 = 0.0f; + acc5 = 0.0f; + acc6 = 0.0f; + acc7 = 0.0f; + + /* Initialize state pointer */ + px = pState; + + /* Initialize coeff pointer */ + pb = (pCoeffs); + + /* This is separated from the others to avoid + * a call to __aeabi_memmove which would be slower + */ + *pStateCurnt++ = *pSrc++; + *pStateCurnt++ = *pSrc++; + *pStateCurnt++ = *pSrc++; + *pStateCurnt++ = *pSrc++; + + /* Read the first seven samples from the state buffer: x[n-numTaps], x[n-numTaps-1], x[n-numTaps-2] */ + x0 = *px++; + x1 = *px++; + x2 = *px++; + x3 = *px++; + x4 = *px++; + x5 = *px++; + x6 = *px++; + + /* Loop unrolling. Process 8 taps at a time. */ + tapCnt = numTaps >> 3u; + + /* Loop over the number of taps. Unroll by a factor of 8. + ** Repeat until we've computed numTaps-8 coefficients. */ + while(tapCnt > 0u) + { + /* Read the b[numTaps-1] coefficient */ + c0 = *(pb++); + + /* Read x[n-numTaps-3] sample */ + x7 = *(px++); + + /* acc0 += b[numTaps-1] * x[n-numTaps] */ + p0 = x0 * c0; + + /* acc1 += b[numTaps-1] * x[n-numTaps-1] */ + p1 = x1 * c0; + + /* acc2 += b[numTaps-1] * x[n-numTaps-2] */ + p2 = x2 * c0; + + /* acc3 += b[numTaps-1] * x[n-numTaps-3] */ + p3 = x3 * c0; + + /* acc4 += b[numTaps-1] * x[n-numTaps-4] */ + p4 = x4 * c0; + + /* acc1 += b[numTaps-1] * x[n-numTaps-5] */ + p5 = x5 * c0; + + /* acc2 += b[numTaps-1] * x[n-numTaps-6] */ + p6 = x6 * c0; + + /* acc3 += b[numTaps-1] * x[n-numTaps-7] */ + p7 = x7 * c0; + + /* Read the b[numTaps-2] coefficient */ + c0 = *(pb++); + + /* Read x[n-numTaps-4] sample */ + x0 = *(px++); + + acc0 += p0; + acc1 += p1; + acc2 += p2; + acc3 += p3; + acc4 += p4; + acc5 += p5; + acc6 += p6; + acc7 += p7; + + + /* Perform the multiply-accumulate */ + p0 = x1 * c0; + p1 = x2 * c0; + p2 = x3 * c0; + p3 = x4 * c0; + p4 = x5 * c0; + p5 = x6 * c0; + p6 = x7 * c0; + p7 = x0 * c0; + + /* Read the b[numTaps-3] coefficient */ + c0 = *(pb++); + + /* Read x[n-numTaps-5] sample */ + x1 = *(px++); + + acc0 += p0; + acc1 += p1; + acc2 += p2; + acc3 += p3; + acc4 += p4; + acc5 += p5; + acc6 += p6; + acc7 += p7; + + /* Perform the multiply-accumulates */ + p0 = x2 * c0; + p1 = x3 * c0; + p2 = x4 * c0; + p3 = x5 * c0; + p4 = x6 * c0; + p5 = x7 * c0; + p6 = x0 * c0; + p7 = x1 * c0; + + /* Read the b[numTaps-4] coefficient */ + c0 = *(pb++); + + /* Read x[n-numTaps-6] sample */ + x2 = *(px++); + + acc0 += p0; + acc1 += p1; + acc2 += p2; + acc3 += p3; + acc4 += p4; + acc5 += p5; + acc6 += p6; + acc7 += p7; + + /* Perform the multiply-accumulates */ + p0 = x3 * c0; + p1 = x4 * c0; + p2 = x5 * c0; + p3 = x6 * c0; + p4 = x7 * c0; + p5 = x0 * c0; + p6 = x1 * c0; + p7 = x2 * c0; + + /* Read the b[numTaps-4] coefficient */ + c0 = *(pb++); + + /* Read x[n-numTaps-6] sample */ + x3 = *(px++); + + acc0 += p0; + acc1 += p1; + acc2 += p2; + acc3 += p3; + acc4 += p4; + acc5 += p5; + acc6 += p6; + acc7 += p7; + + /* Perform the multiply-accumulates */ + p0 = x4 * c0; + p1 = x5 * c0; + p2 = x6 * c0; + p3 = x7 * c0; + p4 = x0 * c0; + p5 = x1 * c0; + p6 = x2 * c0; + p7 = x3 * c0; + + /* Read the b[numTaps-4] coefficient */ + c0 = *(pb++); + + /* Read x[n-numTaps-6] sample */ + x4 = *(px++); + + acc0 += p0; + acc1 += p1; + acc2 += p2; + acc3 += p3; + acc4 += p4; + acc5 += p5; + acc6 += p6; + acc7 += p7; + + /* Perform the multiply-accumulates */ + p0 = x5 * c0; + p1 = x6 * c0; + p2 = x7 * c0; + p3 = x0 * c0; + p4 = x1 * c0; + p5 = x2 * c0; + p6 = x3 * c0; + p7 = x4 * c0; + + /* Read the b[numTaps-4] coefficient */ + c0 = *(pb++); + + /* Read x[n-numTaps-6] sample */ + x5 = *(px++); + + acc0 += p0; + acc1 += p1; + acc2 += p2; + acc3 += p3; + acc4 += p4; + acc5 += p5; + acc6 += p6; + acc7 += p7; + + /* Perform the multiply-accumulates */ + p0 = x6 * c0; + p1 = x7 * c0; + p2 = x0 * c0; + p3 = x1 * c0; + p4 = x2 * c0; + p5 = x3 * c0; + p6 = x4 * c0; + p7 = x5 * c0; + + /* Read the b[numTaps-4] coefficient */ + c0 = *(pb++); + + /* Read x[n-numTaps-6] sample */ + x6 = *(px++); + + acc0 += p0; + acc1 += p1; + acc2 += p2; + acc3 += p3; + acc4 += p4; + acc5 += p5; + acc6 += p6; + acc7 += p7; + + /* Perform the multiply-accumulates */ + p0 = x7 * c0; + p1 = x0 * c0; + p2 = x1 * c0; + p3 = x2 * c0; + p4 = x3 * c0; + p5 = x4 * c0; + p6 = x5 * c0; + p7 = x6 * c0; + + tapCnt--; + + acc0 += p0; + acc1 += p1; + acc2 += p2; + acc3 += p3; + acc4 += p4; + acc5 += p5; + acc6 += p6; + acc7 += p7; + } + + /* If the filter length is not a multiple of 8, compute the remaining filter taps */ + tapCnt = numTaps % 0x8u; + + while(tapCnt > 0u) + { + /* Read coefficients */ + c0 = *(pb++); + + /* Fetch 1 state variable */ + x7 = *(px++); + + /* Perform the multiply-accumulates */ + p0 = x0 * c0; + p1 = x1 * c0; + p2 = x2 * c0; + p3 = x3 * c0; + p4 = x4 * c0; + p5 = x5 * c0; + p6 = x6 * c0; + p7 = x7 * c0; + + /* Reuse the present sample states for next sample */ + x0 = x1; + x1 = x2; + x2 = x3; + x3 = x4; + x4 = x5; + x5 = x6; + x6 = x7; + + acc0 += p0; + acc1 += p1; + acc2 += p2; + acc3 += p3; + acc4 += p4; + acc5 += p5; + acc6 += p6; + acc7 += p7; + + /* Decrement the loop counter */ + tapCnt--; + } + + /* Advance the state pointer by 8 to process the next group of 8 samples */ + pState = pState + 8; + + /* The results in the 8 accumulators, store in the destination buffer. */ + *pDst++ = acc0; + *pDst++ = acc1; + *pDst++ = acc2; + *pDst++ = acc3; + *pDst++ = acc4; + *pDst++ = acc5; + *pDst++ = acc6; + *pDst++ = acc7; + + blkCnt--; + } + + /* If the blockSize is not a multiple of 8, compute any remaining output samples here. + ** No loop unrolling is used. */ + blkCnt = blockSize % 0x8u; + + while(blkCnt > 0u) + { + /* Copy one sample at a time into state buffer */ + *pStateCurnt++ = *pSrc++; + + /* Set the accumulator to zero */ + acc0 = 0.0f; + + /* Initialize state pointer */ + px = pState; + + /* Initialize Coefficient pointer */ + pb = (pCoeffs); + + i = numTaps; + + /* Perform the multiply-accumulates */ + do + { + acc0 += *px++ * *pb++; + i--; + + } while(i > 0u); + + /* The result is store in the destination buffer. */ + *pDst++ = acc0; + + /* Advance state pointer by 1 for the next sample */ + pState = pState + 1; + + blkCnt--; + } + + /* Processing is complete. + ** Now copy the last numTaps - 1 samples to the start of the state buffer. + ** This prepares the state buffer for the next function call. */ + + /* Points to the start of the state buffer */ + pStateCurnt = S->pState; + + tapCnt = (numTaps - 1u) >> 2u; + + /* copy data */ + while(tapCnt > 0u) + { + *pStateCurnt++ = *pState++; + *pStateCurnt++ = *pState++; + *pStateCurnt++ = *pState++; + *pStateCurnt++ = *pState++; + + /* Decrement the loop counter */ + tapCnt--; + } + + /* Calculate remaining number of copies */ + tapCnt = (numTaps - 1u) % 0x4u; + + /* Copy the remaining q31_t data */ + while(tapCnt > 0u) + { + *pStateCurnt++ = *pState++; + + /* Decrement the loop counter */ + tapCnt--; + } +} diff --git a/gui/demos/2_iir_echo.cpp b/gui/demos/2_iir_echo.cpp new file mode 100644 index 0000000..0bd7e9d --- /dev/null +++ b/gui/demos/2_iir_echo.cpp @@ -0,0 +1,22 @@ +adcsample_t *process_data(adcsample_t *samples, unsigned int size) +{ + constexpr float alpha = 0.7; + constexpr unsigned int D = 2000; + + static adcsample_t output[SIZE]; + static adcsample_t prev[D]; // prev[0] = output[0 - D] + + // Do calculations with previous output + for (unsigned int i = 0; i < D; i++) + output[i] = samples[i] + alpha * (prev[i] - 2048); + + // Do calculations with current samples + for (unsigned int i = D; i < size; i++) + output[i] = samples[i] + alpha * (output[i - D] - 2048); + + // Save outputs for next computation + for (unsigned int i = 0; i < D; i++) + prev[i] = output[size - (D - i)]; + + return output; +}