Home

This document is a cache from http://notes.ump.edu.my/fkee/e-Books/C%20Programming%20&%20PC%20interfacing/(ebook-pdf)%20DSP%20-%20Real%20Time%20Digital%20Signal%20Processing.pdf


Real-time digital signal processing: implementations, ... changes in the input signal is limited by its internal clock rate, so that it may be slow to

Document source : notes.ump.edu.my

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 All Pages

where nfft specifies the FFT length, Fs is the sampling frequency, window specifies
the selected window function, and noverlap is the number of samples by which the
segments overlap.
7.3.4 Fast Convolution
As discussed in Chapter 5, FIR filtering is a linear convolution of the finite impulse
response h(n) with the input sequence x(n). If the FIR filter has L coefficients, we need L
real multiplications and L À 1real additions to compute each output y(n). To obtain L
output samples, the number of operations (multiplication and addition) needed is
proportional to L
2
. To reduce the computational requirements, we consider the alter-
nate approach defined in (7.1.26) and (7.1.27), which uses the property that the con-
volution of the sequences x(n) and h(n) is equivalent to multiplying their DFTs X(k)
and H(k).
As described in Section 7.1.3, the linear convolution can be implemented using zero
padding if the data sequence x(n) also has finite duration. To take advantage of efficient
FFT and IFFT algorithms, we use these computational efficient algorithms as illus-
trated in Figure 7.13. The procedure of using FFT to implement linear convolution in a
computationally efficient manner is called the fast convolution. Compared to the direct
implementation of FIR filtering, fast convolution will provide a significant reduction in
computational requirements for higher order FIR filters, thus it is often used to imple-
ment FIR filtering in applications having long data samples.
It is important to note that the fast convolution shown in Figure 7.13 produces the
circular convolution discussed in Section 7.1.3. In order to produce a filter result equal
to a linear convolution, it is necessary to append zeros to the signals in order to
overcome the circular effect. If the data sequence x(n) has finite duration M, the first
step is to pad both sequences with zeros to a length corresponding to an allowable FFT
size N ! L M À 1, where L is the length of the sequence h(n). The FFT is computed
for both sequences, the complex products defined in (7.1.27) are calculated, and the
IFFT is used to obtain the results. The desired linear convolution is contained in the
first L M À 1terms of these results.
The FFT of the zero-padded data requires about N log
2
N complex computations.
Since the filter impulse response h(n) is known as a priori, the FFT of the impulse
response can be pre-calculated and stored. The computation of product
Yk HkXk takes N complex multiplications, and the inverse FFT of the pro-
ducts Y(k) requires another N log
2
N complex multiplications. Therefore the fast con-
volution shown in Figure 7.13 requires about (8N log
2
N 4N) real multiplications.
x(n)
h(n)
X(k)
H(k)
Y(k)
y(n)
FFT
FFT
IFFT
Figure 7.13 Fast convolution algorithm
330
FAST FOURIER TRANSFORM AND ITS APPLICATIONS

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 All Pages






Summary :

7.3.4 Fast Convolution As discussed in Chapter 5, FIR filtering is a linear convolution of the finite impulse response h(n) with the input sequence x(n). The procedure of using FFT to implement linear convolution in a computationally efficient manner is called the fast convolution. The computation of product Yk HkXk takes N complex multiplications, and the inverse FFT of the pro- ducts Y(k) requires another N log 2 N complex multiplications. x(n) h(n) X(k) H(k) Y(k) y(n) FFT FFT IFFT Figure 7.13 Fast convolution algorithm 330 FAST FOURIER TRANSFORM AND ITS APPLICATIONS


Tags : fast,713,linear,fir,sequence,multiplications,complex,figure,data,finite,efficient,response,length





Terms    |    Link pdf-search-files.com    |    Site Map
   |    Content Removal Notice   
   |    Contact   

All books are the property of their respective owners.
Please respect the publisher and the author for their creations if their books copyrighted