<< Chapter < Page Chapter >> Page >
This is an implementation of an autocorrelation-based power spectral density (PSD) estimator. This implementation estimates the PSD of an IIR-filtered pseudo-noise generator.

Reference implementation of a psd estimator

We provide for you in Appendix D and E a complete C implementation of a PSD estimator. The input is an IIR-filtered pseudo-noise (PN)sequence generator and the PSD estimate is based on windowing the autocorrelation with a rectangular window.The code consists of the files lab4b.c , lab4b.h , intrinsics.h , pn.c , iirfilt.c , autocorr.c , c_fft_given_iirc.asm , and the previously-given TI FFT routine. The assembly file c_fft_given_iirc.asm differs from c_fft_given.asm in that the window array has been removed and variables and arrays associated with IIR filtering havebeen added. Note that the multiply functions in the functions are actually compiler directives contained in intrinsics.h . Make sure you know which ones are used and why; note that VPO is not defined by the TI compiler, therefore the corresponding section of the #ifdef statement is not used. Open up the lab4b.pjt project and Rebuild All . Load lab4b.out onto the DSP and run the code. Make sure that anIIR-filtered PN sequence appears on channel 1 and its PSD estimate appears on channel 2.

Does the output match your expectations based on the theory? Does this application illustrate any limitations of the FFT implementation?(Hint: note that most of the values in the FFT input are zero.) The previously-given C implementation uses a similar algorithm as theTI FFT; take a look at the C code for help. What are the limitation(s) of the FFT that show up in this application?

In lab4b.h M sets the number of autocorrelation points that are calculated. What is the maximum valueof M that allows the reference routines to run in real time? In determining this value you may find it useful to connect awave-function generator to the input and copy input on that channel into channel 1 of output. You may limit M to powers of 2 minus 1 .

Appendix a: main routine, header files for psd estimator

lab4b.h

intrinsics.h

lab4b.c

lab4bmain.c

1 #define N 1024 /* Length of output buffers */ 2 #define L N /* Length of input data */ 3 #define logL 10 /* log base 2 of L */ 4 #define M 31 /* Compute 2*M+1 autocorrelation points */ 5 6 /* #define M (L/2-1) */ /* Be sure to use ()'s in this case */ 7 /* or algebraic substitution bugs */ 8 /* can be introduced */ /* Compiler intrinsics for the TI compiler */ /* and the Very Portable Optimizer (VPO) port */ /* to TMS320C54X series DSPs */ /* */ /* Use compile option -DVPO when using VPO */ /* */ /* Copyright September 2005 by Matt Kleffner */ /* under the Creative Commons Attribution License */ /* Works with TMS320C55X series */ #ifndef INTRINSICS_H #define INTRINSICS_H #ifdef VPO long int vpo_l_mul_ii(int w0, int w1); /* fractional multiply without fractional mode (long result) */ #define _l_mul_fract_fb0_ii(w0,w1) \ (vpo_l_mul_ii(w0,w1) << 1) /* fractional multiply with fractional mode already on (long result) */ #define _l_mul_fract_fb1_ii(w0,w1) \ (vpo_l_mul_ii(w0,w1)) /* fractional multiply without fractional mode (int result) */ #define _i_mul_fract_fb0_ii(w0,w1) \ (vpo_l_mul_ii(w0,w1) >> 15) /* fractional multiply with fractional mode already on (int result) */ #define _i_mul_fract_fb1_ii(w0,w1) \ (vpo_l_mul_ii(w0,w1) >> 16) #define _set_fract_bit() vpo_set_fract() #define _reset_fract_bit() vpo_reset_fract() #define _set_ovm_bit() vpo_set_ovm() #define _reset_ovm_bit() vpo_reset_ovm() #define _l_add_shiftl_li(w0,w1) (((int32)(w0))+(((int32)(int16)(w1))<<16) ) #define _l_sub_shiftl_li(w0,w1) (((int32)(w0))-(((int32)(int16)(w1))<<16) ) #else /* fractional multiply without fractional mode (long result) */ #define _l_mul_fract_fb0_ii(w0,w1) \ (((long int)w0 * (long int)w1) << 1) /* fractional multiply with fractional mode already on (long result) */ #define _l_mul_fract_fb1_ii(w0,w1) \ (((long int)w0 * (long int)w1)) /* fractional multiply without fractional mode (int result) */ #define _i_mul_fract_fb0_ii(w0,w1) \ (((long int)w0 * (long int)w1) >> 15) /* fractional multiply with fractional mode already on (int result) */ #define _i_mul_fract_fb1_ii(w0,w1) \ (((long int)w0 * (long int)w1) >> 16) #define _set_fract_bit() asm(" ssbx frct") #define _reset_fract_bit() asm(" rsbx frct") #define _set_ovm_bit() asm(" ssbx ovm") #define _reset_ovm_bit() asm(" rsbx ovm") #endif /* VPO */ #endif /* INTRINSICS_H */

Get Jobilize Job Search Mobile App in your pocket Now!

Get it on Google Play Download on the App Store Now




Source:  OpenStax, Digital signal processing laboratory (ece 420 55x). OpenStax CNX. Jan 18, 2010 Download for free at http://cnx.org/content/col10397/1.10
Google Play and the Google Play logo are trademarks of Google Inc.

Notification Switch

Would you like to follow the 'Digital signal processing laboratory (ece 420 55x)' conversation and receive update notifications?

Ask