| << Chapter < Page | Chapter >> Page > |
Compile your project before looking at the resulting assembly file and investigating the differences between each block. Be sure to reference page 3-32 of the DSP Programmer's Guide to findout what the state of the FRCT and OVM bits are. Run this program on the DSP, halt the program, and compare theoutput values in a memory window. Does each block work properly for all possible values?
A working program can be produced by compiling the C code and
linking assembly modules and the core module. The compilertranslates C code to a relocatable assembly form. The linker
assigns physical addresses on the DSP to the relocatable dataand code segments, resolves
.global references and links runtime libraries.
Close the mathex project and go back to the original Lab 4 project. In the future if there are additional source code files to include in the project, just follow the above instructions. Once you have
completed
lab4.c and
c_fft_given.asm , select
Project->Rebuild All . Load the output file onto the DSP as usual and confirm that valid
FFTs are calculated. Once valid output is obtained, measure howmany clock cycles it takes to compute both the assembly and C FFT.
From your prelab experiments, you should be able to describe
the effect of windowing and zero-padding on FFT spectralanalysis. In your DSP system, experiment with different
inputs, changing
and the type of window. Can you explain what happens
as the input frequency is increased beyond the Nyquist rate? Does the
coincide with what you expect from Matlab? What is
the relationship between the observed spectrum and the DTFT?What would happen if the FFT calculation takes longer than it
takes to fill
inputs with
samples? How long does it take to compute each FFT?
What are the tradeoffs between writing code in C versus assembly?
#include "dsk5510_dual3006cfg.h"
#include "dsk5510.h"
#include "swi_process.h"
#include "dsplib.h"
#define N 1024
#define logN 10
#include "window.h"
/* comment the next line to use DSPLIB fft */
//#define C_FFT
#ifdef C_FFT /* Use C FFT */
/* function defined in lab4fft.c */
void fft(void);
/* FFT data buffers */
int real[N]; /* Real part of data */
int imag[N]; /* Imaginary part of data */
#include "lab4fft.c"
#else /* Use DSPLIB FFT */
/* Function defined by c_fft_given.asm */
void bit_rev(void);
/* FFT data buffers (declared in c_fft_given.asm) */
extern int bit_rev_data[N*2]; /* Data output for bit-reverse function */
extern int fft_data[N*2]; /* In-place FFT & Output array */
#endif /* C_FFT */
// all data processing should be done in SWI_ProcessBuffer
void SWI_ProcessBuffer()
{
static unsigned int mbox_value = 0;
short *psrc, *pdest;
unsigned int i;
mbox_value |= SWI_getmbox();
// buffers are only processed when both transmit and receive are ready
if((mbox_value & DMA_RECEIVE_DONE) && (mbox_value & DMA_TRANSMIT_DONE)) {
mbox_value = 0;
// get buffer pointers
psrc = receive_buffer[receive_buffer_to_process_index];
pdest = transmit_buffer[transmit_buffer_to_fill_index];
// samples are interleaved in input buffer 3-4-1-2
// output buffer is organized 3-4-1-2-3-4-1-2
// The following code would copy input from each input channel to the
// respective output channel:
/*
for (i = 0; i < 1024; i++)
{
pdest[4*i] = psrc[4*i]; //channel 3 output is channel 3 input
pdest[4*i+1] = psrc[4*i+1]; //channel 4 output is channel 4 input
pdest[4*i+2] = psrc[4*i+2]; //channel 1 output is channel 1 input
pdest[4*i+3] = psrc[4*i+3]; //channel 2 output is channel 2 input
}
*/
#ifdef C_FFT /* Use C FFT */
/* I n s e r t c o d e t o f i l l */
/* C F F T b u f f e r s */
#else /* Use DSPLIB FFT */
/* I n s e r t c o d e t o f i l l */
/* a s s e m b l y F F T b u f f e r s */
#endif /* C_FFT */
#ifdef C_FFT /* Use C FFT */
/* Multiply the input signal by the Hamming window. */
/* . . . i n s e r t C / a s m code . . . */
/* Bit-reverse and compute FFT in C */
fft();
/* Now, take absolute value squared of FFT */
/* . . . i n s e r t C / a s m code . . . */
/* Last, set the DC coefficient to -1 for a trigger pulse */
/* . . . i n s e r t C / a s m code . . . */
/* done, wait for next time around! */
#else /* Use DSPLIB FFT */
/* Multiply the input signal by the Hamming window. */
/* . . . i n s e r t C / a s m code . . . */
/* Compute FFT using DSPLIB function */
cfft((DATA *)fft_data,N, SCALE);
/* Bit reverse using assembly function */
bit_rev();
/* Now, take absolute value squared of FFT */
/* . . . i n s e r t C / a s m code . . . */
/* Last, set the DC coefficient to -1 for a trigger pulse */
/* . . . i n s e r t C / a s m code . . . */
/* done, wait for next time around! */
#endif /* C_FFT */
receive_buffer_processed = 1; // flag receive buffer as processed
transmit_buffer_filled = 1; // flag output buffer as full
}
}
Notification Switch
Would you like to follow the 'Digital signal processing laboratory (ece 420 55x)' conversation and receive update notifications?
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|