diff options
| author | Jens Arnold <amiconn@rockbox.org> | 2007-10-21 13:46:26 +0000 |
|---|---|---|
| committer | Jens Arnold <amiconn@rockbox.org> | 2007-10-21 13:46:26 +0000 |
| commit | c1051549b93d1905602ba25409cbd8f4a607c3b5 (patch) | |
| tree | 303c0129884c3c32c6ca1e425b8edd4c9caf3ebc /firmware/target/arm/sandisk/sansa-c200/lcd-as-c200.S | |
| parent | e1a91b9138148bd33ee0bae72a7367b6282aca7f (diff) | |
| download | rockbox-c1051549b93d1905602ba25409cbd8f4a607c3b5.tar.gz rockbox-c1051549b93d1905602ba25409cbd8f4a607c3b5.zip | |
Implement YUV dithering for c200, and enable the option in mpegplayer.
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@15246 a1c6a512-1295-4272-9138-f99709370657
Diffstat (limited to 'firmware/target/arm/sandisk/sansa-c200/lcd-as-c200.S')
| -rw-r--r-- | firmware/target/arm/sandisk/sansa-c200/lcd-as-c200.S | 306 |
1 files changed, 306 insertions, 0 deletions
diff --git a/firmware/target/arm/sandisk/sansa-c200/lcd-as-c200.S b/firmware/target/arm/sandisk/sansa-c200/lcd-as-c200.S index 9b0cd75bae..aaeb8439b6 100644 --- a/firmware/target/arm/sandisk/sansa-c200/lcd-as-c200.S +++ b/firmware/target/arm/sandisk/sansa-c200/lcd-as-c200.S | |||
| @@ -241,3 +241,309 @@ lcd_write_yuv420_lines: | |||
| 241 | .ltorg @ dump constant pool | 241 | .ltorg @ dump constant pool |
| 242 | .size lcd_write_yuv420_lines, .-lcd_write_yuv420_lines | 242 | .size lcd_write_yuv420_lines, .-lcd_write_yuv420_lines |
| 243 | 243 | ||
| 244 | /**************************************************************************** | ||
| 245 | * void lcd_write_yuv_420_lines_odither(unsigned char const * const src[3], | ||
| 246 | * int width, | ||
| 247 | * int stride, | ||
| 248 | * int x_screen, | ||
| 249 | * int y_screen); | ||
| 250 | * | ||
| 251 | * |R| |1.000000 -0.000001 1.402000| |Y'| | ||
| 252 | * |G| = |1.000000 -0.334136 -0.714136| |Pb| | ||
| 253 | * |B| |1.000000 1.772000 0.000000| |Pr| | ||
| 254 | * Red scaled at twice g & b but at same precision to place it in correct | ||
| 255 | * bit position after multiply and leave instruction count lower. | ||
| 256 | * |R| |258 0 408| |Y' - 16| | ||
| 257 | * |G| = |149 -49 -104| |Cb - 128| | ||
| 258 | * |B| |149 258 0| |Cr - 128| | ||
| 259 | * | ||
| 260 | * Write four RGB565 pixels in the following order on each loop: | ||
| 261 | * 1 3 + > down | ||
| 262 | * 2 4 \/ left | ||
| 263 | * | ||
| 264 | * Kernel pattern (raw|rotated|use order): | ||
| 265 | * 5 3 4 2 2 6 3 7 row0 row2 > down | ||
| 266 | * 1 7 0 6 | 4 0 5 1 | 2 4 6 0 3 5 7 1 col0 left | ||
| 267 | * 4 2 5 3 | 3 7 2 6 | 3 5 7 1 2 4 6 0 col2 \/ | ||
| 268 | * 0 6 1 7 5 1 4 0 | ||
| 269 | */ | ||
| 270 | .section .icode, "ax", %progbits | ||
| 271 | .align 2 | ||
| 272 | .global lcd_write_yuv420_lines_odither | ||
| 273 | .type lcd_write_yuv420_lines_odither, %function | ||
| 274 | lcd_write_yuv420_lines_odither: | ||
| 275 | @ r0 = yuv_src | ||
| 276 | @ r1 = width | ||
| 277 | @ r2 = stride | ||
| 278 | @ r3 = x_screen | ||
| 279 | @ [sp] = y_screen | ||
| 280 | stmfd sp!, { r4-r12, lr } @ save non-scratch | ||
| 281 | ldmia r0, { r4, r5, r6 } @ r4 = yuv_src[0] = Y'_p | ||
| 282 | @ r5 = yuv_src[1] = Cb_p | ||
| 283 | @ r6 = yuv_src[2] = Cr_p | ||
| 284 | @ | ||
| 285 | sub r2, r2, #1 @ | ||
| 286 | ldr r14, [sp, #40] @ Line up pattern and kernel quadrant | ||
| 287 | eor r14, r14, r3 @ | ||
| 288 | and r14, r14, #0x2 @ | ||
| 289 | mov r14, r14, lsl #6 @ 0x00 or 0x80 | ||
| 290 | mov r3, #0x70000000 @ | ||
| 291 | orr r3, r3, #0x3000 @ r3 = LCD1_BASE | ||
| 292 | 10: @ loop line @ | ||
| 293 | @ | ||
| 294 | ldrb r7, [r4], #1 @ r7 = *Y'_p++; | ||
| 295 | ldrb r8, [r5], #1 @ r8 = *Cb_p++; | ||
| 296 | ldrb r9, [r6], #1 @ r9 = *Cr_p++; | ||
| 297 | @ | ||
| 298 | eor r14, r14, #0x80 @ flip pattern quadrant | ||
| 299 | @ | ||
| 300 | sub r7, r7, #16 @ r7 = Y = (Y' - 16)*149 | ||
| 301 | add r12, r7, r7, asl #2 @ | ||
| 302 | add r12, r12, r12, asl #4 @ | ||
| 303 | add r7, r12, r7, asl #6 @ | ||
| 304 | @ | ||
| 305 | sub r8, r8, #128 @ Cb -= 128 | ||
| 306 | sub r9, r9, #128 @ Cr -= 128 | ||
| 307 | @ | ||
| 308 | add r10, r8, r8, asl #4 @ r10 = guv = Cr*104 + Cb*49 | ||
| 309 | add r10, r10, r8, asl #5 @ | ||
| 310 | add r10, r10, r9, asl #3 @ | ||
| 311 | add r10, r10, r9, asl #5 @ | ||
| 312 | add r10, r10, r9, asl #6 @ | ||
| 313 | @ | ||
| 314 | mov r8, r8, asl #1 @ r8 = bu = Cb*258 | ||
| 315 | add r8, r8, r8, asl #7 @ | ||
| 316 | @ | ||
| 317 | add r9, r9, r9, asl #1 @ r9 = rv = Cr*408 | ||
| 318 | add r9, r9, r9, asl #4 @ | ||
| 319 | mov r9, r9, asl #3 @ | ||
| 320 | @ | ||
| 321 | @ compute R, G, and B | ||
| 322 | add r0, r8, r7 @ r0 = b' = Y + bu | ||
| 323 | add r11, r9, r7, asl #1 @ r11 = r' = Y*2 + rv | ||
| 324 | rsb r7, r10, r7 @ r7 = g' = Y + guv | ||
| 325 | @ | ||
| 326 | @ r8 = bu, r9 = rv, r10 = guv | ||
| 327 | @ | ||
| 328 | sub r12, r0, r0, lsr #5 @ r0 = 31/32*b + b/256 | ||
| 329 | add r0, r12, r0, lsr #8 @ | ||
| 330 | @ | ||
| 331 | sub r12, r11, r11, lsr #5 @ r11 = 31/32*r + r/256 | ||
| 332 | add r11, r12, r11, lsr #8 @ | ||
| 333 | @ | ||
| 334 | sub r12, r7, r7, lsr #6 @ r7 = 63/64*g + g/256 | ||
| 335 | add r7, r12, r7, lsr #8 @ | ||
| 336 | @ | ||
| 337 | add r12, r14, #0x100 @ | ||
| 338 | @ | ||
| 339 | add r0, r0, r12 @ b = r0 + delta | ||
| 340 | add r11, r11, r12, lsl #1 @ r = r11 + delta*2 | ||
| 341 | add r7, r7, r12, lsr #1 @ g = r7 + delta/2 | ||
| 342 | @ | ||
| 343 | orr r12, r0, r11, asr #1 @ check if clamping is needed... | ||
| 344 | orr r12, r12, r7 @ ...at all | ||
| 345 | movs r12, r12, asr #15 @ | ||
| 346 | beq 15f @ no clamp @ | ||
| 347 | movs r12, r0, asr #15 @ clamp b | ||
| 348 | mvnne r0, r12, lsr #15 @ | ||
| 349 | andne r0, r0, #0x7c00 @ mask b only if clamped | ||
| 350 | movs r12, r11, asr #16 @ clamp r | ||
| 351 | mvnne r11, r12, lsr #16 @ | ||
| 352 | movs r12, r7, asr #15 @ clamp g | ||
| 353 | mvnne r7, r12, lsr #15 @ | ||
| 354 | 15: @ no clamp @ | ||
| 355 | @ | ||
| 356 | ldrb r12, [r4, r2] @ r12 = Y' = *(Y'_p + stride) | ||
| 357 | @ | ||
| 358 | |||
| 359 | and r11, r11, #0xf800 @ pack pixel | ||
| 360 | mov r11, r11, lsr #8 | ||
| 361 | and r7, r7, #0x7e00 | ||
| 362 | orr r11, r11, r7, lsr #12 | ||
| 363 | mov r7, r7, lsr#4 | ||
| 364 | orr r0, r7, r0, lsr #10 | ||
| 365 | 1: @ busy @ | ||
| 366 | ldr r7, [r3] @ r7 = LCD1_BASE | ||
| 367 | tst r7, #LCD1_BUSY_MASK @ bridge busy? | ||
| 368 | bne 1b @ | ||
| 369 | str r11, [r3, #0x10] @ send MSB | ||
| 370 | 1: @ busy @ | ||
| 371 | ldr r7, [r3] @ r7 = LCD1_BASE | ||
| 372 | tst r7, #LCD1_BUSY_MASK @ bridge busy? | ||
| 373 | bne 1b @ | ||
| 374 | str r0, [r3, #0x10] @ send LSB | ||
| 375 | @ | ||
| 376 | sub r7, r12, #16 @ r7 = Y = (Y' - 16)*149 | ||
| 377 | add r12, r7, r7, asl #2 @ | ||
| 378 | add r12, r12, r12, asl #4 @ | ||
| 379 | add r7, r12, r7, asl #6 @ | ||
| 380 | @ compute R, G, and B | ||
| 381 | add r0, r8, r7 @ r0 = b' = Y + bu | ||
| 382 | add r11, r9, r7, asl #1 @ r11 = r' = Y*2 + rv | ||
| 383 | rsb r7, r10, r7 @ r7 = g' = Y + guv | ||
| 384 | @ | ||
| 385 | sub r12, r0, r0, lsr #5 @ r0 = 31/32*b' + b'/256 | ||
| 386 | add r0, r12, r0, lsr #8 @ | ||
| 387 | @ | ||
| 388 | sub r12, r11, r11, lsr #5 @ r11 = 31/32*r' + r'/256 | ||
| 389 | add r11, r12, r11, lsr #8 @ | ||
| 390 | @ | ||
| 391 | sub r12, r7, r7, lsr #6 @ r7 = 63/64*g' + g'/256 | ||
| 392 | add r7, r12, r7, lsr #8 @ | ||
| 393 | @ | ||
| 394 | add r12, r14, #0x200 @ | ||
| 395 | @ | ||
| 396 | add r0, r0, r12 @ b = r0 + delta | ||
| 397 | add r11, r11, r12, lsl #1 @ r = r11 + delta*2 | ||
| 398 | add r7, r7, r12, lsr #1 @ g = r7 + delta/2 | ||
| 399 | @ | ||
| 400 | orr r12, r0, r11, asr #1 @ check if clamping is needed... | ||
| 401 | orr r12, r12, r7 @ ...at all | ||
| 402 | movs r12, r12, asr #15 @ | ||
| 403 | beq 15f @ no clamp @ | ||
| 404 | movs r12, r0, asr #15 @ clamp b | ||
| 405 | mvnne r0, r12, lsr #15 @ | ||
| 406 | andne r0, r0, #0x7c00 @ mask b only if clamped | ||
| 407 | movs r12, r11, asr #16 @ clamp r | ||
| 408 | mvnne r11, r12, lsr #16 @ | ||
| 409 | movs r12, r7, asr #15 @ clamp g | ||
| 410 | mvnne r7, r12, lsr #15 @ | ||
| 411 | 15: @ no clamp @ | ||
| 412 | @ | ||
| 413 | ldrb r12, [r4], #1 @ r12 = Y' = *(Y'_p++) | ||
| 414 | |||
| 415 | and r11, r11, #0xf800 @ pack pixel | ||
| 416 | mov r11, r11, lsr #8 | ||
| 417 | and r7, r7, #0x7e00 | ||
| 418 | orr r11, r11, r7, lsr #12 | ||
| 419 | mov r7, r7, lsr#4 | ||
| 420 | orr r0, r7, r0, lsr #10 | ||
| 421 | 1: @ busy @ | ||
| 422 | ldr r7, [r3] @ r7 = LCD1_BASE | ||
| 423 | tst r7, #LCD1_BUSY_MASK @ bridge busy? | ||
| 424 | bne 1b @ | ||
| 425 | str r11, [r3, #0x10] @ send MSB | ||
| 426 | 1: @ busy @ | ||
| 427 | ldr r7, [r3] @ r7 = LCD1_BASE | ||
| 428 | tst r7, #LCD1_BUSY_MASK @ bridge busy? | ||
| 429 | bne 1b @ | ||
| 430 | str r0, [r3, #0x10] @ send LSB | ||
| 431 | |||
| 432 | sub r7, r12, #16 @ r7 = Y = (Y' - 16)*149 | ||
| 433 | add r12, r7, r7, asl #2 @ | ||
| 434 | add r12, r12, r12, asl #4 @ | ||
| 435 | add r7, r12, r7, asl #6 @ | ||
| 436 | @ compute R, G, and B | ||
| 437 | add r0, r8, r7 @ r0 = b' = Y + bu | ||
| 438 | add r11, r9, r7, asl #1 @ r11 = r' = Y*2 + rv | ||
| 439 | rsb r7, r10, r7 @ r7 = g' = Y + guv | ||
| 440 | @ | ||
| 441 | @ r8 = bu, r9 = rv, r10 = guv | ||
| 442 | @ | ||
| 443 | sub r12, r0, r0, lsr #5 @ r0 = 31/32*b' + b'/256 | ||
| 444 | add r0, r12, r0, lsr #8 @ | ||
| 445 | @ | ||
| 446 | sub r12, r11, r11, lsr #5 @ r11 = 31/32*r' + r'/256 | ||
| 447 | add r11, r12, r11, lsr #8 @ | ||
| 448 | @ | ||
| 449 | sub r12, r7, r7, lsr #6 @ r7 = 63/64*g' + g'/256 | ||
| 450 | add r7, r12, r7, lsr #8 @ | ||
| 451 | @ | ||
| 452 | add r12, r14, #0x300 @ | ||
| 453 | @ | ||
| 454 | add r0, r0, r12 @ b = r0 + delta | ||
| 455 | add r11, r11, r12, lsl #1 @ r = r11 + delta*2 | ||
| 456 | add r7, r7, r12, lsr #1 @ g = r7 + delta/2 | ||
| 457 | @ | ||
| 458 | orr r12, r0, r11, asr #1 @ check if clamping is needed... | ||
| 459 | orr r12, r12, r7 @ ...at all | ||
| 460 | movs r12, r12, asr #15 @ | ||
| 461 | beq 15f @ no clamp @ | ||
| 462 | movs r12, r0, asr #15 @ clamp b | ||
| 463 | mvnne r0, r12, lsr #15 @ | ||
| 464 | andne r0, r0, #0x7c00 @ mask b only if clamped | ||
| 465 | movs r12, r11, asr #16 @ clamp r | ||
| 466 | mvnne r11, r12, lsr #16 @ | ||
| 467 | movs r12, r7, asr #15 @ clamp g | ||
| 468 | mvnne r7, r12, lsr #15 @ | ||
| 469 | 15: @ no clamp @ | ||
| 470 | @ | ||
| 471 | ldrb r12, [r4, r2] @ r12 = Y' = *(Y'_p + stride) | ||
| 472 | |||
| 473 | and r11, r11, #0xf800 @ pack pixel | ||
| 474 | mov r11, r11, lsr #8 | ||
| 475 | and r7, r7, #0x7e00 | ||
| 476 | orr r11, r11, r7, lsr #12 | ||
| 477 | mov r7, r7, lsr#4 | ||
| 478 | orr r0, r7, r0, lsr #10 | ||
| 479 | 1: @ busy @ | ||
| 480 | ldr r7, [r3] @ r7 = LCD1_BASE | ||
| 481 | tst r7, #LCD1_BUSY_MASK @ bridge busy? | ||
| 482 | bne 1b @ | ||
| 483 | str r11, [r3, #0x10] @ send MSB | ||
| 484 | 1: @ busy @ | ||
| 485 | ldr r7, [r3] @ r7 = LCD1_BASE | ||
| 486 | tst r7, #LCD1_BUSY_MASK @ bridge busy? | ||
| 487 | bne 1b @ | ||
| 488 | str r0, [r3, #0x10] @ send LSB | ||
| 489 | |||
| 490 | sub r7, r12, #16 @ r7 = Y = (Y' - 16)*149 | ||
| 491 | add r12, r7, r7, asl #2 @ | ||
| 492 | add r12, r12, r12, asl #4 @ | ||
| 493 | add r7, r12, r7, asl #6 @ | ||
| 494 | @ compute R, G, and B | ||
| 495 | add r0, r8, r7 @ r0 = b' = Y + bu | ||
| 496 | add r11, r9, r7, asl #1 @ r11 = r' = Y*2 + rv | ||
| 497 | rsb r7, r10, r7 @ r7 = g' = Y + guv | ||
| 498 | @ | ||
| 499 | sub r12, r0, r0, lsr #5 @ r0 = 31/32*b + b/256 | ||
| 500 | add r0, r12, r0, lsr #8 @ | ||
| 501 | @ | ||
| 502 | sub r12, r11, r11, lsr #5 @ r11 = 31/32*r + r/256 | ||
| 503 | add r11, r12, r11, lsr #8 @ | ||
| 504 | @ | ||
| 505 | sub r12, r7, r7, lsr #6 @ r7 = 63/64*g + g/256 | ||
| 506 | add r7, r12, r7, lsr #8 @ | ||
| 507 | @ | ||
| 508 | @ This element is zero - use r14 @ | ||
| 509 | @ | ||
| 510 | add r0, r0, r14 @ b = r0 + delta | ||
| 511 | add r11, r11, r14, lsl #1 @ r = r11 + delta*2 | ||
| 512 | add r7, r7, r14, lsr #1 @ g = r7 + delta/2 | ||
| 513 | @ | ||
| 514 | orr r12, r0, r11, asr #1 @ check if clamping is needed... | ||
| 515 | orr r12, r12, r7 @ ...at all | ||
| 516 | movs r12, r12, asr #15 @ | ||
| 517 | beq 15f @ no clamp @ | ||
| 518 | movs r12, r0, asr #15 @ clamp b | ||
| 519 | mvnne r0, r12, lsr #15 @ | ||
| 520 | andne r0, r0, #0x7c00 @ mask b only if clamped | ||
| 521 | movs r12, r11, asr #16 @ clamp r | ||
| 522 | mvnne r11, r12, lsr #16 @ | ||
| 523 | movs r12, r7, asr #15 @ clamp g | ||
| 524 | mvnne r7, r12, lsr #15 @ | ||
| 525 | 15: @ no clamp @ | ||
| 526 | |||
| 527 | and r11, r11, #0xf800 @ pack pixel | ||
| 528 | mov r11, r11, lsr #8 | ||
| 529 | and r7, r7, #0x7e00 | ||
| 530 | orr r11, r11, r7, lsr #12 | ||
| 531 | mov r7, r7, lsr#4 | ||
| 532 | orr r0, r7, r0, lsr #10 | ||
| 533 | 1: @ busy @ | ||
| 534 | ldr r7, [r3] @ r7 = LCD1_BASE | ||
| 535 | tst r7, #LCD1_BUSY_MASK @ bridge busy? | ||
| 536 | bne 1b @ | ||
| 537 | str r11, [r3, #0x10] @ send MSB | ||
| 538 | 1: @ busy @ | ||
| 539 | ldr r7, [r3] @ r7 = LCD1_BASE | ||
| 540 | tst r7, #LCD1_BUSY_MASK @ bridge busy? | ||
| 541 | bne 1b @ | ||
| 542 | str r0, [r3, #0x10] @ send LSB | ||
| 543 | |||
| 544 | subs r1, r1, #2 @ subtract block from width | ||
| 545 | bgt 10b @ loop line @ | ||
| 546 | @ | ||
| 547 | ldmfd sp!, { r4-r12, pc } @ restore registers and return | ||
| 548 | .ltorg @ dump constant pool | ||
| 549 | .size lcd_write_yuv420_lines_odither, .-lcd_write_yuv420_lines_odither | ||