You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.

666 lines
30KB

  1. /*
  2. * Copyright (C) 2001-2003 Michael Niedermayer <michaelni@gmx.at>
  3. *
  4. * This file is part of FFmpeg.
  5. *
  6. * FFmpeg is free software; you can redistribute it and/or
  7. * modify it under the terms of the GNU Lesser General Public
  8. * License as published by the Free Software Foundation; either
  9. * version 2.1 of the License, or (at your option) any later version.
  10. *
  11. * FFmpeg is distributed in the hope that it will be useful,
  12. * but WITHOUT ANY WARRANTY; without even the implied warranty of
  13. * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
  14. * Lesser General Public License for more details.
  15. *
  16. * You should have received a copy of the GNU Lesser General Public
  17. * License along with FFmpeg; if not, write to the Free Software
  18. * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
  19. */
  20. #ifndef SWSCALE_SWSCALE_INTERNAL_H
  21. #define SWSCALE_SWSCALE_INTERNAL_H
  22. #include "config.h"
  23. #if HAVE_ALTIVEC_H
  24. #include <altivec.h>
  25. #endif
  26. #include "libavutil/avutil.h"
  27. #define STR(s) AV_TOSTRING(s) //AV_STRINGIFY is too long
  28. #define FAST_BGR2YV12 //use 7-bit instead of 15-bit coefficients
  29. #define MAX_FILTER_SIZE 256
  30. #define DITHER1XBPP
  31. #if HAVE_BIGENDIAN
  32. #define ALT32_CORR (-1)
  33. #else
  34. #define ALT32_CORR 1
  35. #endif
  36. #if ARCH_X86_64
  37. # define APCK_PTR2 8
  38. # define APCK_COEF 16
  39. # define APCK_SIZE 24
  40. #else
  41. # define APCK_PTR2 4
  42. # define APCK_COEF 8
  43. # define APCK_SIZE 16
  44. #endif
  45. struct SwsContext;
  46. typedef int (*SwsFunc)(struct SwsContext *context, const uint8_t* src[],
  47. int srcStride[], int srcSliceY, int srcSliceH,
  48. uint8_t* dst[], int dstStride[]);
  49. /**
  50. * Write one line of horizontally scaled Y/U/V/A to planar output
  51. * without any additional vertical scaling (or point-scaling).
  52. *
  53. * @param c SWS scaling context
  54. * @param lumSrc scaled luma (Y) source data, 15bit for 8bit output
  55. * @param chrUSrc scaled chroma (U) source data, 15bit for 8bit output
  56. * @param chrVSrc scaled chroma (V) source data, 15bit for 8bit output
  57. * @param alpSrc scaled alpha (A) source data, 15bit for 8bit output
  58. * @param dest pointer to the 4 output planes (Y/U/V/A)
  59. * @param dstW width of dest[0], dest[3], lumSrc and alpSrc in pixels
  60. * @param chrDstW width of dest[1], dest[2], chrUSrc and chrVSrc
  61. */
  62. typedef void (*yuv2planar1_fn) (struct SwsContext *c,
  63. const int16_t *lumSrc, const int16_t *chrUSrc,
  64. const int16_t *chrVSrc, const int16_t *alpSrc,
  65. uint8_t *dest[4], int dstW, int chrDstW,
  66. const uint8_t *lumDither, const uint8_t *chrDither);
  67. /**
  68. * Write one line of horizontally scaled Y/U/V/A to planar output
  69. * with multi-point vertical scaling between input pixels.
  70. *
  71. * @param c SWS scaling context
  72. * @param lumFilter vertical luma/alpha scaling coefficients, 12bit [0,4096]
  73. * @param lumSrc scaled luma (Y) source data, 15bit for 8bit output
  74. * @param lumFilterSize number of vertical luma/alpha input lines to scale
  75. * @param chrFilter vertical chroma scaling coefficients, 12bit [0,4096]
  76. * @param chrUSrc scaled chroma (U) source data, 15bit for 8bit output
  77. * @param chrVSrc scaled chroma (V) source data, 15bit for 8bit output
  78. * @param chrFilterSize number of vertical chroma input lines to scale
  79. * @param alpSrc scaled alpha (A) source data, 15bit for 8bit output
  80. * @param dest pointer to the 4 output planes (Y/U/V/A)
  81. * @param dstW width of dest[0], dest[3], lumSrc and alpSrc in pixels
  82. * @param chrDstW width of dest[1], dest[2], chrUSrc and chrVSrc
  83. */
  84. typedef void (*yuv2planarX_fn) (struct SwsContext *c, const int16_t *lumFilter,
  85. const int16_t **lumSrc, int lumFilterSize,
  86. const int16_t *chrFilter, const int16_t **chrUSrc,
  87. const int16_t **chrVSrc, int chrFilterSize,
  88. const int16_t **alpSrc, uint8_t *dest[4],
  89. int dstW, int chrDstW, const uint8_t *lumDither, const uint8_t *chrDither);
  90. /**
  91. * Write one line of horizontally scaled Y/U/V/A to packed-pixel YUV/RGB
  92. * output without any additional vertical scaling (or point-scaling). Note
  93. * that this function may do chroma scaling, see the "uvalpha" argument.
  94. *
  95. * @param c SWS scaling context
  96. * @param lumSrc scaled luma (Y) source data, 15bit for 8bit output
  97. * @param chrUSrc scaled chroma (U) source data, 15bit for 8bit output
  98. * @param chrVSrc scaled chroma (V) source data, 15bit for 8bit output
  99. * @param alpSrc scaled alpha (A) source data, 15bit for 8bit output
  100. * @param dest pointer to the output plane
  101. * @param dstW width of lumSrc and alpSrc in pixels, number of pixels
  102. * to write into dest[]
  103. * @param uvalpha chroma scaling coefficient for the second line of chroma
  104. * pixels, either 2048 or 0. If 0, one chroma input is used
  105. * for 2 output pixels (or if the SWS_FLAG_FULL_CHR_INT flag
  106. * is set, it generates 1 output pixel). If 2048, two chroma
  107. * input pixels should be averaged for 2 output pixels (this
  108. * only happens if SWS_FLAG_FULL_CHR_INT is not set)
  109. * @param y vertical line number for this output. This does not need
  110. * to be used to calculate the offset in the destination,
  111. * but can be used to generate comfort noise using dithering
  112. * for some output formats.
  113. */
  114. typedef void (*yuv2packed1_fn) (struct SwsContext *c, const int16_t *lumSrc,
  115. const int16_t *chrUSrc[2], const int16_t *chrVSrc[2],
  116. const int16_t *alpSrc, uint8_t *dest,
  117. int dstW, int uvalpha, int y);
  118. /**
  119. * Write one line of horizontally scaled Y/U/V/A to packed-pixel YUV/RGB
  120. * output by doing bilinear scaling between two input lines.
  121. *
  122. * @param c SWS scaling context
  123. * @param lumSrc scaled luma (Y) source data, 15bit for 8bit output
  124. * @param chrUSrc scaled chroma (U) source data, 15bit for 8bit output
  125. * @param chrVSrc scaled chroma (V) source data, 15bit for 8bit output
  126. * @param alpSrc scaled alpha (A) source data, 15bit for 8bit output
  127. * @param dest pointer to the output plane
  128. * @param dstW width of lumSrc and alpSrc in pixels, number of pixels
  129. * to write into dest[]
  130. * @param yalpha luma/alpha scaling coefficients for the second input line.
  131. * The first line's coefficients can be calculated by using
  132. * 4096 - yalpha
  133. * @param uvalpha chroma scaling coefficient for the second input line. The
  134. * first line's coefficients can be calculated by using
  135. * 4096 - uvalpha
  136. * @param y vertical line number for this output. This does not need
  137. * to be used to calculate the offset in the destination,
  138. * but can be used to generate comfort noise using dithering
  139. * for some output formats.
  140. */
  141. typedef void (*yuv2packed2_fn) (struct SwsContext *c, const int16_t *lumSrc[2],
  142. const int16_t *chrUSrc[2], const int16_t *chrVSrc[2],
  143. const int16_t *alpSrc[2], uint8_t *dest,
  144. int dstW, int yalpha, int uvalpha, int y);
  145. /**
  146. * Write one line of horizontally scaled Y/U/V/A to packed-pixel YUV/RGB
  147. * output by doing multi-point vertical scaling between input pixels.
  148. *
  149. * @param c SWS scaling context
  150. * @param lumFilter vertical luma/alpha scaling coefficients, 12bit [0,4096]
  151. * @param lumSrc scaled luma (Y) source data, 15bit for 8bit output
  152. * @param lumFilterSize number of vertical luma/alpha input lines to scale
  153. * @param chrFilter vertical chroma scaling coefficients, 12bit [0,4096]
  154. * @param chrUSrc scaled chroma (U) source data, 15bit for 8bit output
  155. * @param chrVSrc scaled chroma (V) source data, 15bit for 8bit output
  156. * @param chrFilterSize number of vertical chroma input lines to scale
  157. * @param alpSrc scaled alpha (A) source data, 15bit for 8bit output
  158. * @param dest pointer to the output plane
  159. * @param dstW width of lumSrc and alpSrc in pixels, number of pixels
  160. * to write into dest[]
  161. * @param y vertical line number for this output. This does not need
  162. * to be used to calculate the offset in the destination,
  163. * but can be used to generate comfort noise using dithering
  164. * or some output formats.
  165. */
  166. typedef void (*yuv2packedX_fn) (struct SwsContext *c, const int16_t *lumFilter,
  167. const int16_t **lumSrc, int lumFilterSize,
  168. const int16_t *chrFilter, const int16_t **chrUSrc,
  169. const int16_t **chrVSrc, int chrFilterSize,
  170. const int16_t **alpSrc, uint8_t *dest,
  171. int dstW, int y);
  172. /* This struct should be aligned on at least a 32-byte boundary. */
  173. typedef struct SwsContext {
  174. /**
  175. * info on struct for av_log
  176. */
  177. const AVClass *av_class;
  178. /**
  179. * Note that src, dst, srcStride, dstStride will be copied in the
  180. * sws_scale() wrapper so they can be freely modified here.
  181. */
  182. SwsFunc swScale;
  183. int srcW; ///< Width of source luma/alpha planes.
  184. int srcH; ///< Height of source luma/alpha planes.
  185. int dstH; ///< Height of destination luma/alpha planes.
  186. int chrSrcW; ///< Width of source chroma planes.
  187. int chrSrcH; ///< Height of source chroma planes.
  188. int chrDstW; ///< Width of destination chroma planes.
  189. int chrDstH; ///< Height of destination chroma planes.
  190. int lumXInc, chrXInc;
  191. int lumYInc, chrYInc;
  192. enum PixelFormat dstFormat; ///< Destination pixel format.
  193. enum PixelFormat srcFormat; ///< Source pixel format.
  194. int dstFormatBpp; ///< Number of bits per pixel of the destination pixel format.
  195. int srcFormatBpp; ///< Number of bits per pixel of the source pixel format.
  196. int chrSrcHSubSample; ///< Binary logarithm of horizontal subsampling factor between luma/alpha and chroma planes in source image.
  197. int chrSrcVSubSample; ///< Binary logarithm of vertical subsampling factor between luma/alpha and chroma planes in source image.
  198. int chrDstHSubSample; ///< Binary logarithm of horizontal subsampling factor between luma/alpha and chroma planes in destination image.
  199. int chrDstVSubSample; ///< Binary logarithm of vertical subsampling factor between luma/alpha and chroma planes in destination image.
  200. int vChrDrop; ///< Binary logarithm of extra vertical subsampling factor in source image chroma planes specified by user.
  201. int sliceDir; ///< Direction that slices are fed to the scaler (1 = top-to-bottom, -1 = bottom-to-top).
  202. double param[2]; ///< Input parameters for scaling algorithms that need them.
  203. uint32_t pal_yuv[256];
  204. uint32_t pal_rgb[256];
  205. /**
  206. * @name Scaled horizontal lines ring buffer.
  207. * The horizontal scaler keeps just enough scaled lines in a ring buffer
  208. * so they may be passed to the vertical scaler. The pointers to the
  209. * allocated buffers for each line are duplicated in sequence in the ring
  210. * buffer to simplify indexing and avoid wrapping around between lines
  211. * inside the vertical scaler code. The wrapping is done before the
  212. * vertical scaler is called.
  213. */
  214. //@{
  215. int16_t **lumPixBuf; ///< Ring buffer for scaled horizontal luma plane lines to be fed to the vertical scaler.
  216. int16_t **chrUPixBuf; ///< Ring buffer for scaled horizontal chroma plane lines to be fed to the vertical scaler.
  217. int16_t **chrVPixBuf; ///< Ring buffer for scaled horizontal chroma plane lines to be fed to the vertical scaler.
  218. int16_t **alpPixBuf; ///< Ring buffer for scaled horizontal alpha plane lines to be fed to the vertical scaler.
  219. int vLumBufSize; ///< Number of vertical luma/alpha lines allocated in the ring buffer.
  220. int vChrBufSize; ///< Number of vertical chroma lines allocated in the ring buffer.
  221. int lastInLumBuf; ///< Last scaled horizontal luma/alpha line from source in the ring buffer.
  222. int lastInChrBuf; ///< Last scaled horizontal chroma line from source in the ring buffer.
  223. int lumBufIndex; ///< Index in ring buffer of the last scaled horizontal luma/alpha line from source.
  224. int chrBufIndex; ///< Index in ring buffer of the last scaled horizontal chroma line from source.
  225. //@}
  226. uint8_t *formatConvBuffer;
  227. /**
  228. * @name Horizontal and vertical filters.
  229. * To better understand the following fields, here is a pseudo-code of
  230. * their usage in filtering a horizontal line:
  231. * @code
  232. * for (i = 0; i < width; i++) {
  233. * dst[i] = 0;
  234. * for (j = 0; j < filterSize; j++)
  235. * dst[i] += src[ filterPos[i] + j ] * filter[ filterSize * i + j ];
  236. * dst[i] >>= FRAC_BITS; // The actual implementation is fixed-point.
  237. * }
  238. * @endcode
  239. */
  240. //@{
  241. int16_t *hLumFilter; ///< Array of horizontal filter coefficients for luma/alpha planes.
  242. int16_t *hChrFilter; ///< Array of horizontal filter coefficients for chroma planes.
  243. int16_t *vLumFilter; ///< Array of vertical filter coefficients for luma/alpha planes.
  244. int16_t *vChrFilter; ///< Array of vertical filter coefficients for chroma planes.
  245. int16_t *hLumFilterPos; ///< Array of horizontal filter starting positions for each dst[i] for luma/alpha planes.
  246. int16_t *hChrFilterPos; ///< Array of horizontal filter starting positions for each dst[i] for chroma planes.
  247. int16_t *vLumFilterPos; ///< Array of vertical filter starting positions for each dst[i] for luma/alpha planes.
  248. int16_t *vChrFilterPos; ///< Array of vertical filter starting positions for each dst[i] for chroma planes.
  249. int hLumFilterSize; ///< Horizontal filter size for luma/alpha pixels.
  250. int hChrFilterSize; ///< Horizontal filter size for chroma pixels.
  251. int vLumFilterSize; ///< Vertical filter size for luma/alpha pixels.
  252. int vChrFilterSize; ///< Vertical filter size for chroma pixels.
  253. //@}
  254. int lumMmx2FilterCodeSize; ///< Runtime-generated MMX2 horizontal fast bilinear scaler code size for luma/alpha planes.
  255. int chrMmx2FilterCodeSize; ///< Runtime-generated MMX2 horizontal fast bilinear scaler code size for chroma planes.
  256. uint8_t *lumMmx2FilterCode; ///< Runtime-generated MMX2 horizontal fast bilinear scaler code for luma/alpha planes.
  257. uint8_t *chrMmx2FilterCode; ///< Runtime-generated MMX2 horizontal fast bilinear scaler code for chroma planes.
  258. int canMMX2BeUsed;
  259. int dstY; ///< Last destination vertical line output from last slice.
  260. int flags; ///< Flags passed by the user to select scaler algorithm, optimizations, subsampling, etc...
  261. void * yuvTable; // pointer to the yuv->rgb table start so it can be freed()
  262. uint8_t * table_rV[256];
  263. uint8_t * table_gU[256];
  264. int table_gV[256];
  265. uint8_t * table_bU[256];
  266. //Colorspace stuff
  267. int contrast, brightness, saturation; // for sws_getColorspaceDetails
  268. int srcColorspaceTable[4];
  269. int dstColorspaceTable[4];
  270. int srcRange; ///< 0 = MPG YUV range, 1 = JPG YUV range (source image).
  271. int dstRange; ///< 0 = MPG YUV range, 1 = JPG YUV range (destination image).
  272. int yuv2rgb_y_offset;
  273. int yuv2rgb_y_coeff;
  274. int yuv2rgb_v2r_coeff;
  275. int yuv2rgb_v2g_coeff;
  276. int yuv2rgb_u2g_coeff;
  277. int yuv2rgb_u2b_coeff;
  278. #define RED_DITHER "0*8"
  279. #define GREEN_DITHER "1*8"
  280. #define BLUE_DITHER "2*8"
  281. #define Y_COEFF "3*8"
  282. #define VR_COEFF "4*8"
  283. #define UB_COEFF "5*8"
  284. #define VG_COEFF "6*8"
  285. #define UG_COEFF "7*8"
  286. #define Y_OFFSET "8*8"
  287. #define U_OFFSET "9*8"
  288. #define V_OFFSET "10*8"
  289. #define LUM_MMX_FILTER_OFFSET "11*8"
  290. #define CHR_MMX_FILTER_OFFSET "11*8+4*4*256"
  291. #define DSTW_OFFSET "11*8+4*4*256*2" //do not change, it is hardcoded in the ASM
  292. #define ESP_OFFSET "11*8+4*4*256*2+8"
  293. #define VROUNDER_OFFSET "11*8+4*4*256*2+16"
  294. #define U_TEMP "11*8+4*4*256*2+24"
  295. #define V_TEMP "11*8+4*4*256*2+32"
  296. #define Y_TEMP "11*8+4*4*256*2+40"
  297. #define ALP_MMX_FILTER_OFFSET "11*8+4*4*256*2+48"
  298. #define UV_OFF "11*8+4*4*256*3+48"
  299. #define UV_OFFx2 "11*8+4*4*256*3+56"
  300. #define DITHER16 "11*8+4*4*256*3+64"
  301. #define DITHER32 "11*8+4*4*256*3+64+16"
  302. DECLARE_ALIGNED(8, uint64_t, redDither);
  303. DECLARE_ALIGNED(8, uint64_t, greenDither);
  304. DECLARE_ALIGNED(8, uint64_t, blueDither);
  305. DECLARE_ALIGNED(8, uint64_t, yCoeff);
  306. DECLARE_ALIGNED(8, uint64_t, vrCoeff);
  307. DECLARE_ALIGNED(8, uint64_t, ubCoeff);
  308. DECLARE_ALIGNED(8, uint64_t, vgCoeff);
  309. DECLARE_ALIGNED(8, uint64_t, ugCoeff);
  310. DECLARE_ALIGNED(8, uint64_t, yOffset);
  311. DECLARE_ALIGNED(8, uint64_t, uOffset);
  312. DECLARE_ALIGNED(8, uint64_t, vOffset);
  313. int32_t lumMmxFilter[4*MAX_FILTER_SIZE];
  314. int32_t chrMmxFilter[4*MAX_FILTER_SIZE];
  315. int dstW; ///< Width of destination luma/alpha planes.
  316. DECLARE_ALIGNED(8, uint64_t, esp);
  317. DECLARE_ALIGNED(8, uint64_t, vRounder);
  318. DECLARE_ALIGNED(8, uint64_t, u_temp);
  319. DECLARE_ALIGNED(8, uint64_t, v_temp);
  320. DECLARE_ALIGNED(8, uint64_t, y_temp);
  321. int32_t alpMmxFilter[4*MAX_FILTER_SIZE];
  322. DECLARE_ALIGNED(8, ptrdiff_t, uv_off); ///< offset (in pixels) between u and v planes
  323. DECLARE_ALIGNED(8, ptrdiff_t, uv_offx2); ///< offset (in bytes) between u and v planes
  324. uint16_t dither16[8];
  325. uint32_t dither32[8];
  326. #if HAVE_ALTIVEC
  327. vector signed short CY;
  328. vector signed short CRV;
  329. vector signed short CBU;
  330. vector signed short CGU;
  331. vector signed short CGV;
  332. vector signed short OY;
  333. vector unsigned short CSHIFT;
  334. vector signed short *vYCoeffsBank, *vCCoeffsBank;
  335. #endif
  336. #if ARCH_BFIN
  337. DECLARE_ALIGNED(4, uint32_t, oy);
  338. DECLARE_ALIGNED(4, uint32_t, oc);
  339. DECLARE_ALIGNED(4, uint32_t, zero);
  340. DECLARE_ALIGNED(4, uint32_t, cy);
  341. DECLARE_ALIGNED(4, uint32_t, crv);
  342. DECLARE_ALIGNED(4, uint32_t, rmask);
  343. DECLARE_ALIGNED(4, uint32_t, cbu);
  344. DECLARE_ALIGNED(4, uint32_t, bmask);
  345. DECLARE_ALIGNED(4, uint32_t, cgu);
  346. DECLARE_ALIGNED(4, uint32_t, cgv);
  347. DECLARE_ALIGNED(4, uint32_t, gmask);
  348. #endif
  349. #if HAVE_VIS
  350. DECLARE_ALIGNED(8, uint64_t, sparc_coeffs)[10];
  351. #endif
  352. /* function pointers for swScale() */
  353. yuv2planar1_fn yuv2yuv1;
  354. yuv2planarX_fn yuv2yuvX;
  355. yuv2packed1_fn yuv2packed1;
  356. yuv2packed2_fn yuv2packed2;
  357. yuv2packedX_fn yuv2packedX;
  358. void (*lumToYV12)(uint8_t *dst, const uint8_t *src,
  359. int width, uint32_t *pal); ///< Unscaled conversion of luma plane to YV12 for horizontal scaler.
  360. void (*alpToYV12)(uint8_t *dst, const uint8_t *src,
  361. int width, uint32_t *pal); ///< Unscaled conversion of alpha plane to YV12 for horizontal scaler.
  362. void (*chrToYV12)(uint8_t *dstU, uint8_t *dstV,
  363. const uint8_t *src1, const uint8_t *src2,
  364. int width, uint32_t *pal); ///< Unscaled conversion of chroma planes to YV12 for horizontal scaler.
  365. /**
  366. * Scale one horizontal line of input data using a bilinear filter
  367. * to produce one line of output data. Compared to SwsContext->hScale(),
  368. * please take note of the following caveats when using these:
  369. * - Scaling is done using only 7bit instead of 14bit coefficients.
  370. * - You can use no more than 5 input pixels to produce 4 output
  371. * pixels. Therefore, this filter should not be used for downscaling
  372. * by more than ~20% in width (because that equals more than 5/4th
  373. * downscaling and thus more than 5 pixels input per 4 pixels output).
  374. * - In general, bilinear filters create artifacts during downscaling
  375. * (even when <20%), because one output pixel will span more than one
  376. * input pixel, and thus some pixels will need edges of both neighbor
  377. * pixels to interpolate the output pixel. Since you can use at most
  378. * two input pixels per output pixel in bilinear scaling, this is
  379. * impossible and thus downscaling by any size will create artifacts.
  380. * To enable this type of scaling, set SWS_FLAG_FAST_BILINEAR
  381. * in SwsContext->flags.
  382. */
  383. /** @{ */
  384. void (*hyscale_fast)(struct SwsContext *c,
  385. int16_t *dst, int dstWidth,
  386. const uint8_t *src, int srcW, int xInc);
  387. void (*hcscale_fast)(struct SwsContext *c,
  388. int16_t *dst1, int16_t *dst2, int dstWidth,
  389. const uint8_t *src1, const uint8_t *src2,
  390. int srcW, int xInc);
  391. /** @} */
  392. /**
  393. * Scale one horizontal line of input data using a filter over the input
  394. * lines, to produce one (differently sized) line of output data.
  395. *
  396. * @param dst pointer to destination buffer for horizontally scaled
  397. * data. If the scaling depth (SwsContext->scalingBpp) is
  398. * 8, data will be 15bpp in 16bits (int16_t) width. If
  399. * scaling depth is 16, data will be 19bpp in 32bpp
  400. * (int32_t) width.
  401. * @param dstW width of destination image
  402. * @param src pointer to source data to be scaled. If scaling depth
  403. * is 8, this is 8bpp in 8bpp (uint8_t) width. If scaling
  404. * depth is 16, this is 16bpp in 16bpp (uint16_t) depth.
  405. * @param filter filter coefficients to be used per output pixel for
  406. * scaling. This contains 14bpp filtering coefficients.
  407. * Guaranteed to contain dstW * filterSize entries.
  408. * @param filterPos position of the first input pixel to be used for
  409. * each output pixel during scaling. Guaranteed to
  410. * contain dstW entries.
  411. * @param filterSize the number of input coefficients to be used (and
  412. * thus the number of input pixels to be used) for
  413. * creating a single output pixel. Is aligned to 4
  414. * (and input coefficients thus padded with zeroes)
  415. * to simplify creating SIMD code.
  416. */
  417. void (*hScale)(int16_t *dst, int dstW, const uint8_t *src,
  418. const int16_t *filter, const int16_t *filterPos,
  419. int filterSize);
  420. void (*hScale16)(int16_t *dst, int dstW, const uint16_t *src, int srcW,
  421. int xInc, const int16_t *filter, const int16_t *filterPos,
  422. long filterSize, int shift);
  423. void (*lumConvertRange)(int16_t *dst, int width); ///< Color range conversion function for luma plane if needed.
  424. void (*chrConvertRange)(int16_t *dst1, int16_t *dst2, int width); ///< Color range conversion function for chroma planes if needed.
  425. int needs_hcscale; ///< Set if there are chroma planes to be converted.
  426. } SwsContext;
  427. //FIXME check init (where 0)
  428. SwsFunc ff_yuv2rgb_get_func_ptr(SwsContext *c);
  429. int ff_yuv2rgb_c_init_tables(SwsContext *c, const int inv_table[4],
  430. int fullRange, int brightness,
  431. int contrast, int saturation);
  432. void ff_yuv2rgb_init_tables_altivec(SwsContext *c, const int inv_table[4],
  433. int brightness, int contrast, int saturation);
  434. void updateMMXDitherTables(SwsContext *c, int dstY, int lumBufIndex, int chrBufIndex,
  435. int lastInLumBuf, int lastInChrBuf);
  436. SwsFunc ff_yuv2rgb_init_mmx(SwsContext *c);
  437. SwsFunc ff_yuv2rgb_init_vis(SwsContext *c);
  438. SwsFunc ff_yuv2rgb_init_mlib(SwsContext *c);
  439. SwsFunc ff_yuv2rgb_init_altivec(SwsContext *c);
  440. SwsFunc ff_yuv2rgb_get_func_ptr_bfin(SwsContext *c);
  441. void ff_bfin_get_unscaled_swscale(SwsContext *c);
  442. #if FF_API_SWS_FORMAT_NAME
  443. /**
  444. * @deprecated Use av_get_pix_fmt_name() instead.
  445. */
  446. attribute_deprecated
  447. const char *sws_format_name(enum PixelFormat format);
  448. #endif
  449. //FIXME replace this with something faster
  450. #define is16BPS(x) ( \
  451. (x)==PIX_FMT_GRAY16BE \
  452. || (x)==PIX_FMT_GRAY16LE \
  453. || (x)==PIX_FMT_BGR48BE \
  454. || (x)==PIX_FMT_BGR48LE \
  455. || (x)==PIX_FMT_RGB48BE \
  456. || (x)==PIX_FMT_RGB48LE \
  457. || (x)==PIX_FMT_YUV420P16LE \
  458. || (x)==PIX_FMT_YUV422P16LE \
  459. || (x)==PIX_FMT_YUV444P16LE \
  460. || (x)==PIX_FMT_YUV420P16BE \
  461. || (x)==PIX_FMT_YUV422P16BE \
  462. || (x)==PIX_FMT_YUV444P16BE \
  463. )
  464. #define isNBPS(x) ( \
  465. (x)==PIX_FMT_YUV420P9LE \
  466. || (x)==PIX_FMT_YUV420P9BE \
  467. || (x)==PIX_FMT_YUV444P9BE \
  468. || (x)==PIX_FMT_YUV444P9LE \
  469. || (x)==PIX_FMT_YUV422P10BE \
  470. || (x)==PIX_FMT_YUV422P10LE \
  471. || (x)==PIX_FMT_YUV444P10BE \
  472. || (x)==PIX_FMT_YUV444P10LE \
  473. || (x)==PIX_FMT_YUV420P10LE \
  474. || (x)==PIX_FMT_YUV420P10BE \
  475. || (x)==PIX_FMT_YUV422P10LE \
  476. || (x)==PIX_FMT_YUV422P10BE \
  477. )
  478. #define is9_OR_10BPS isNBPS //for ronald
  479. #define isBE(x) ((x)&1)
  480. #define isPlanar8YUV(x) ( \
  481. (x)==PIX_FMT_YUV410P \
  482. || (x)==PIX_FMT_YUV420P \
  483. || (x)==PIX_FMT_YUVA420P \
  484. || (x)==PIX_FMT_YUV411P \
  485. || (x)==PIX_FMT_YUV422P \
  486. || (x)==PIX_FMT_YUV444P \
  487. || (x)==PIX_FMT_YUV440P \
  488. || (x)==PIX_FMT_NV12 \
  489. || (x)==PIX_FMT_NV21 \
  490. )
  491. #define isPlanarYUV(x) ( \
  492. isPlanar8YUV(x) \
  493. || (x)==PIX_FMT_YUV420P9LE \
  494. || (x)==PIX_FMT_YUV444P9LE \
  495. || (x)==PIX_FMT_YUV420P10LE \
  496. || (x)==PIX_FMT_YUV422P10LE \
  497. || (x)==PIX_FMT_YUV444P10LE \
  498. || (x)==PIX_FMT_YUV420P16LE \
  499. || (x)==PIX_FMT_YUV422P10LE \
  500. || (x)==PIX_FMT_YUV422P16LE \
  501. || (x)==PIX_FMT_YUV444P16LE \
  502. || (x)==PIX_FMT_YUV420P9BE \
  503. || (x)==PIX_FMT_YUV444P9BE \
  504. || (x)==PIX_FMT_YUV420P10BE \
  505. || (x)==PIX_FMT_YUV422P10BE \
  506. || (x)==PIX_FMT_YUV444P10BE \
  507. || (x)==PIX_FMT_YUV420P16BE \
  508. || (x)==PIX_FMT_YUV422P10BE \
  509. || (x)==PIX_FMT_YUV422P16BE \
  510. || (x)==PIX_FMT_YUV444P16BE \
  511. )
  512. #define isYUV(x) ( \
  513. (x)==PIX_FMT_UYVY422 \
  514. || (x)==PIX_FMT_YUYV422 \
  515. || isPlanarYUV(x) \
  516. )
  517. #define isGray(x) ( \
  518. (x)==PIX_FMT_GRAY8 \
  519. || (x)==PIX_FMT_GRAY8A \
  520. || (x)==PIX_FMT_GRAY16BE \
  521. || (x)==PIX_FMT_GRAY16LE \
  522. )
  523. #define isGray16(x) ( \
  524. (x)==PIX_FMT_GRAY16BE \
  525. || (x)==PIX_FMT_GRAY16LE \
  526. )
  527. #define isRGBinInt(x) ( \
  528. (x)==PIX_FMT_RGB48BE \
  529. || (x)==PIX_FMT_RGB48LE \
  530. || (x)==PIX_FMT_RGB32 \
  531. || (x)==PIX_FMT_RGB32_1 \
  532. || (x)==PIX_FMT_RGB24 \
  533. || (x)==PIX_FMT_RGB565BE \
  534. || (x)==PIX_FMT_RGB565LE \
  535. || (x)==PIX_FMT_RGB555BE \
  536. || (x)==PIX_FMT_RGB555LE \
  537. || (x)==PIX_FMT_RGB444BE \
  538. || (x)==PIX_FMT_RGB444LE \
  539. || (x)==PIX_FMT_RGB8 \
  540. || (x)==PIX_FMT_RGB4 \
  541. || (x)==PIX_FMT_RGB4_BYTE \
  542. || (x)==PIX_FMT_MONOBLACK \
  543. || (x)==PIX_FMT_MONOWHITE \
  544. )
  545. #define isBGRinInt(x) ( \
  546. (x)==PIX_FMT_BGR48BE \
  547. || (x)==PIX_FMT_BGR48LE \
  548. || (x)==PIX_FMT_BGR32 \
  549. || (x)==PIX_FMT_BGR32_1 \
  550. || (x)==PIX_FMT_BGR24 \
  551. || (x)==PIX_FMT_BGR565BE \
  552. || (x)==PIX_FMT_BGR565LE \
  553. || (x)==PIX_FMT_BGR555BE \
  554. || (x)==PIX_FMT_BGR555LE \
  555. || (x)==PIX_FMT_BGR444BE \
  556. || (x)==PIX_FMT_BGR444LE \
  557. || (x)==PIX_FMT_BGR8 \
  558. || (x)==PIX_FMT_BGR4 \
  559. || (x)==PIX_FMT_BGR4_BYTE \
  560. || (x)==PIX_FMT_MONOBLACK \
  561. || (x)==PIX_FMT_MONOWHITE \
  562. )
  563. #define isRGBinBytes(x) ( \
  564. (x)==PIX_FMT_RGB48BE \
  565. || (x)==PIX_FMT_RGB48LE \
  566. || (x)==PIX_FMT_RGBA \
  567. || (x)==PIX_FMT_ARGB \
  568. || (x)==PIX_FMT_RGB24 \
  569. )
  570. #define isBGRinBytes(x) ( \
  571. (x)==PIX_FMT_BGR48BE \
  572. || (x)==PIX_FMT_BGR48LE \
  573. || (x)==PIX_FMT_BGRA \
  574. || (x)==PIX_FMT_ABGR \
  575. || (x)==PIX_FMT_BGR24 \
  576. )
  577. #define isAnyRGB(x) ( \
  578. isRGBinInt(x) \
  579. || isBGRinInt(x) \
  580. )
  581. #define isALPHA(x) ( \
  582. (x)==PIX_FMT_BGR32 \
  583. || (x)==PIX_FMT_BGR32_1 \
  584. || (x)==PIX_FMT_RGB32 \
  585. || (x)==PIX_FMT_RGB32_1 \
  586. || (x)==PIX_FMT_PAL8 \
  587. || (x)==PIX_FMT_GRAY8A \
  588. || (x)==PIX_FMT_YUVA420P \
  589. )
  590. #define isPacked(x) ( \
  591. (x)==PIX_FMT_PAL8 \
  592. || (x)==PIX_FMT_YUYV422 \
  593. || (x)==PIX_FMT_UYVY422 \
  594. || (x)==PIX_FMT_Y400A \
  595. || isAnyRGB(x) \
  596. )
  597. #define usePal(x) ((av_pix_fmt_descriptors[x].flags & PIX_FMT_PAL) || (x) == PIX_FMT_GRAY8A)
  598. extern const uint64_t ff_dither4[2];
  599. extern const uint64_t ff_dither8[2];
  600. extern const uint8_t dithers[8][8][8];
  601. extern const uint16_t dither_scale[15][16];
  602. extern const AVClass sws_context_class;
  603. /**
  604. * Sets c->swScale to an unscaled converter if one exists for the specific
  605. * source and destination formats, bit depths, flags, etc.
  606. */
  607. void ff_get_unscaled_swscale(SwsContext *c);
  608. void ff_swscale_get_unscaled_altivec(SwsContext *c);
  609. /**
  610. * Returns function pointer to fastest main scaler path function depending
  611. * on architecture and available optimizations.
  612. */
  613. SwsFunc ff_getSwsFunc(SwsContext *c);
  614. void ff_sws_init_swScale_altivec(SwsContext *c);
  615. void ff_sws_init_swScale_mmx(SwsContext *c);
  616. #endif /* SWSCALE_SWSCALE_INTERNAL_H */