gstvideobox.c 91.8 KB
Newer Older
1
/* GStreamer
2 3 4
 * Copyright (C) 1999 Erik Walthinsen <omega@cse.ogi.edu>
 * Copyright (C) 2006 Tim-Philipp Müller <tim centricular net>
 * Copyright (C) 2010 Sebastian Dröge <sebastian.droege@collabora.co.uk>
5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
 *
 * This library is free software; you can redistribute it and/or
 * modify it under the terms of the GNU Library General Public
 * License as published by the Free Software Foundation; either
 * version 2 of the License, or (at your option) any later version.
 *
 * This library is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
 * Library General Public License for more details.
 *
 * You should have received a copy of the GNU Library General Public
 * License along with this library; if not, write to the
 * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
 * Boston, MA 02111-1307, USA.
 */
21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38
/**
 * SECTION:element-videobox
 * @see_also: #GstVideoCrop
 *
 * This plugin crops or enlarges the image. It takes 4 values as input, a
 * top, bottom, left and right offset. Positive values will crop that much
 * pixels from the respective border of the image, negative values will add
 * that much pixels. When pixels are added, you can specify their color. 
 * Some predefined colors are usable with an enum property.
 * 
 * The plugin is alpha channel aware and will try to negotiate with a format
 * that supports alpha channels first. When alpha channel is active two
 * other properties, alpha and border_alpha can be used to set the alpha
 * values of the inner picture and the border respectively. an alpha value of
 * 0.0 means total transparency, 1.0 is opaque.
 * 
 * The videobox plugin has many uses such as doing a mosaic of pictures, 
 * letterboxing video, cutting out pieces of video, picture in picture, etc..
39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54
 *
 * Setting autocrop to true changes the behavior of the plugin so that
 * caps determine crop properties rather than the other way around: given
 * input and output dimensions, the crop values are selected so that the
 * smaller frame is effectively centered in the larger frame.  This
 * involves either cropping or padding.
 * 
 * If you use autocrop there is little point in setting the other
 * properties manually because they will be overriden if the caps change,
 * but nothing stops you from doing so.
 * 
 * Sample pipeline:
 * |[
 * gst-launch videotestsrc ! videobox autocrop=true ! \
 *   "video/x-raw-yuv, width=600, height=400" ! ffmpegcolorspace ! ximagesink
 * ]|
55
 */
56

57 58 59
#ifdef HAVE_CONFIG_H
#include "config.h"
#endif
60 61 62

#include "gstvideobox.h"

63
#include <math.h>
64
#include <liboil/liboil.h>
65 66
#include <string.h>

67 68
#include <gst/controller/gstcontroller.h>

69
GST_DEBUG_CATEGORY_STATIC (videobox_debug);
70 71
#define GST_CAT_DEFAULT videobox_debug

72
/* From videotestsrc.c */
73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92
static const guint8 yuv_sdtv_colors_Y[VIDEO_BOX_FILL_LAST] =
    { 16, 145, 41, 81, 210, 235 };
static const guint8 yuv_sdtv_colors_U[VIDEO_BOX_FILL_LAST] =
    { 128, 54, 240, 90, 16, 128 };
static const guint8 yuv_sdtv_colors_V[VIDEO_BOX_FILL_LAST] =
    { 128, 34, 110, 240, 146, 128 };

static const guint8 yuv_hdtv_colors_Y[VIDEO_BOX_FILL_LAST] =
    { 16, 173, 32, 63, 219, 235 };
static const guint8 yuv_hdtv_colors_U[VIDEO_BOX_FILL_LAST] =
    { 128, 42, 240, 102, 16, 128 };
static const guint8 yuv_hdtv_colors_V[VIDEO_BOX_FILL_LAST] =
    { 128, 26, 118, 240, 138, 128 };

static const guint8 rgb_colors_R[VIDEO_BOX_FILL_LAST] =
    { 0, 0, 0, 255, 255, 255 };
static const guint8 rgb_colors_G[VIDEO_BOX_FILL_LAST] =
    { 0, 255, 0, 0, 255, 255 };
static const guint8 rgb_colors_B[VIDEO_BOX_FILL_LAST] =
    { 0, 0, 255, 0, 0, 255 };
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

/* Generated by -bad/ext/cog/generate_tables */
static const int cog_ycbcr_to_rgb_matrix_8bit_hdtv[] = {
  298, 0, 459, -63514,
  298, -55, -136, 19681,
  298, 541, 0, -73988,
};

static const int cog_ycbcr_to_rgb_matrix_8bit_sdtv[] = {
  298, 0, 409, -57068,
  298, -100, -208, 34707,
  298, 516, 0, -70870,
};

static const gint cog_rgb_to_ycbcr_matrix_8bit_hdtv[] = {
  47, 157, 16, 4096,
  -26, -87, 112, 32768,
  112, -102, -10, 32768,
};

static const gint cog_rgb_to_ycbcr_matrix_8bit_sdtv[] = {
  66, 129, 25, 4096,
  -38, -74, 112, 32768,
  112, -94, -18, 32768,
};

static const gint cog_ycbcr_sdtv_to_ycbcr_hdtv_matrix_8bit[] = {
  256, -30, -53, 10600,
  0, 261, 29, -4367,
  0, 19, 262, -3289,
};

static const gint cog_ycbcr_hdtv_to_ycbcr_sdtv_matrix_8bit[] = {
  256, 25, 49, -9536,
  0, 253, -28, 3958,
  0, -19, 252, 2918,
};

131 132 133 134 135 136
static const gint cog_identity_matrix_8bit[] = {
  256, 0, 0, 0,
  0, 256, 0, 0,
  0, 0, 256, 0,
};

137 138 139
#define APPLY_MATRIX(m,o,v1,v2,v3) ((m[o*4] * v1 + m[o*4+1] * v2 + m[o*4+2] * v3 + m[o*4+3]) >> 8)

static void
140 141
fill_ayuv (GstVideoBoxFill fill_type, guint b_alpha, GstVideoFormat format,
    guint8 * dest, gboolean sdtv, gint width, gint height)
142 143 144
{
  guint32 empty_pixel;

145 146
  b_alpha = CLAMP (b_alpha, 0, 255);

147 148 149 150 151 152 153 154 155
  if (sdtv)
    empty_pixel = GUINT32_FROM_BE ((b_alpha << 24) |
        (yuv_sdtv_colors_Y[fill_type] << 16) |
        (yuv_sdtv_colors_U[fill_type] << 8) | yuv_sdtv_colors_V[fill_type]);
  else
    empty_pixel = GUINT32_FROM_BE ((b_alpha << 24) |
        (yuv_hdtv_colors_Y[fill_type] << 16) |
        (yuv_hdtv_colors_U[fill_type] << 8) | yuv_hdtv_colors_V[fill_type]);

156
  oil_splat_u32_ns ((guint32 *) dest, &empty_pixel, width * height);
157 158 159
}

static void
160 161 162 163 164
copy_ayuv_ayuv (guint i_alpha, GstVideoFormat dest_format, guint8 * dest,
    gboolean dest_sdtv, gint dest_width, gint dest_height, gint dest_x,
    gint dest_y, GstVideoFormat src_format, const guint8 * src,
    gboolean src_sdtv, gint src_width, gint src_height, gint src_x, gint src_y,
    gint w, gint h)
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
{
  gint i, j;
  gint src_stride = 4 * src_width;
  gint dest_stride = 4 * dest_width;

  dest = dest + dest_y * dest_width * 4 + dest_x * 4;
  src = src + src_y * src_width * 4 + src_x * 4;

  w *= 4;

  if (dest_sdtv != src_sdtv) {
    gint matrix[12];
    gint y, u, v;

    memcpy (matrix,
        dest_sdtv ? cog_ycbcr_hdtv_to_ycbcr_sdtv_matrix_8bit :
        cog_ycbcr_sdtv_to_ycbcr_hdtv_matrix_8bit, 12 * sizeof (gint));

    for (i = 0; i < h; i++) {
      for (j = 0; j < w; j += 4) {
        dest[j] = (src[j] * i_alpha) >> 8;
        y = src[j + 1];
        u = src[j + 2];
        v = src[j + 3];
        dest[j + 1] = APPLY_MATRIX (matrix, 0, y, u, v);
        dest[j + 2] = APPLY_MATRIX (matrix, 1, y, u, v);
        dest[j + 3] = APPLY_MATRIX (matrix, 2, y, u, v);
      }
      dest += dest_stride;
      src += src_stride;
    }
  } else {
    for (i = 0; i < h; i++) {
      for (j = 0; j < w; j += 4) {
        dest[j] = (src[j] * i_alpha) >> 8;
        dest[j + 1] = src[j + 1];
        dest[j + 2] = src[j + 2];
        dest[j + 3] = src[j + 3];
      }
      dest += dest_stride;
      src += src_stride;
    }
  }
}

static void
211 212 213 214 215
copy_ayuv_i420 (guint i_alpha, GstVideoFormat dest_format, guint8 * dest,
    gboolean dest_sdtv, gint dest_width, gint dest_height, gint dest_x,
    gint dest_y, GstVideoFormat src_format, const guint8 * src,
    gboolean src_sdtv, gint src_width, gint src_height, gint src_x, gint src_y,
    gint w, gint h)
216 217
{
  gint i, j;
218
  guint8 *destY, *destY2, *destU, *destV;
219
  gint dest_strideY, dest_strideUV;
220 221 222 223 224 225 226
  const guint8 *src2;
  gint src_stride;
  gint y_idx, uv_idx;
  gint y1, y2, y3, y4;
  gint u1, u2, u3, u4;
  gint v1, v2, v3, v4;
  gint matrix[12];
227

228 229
  dest_strideY = gst_video_format_get_row_stride (dest_format, 0, dest_width);
  dest_strideUV = gst_video_format_get_row_stride (dest_format, 1, dest_width);
230

231 232
  src_stride = 4 * src_width;

233
  destY =
234
      dest + gst_video_format_get_component_offset (dest_format, 0,
235 236
      dest_width, dest_height);
  destU =
237
      dest + gst_video_format_get_component_offset (dest_format, 1,
238 239
      dest_width, dest_height);
  destV =
240
      dest + gst_video_format_get_component_offset (dest_format, 2,
241 242 243
      dest_width, dest_height);

  destY = destY + dest_y * dest_strideY + dest_x;
244
  destY2 = (dest_y < dest_height) ? destY + dest_strideY : destY;
245 246 247
  destU = destU + (dest_y / 2) * dest_strideUV + dest_x / 2;
  destV = destV + (dest_y / 2) * dest_strideUV + dest_x / 2;

248 249
  src = src + src_y * src_stride + src_x * 4;
  src2 = (src_y < src_height) ? src + src_stride : src;
250

251 252
  h = dest_y + h;
  w = dest_x + w;
253

254
  if (src_sdtv != dest_sdtv)
255 256 257
    memcpy (matrix,
        dest_sdtv ? cog_ycbcr_hdtv_to_ycbcr_sdtv_matrix_8bit :
        cog_ycbcr_sdtv_to_ycbcr_hdtv_matrix_8bit, 12 * sizeof (gint));
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
  else
    memcpy (matrix, cog_identity_matrix_8bit, 12 * sizeof (gint));

  /* 1. Handle the first destination scanline specially if it
   *    doesn't start at the macro pixel boundary, i.e. blend
   *    with the background! */
  if (dest_y % 2 == 1) {
    /* 1.1. Handle the first destination pixel if it doesn't
     *      start at the macro pixel boundary, i.e. blend with
     *      the background! */
    if (dest_x % 2 == 1) {
      y1 = src[4 * 0 + 1];
      u1 = src[4 * 0 + 2];
      v1 = src[4 * 0 + 3];

      destY[0] = CLAMP (APPLY_MATRIX (matrix, 0, y1, u1, v1), 0, 255);
      destU[0] =
          CLAMP ((3 * destU[0] + APPLY_MATRIX (matrix, 1, y1, u1, v1)) / 4, 0,
          255);
      destV[0] =
          CLAMP ((3 * destV[0] + APPLY_MATRIX (matrix, 2, y1, u1, v1)) / 4, 0,
          255);

      j = dest_x + 1;
      y_idx = uv_idx = 1;
    } else {
      j = dest_x;
      y_idx = uv_idx = 0;
    }
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
    /* 1.2. Copy all macro pixels from the source to the destination
     *      but blend with the background because we're only filling
     *      the lower part of the macro pixels. */
    for (; j < w - 1; j += 2) {
      y1 = src[4 * y_idx + 1];
      y2 = src[4 * y_idx + 4 + 1];

      u1 = src[4 * y_idx + 2];
      u2 = src[4 * y_idx + 4 + 2];

      v1 = src[4 * y_idx + 3];
      v2 = src[4 * y_idx + 4 + 3];

      destY[y_idx] = CLAMP (APPLY_MATRIX (matrix, 0, y1, u1, v1), 0, 255);
      destY[y_idx + 1] = CLAMP (APPLY_MATRIX (matrix, 0, y2, u2, v2), 0, 255);
      destU[uv_idx] = CLAMP (
          (2 * destU[uv_idx] + APPLY_MATRIX (matrix, 1, y1, u1,
                  v1) + APPLY_MATRIX (matrix, 1, y2, u2, v2)) / 4, 0, 255);
      destV[uv_idx] = CLAMP (
          (2 * destV[uv_idx] + APPLY_MATRIX (matrix, 2, y1, u1,
                  v1) + APPLY_MATRIX (matrix, 2, y2, u2, v2)) / 4, 0, 255);

      y_idx += 2;
      uv_idx++;
    }
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
    /* 1.3. Now copy the last pixel if one exists and blend it
     *      with the background because we only fill part of
     *      the macro pixel. In case this is the last pixel of
     *      the destination we will a larger part. */
    if (j == w - 1 && j == dest_width - 1) {
      y1 = src[4 * y_idx + 1];
      u1 = src[4 * y_idx + 2];
      v1 = src[4 * y_idx + 3];

      destY[y_idx] = CLAMP (APPLY_MATRIX (matrix, 0, y1, u1, v1), 0, 255);
      destU[uv_idx] = CLAMP (
          (destU[uv_idx] + APPLY_MATRIX (matrix, 1, y1, u1, v1)) / 2, 0, 255);
      destV[uv_idx] = CLAMP (
          (destV[uv_idx] + APPLY_MATRIX (matrix, 2, y1, u1, v1)) / 2, 0, 255);
    } else if (j == w - 1) {
      y1 = src[4 * y_idx + 1];
      u1 = src[4 * y_idx + 2];
      v1 = src[4 * y_idx + 3];

      destY[y_idx] = CLAMP (APPLY_MATRIX (matrix, 0, y1, u1, v1), 0, 255);
      destU[uv_idx] = CLAMP (
          (3 * destU[uv_idx] + APPLY_MATRIX (matrix, 1, y1, u1, v1)) / 4, 0,
          255);
      destV[uv_idx] =
          CLAMP ((3 * destV[uv_idx] + APPLY_MATRIX (matrix, 2, y1, u1, v1)) / 4,
          0, 255);
    }
341

342 343 344 345 346 347 348 349 350 351
    destY += dest_strideY;
    destY2 += dest_strideY;
    destU += dest_strideUV;
    destV += dest_strideUV;
    src += src_stride;
    src2 += src_stride;
    i = dest_y + 1;
  } else {
    i = dest_y;
  }
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
  /* 2. Copy all macro pixel scanlines, the destination scanline
   *    now starts at macro pixel boundary. */
  for (; i < h - 1; i += 2) {
    /* 2.1. Handle the first destination pixel if it doesn't
     *      start at the macro pixel boundary, i.e. blend with
     *      the background! */
    if (dest_x % 2 == 1) {
      y1 = src[4 * 0 + 1];
      y2 = src2[4 * 0 + 1];
      u1 = src[4 * 0 + 2];
      u2 = src2[4 * 0 + 2];
      v1 = src[4 * 0 + 3];
      v2 = src2[4 * 0 + 3];

      destY[0] = CLAMP (APPLY_MATRIX (matrix, 0, y1, u1, v1), 0, 255);
      destY2[0] = CLAMP (APPLY_MATRIX (matrix, 0, y2, u2, v2), 0, 255);
      destU[0] = CLAMP (
          (2 * destU[0] + APPLY_MATRIX (matrix, 1, y1, u1,
                  v1) + APPLY_MATRIX (matrix, 1, y2, u2, v2)) / 4, 0, 255);
      destV[0] = CLAMP (
          (2 * destV[0] + APPLY_MATRIX (matrix, 2, y1, u1,
                  v1) + APPLY_MATRIX (matrix, 2, y2, u2, v2)) / 4, 0, 255);
      j = dest_x + 1;
      y_idx = uv_idx = 1;
    } else {
      j = dest_x;
      y_idx = uv_idx = 0;
    }
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
    /* 2.2. Copy all macro pixels from the source to the destination.
     *      All pixels now start at macro pixel boundary, i.e. no
     *      blending with the background is necessary. */
    for (; j < w - 1; j += 2) {
      y1 = src[4 * y_idx + 1];
      y2 = src[4 * y_idx + 4 + 1];
      y3 = src2[4 * y_idx + 1];
      y4 = src2[4 * y_idx + 4 + 1];

      u1 = src[4 * y_idx + 2];
      u2 = src[4 * y_idx + 4 + 2];
      u3 = src2[4 * y_idx + 2];
      u4 = src2[4 * y_idx + 4 + 2];

      v1 = src[4 * y_idx + 3];
      v2 = src[4 * y_idx + 4 + 3];
      v3 = src2[4 * y_idx + 3];
      v4 = src2[4 * y_idx + 4 + 3];

      destY[y_idx] = CLAMP (APPLY_MATRIX (matrix, 0, y1, u1, v1), 0, 255);
      destY[y_idx + 1] = CLAMP (APPLY_MATRIX (matrix, 0, y2, u2, v2), 0, 255);
      destY2[y_idx] = CLAMP (APPLY_MATRIX (matrix, 0, y3, u3, v3), 0, 255);
      destY2[y_idx + 1] = CLAMP (APPLY_MATRIX (matrix, 0, y4, u4, v4), 0, 255);

      destU[uv_idx] = CLAMP (
          (APPLY_MATRIX (matrix, 1, y1, u1, v1) + APPLY_MATRIX (matrix, 1, y2,
                  u2, v2) + APPLY_MATRIX (matrix, 1, y3, u3,
                  v3) + APPLY_MATRIX (matrix, 1, y4, u4, v4)) / 4, 0, 255);
      destV[uv_idx] = CLAMP (
          (APPLY_MATRIX (matrix, 2, y1, u1, v1) + APPLY_MATRIX (matrix, 2, y2,
                  u2, v2) + APPLY_MATRIX (matrix, 2, y3, u3,
                  v3) + APPLY_MATRIX (matrix, 2, y4, u4, v4)) / 4, 0, 255);

      y_idx += 2;
      uv_idx++;
    }
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
    /* 2.3. Now copy the last pixel if one exists and blend it
     *      with the background because we only fill part of
     *      the macro pixel. In case this is the last pixel of
     *      the destination we will a larger part. */
    if (j == w - 1 && j == dest_width - 1) {
      y1 = src[4 * y_idx + 1];
      y2 = src2[4 * y_idx + 1];

      u1 = src[4 * y_idx + 2];
      u2 = src2[4 * y_idx + 2];

      v1 = src[4 * y_idx + 3];
      v2 = src2[4 * y_idx + 3];

      destY[y_idx] = CLAMP (APPLY_MATRIX (matrix, 0, y1, u1, v1), 0, 255);
      destY2[y_idx] = CLAMP (APPLY_MATRIX (matrix, 0, y2, u2, v2), 0, 255);
      destU[uv_idx] = CLAMP (
          (APPLY_MATRIX (matrix, 1, y1, u1, v1) + APPLY_MATRIX (matrix, 2, y2,
                  u2, v2)) / 2, 0, 255);
      destV[uv_idx] = CLAMP (
          (APPLY_MATRIX (matrix, 1, y1, u1, v1) + APPLY_MATRIX (matrix, 2, y2,
                  u2, v2)) / 2, 0, 255);
    } else if (j == w - 1) {
      y1 = src[4 * y_idx + 1];
      y2 = src2[4 * y_idx + 1];

      u1 = src[4 * y_idx + 2];
      u2 = src2[4 * y_idx + 2];

      v1 = src[4 * y_idx + 3];
      v2 = src2[4 * y_idx + 3];

      destY[y_idx] = CLAMP (APPLY_MATRIX (matrix, 0, y1, u1, v1), 0, 255);
      destY2[y_idx] = CLAMP (APPLY_MATRIX (matrix, 0, y2, u2, v2), 0, 255);
      destU[uv_idx] = CLAMP (
          (2 * destU[uv_idx] + APPLY_MATRIX (matrix, 1, y1, u1,
                  v1) + APPLY_MATRIX (matrix, 2, y2, u2, v2)) / 4, 0, 255);
      destV[uv_idx] = CLAMP (
          (2 * destV[uv_idx] + APPLY_MATRIX (matrix, 1, y1, u1,
                  v1) + APPLY_MATRIX (matrix, 2, y2, u2, v2)) / 4, 0, 255);
    }
460

461 462 463 464 465 466 467
    destY += 2 * dest_strideY;
    destY2 += 2 * dest_strideY;
    destU += dest_strideUV;
    destV += dest_strideUV;
    src += 2 * src_stride;
    src2 += 2 * src_stride;
  }
468

469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491
  /* 3. Handle the last scanline if one exists. This again
   *    doesn't start at macro pixel boundary but should
   *    only fill the upper part of the macro pixels. */
  if (i == h - 1 && i == dest_height - 1) {
    /* 3.1. Handle the first destination pixel if it doesn't
     *      start at the macro pixel boundary, i.e. blend with
     *      the background! */
    if (dest_x % 2 == 1) {
      y1 = src[4 * 0 + 1];
      u1 = src[4 * 0 + 2];
      v1 = src[4 * 0 + 3];

      destY[0] = CLAMP (APPLY_MATRIX (matrix, 0, y1, u1, v1), 0, 255);
      destU[0] =
          CLAMP ((destU[0] + APPLY_MATRIX (matrix, 1, y1, u1, v1)) / 2, 0, 255);
      destV[0] =
          CLAMP ((destV[0] + APPLY_MATRIX (matrix, 2, y1, u1, v1)) / 2, 0, 255);

      j = dest_x + 1;
      y_idx = uv_idx = 1;
    } else {
      j = dest_x;
      y_idx = uv_idx = 0;
492 493
    }

494 495 496 497 498 499
    /* 3.2. Copy all macro pixels from the source to the destination
     *      but blend with the background because we're only filling
     *      the upper part of the macro pixels. */
    for (; j < w - 1; j += 2) {
      y1 = src[4 * y_idx + 1];
      y2 = src[4 * y_idx + 4 + 1];
500

501 502
      u1 = src[4 * y_idx + 2];
      u2 = src[4 * y_idx + 4 + 2];
503

504 505
      v1 = src[4 * y_idx + 3];
      v2 = src[4 * y_idx + 4 + 3];
506

507 508
      destY[y_idx] = CLAMP (APPLY_MATRIX (matrix, 0, y1, u1, v1), 0, 255);
      destY[y_idx + 1] = CLAMP (APPLY_MATRIX (matrix, 0, y2, u2, v2), 0, 255);
509

510 511 512 513 514 515
      destU[uv_idx] = CLAMP (
          (2 * destU[uv_idx] + APPLY_MATRIX (matrix, 1, y1, u1,
                  v1) + APPLY_MATRIX (matrix, 1, y2, u2, v2)) / 4, 0, 255);
      destV[uv_idx] = CLAMP (
          (2 * destV[uv_idx] + APPLY_MATRIX (matrix, 2, y1, u1,
                  v1) + APPLY_MATRIX (matrix, 2, y2, u2, v2)) / 4, 0, 255);
516

517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624
      y_idx += 2;
      uv_idx++;
    }

    /* 3.3. Now copy the last pixel if one exists and blend it
     *      with the background because we only fill part of
     *      the macro pixel. In case this is the last pixel of
     *      the destination we will a larger part. */
    if (j == w - 1 && j == dest_width - 1) {
      y1 = src[4 * y_idx + 1];
      u1 = src[4 * y_idx + 2];
      v1 = src[4 * y_idx + 3];

      destY[y_idx] = CLAMP (APPLY_MATRIX (matrix, 0, y1, u1, v1), 0, 255);
      destU[uv_idx] = CLAMP (
          (destU[uv_idx] + APPLY_MATRIX (matrix, 1, y1, u1, v1)) / 2, 0, 255);
      destV[uv_idx] = CLAMP (
          (destV[uv_idx] + APPLY_MATRIX (matrix, 1, y1, u1, v1)) / 2, 0, 255);
    } else if (j == w - 1) {
      y1 = src[4 * y_idx + 1];
      u1 = src[4 * y_idx + 2];
      v1 = src[4 * y_idx + 3];

      destY[y_idx] = CLAMP (APPLY_MATRIX (matrix, 0, y1, u1, v1), 0, 255);
      destU[uv_idx] = CLAMP (
          (3 * destU[uv_idx] + APPLY_MATRIX (matrix, 1, y1, u1, v1)) / 4, 0,
          255);
      destV[uv_idx] =
          CLAMP ((3 * destV[uv_idx] + APPLY_MATRIX (matrix, 1, y1, u1, v1)) / 4,
          0, 255);
    }
  } else if (i == h - 1) {
    /* 3.1. Handle the first destination pixel if it doesn't
     *      start at the macro pixel boundary, i.e. blend with
     *      the background! */
    if (dest_x % 2 == 1) {
      y1 = src[4 * 0 + 1];
      u1 = src[4 * 0 + 2];
      v1 = src[4 * 0 + 3];

      destY[0] = CLAMP (APPLY_MATRIX (matrix, 0, y1, u1, v1), 0, 255);
      destU[0] =
          CLAMP ((3 * destU[0] + APPLY_MATRIX (matrix, 1, y1, u1, v1)) / 4, 0,
          255);
      destV[0] =
          CLAMP ((3 * destV[0] + APPLY_MATRIX (matrix, 2, y1, u1, v1)) / 4, 0,
          255);

      j = dest_x + 1;
      y_idx = uv_idx = 1;
    } else {
      j = dest_x;
      y_idx = uv_idx = 0;
    }

    /* 3.2. Copy all macro pixels from the source to the destination
     *      but blend with the background because we're only filling
     *      the upper part of the macro pixels. */
    for (; j < w - 1; j += 2) {
      y1 = src[4 * y_idx + 1];
      y2 = src[4 * y_idx + 4 + 1];

      u1 = src[4 * y_idx + 2];
      u2 = src[4 * y_idx + 4 + 2];

      v1 = src[4 * y_idx + 3];
      v2 = src[4 * y_idx + 4 + 3];

      destY[y_idx] = CLAMP (APPLY_MATRIX (matrix, 0, y1, u1, v1), 0, 255);
      destY[y_idx + 1] = CLAMP (APPLY_MATRIX (matrix, 0, y2, u2, v2), 0, 255);

      destU[uv_idx] = CLAMP (
          (2 * destU[uv_idx] + APPLY_MATRIX (matrix, 1, y1, u1,
                  v1) + APPLY_MATRIX (matrix, 1, y2, u2, v2)) / 4, 0, 255);
      destV[uv_idx] = CLAMP (
          (2 * destV[uv_idx] + APPLY_MATRIX (matrix, 2, y1, u1,
                  v1) + APPLY_MATRIX (matrix, 2, y2, u2, v2)) / 4, 0, 255);

      y_idx += 2;
      uv_idx++;
    }

    /* 3.3. Now copy the last pixel if one exists and blend it
     *      with the background because we only fill part of
     *      the macro pixel. In case this is the last pixel of
     *      the destination we will a larger part. */
    if (j == w - 1 && j == dest_width - 1) {
      y1 = src[4 * y_idx + 1];
      u1 = src[4 * y_idx + 2];
      v1 = src[4 * y_idx + 3];

      destY[y_idx] = CLAMP (APPLY_MATRIX (matrix, 0, y1, u1, v1), 0, 255);
      destU[uv_idx] = CLAMP (
          (destU[uv_idx] + APPLY_MATRIX (matrix, 1, y1, u1, v1)) / 2, 0, 255);
      destV[uv_idx] = CLAMP (
          (destV[uv_idx] + APPLY_MATRIX (matrix, 1, y1, u1, v1)) / 2, 0, 255);
    } else if (j == w - 1) {
      y1 = src[4 * y_idx + 1];
      u1 = src[4 * y_idx + 2];
      v1 = src[4 * y_idx + 3];

      destY[y_idx] = CLAMP (APPLY_MATRIX (matrix, 0, y1, u1, v1), 0, 255);
      destU[uv_idx] = CLAMP (
          (3 * destU[uv_idx] + APPLY_MATRIX (matrix, 1, y1, u1, v1)) / 4, 0,
          255);
      destV[uv_idx] =
          CLAMP ((3 * destV[uv_idx] + APPLY_MATRIX (matrix, 1, y1, u1, v1)) / 4,
          0, 255);
625 626 627 628 629
    }
  }
}

static void
630 631
fill_i420 (GstVideoBoxFill fill_type, guint b_alpha, GstVideoFormat format,
    guint8 * dest, gboolean sdtv, gint width, gint height)
632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654
{
  guint8 empty_pixel[3];
  guint8 *destY, *destU, *destV;
  gint strideY, strideUV;
  gint heightY, heightUV;

  if (sdtv) {
    empty_pixel[0] = yuv_sdtv_colors_Y[fill_type];
    empty_pixel[1] = yuv_sdtv_colors_U[fill_type];
    empty_pixel[2] = yuv_sdtv_colors_V[fill_type];
  } else {
    empty_pixel[0] = yuv_hdtv_colors_Y[fill_type];
    empty_pixel[1] = yuv_hdtv_colors_U[fill_type];
    empty_pixel[2] = yuv_hdtv_colors_V[fill_type];
  }

  strideY = gst_video_format_get_row_stride (GST_VIDEO_FORMAT_I420, 0, width);
  strideUV = gst_video_format_get_row_stride (GST_VIDEO_FORMAT_I420, 1, width);

  destY =
      dest + gst_video_format_get_component_offset (GST_VIDEO_FORMAT_I420, 0,
      width, height);
  destU =
655
      dest + gst_video_format_get_component_offset (format, 1, width, height);
656
  destV =
657
      dest + gst_video_format_get_component_offset (format, 2, width, height);
658 659 660 661 662 663

  heightY =
      gst_video_format_get_component_height (GST_VIDEO_FORMAT_I420, 0, height);
  heightUV =
      gst_video_format_get_component_height (GST_VIDEO_FORMAT_I420, 1, height);

664 665 666
  oil_splat_u8_ns (destY, &empty_pixel[0], strideY * heightY);
  oil_splat_u8_ns (destU, &empty_pixel[1], strideUV * heightUV);
  oil_splat_u8_ns (destV, &empty_pixel[2], strideUV * heightUV);
667 668 669
}

static void
670 671 672 673 674
copy_i420_i420 (guint i_alpha, GstVideoFormat dest_format, guint8 * dest,
    gboolean dest_sdtv, gint dest_width, gint dest_height, gint dest_x,
    gint dest_y, GstVideoFormat src_format, const guint8 * src,
    gboolean src_sdtv, gint src_width, gint src_height, gint src_x, gint src_y,
    gint w, gint h)
675
{
676
  gint i, j;
677 678
  guint8 *destY, *destU, *destV;
  const guint8 *srcY, *srcU, *srcV;
679 680
  guint8 *destY2;
  const guint8 *srcY2, *srcU2, *srcV2;
681 682
  gint dest_strideY, dest_strideUV;
  gint src_strideY, src_strideUV;
683 684 685 686 687 688
  gint src_y_idx, src_uv_idx;
  gint dest_y_idx, dest_uv_idx;
  gint matrix[12];
  gint y1, y2, y3, y4;
  gint u1, u2, u3, u4;
  gint v1, v2, v3, v4;
689 690 691 692 693 694 695 696 697 698 699 700 701 702

  dest_strideY =
      gst_video_format_get_row_stride (GST_VIDEO_FORMAT_I420, 0, dest_width);
  dest_strideUV =
      gst_video_format_get_row_stride (GST_VIDEO_FORMAT_I420, 1, dest_width);
  src_strideY =
      gst_video_format_get_row_stride (GST_VIDEO_FORMAT_I420, 0, src_width);
  src_strideUV =
      gst_video_format_get_row_stride (GST_VIDEO_FORMAT_I420, 1, src_width);

  destY =
      dest + gst_video_format_get_component_offset (GST_VIDEO_FORMAT_I420, 0,
      dest_width, dest_height);
  destU =
703
      dest + gst_video_format_get_component_offset (dest_format, 1,
704 705
      dest_width, dest_height);
  destV =
706
      dest + gst_video_format_get_component_offset (dest_format, 2,
707 708 709 710 711 712
      dest_width, dest_height);

  srcY =
      src + gst_video_format_get_component_offset (GST_VIDEO_FORMAT_I420, 0,
      src_width, src_height);
  srcU =
713
      src + gst_video_format_get_component_offset (src_format, 1,
714 715
      src_width, src_height);
  srcV =
716
      src + gst_video_format_get_component_offset (src_format, 2,
717 718 719 720 721 722 723 724 725 726 727
      src_width, src_height);


  destY = destY + dest_y * dest_strideY + dest_x;
  destU = destU + (dest_y / 2) * dest_strideUV + dest_x / 2;
  destV = destV + (dest_y / 2) * dest_strideUV + dest_x / 2;

  srcY = srcY + src_y * src_strideY + src_x;
  srcU = srcU + (src_y / 2) * src_strideUV + src_x / 2;
  srcV = srcV + (src_y / 2) * src_strideUV + src_x / 2;

728 729
  destY2 = destY + dest_strideY;
  srcY2 = srcY + src_strideY;
730

731 732
  h = dest_y + h;
  w = dest_x + w;
733

734
  if (src_sdtv != dest_sdtv)
735 736 737
    memcpy (matrix,
        dest_sdtv ? cog_ycbcr_hdtv_to_ycbcr_sdtv_matrix_8bit :
        cog_ycbcr_sdtv_to_ycbcr_hdtv_matrix_8bit, 12 * sizeof (gint));
738 739
  else
    memcpy (matrix, cog_identity_matrix_8bit, 12 * sizeof (gint));
740

741 742 743 744 745 746 747 748 749 750 751
  /* 1. Handle the first destination scanline specially if it
   *    doesn't start at the macro pixel boundary, i.e. blend
   *    with the background! */
  if (dest_y % 2 == 1) {
    /* 1.1. Handle the first destination pixel if it doesn't
     *      start at the macro pixel boundary, i.e. blend with
     *      the background! */
    if (dest_x % 2 == 1) {
      y1 = srcY[0];
      u1 = srcU[0];
      v1 = srcV[0];
752

753 754 755 756 757 758 759
      destY[0] = CLAMP (APPLY_MATRIX (matrix, 0, y1, u1, v1), 0, 255);
      destU[0] =
          CLAMP ((3 * destU[0] + APPLY_MATRIX (matrix, 1, y1, u1, v1)) / 4, 0,
          255);
      destV[0] =
          CLAMP ((3 * destV[0] + APPLY_MATRIX (matrix, 2, y1, u1, v1)) / 4, 0,
          255);
760

761 762 763 764 765 766 767 768
      j = dest_x + 1;
      src_y_idx = dest_y_idx = dest_uv_idx = 1;
      src_uv_idx = (src_x % 2) + 1;
    } else {
      j = dest_x;
      src_y_idx = dest_y_idx = dest_uv_idx = 0;
      src_uv_idx = (src_x % 2);
    }
769

770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792
    /* 1.2. Copy all macro pixels from the source to the destination
     *      but blend with the background because we're only filling
     *      the lower part of the macro pixels. */
    for (; j < w - 1; j += 2) {
      y1 = srcY[src_y_idx];
      y2 = srcY[src_y_idx + 1];

      u1 = srcU[src_uv_idx / 2];
      v1 = srcV[src_uv_idx / 2];
      src_uv_idx++;
      u2 = srcU[src_uv_idx / 2];
      v2 = srcV[src_uv_idx / 2];
      src_uv_idx++;

      destY[dest_y_idx] = CLAMP (APPLY_MATRIX (matrix, 0, y1, u1, v1), 0, 255);
      destY[dest_y_idx + 1] =
          CLAMP (APPLY_MATRIX (matrix, 0, y2, u2, v2), 0, 255);
      destU[dest_uv_idx] =
          CLAMP ((2 * destU[dest_uv_idx] + APPLY_MATRIX (matrix, 1, y1, u1,
                  v1) + APPLY_MATRIX (matrix, 1, y2, u2, v2)) / 4, 0, 255);
      destV[dest_uv_idx] =
          CLAMP ((2 * destV[dest_uv_idx] + APPLY_MATRIX (matrix, 2, y1, u1,
                  v1) + APPLY_MATRIX (matrix, 2, y2, u2, v2)) / 4, 0, 255);
793

794 795 796 797
      dest_y_idx += 2;
      src_y_idx += 2;
      dest_uv_idx++;
    }
798

799 800 801 802 803 804 805 806
    /* 1.3. Now copy the last pixel if one exists and blend it
     *      with the background because we only fill part of
     *      the macro pixel. In case this is the last pixel of
     *      the destination we will a larger part. */
    if (j == w - 1 && j == dest_width - 1) {
      y1 = srcY[src_y_idx];
      u1 = srcU[src_uv_idx / 2];
      v1 = srcV[src_uv_idx / 2];
807

808 809 810 811 812 813 814 815 816 817 818
      destY[dest_y_idx] = CLAMP (APPLY_MATRIX (matrix, 0, y1, u1, v1), 0, 255);
      destU[dest_uv_idx] = CLAMP (
          (destU[dest_uv_idx] + APPLY_MATRIX (matrix, 1, y1, u1, v1)) / 2, 0,
          255);
      destV[dest_uv_idx] =
          CLAMP ((destV[dest_uv_idx] + APPLY_MATRIX (matrix, 2, y1, u1,
                  v1)) / 2, 0, 255);
    } else if (j == w - 1) {
      y1 = srcY[src_y_idx];
      u1 = srcU[src_uv_idx / 2];
      v1 = srcV[src_uv_idx / 2];
819

820 821 822 823 824 825 826 827
      destY[dest_y_idx] = CLAMP (APPLY_MATRIX (matrix, 0, y1, u1, v1), 0, 255);
      destU[dest_uv_idx] = CLAMP (
          (3 * destU[dest_uv_idx] + APPLY_MATRIX (matrix, 1, y1, u1, v1)) / 4,
          0, 255);
      destV[dest_uv_idx] =
          CLAMP ((3 * destV[dest_uv_idx] + APPLY_MATRIX (matrix, 2, y1, u1,
                  v1)) / 4, 0, 255);
    }
828

829 830 831 832 833 834 835 836
    destY += dest_strideY;
    destY2 += dest_strideY;
    destU += dest_strideUV;
    destV += dest_strideUV;
    srcY += src_strideY;
    srcY2 += src_strideY;
    src_y++;
    if (src_y % 2 == 0) {
837 838 839
      srcU += src_strideUV;
      srcV += src_strideUV;
    }
840
    i = dest_y + 1;
841
  } else {
842 843 844 845 846 847 848 849 850 851 852 853 854 855 856
    i = dest_y;
  }

  /* 2. Copy all macro pixel scanlines, the destination scanline
   *    now starts at macro pixel boundary. */
  for (; i < h - 1; i += 2) {
    /* 2.1. Handle the first destination pixel if it doesn't
     *      start at the macro pixel boundary, i.e. blend with
     *      the background! */

    srcU2 = srcU;
    srcV2 = srcV;
    if (src_y % 2 == 1) {
      srcU2 += src_strideUV;
      srcV2 += src_strideUV;
857 858
    }

859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148
    if (dest_x % 2 == 1) {
      y1 = srcY[0];
      y2 = srcY2[0];
      u1 = srcU[0];
      v1 = srcV[0];
      u2 = srcU2[0];
      v2 = srcV2[0];

      destY[0] = CLAMP (APPLY_MATRIX (matrix, 0, y1, u1, v1), 0, 255);
      destY2[0] = CLAMP (APPLY_MATRIX (matrix, 0, y2, u2, v2), 0, 255);
      destU[0] = CLAMP (
          (2 * destU[0] + APPLY_MATRIX (matrix, 1, y1, u1,
                  v1) + APPLY_MATRIX (matrix, 1, y2, u2, v2)) / 4, 0, 255);
      destV[0] = CLAMP (
          (2 * destV[0] + APPLY_MATRIX (matrix, 2, y1, u1,
                  v1) + APPLY_MATRIX (matrix, 2, y2, u2, v2)) / 4, 0, 255);
      j = dest_x + 1;
      src_y_idx = dest_y_idx = dest_uv_idx = 1;
      src_uv_idx = (src_x % 2) + 1;
    } else {
      j = dest_x;
      src_y_idx = dest_y_idx = dest_uv_idx = 0;
      src_uv_idx = (src_x % 2);
    }

    /* 2.2. Copy all macro pixels from the source to the destination.
     *      All pixels now start at macro pixel boundary, i.e. no
     *      blending with the background is necessary. */
    for (; j < w - 1; j += 2) {
      y1 = srcY[src_y_idx];
      y2 = srcY[src_y_idx + 1];
      y3 = srcY2[src_y_idx];
      y4 = srcY2[src_y_idx + 1];

      u1 = srcU[src_uv_idx / 2];
      u3 = srcU2[src_uv_idx / 2];
      v1 = srcV[src_uv_idx / 2];
      v3 = srcV2[src_uv_idx / 2];
      src_uv_idx++;
      u2 = srcU[src_uv_idx / 2];
      u4 = srcU2[src_uv_idx / 2];
      v2 = srcV[src_uv_idx / 2];
      v4 = srcV2[src_uv_idx / 2];
      src_uv_idx++;

      destY[dest_y_idx] = CLAMP (APPLY_MATRIX (matrix, 0, y1, u1, v1), 0, 255);
      destY[dest_y_idx + 1] =
          CLAMP (APPLY_MATRIX (matrix, 0, y2, u2, v2), 0, 255);
      destY2[dest_y_idx] = CLAMP (APPLY_MATRIX (matrix, 0, y3, u3, v3), 0, 255);
      destY2[dest_y_idx + 1] =
          CLAMP (APPLY_MATRIX (matrix, 0, y4, u4, v4), 0, 255);

      destU[dest_uv_idx] = CLAMP (
          (APPLY_MATRIX (matrix, 1, y1, u1, v1) + APPLY_MATRIX (matrix, 1, y2,
                  u2, v2) + APPLY_MATRIX (matrix, 1, y3, u3,
                  v3) + APPLY_MATRIX (matrix, 1, y4, u4, v4)) / 4, 0, 255);
      destV[dest_uv_idx] = CLAMP (
          (APPLY_MATRIX (matrix, 2, y1, u1, v1) + APPLY_MATRIX (matrix, 2, y2,
                  u2, v2) + APPLY_MATRIX (matrix, 2, y3, u3,
                  v3) + APPLY_MATRIX (matrix, 2, y4, u4, v4)) / 4, 0, 255);

      dest_y_idx += 2;
      src_y_idx += 2;
      dest_uv_idx++;
    }

    /* 2.3. Now copy the last pixel if one exists and blend it
     *      with the background because we only fill part of
     *      the macro pixel. In case this is the last pixel of
     *      the destination we will a larger part. */
    if (j == w - 1 && j == dest_width - 1) {
      y1 = srcY[src_y_idx];
      y2 = srcY2[src_y_idx];

      u1 = srcU[src_uv_idx / 2];
      u2 = srcU2[src_uv_idx / 2];

      v1 = srcV[src_uv_idx / 2];
      v2 = srcV2[src_uv_idx / 2];

      destY[dest_y_idx] = CLAMP (APPLY_MATRIX (matrix, 0, y1, u1, v1), 0, 255);
      destY2[dest_y_idx] = CLAMP (APPLY_MATRIX (matrix, 0, y2, u2, v2), 0, 255);
      destU[dest_uv_idx] = CLAMP (
          (APPLY_MATRIX (matrix, 1, y1, u1, v1) + APPLY_MATRIX (matrix, 2, y2,
                  u2, v2)) / 2, 0, 255);
      destV[dest_uv_idx] = CLAMP (
          (APPLY_MATRIX (matrix, 1, y1, u1, v1) + APPLY_MATRIX (matrix, 2, y2,
                  u2, v2)) / 2, 0, 255);
    } else if (j == w - 1) {
      y1 = srcY[src_y_idx];
      y2 = srcY2[src_y_idx];

      u1 = srcU[src_uv_idx / 2];
      u2 = srcU2[src_uv_idx / 2];

      v1 = srcV[src_uv_idx / 2];
      v2 = srcV2[src_uv_idx / 2];

      destY[dest_y_idx] = CLAMP (APPLY_MATRIX (matrix, 0, y1, u1, v1), 0, 255);
      destY2[dest_y_idx] = CLAMP (APPLY_MATRIX (matrix, 0, y2, u2, v2), 0, 255);
      destU[dest_uv_idx] = CLAMP (
          (2 * destU[dest_uv_idx] + APPLY_MATRIX (matrix, 1, y1, u1,
                  v1) + APPLY_MATRIX (matrix, 2, y2, u2, v2)) / 4, 0, 255);
      destV[dest_uv_idx] = CLAMP (
          (2 * destV[dest_uv_idx] + APPLY_MATRIX (matrix, 1, y1, u1,
                  v1) + APPLY_MATRIX (matrix, 2, y2, u2, v2)) / 4, 0, 255);
    }

    destY += 2 * dest_strideY;
    destY2 += 2 * dest_strideY;
    destU += dest_strideUV;
    destV += dest_strideUV;
    srcY += 2 * src_strideY;
    srcY2 += 2 * src_strideY;

    src_y += 2;
    srcU += src_strideUV;
    srcV += src_strideUV;
  }

  /* 3. Handle the last scanline if one exists. This again
   *    doesn't start at macro pixel boundary but should
   *    only fill the upper part of the macro pixels. */
  if (i == h - 1 && i == dest_height - 1) {
    /* 3.1. Handle the first destination pixel if it doesn't
     *      start at the macro pixel boundary, i.e. blend with
     *      the background! */
    if (dest_x % 2 == 1) {
      y1 = srcY[0];
      u1 = srcU[0];
      v1 = srcV[0];

      destY[0] = CLAMP (APPLY_MATRIX (matrix, 0, y1, u1, v1), 0, 255);
      destU[0] =
          CLAMP ((destU[0] + APPLY_MATRIX (matrix, 1, y1, u1, v1)) / 2, 0, 255);
      destV[0] =
          CLAMP ((destV[0] + APPLY_MATRIX (matrix, 2, y1, u1, v1)) / 2, 0, 255);

      j = dest_x + 1;
      src_y_idx = dest_y_idx = dest_uv_idx = 1;
      src_uv_idx = (src_x % 2) + 1;
    } else {
      j = dest_x;
      src_y_idx = dest_y_idx = dest_uv_idx = 0;
      src_uv_idx = (src_x % 2);
    }

    /* 3.2. Copy all macro pixels from the source to the destination
     *      but blend with the background because we're only filling
     *      the upper part of the macro pixels. */
    for (; j < w - 1; j += 2) {
      y1 = srcY[src_y_idx];
      y2 = srcY[src_y_idx + 1];

      u1 = srcU[src_uv_idx / 2];
      v1 = srcV[src_uv_idx / 2];
      src_uv_idx++;
      u2 = srcU[src_uv_idx / 2];
      v2 = srcV[src_uv_idx / 2];
      src_uv_idx++;

      destY[dest_y_idx] = CLAMP (APPLY_MATRIX (matrix, 0, y1, u1, v1), 0, 255);
      destY[dest_y_idx + 1] =
          CLAMP (APPLY_MATRIX (matrix, 0, y2, u2, v2), 0, 255);

      destU[dest_uv_idx] = CLAMP (
          (2 * destU[dest_uv_idx] + APPLY_MATRIX (matrix, 1, y1, u1,
                  v1) + APPLY_MATRIX (matrix, 1, y2, u2, v2)) / 4, 0, 255);
      destV[dest_uv_idx] = CLAMP (
          (2 * destV[dest_uv_idx] + APPLY_MATRIX (matrix, 2, y1, u1,
                  v1) + APPLY_MATRIX (matrix, 2, y2, u2, v2)) / 4, 0, 255);

      dest_y_idx += 2;
      src_y_idx += 2;
      dest_uv_idx++;
    }

    /* 3.3. Now copy the last pixel if one exists and blend it
     *      with the background because we only fill part of
     *      the macro pixel. In case this is the last pixel of
     *      the destination we will a larger part. */
    if (j == w - 1 && j == dest_width - 1) {
      y1 = srcY[src_y_idx];
      u1 = srcU[src_uv_idx / 2];
      v1 = srcV[src_uv_idx / 2];

      destY[dest_y_idx] = CLAMP (APPLY_MATRIX (matrix, 0, y1, u1, v1), 0, 255);
      destU[dest_uv_idx] = CLAMP (
          (destU[dest_uv_idx] + APPLY_MATRIX (matrix, 1, y1, u1, v1)) / 2, 0,
          255);
      destV[dest_uv_idx] =
          CLAMP ((destV[dest_uv_idx] + APPLY_MATRIX (matrix, 1, y1, u1,
                  v1)) / 2, 0, 255);
    } else if (j == w - 1) {
      y1 = srcY[src_y_idx];
      u1 = srcU[src_uv_idx / 2];
      v1 = srcV[src_uv_idx / 2];

      destY[dest_y_idx] = CLAMP (APPLY_MATRIX (matrix, 0, y1, u1, v1), 0, 255);
      destU[dest_uv_idx] = CLAMP (
          (3 * destU[dest_uv_idx] + APPLY_MATRIX (matrix, 1, y1, u1, v1)) / 4,
          0, 255);
      destV[dest_uv_idx] =
          CLAMP ((3 * destV[dest_uv_idx] + APPLY_MATRIX (matrix, 1, y1, u1,
                  v1)) / 4, 0, 255);
    }
  } else if (i == h - 1) {
    /* 3.1. Handle the first destination pixel if it doesn't
     *      start at the macro pixel boundary, i.e. blend with
     *      the background! */
    if (dest_x % 2 == 1) {
      y1 = srcY[0];
      u1 = srcU[0];
      v1 = srcV[0];

      destY[0] = CLAMP (APPLY_MATRIX (matrix, 0, y1, u1, v1), 0, 255);
      destU[0] =
          CLAMP ((3 * destU[0] + APPLY_MATRIX (matrix, 1, y1, u1, v1)) / 4, 0,
          255);
      destV[0] =
          CLAMP ((3 * destV[0] + APPLY_MATRIX (matrix, 2, y1, u1, v1)) / 4, 0,
          255);

      j = dest_x + 1;
      src_y_idx = dest_y_idx = dest_uv_idx = 1;
      src_uv_idx = (src_x % 2) + 1;
    } else {
      j = dest_x;
      src_y_idx = dest_y_idx = dest_uv_idx = 0;
      src_uv_idx = (src_x % 2);
    }

    /* 3.2. Copy all macro pixels from the source to the destination
     *      but blend with the background because we're only filling
     *      the upper part of the macro pixels. */
    for (; j < w - 1; j += 2) {
      y1 = srcY[src_y_idx];
      y2 = srcY[src_y_idx + 1];

      u1 = srcU[src_uv_idx / 2];
      v1 = srcV[src_uv_idx / 2];
      src_uv_idx++;
      u2 = srcU[src_uv_idx / 2];
      v2 = srcV[src_uv_idx / 2];
      src_uv_idx++;

      destY[dest_y_idx] = CLAMP (APPLY_MATRIX (matrix, 0, y1, u1, v1), 0, 255);
      destY[dest_y_idx + 1] =
          CLAMP (APPLY_MATRIX (matrix, 0, y2, u2, v2), 0, 255);

      destU[dest_uv_idx] = CLAMP (
          (2 * destU[dest_uv_idx] + APPLY_MATRIX (matrix, 1, y1, u1,
                  v1) + APPLY_MATRIX (matrix, 1, y2, u2, v2)) / 4, 0, 255);
      destV[dest_uv_idx] = CLAMP (
          (2 * destV[dest_uv_idx] + APPLY_MATRIX (matrix, 2, y1, u1,
                  v1) + APPLY_MATRIX (matrix, 2, y2, u2, v2)) / 4, 0, 255);

      dest_y_idx += 2;
      src_y_idx += 2;
      dest_uv_idx++;
    }

    /* 3.3. Now copy the last pixel if one exists and blend it
     *      with the background because we only fill part of
     *      the macro pixel. In case this is the last pixel of
     *      the destination we will a larger part. */
    if (j == w - 1 && j == dest_width - 1) {
      y1 = srcY[src_y_idx];
      u1 = srcU[src_uv_idx / 2];
      v1 = srcV[src_uv_idx / 2];

      destY[dest_y_idx] = CLAMP (APPLY_MATRIX (matrix, 0, y1, u1, v1), 0, 255);
      destU[dest_uv_idx] = CLAMP (
          (destU[dest_uv_idx] + APPLY_MATRIX (matrix, 1, y1, u1, v1)) / 2, 0,
          255);
      destV[dest_uv_idx] =
          CLAMP ((destV[dest_uv_idx] + APPLY_MATRIX (matrix, 1, y1, u1,
                  v1)) / 2, 0, 255);
    } else if (j == w - 1) {
      y1 = srcY[src_y_idx];
      u1 = srcU[src_uv_idx / 2];
      v1 = srcV[src_uv_idx / 2];

      destY[dest_y_idx] = CLAMP (APPLY_MATRIX (matrix, 0, y1, u1, v1), 0, 255);
      destU[dest_uv_idx] = CLAMP (
          (3 * destU[dest_uv_idx] + APPLY_MATRIX (matrix, 1, y1, u1, v1)) / 4,
          0, 255);
      destV[dest_uv_idx] =
          CLAMP ((3 * destV[dest_uv_idx] + APPLY_MATRIX (matrix, 1, y1, u1,
                  v1)) / 4, 0, 255);
1149 1150 1151 1152 1153
    }
  }
}

static void
1154 1155 1156 1157 1158
copy_i420_ayuv (guint i_alpha, GstVideoFormat dest_format, guint8 * dest,
    gboolean dest_sdtv, gint dest_width, gint dest_height, gint dest_x,
    gint dest_y, GstVideoFormat src_format, const guint8 * src,
    gboolean src_sdtv, gint src_width, gint src_height, gint src_x, gint src_y,
    gint w, gint h)
1159 1160 1161
{
  const guint8 *srcY, *srcU, *srcV;
  gint src_strideY, src_strideUV;
1162
  gint dest_stride;
1163 1164 1165 1166 1167 1168 1169 1170 1171 1172

  src_strideY =
      gst_video_format_get_row_stride (GST_VIDEO_FORMAT_I420, 0, src_width);
  src_strideUV =
      gst_video_format_get_row_stride (GST_VIDEO_FORMAT_I420, 1, src_width);

  srcY =
      src + gst_video_format_get_component_offset (GST_VIDEO_FORMAT_I420, 0,
      src_width, src_height);
  srcU =
1173
      src + gst_video_format_get_component_offset (src_format, 1,
1174 1175
      src_width, src_height);
  srcV =
1176
      src + gst_video_format_get_component_offset (src_format, 2,
1177 1178
      src_width, src_height);

1179
  dest_stride = dest_width * 4;
1180

1181
  dest = dest + dest_y * dest_stride + dest_x * 4;
1182 1183 1184 1185 1186

  srcY = srcY + src_y * src_strideY + src_x;
  srcU = srcU + (src_y / 2) * src_strideUV + src_x / 2;
  srcV = srcV + (src_y / 2) * src_strideUV + src_x / 2;

1187 1188
  i_alpha = CLAMP (i_alpha, 0, 255);

1189
  if (src_sdtv != dest_sdtv) {
1190 1191 1192
    gint i, j, uv_idx;
    gint y, u, v;
    gint y1, u1, v1;
1193 1194 1195 1196 1197 1198
    gint matrix[12];

    memcpy (matrix,
        dest_sdtv ? cog_ycbcr_hdtv_to_ycbcr_sdtv_matrix_8bit :
        cog_ycbcr_sdtv_to_ycbcr_hdtv_matrix_8bit, 12 * sizeof (gint));

1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212
    for (i = 0; i < h; i++) {
      for (j = 0, uv_idx = src_x % 2; j < w; j++, uv_idx++) {
        y = srcY[j];
        u = srcU[uv_idx / 2];
        v = srcV[uv_idx / 2];

        y1 = APPLY_MATRIX (matrix, 0, y, u, v);
        u1 = APPLY_MATRIX (matrix, 1, y, u, v);
        v1 = APPLY_MATRIX (matrix, 2, y, u, v);

        dest[4 * j + 0] = i_alpha;
        dest[4 * j + 1] = y1;
        dest[4 * j + 2] = u1;
        dest[4 * j + 3] = v1;
1213
      }
1214
      dest += dest_stride;
1215

1216
      src_y++;
1217
      srcY += src_strideY;
1218 1219 1220 1221
      if (src_y % 2 == 0) {
        srcU += src_strideUV;
        srcV += src_strideUV;
      }
1222 1223
    }
  } else {
1224 1225
    gint i, j, uv_idx;
    gint y, u, v;
1226

1227 1228 1229 1230 1231 1232 1233 1234 1235 1236
    for (i = 0; i < h; i++) {
      for (j = 0, uv_idx = src_x % 2; j < w; j++, uv_idx++) {
        y = srcY[j];
        u = srcU[uv_idx / 2];
        v = srcV[uv_idx / 2];

        dest[4 * j + 0] = i_alpha;
        dest[4 * j + 1] = y;
        dest[4 * j + 2] = u;
        dest[4 * j + 3] = v;
1237
      }
1238
      dest += dest_stride;
1239

1240
      src_y++;
1241
      srcY += src_strideY;
1242 1243 1244 1245
      if (src_y % 2 == 0) {
        srcU += src_strideUV;
        srcV += src_strideUV;
      }
1246 1247 1248 1249
    }
  }
}

1250 1251 1252 1253 1254 1255
static void
fill_rgb32 (GstVideoBoxFill fill_type, guint b_alpha, GstVideoFormat format,
    guint8 * dest, gboolean sdtv, gint width, gint height)
{
  guint32 empty_pixel;
  gint p[4];
1256

1257 1258 1259 1260
  p[0] = gst_video_format_get_component_offset (format, 3, width, height);
  p[1] = gst_video_format_get_component_offset (format, 0, width, height);
  p[2] = gst_video_format_get_component_offset (format, 1, width, height);
  p[3] = gst_video_format_get_component_offset (format, 2, width, height);
1261

1262 1263
  b_alpha = CLAMP (b_alpha, 0, 255);

1264
  empty_pixel = GUINT32_FROM_LE ((b_alpha << (p[0] * 8)) |
1265 1266 1267
      (rgb_colors_R[fill_type] << (p[1] * 8)) |
      (rgb_colors_G[fill_type] << (p[2] * 8)) |
      (rgb_colors_B[fill_type] << (p[3] * 8)));
1268 1269 1270 1271

  oil_splat_u32_ns ((guint32 *) dest, &empty_pixel, width * height);
}

1272 1273 1274 1275 1276 1277 1278 1279
static void
fill_rgb24 (GstVideoBoxFill fill_type, guint b_alpha, GstVideoFormat format,
    guint8 * dest, gboolean sdtv, gint width, gint height)
{
  gint dest_stride = GST_ROUND_UP_4 (width * 3);
  gint p[4];
  gint i, j;

1280 1281 1282 1283
  p[0] = gst_video_format_get_component_offset (format, 3, width, height);
  p[1] = gst_video_format_get_component_offset (format, 0, width, height);
  p[2] = gst_video_format_get_component_offset (format, 1, width, height);
  p[3] = gst_video_format_get_component_offset (format, 2, width, height);
1284 1285 1286

  for (i = 0; i < height; i++) {
    for (j = 0; j < width; j++) {
Sebastian Dröge's avatar
Sebastian Dröge committed
1287 1288 1289
      dest[3 * j + p[1]] = rgb_colors_R[fill_type];
      dest[3 * j + p[2]] = rgb_colors_G[fill_type];
      dest[3 * j + p[3]] = rgb_colors_B[fill_type];
1290 1291 1292 1293 1294
    }
    dest += dest_stride;
  }
}

1295 1296 1297 1298 1299 1300 1301 1302
static void
copy_rgb32 (guint i_alpha, GstVideoFormat dest_format, guint8 * dest,
    gboolean dest_sdtv, gint dest_width, gint dest_height, gint dest_x,
    gint dest_y, GstVideoFormat src_format, const guint8 * src,
    gboolean src_sdtv, gint src_width, gint src_height, gint src_x, gint src_y,
    gint w, gint h)
{
  gint i, j;
1303
  gint src_stride, dest_stride;
1304
  gboolean in_alpha, out_alpha;
1305
  gint in_bpp, out_bpp;
1306 1307
  gint p_out[4];
  gint p_in[4];
1308 1309 1310 1311
  gboolean packed_out = (dest_format == GST_VIDEO_FORMAT_RGB
      || dest_format == GST_VIDEO_FORMAT_BGR);
  gboolean packed_in = (src_format == GST_VIDEO_FORMAT_RGB
      || src_format == GST_VIDEO_FORMAT_BGR);
1312

1313 1314 1315 1316
  src_stride = (packed_in) ? GST_ROUND_UP_4 (3 * src_width) : 4 * src_width;
  dest_stride = (packed_out) ? GST_ROUND_UP_4 (3 * dest_width) : 4 * dest_width;
  in_bpp = (packed_in) ? 3 : 4;
  out_bpp = (packed_out) ? 3 : 4;
1317

1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344
  out_alpha = gst_video_format_has_alpha (dest_format);
  p_out[0] =
      gst_video_format_get_component_offset (dest_format, 3, dest_width,
      dest_height);
  p_out[1] =
      gst_video_format_get_component_offset (dest_format, 0, dest_width,
      dest_height);
  p_out[2] =
      gst_video_format_get_component_offset (dest_format, 1, dest_width,
      dest_height);
  p_out[3] =
      gst_video_format_get_component_offset (dest_format, 2, dest_width,
      dest_height);

  in_alpha = gst_video_format_has_alpha (src_format);
  p_in[0] =
      gst_video_format_get_component_offset (src_format, 3, src_width,
      src_height);
  p_in[1] =
      gst_video_format_get_component_offset (src_format, 0, src_width,
      src_height);
  p_in[2] =
      gst_video_format_get_component_offset (src_format, 1, src_width,
      src_height);
  p_in[3] =
      gst_video_format_get_component_offset (src_format, 2, src_width,
      src_height);
1345

1346 1347
  dest = dest + dest_y * dest_stride + dest_x * out_bpp;
  src = src + src_y * src_stride + src_x * in_bpp;
1348 1349

  if (in_alpha && out_alpha) {
1350
    w *= 4;
1351 1352 1353 1354 1355 1356 1357 1358 1359 1360
    for (i = 0; i < h; i++) {
      for (j = 0; j < w; j += 4) {
        dest[j + p_out[0]] = (src[j + p_in[0]] * i_alpha) >> 8;
        dest[j + p_out[1]] = src[j + p_in[1]];
        dest[j + p_out[2]] = src[j + p_in[2]];
        dest[j + p_out[3]] = src[j + p_in[3]];
      }
      dest += dest_stride;
      src += src_stride;
    }
1361 1362
  } else if (out_alpha && !packed_in) {
    w *= 4;
1363 1364
    i_alpha = CLAMP (i_alpha, 0, 255);

1365 1366
    for (i = 0; i < h; i++) {
      for (j = 0; j < w; j += 4) {
1367
        dest[j + p_out[0]] = i_alpha;
1368 1369 1370 1371 1372 1373 1374
        dest[j + p_out[1]] = src[j + p_in[1]];
        dest[j + p_out[2]] = src[j + p_in[2]];
        dest[j + p_out[3]] = src[j + p_in[3]];
      }
      dest += dest_stride;
      src += src_stride;
    }
1375
  } else if (out_alpha && packed_in) {
1376 1377
    i_alpha = CLAMP (i_alpha, 0, 255);

1378 1379
    for (i = 0; i < h; i++) {
      for (j = 0; j < w; j++) {
1380
        dest[4 * j + p_out[0]] = i_alpha;
1381 1382 1383 1384 1385 1386 1387
        dest[4 * j + p_out[1]] = src[in_bpp * j + p_in[1]];
        dest[4 * j + p_out[2]] = src[in_bpp * j + p_in[2]];
        dest[4 * j + p_out[3]] = src[in_bpp * j + p_in[3]];
      }
      dest += dest_stride;
      src += src_stride;
    }
1388 1389
  } else if (!packed_out && !packed_in) {
    w *= 4;
1390 1391 1392 1393 1394 1395 1396 1397 1398
    for (i = 0; i < h; i++) {
      for (j = 0; j < w; j += 4) {
        dest[j + p_out[1]] = src[j + p_in[1]];
        dest[j + p_out[2]] = src[j + p_in[2]];
        dest[j + p_out[3]] = src[j + p_in[3]];
      }
      dest += dest_stride;
      src += src_stride;
    }
1399 1400 1401 1402 1403 1404 1405 1406 1407 1408
  } else {
    for (i = 0; i < h; i++) {
      for (j = 0; j < w; j++) {
        dest[out_bpp * j + p_out[1]] = src[in_bpp * j + p_in[1]];
        dest[out_bpp * j + p_out[2]] = src[in_bpp * j + p_in[2]];
        dest[out_bpp * j + p_out[3]] = src[in_bpp * j + p_in[3]];
      }
      dest += dest_stride;
      src += src_stride;
    }
1409 1410 1411
  }
}

1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434
static void
copy_rgb32_ayuv (guint i_alpha, GstVideoFormat dest_format, guint8 * dest,
    gboolean dest_sdtv, gint dest_width, gint dest_height, gint dest_x,
    gint dest_y, GstVideoFormat src_format, const guint8 * src,
    gboolean src_sdtv, gint src_width, gint src_height, gint src_x, gint src_y,
    gint w, gint h)
{
  gint i, j;
  gint src_stride, dest_stride;
  gboolean in_alpha;
  gint in_bpp;
  gint p_in[4];
  gboolean packed_in = (src_format == GST_VIDEO_FORMAT_RGB
      || src_format == GST_VIDEO_FORMAT_BGR);
  gint matrix[12];
  gint a;
  gint y, u, v;
  gint r, g, b;

  src_stride = (packed_in) ? GST_ROUND_UP_4 (3 * src_width) : 4 * src_width;
  dest_stride = 4 * dest_width;
  in_bpp = (packed_in) ? 3 : 4;

1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447
  in_alpha = gst_video_format_has_alpha (src_format);
  p_in[0] =
      gst_video_format_get_component_offset (src_format, 3, src_width,
      src_height);
  p_in[1] =
      gst_video_format_get_component_offset (src_format, 0, src_width,
      src_height);
  p_in[2] =
      gst_video_format_get_component_offset (src_format, 1, src_width,
      src_height);
  p_in[3] =
      gst_video_format_get_component_offset (src_format, 2, src_width,
      src_height);
1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478

  memcpy (matrix,
      (dest_sdtv) ? cog_rgb_to_ycbcr_matrix_8bit_sdtv :
      cog_rgb_to_ycbcr_matrix_8bit_hdtv, 12 * sizeof (gint));

  dest = dest + dest_y * dest_stride + dest_x * 4;
  src = src + src_y * src_stride + src_x * in_bpp;

  if (in_alpha) {
    w *= 4;
    for (i = 0; i < h; i++) {
      for (j = 0; j < w; j += 4) {
        a = (src[j + p_in[0]] * i_alpha) >> 8;
        r = src[j + p_in[1]];
        g = src[j + p_in[2]];
        b = src[j + p_in[3]];

        y = APPLY_MATRIX (matrix, 0, r, g, b);
        u = APPLY_MATRIX (matrix, 1, r, g, b);
        v = APPLY_MATRIX (matrix, 2, r, g, b);

        dest[j + 0] = a;
        dest[j + 1] = CLAMP (y, 0, 255);
        dest[j + 2] = CLAMP (u, 0, 255);
        dest[j + 3] = CLAMP (v, 0, 255);
      }
      dest += dest_stride;
      src += src_stride;
    }
  } else if (!packed_in) {
    w *= 4;
1479 1480
    i_alpha = CLAMP (i_alpha, 0, 255);

1481 1482
    for (i = 0; i < h; i++) {
      for (j = 0; j < w; j += 4) {
1483
        a = i_alpha;
1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498 1499 1500
        r = src[j + p_in[1]];
        g = src[j + p_in[2]];
        b = src[j + p_in[3]];

        y = APPLY_MATRIX (matrix, 0, r, g, b);
        u = APPLY_MATRIX (matrix, 1, r, g, b);
        v = APPLY_MATRIX (matrix, 2, r, g, b);

        dest[j + 0] = a;
        dest[j + 1] = CLAMP (y, 0, 255);
        dest[j + 2] = CLAMP (u, 0, 255);
        dest[j + 3] = CLAMP (v, 0, 255);
      }
      dest += dest_stride;
      src += src_stride;
    }
  } else {
1501 1502
    i_alpha = CLAMP (i_alpha, 0, 255);

1503 1504
    for (i = 0; i < h; i++) {
      for (j = 0; j < w; j++) {
1505
        a = i_alpha;
1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547
        r = src[in_bpp * j + p_in[1]];
        g = src[in_bpp * j + p_in[2]];
        b = src[in_bpp * j + p_in[3]];

        y = APPLY_MATRIX (matrix, 0, r, g, b);
        u = APPLY_MATRIX (matrix, 1, r, g, b);
        v = APPLY_MATRIX (matrix, 2, r, g, b);

        dest[4 * j + 0] = a;
        dest[4 * j + 1] = CLAMP (y, 0, 255);
        dest[4 * j + 2] = CLAMP (u, 0, 255);
        dest[4 * j + 3] = CLAMP (v, 0, 255);
      }
      dest += dest_stride;
      src += src_stride;
    }
  }
}

static void
copy_ayuv_rgb32 (guint i_alpha, GstVideoFormat dest_format, guint8 * dest,
    gboolean dest_sdtv, gint dest_width, gint dest_height, gint dest_x,
    gint dest_y, GstVideoFormat src_format, const guint8 * src,
    gboolean src_sdtv, gint src_width, gint src_height, gint src_x, gint src_y,
    gint w, gint h)
{
  gint i, j;
  gint src_stride, dest_stride;
  gboolean out_alpha;
  gint out_bpp;
  gint p_out[4];
  gboolean packed_out = (dest_format == GST_VIDEO_FORMAT_RGB
      || dest_format == GST_VIDEO_FORMAT_BGR);
  gint matrix[12];
  gint a;
  gint y, u, v;
  gint r, g, b;

  dest_stride = (packed_out) ? GST_ROUND_UP_4 (3 * dest_width) : 4 * dest_width;
  src_stride = 4 * src_width;
  out_bpp = (packed_out) ? 3 : 4;

1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559 1560
  out_alpha = gst_video_format_has_alpha (dest_format);
  p_out[0] =
      gst_video_format_get_component_offset (dest_format, 3, dest_width,
      dest_height);
  p_out[1] =
      gst_video_format_get_component_offset (dest_format, 0, dest_width,
      dest_height);
  p_out[2] =
      gst_video_format_get_component_offset (dest_format, 1, dest_width,
      dest_height);
  p_out[3] =
      gst_video_format_get_component_offset (dest_format, 2, dest_width,
      dest_height);
1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629

  memcpy (matrix,
      (src_sdtv) ? cog_ycbcr_to_rgb_matrix_8bit_sdtv :
      cog_ycbcr_to_rgb_matrix_8bit_hdtv, 12 * sizeof (gint));

  dest = dest + dest_y * dest_stride + dest_x * out_bpp;
  src = src + src_y * src_stride + src_x * 4;

  if (out_alpha) {
    w *= 4;
    for (i = 0; i < h; i++) {
      for (j = 0; j < w; j += 4) {
        a = (src[j + 0] * i_alpha) >> 8;
        y = src[j + 1];
        u = src[j + 2];
        v = src[j + 3];

        r = APPLY_MATRIX (matrix, 0, y, u, v);
        g = APPLY_MATRIX (matrix, 1, y, u, v);
        b = APPLY_MATRIX (matrix, 2, y, u, v);

        dest[j + p_out[0]] = a;
        dest[j + p_out[1]] = CLAMP (r, 0, 255);
        dest[j + p_out[2]] = CLAMP (g, 0, 255);
        dest[j + p_out[3]] = CLAMP (b, 0, 255);
      }
      dest += dest_stride;
      src += src_stride;
    }
  } else if (!packed_out) {
    w *= 4;
    for (i = 0; i < h; i++) {
      for (j = 0; j < w; j += 4) {
        y = src[j + 1];
        u = src[j + 2];
        v = src[j + 3];

        r = APPLY_MATRIX (matrix, 0, y, u, v);
        g = APPLY_MATRIX (matrix, 1, y, u, v);
        b = APPLY_MATRIX (matrix, 2, y, u, v);

        dest[j + p_out[1]] = CLAMP (r, 0, 255);
        dest[j + p_out[2]] = CLAMP (g, 0, 255);
        dest[j + p_out[3]] = CLAMP (b, 0, 255);
      }
      dest += dest_stride;
      src += src_stride;
    }
  } else {
    for (i = 0; i < h; i++) {
      for (j = 0; j < w; j++) {
        y = src[4 * j + 1];
        u = src[4 * j + 2];
        v = src[4 * j + 3];

        r = APPLY_MATRIX (matrix, 0, y, u, v);
        g = APPLY_MATRIX (matrix, 1, y, u, v);
        b = APPLY_MATRIX (matrix, 2, y, u, v);

        dest[out_bpp * j + p_out[1]] = CLAMP (r, 0, 255);
        dest[out_bpp * j + p_out[2]] = CLAMP (g, 0, 255);
        dest[out_bpp * j + p_out[3]] = CLAMP (b, 0, 255);
      }
      dest += dest_stride;
      src += src_stride;
    }
  }
}

1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691 1692
static void
fill_gray (GstVideoBoxFill fill_type, guint b_alpha, GstVideoFormat format,
    guint8 * dest, gboolean sdtv, gint width, gint height)
{
  gint i, j;
  gint dest_stride;

  if (format == GST_VIDEO_FORMAT_GRAY8) {
    guint8 val = yuv_sdtv_colors_Y[fill_type];

    dest_stride = GST_ROUND_UP_4 (width);
    for (i = 0; i < height; i++) {
      oil_splat_u8_ns (dest, &val, width);
      dest += dest_stride;
    }
  } else {
    guint16 val = yuv_sdtv_colors_Y[fill_type] << 8;

    dest_stride = GST_ROUND_UP_4 (width * 2);
    if (format == GST_VIDEO_FORMAT_GRAY16_BE) {
      for (i = 0; i < height; i++) {
        for (j = 0; j < width; j++) {
          GST_WRITE_UINT16_BE (dest + 2 * j, val);
        }
        dest += dest_stride;
      }
    } else {
      for (i = 0; i < height; i++) {
        for (j = 0; j < width; j++) {
          GST_WRITE_UINT16_LE (dest + 2 * j, val);
        }
        dest += dest_stride;
      }
    }
  }
}

static void
copy_packed_simple (guint i_alpha, GstVideoFormat dest_format, guint8 * dest,
    gboolean dest_sdtv, gint dest_width, gint dest_height, gint dest_x,
    gint dest_y, GstVideoFormat src_format, const guint8 * src,
    gboolean src_sdtv, gint src_width, gint src_height, gint src_x, gint src_y,
    gint w, gint h)
{
  gint i;
  gint src_stride, dest_stride;
  gint pixel_stride, row_size;

  src_stride = gst_video_format_get_row_stride (src_format, 0, src_width);
  dest_stride = gst_video_format_get_row_stride (dest_format, 0, dest_width);
  pixel_stride = gst_video_format_get_pixel_stride (dest_format, 0);
  row_size = w * pixel_stride;

  dest = dest + dest_y * dest_stride + dest_x * pixel_stride;
  src = src + src_y * src_stride + src_x * pixel_stride;

  for (i = 0; i < h; i++) {
    oil_copy_u8 (dest, src, row_size);
    dest += dest_stride;
    src += src_stride;
  }
}

1693 1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715
static void
fill_yuy2 (GstVideoBoxFill fill_type, guint b_alpha, GstVideoFormat format,
    guint8 * dest, gboolean sdtv, gint width, gint height)
{
  guint8 y, u, v