summaryrefslogtreecommitdiffstats
path: root/media/libvpx/vp9/common/vp9_convolve.c
blob: 90e337fd66e13b8f649dd232e4a64a0c3a2341ae (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
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
/*
 *  Copyright (c) 2013 The WebM project authors. All Rights Reserved.
 *
 *  Use of this source code is governed by a BSD-style license
 *  that can be found in the LICENSE file in the root of the source
 *  tree. An additional intellectual property rights grant can be found
 *  in the file PATENTS.  All contributing project authors may
 *  be found in the AUTHORS file in the root of the source tree.
 */

#include <assert.h>

#include "./vpx_config.h"
#include "./vp9_rtcd.h"
#include "vp9/common/vp9_common.h"
#include "vp9/common/vp9_convolve.h"
#include "vp9/common/vp9_filter.h"
#include "vpx/vpx_integer.h"
#include "vpx_ports/mem.h"

static void convolve_horiz(const uint8_t *src, ptrdiff_t src_stride,
                           uint8_t *dst, ptrdiff_t dst_stride,
                           const InterpKernel *x_filters,
                           int x0_q4, int x_step_q4, int w, int h) {
  int x, y;
  src -= SUBPEL_TAPS / 2 - 1;
  for (y = 0; y < h; ++y) {
    int x_q4 = x0_q4;
    for (x = 0; x < w; ++x) {
      const uint8_t *const src_x = &src[x_q4 >> SUBPEL_BITS];
      const int16_t *const x_filter = x_filters[x_q4 & SUBPEL_MASK];
      int k, sum = 0;
      for (k = 0; k < SUBPEL_TAPS; ++k)
        sum += src_x[k] * x_filter[k];
      dst[x] = clip_pixel(ROUND_POWER_OF_TWO(sum, FILTER_BITS));
      x_q4 += x_step_q4;
    }
    src += src_stride;
    dst += dst_stride;
  }
}

static void convolve_avg_horiz(const uint8_t *src, ptrdiff_t src_stride,
                               uint8_t *dst, ptrdiff_t dst_stride,
                               const InterpKernel *x_filters,
                               int x0_q4, int x_step_q4, int w, int h) {
  int x, y;
  src -= SUBPEL_TAPS / 2 - 1;
  for (y = 0; y < h; ++y) {
    int x_q4 = x0_q4;
    for (x = 0; x < w; ++x) {
      const uint8_t *const src_x = &src[x_q4 >> SUBPEL_BITS];
      const int16_t *const x_filter = x_filters[x_q4 & SUBPEL_MASK];
      int k, sum = 0;
      for (k = 0; k < SUBPEL_TAPS; ++k)
        sum += src_x[k] * x_filter[k];
      dst[x] = ROUND_POWER_OF_TWO(dst[x] +
          clip_pixel(ROUND_POWER_OF_TWO(sum, FILTER_BITS)), 1);
      x_q4 += x_step_q4;
    }
    src += src_stride;
    dst += dst_stride;
  }
}

static void convolve_vert(const uint8_t *src, ptrdiff_t src_stride,
                          uint8_t *dst, ptrdiff_t dst_stride,
                          const InterpKernel *y_filters,
                          int y0_q4, int y_step_q4, int w, int h) {
  int x, y;
  src -= src_stride * (SUBPEL_TAPS / 2 - 1);

  for (x = 0; x < w; ++x) {
    int y_q4 = y0_q4;
    for (y = 0; y < h; ++y) {
      const unsigned char *src_y = &src[(y_q4 >> SUBPEL_BITS) * src_stride];
      const int16_t *const y_filter = y_filters[y_q4 & SUBPEL_MASK];
      int k, sum = 0;
      for (k = 0; k < SUBPEL_TAPS; ++k)
        sum += src_y[k * src_stride] * y_filter[k];
      dst[y * dst_stride] = clip_pixel(ROUND_POWER_OF_TWO(sum, FILTER_BITS));
      y_q4 += y_step_q4;
    }
    ++src;
    ++dst;
  }
}

static void convolve_avg_vert(const uint8_t *src, ptrdiff_t src_stride,
                              uint8_t *dst, ptrdiff_t dst_stride,
                              const InterpKernel *y_filters,
                              int y0_q4, int y_step_q4, int w, int h) {
  int x, y;
  src -= src_stride * (SUBPEL_TAPS / 2 - 1);

  for (x = 0; x < w; ++x) {
    int y_q4 = y0_q4;
    for (y = 0; y < h; ++y) {
      const unsigned char *src_y = &src[(y_q4 >> SUBPEL_BITS) * src_stride];
      const int16_t *const y_filter = y_filters[y_q4 & SUBPEL_MASK];
      int k, sum = 0;
      for (k = 0; k < SUBPEL_TAPS; ++k)
        sum += src_y[k * src_stride] * y_filter[k];
      dst[y * dst_stride] = ROUND_POWER_OF_TWO(dst[y * dst_stride] +
          clip_pixel(ROUND_POWER_OF_TWO(sum, FILTER_BITS)), 1);
      y_q4 += y_step_q4;
    }
    ++src;
    ++dst;
  }
}

static void convolve(const uint8_t *src, ptrdiff_t src_stride,
                     uint8_t *dst, ptrdiff_t dst_stride,
                     const InterpKernel *const x_filters,
                     int x0_q4, int x_step_q4,
                     const InterpKernel *const y_filters,
                     int y0_q4, int y_step_q4,
                     int w, int h) {
  // Note: Fixed size intermediate buffer, temp, places limits on parameters.
  // 2d filtering proceeds in 2 steps:
  //   (1) Interpolate horizontally into an intermediate buffer, temp.
  //   (2) Interpolate temp vertically to derive the sub-pixel result.
  // Deriving the maximum number of rows in the temp buffer (135):
  // --Smallest scaling factor is x1/2 ==> y_step_q4 = 32 (Normative).
  // --Largest block size is 64x64 pixels.
  // --64 rows in the downscaled frame span a distance of (64 - 1) * 32 in the
  //   original frame (in 1/16th pixel units).
  // --Must round-up because block may be located at sub-pixel position.
  // --Require an additional SUBPEL_TAPS rows for the 8-tap filter tails.
  // --((64 - 1) * 32 + 15) >> 4 + 8 = 135.
  uint8_t temp[135 * 64];
  int intermediate_height =
          (((h - 1) * y_step_q4 + y0_q4) >> SUBPEL_BITS) + SUBPEL_TAPS;

  assert(w <= 64);
  assert(h <= 64);
  assert(y_step_q4 <= 32);
  assert(x_step_q4 <= 32);

  convolve_horiz(src - src_stride * (SUBPEL_TAPS / 2 - 1), src_stride, temp, 64,
                 x_filters, x0_q4, x_step_q4, w, intermediate_height);
  convolve_vert(temp + 64 * (SUBPEL_TAPS / 2 - 1), 64, dst, dst_stride,
                y_filters, y0_q4, y_step_q4, w, h);
}

static const InterpKernel *get_filter_base(const int16_t *filter) {
  // NOTE: This assumes that the filter table is 256-byte aligned.
  // TODO(agrange) Modify to make independent of table alignment.
  return (const InterpKernel *)(((intptr_t)filter) & ~((intptr_t)0xFF));
}

static int get_filter_offset(const int16_t *f, const InterpKernel *base) {
  return (int)((const InterpKernel *)(intptr_t)f - base);
}

void vp9_convolve8_horiz_c(const uint8_t *src, ptrdiff_t src_stride,
                           uint8_t *dst, ptrdiff_t dst_stride,
                           const int16_t *filter_x, int x_step_q4,
                           const int16_t *filter_y, int y_step_q4,
                           int w, int h) {
  const InterpKernel *const filters_x = get_filter_base(filter_x);
  const int x0_q4 = get_filter_offset(filter_x, filters_x);

  (void)filter_y;
  (void)y_step_q4;

  convolve_horiz(src, src_stride, dst, dst_stride, filters_x,
                 x0_q4, x_step_q4, w, h);
}

void vp9_convolve8_avg_horiz_c(const uint8_t *src, ptrdiff_t src_stride,
                               uint8_t *dst, ptrdiff_t dst_stride,
                               const int16_t *filter_x, int x_step_q4,
                               const int16_t *filter_y, int y_step_q4,
                               int w, int h) {
  const InterpKernel *const filters_x = get_filter_base(filter_x);
  const int x0_q4 = get_filter_offset(filter_x, filters_x);

  (void)filter_y;
  (void)y_step_q4;

  convolve_avg_horiz(src, src_stride, dst, dst_stride, filters_x,
                     x0_q4, x_step_q4, w, h);
}

void vp9_convolve8_vert_c(const uint8_t *src, ptrdiff_t src_stride,
                          uint8_t *dst, ptrdiff_t dst_stride,
                          const int16_t *filter_x, int x_step_q4,
                          const int16_t *filter_y, int y_step_q4,
                          int w, int h) {
  const InterpKernel *const filters_y = get_filter_base(filter_y);
  const int y0_q4 = get_filter_offset(filter_y, filters_y);

  (void)filter_x;
  (void)x_step_q4;

  convolve_vert(src, src_stride, dst, dst_stride, filters_y,
                y0_q4, y_step_q4, w, h);
}

void vp9_convolve8_avg_vert_c(const uint8_t *src, ptrdiff_t src_stride,
                              uint8_t *dst, ptrdiff_t dst_stride,
                              const int16_t *filter_x, int x_step_q4,
                              const int16_t *filter_y, int y_step_q4,
                              int w, int h) {
  const InterpKernel *const filters_y = get_filter_base(filter_y);
  const int y0_q4 = get_filter_offset(filter_y, filters_y);

  (void)filter_x;
  (void)x_step_q4;

  convolve_avg_vert(src, src_stride, dst, dst_stride, filters_y,
                    y0_q4, y_step_q4, w, h);
}

void vp9_convolve8_c(const uint8_t *src, ptrdiff_t src_stride,
                     uint8_t *dst, ptrdiff_t dst_stride,
                     const int16_t *filter_x, int x_step_q4,
                     const int16_t *filter_y, int y_step_q4,
                     int w, int h) {
  const InterpKernel *const filters_x = get_filter_base(filter_x);
  const int x0_q4 = get_filter_offset(filter_x, filters_x);

  const InterpKernel *const filters_y = get_filter_base(filter_y);
  const int y0_q4 = get_filter_offset(filter_y, filters_y);

  convolve(src, src_stride, dst, dst_stride,
           filters_x, x0_q4, x_step_q4,
           filters_y, y0_q4, y_step_q4, w, h);
}

void vp9_convolve8_avg_c(const uint8_t *src, ptrdiff_t src_stride,
                         uint8_t *dst, ptrdiff_t dst_stride,
                         const int16_t *filter_x, int x_step_q4,
                         const int16_t *filter_y, int y_step_q4,
                         int w, int h) {
  /* Fixed size intermediate buffer places limits on parameters. */
  DECLARE_ALIGNED(16, uint8_t, temp[64 * 64]);
  assert(w <= 64);
  assert(h <= 64);

  vp9_convolve8_c(src, src_stride, temp, 64,
                  filter_x, x_step_q4, filter_y, y_step_q4, w, h);
  vp9_convolve_avg_c(temp, 64, dst, dst_stride, NULL, 0, NULL, 0, w, h);
}

void vp9_convolve_copy_c(const uint8_t *src, ptrdiff_t src_stride,
                         uint8_t *dst, ptrdiff_t dst_stride,
                         const int16_t *filter_x, int filter_x_stride,
                         const int16_t *filter_y, int filter_y_stride,
                         int w, int h) {
  int r;

  (void)filter_x;  (void)filter_x_stride;
  (void)filter_y;  (void)filter_y_stride;

  for (r = h; r > 0; --r) {
    memcpy(dst, src, w);
    src += src_stride;
    dst += dst_stride;
  }
}

void vp9_convolve_avg_c(const uint8_t *src, ptrdiff_t src_stride,
                        uint8_t *dst, ptrdiff_t dst_stride,
                        const int16_t *filter_x, int filter_x_stride,
                        const int16_t *filter_y, int filter_y_stride,
                        int w, int h) {
  int x, y;

  (void)filter_x;  (void)filter_x_stride;
  (void)filter_y;  (void)filter_y_stride;

  for (y = 0; y < h; ++y) {
    for (x = 0; x < w; ++x)
      dst[x] = ROUND_POWER_OF_TWO(dst[x] + src[x], 1);

    src += src_stride;
    dst += dst_stride;
  }
}

#if CONFIG_VP9_HIGHBITDEPTH
static void highbd_convolve_horiz(const uint8_t *src8, ptrdiff_t src_stride,
                                  uint8_t *dst8, ptrdiff_t dst_stride,
                                  const InterpKernel *x_filters,
                                  int x0_q4, int x_step_q4,
                                  int w, int h, int bd) {
  int x, y;
  uint16_t *src = CONVERT_TO_SHORTPTR(src8);
  uint16_t *dst = CONVERT_TO_SHORTPTR(dst8);
  src -= SUBPEL_TAPS / 2 - 1;
  for (y = 0; y < h; ++y) {
    int x_q4 = x0_q4;
    for (x = 0; x < w; ++x) {
      const uint16_t *const src_x = &src[x_q4 >> SUBPEL_BITS];
      const int16_t *const x_filter = x_filters[x_q4 & SUBPEL_MASK];
      int k, sum = 0;
      for (k = 0; k < SUBPEL_TAPS; ++k)
        sum += src_x[k] * x_filter[k];
      dst[x] = clip_pixel_highbd(ROUND_POWER_OF_TWO(sum, FILTER_BITS), bd);
      x_q4 += x_step_q4;
    }
    src += src_stride;
    dst += dst_stride;
  }
}

static void highbd_convolve_avg_horiz(const uint8_t *src8, ptrdiff_t src_stride,
                                      uint8_t *dst8, ptrdiff_t dst_stride,
                                      const InterpKernel *x_filters,
                                      int x0_q4, int x_step_q4,
                                      int w, int h, int bd) {
  int x, y;
  uint16_t *src = CONVERT_TO_SHORTPTR(src8);
  uint16_t *dst = CONVERT_TO_SHORTPTR(dst8);
  src -= SUBPEL_TAPS / 2 - 1;
  for (y = 0; y < h; ++y) {
    int x_q4 = x0_q4;
    for (x = 0; x < w; ++x) {
      const uint16_t *const src_x = &src[x_q4 >> SUBPEL_BITS];
      const int16_t *const x_filter = x_filters[x_q4 & SUBPEL_MASK];
      int k, sum = 0;
      for (k = 0; k < SUBPEL_TAPS; ++k)
        sum += src_x[k] * x_filter[k];
      dst[x] = ROUND_POWER_OF_TWO(dst[x] +
          clip_pixel_highbd(ROUND_POWER_OF_TWO(sum, FILTER_BITS), bd), 1);
      x_q4 += x_step_q4;
    }
    src += src_stride;
    dst += dst_stride;
  }
}

static void highbd_convolve_vert(const uint8_t *src8, ptrdiff_t src_stride,
                                 uint8_t *dst8, ptrdiff_t dst_stride,
                                 const InterpKernel *y_filters,
                                 int y0_q4, int y_step_q4, int w, int h,
                                 int bd) {
  int x, y;
  uint16_t *src = CONVERT_TO_SHORTPTR(src8);
  uint16_t *dst = CONVERT_TO_SHORTPTR(dst8);
  src -= src_stride * (SUBPEL_TAPS / 2 - 1);
  for (x = 0; x < w; ++x) {
    int y_q4 = y0_q4;
    for (y = 0; y < h; ++y) {
      const uint16_t *src_y = &src[(y_q4 >> SUBPEL_BITS) * src_stride];
      const int16_t *const y_filter = y_filters[y_q4 & SUBPEL_MASK];
      int k, sum = 0;
      for (k = 0; k < SUBPEL_TAPS; ++k)
        sum += src_y[k * src_stride] * y_filter[k];
      dst[y * dst_stride] = clip_pixel_highbd(
          ROUND_POWER_OF_TWO(sum, FILTER_BITS), bd);
      y_q4 += y_step_q4;
    }
    ++src;
    ++dst;
  }
}

static void highbd_convolve_avg_vert(const uint8_t *src8, ptrdiff_t src_stride,
                                     uint8_t *dst8, ptrdiff_t dst_stride,
                                     const InterpKernel *y_filters,
                                     int y0_q4, int y_step_q4, int w, int h,
                                     int bd) {
  int x, y;
  uint16_t *src = CONVERT_TO_SHORTPTR(src8);
  uint16_t *dst = CONVERT_TO_SHORTPTR(dst8);
  src -= src_stride * (SUBPEL_TAPS / 2 - 1);
  for (x = 0; x < w; ++x) {
    int y_q4 = y0_q4;
    for (y = 0; y < h; ++y) {
      const uint16_t *src_y = &src[(y_q4 >> SUBPEL_BITS) * src_stride];
      const int16_t *const y_filter = y_filters[y_q4 & SUBPEL_MASK];
      int k, sum = 0;
      for (k = 0; k < SUBPEL_TAPS; ++k)
        sum += src_y[k * src_stride] * y_filter[k];
      dst[y * dst_stride] = ROUND_POWER_OF_TWO(dst[y * dst_stride] +
          clip_pixel_highbd(ROUND_POWER_OF_TWO(sum, FILTER_BITS), bd), 1);
      y_q4 += y_step_q4;
    }
    ++src;
    ++dst;
  }
}

static void highbd_convolve(const uint8_t *src, ptrdiff_t src_stride,
                            uint8_t *dst, ptrdiff_t dst_stride,
                            const InterpKernel *const x_filters,
                            int x0_q4, int x_step_q4,
                            const InterpKernel *const y_filters,
                            int y0_q4, int y_step_q4,
                            int w, int h, int bd) {
  // Note: Fixed size intermediate buffer, temp, places limits on parameters.
  // 2d filtering proceeds in 2 steps:
  //   (1) Interpolate horizontally into an intermediate buffer, temp.
  //   (2) Interpolate temp vertically to derive the sub-pixel result.
  // Deriving the maximum number of rows in the temp buffer (135):
  // --Smallest scaling factor is x1/2 ==> y_step_q4 = 32 (Normative).
  // --Largest block size is 64x64 pixels.
  // --64 rows in the downscaled frame span a distance of (64 - 1) * 32 in the
  //   original frame (in 1/16th pixel units).
  // --Must round-up because block may be located at sub-pixel position.
  // --Require an additional SUBPEL_TAPS rows for the 8-tap filter tails.
  // --((64 - 1) * 32 + 15) >> 4 + 8 = 135.
  uint16_t temp[64 * 135];
  int intermediate_height =
          (((h - 1) * y_step_q4 + y0_q4) >> SUBPEL_BITS) + SUBPEL_TAPS;

  assert(w <= 64);
  assert(h <= 64);
  assert(y_step_q4 <= 32);
  assert(x_step_q4 <= 32);

  highbd_convolve_horiz(src - src_stride * (SUBPEL_TAPS / 2 - 1),
                        src_stride, CONVERT_TO_BYTEPTR(temp), 64,
                        x_filters, x0_q4, x_step_q4, w,
                        intermediate_height, bd);
  highbd_convolve_vert(CONVERT_TO_BYTEPTR(temp) + 64 * (SUBPEL_TAPS / 2 - 1),
                       64, dst, dst_stride, y_filters, y0_q4, y_step_q4,
                       w, h, bd);
}


void vp9_highbd_convolve8_horiz_c(const uint8_t *src, ptrdiff_t src_stride,
                                  uint8_t *dst, ptrdiff_t dst_stride,
                                  const int16_t *filter_x, int x_step_q4,
                                  const int16_t *filter_y, int y_step_q4,
                                  int w, int h, int bd) {
  const InterpKernel *const filters_x = get_filter_base(filter_x);
  const int x0_q4 = get_filter_offset(filter_x, filters_x);
  (void)filter_y;
  (void)y_step_q4;

  highbd_convolve_horiz(src, src_stride, dst, dst_stride, filters_x,
                        x0_q4, x_step_q4, w, h, bd);
}

void vp9_highbd_convolve8_avg_horiz_c(const uint8_t *src, ptrdiff_t src_stride,
                                      uint8_t *dst, ptrdiff_t dst_stride,
                                      const int16_t *filter_x, int x_step_q4,
                                      const int16_t *filter_y, int y_step_q4,
                                      int w, int h, int bd) {
  const InterpKernel *const filters_x = get_filter_base(filter_x);
  const int x0_q4 = get_filter_offset(filter_x, filters_x);
  (void)filter_y;
  (void)y_step_q4;

  highbd_convolve_avg_horiz(src, src_stride, dst, dst_stride, filters_x,
                            x0_q4, x_step_q4, w, h, bd);
}

void vp9_highbd_convolve8_vert_c(const uint8_t *src, ptrdiff_t src_stride,
                                 uint8_t *dst, ptrdiff_t dst_stride,
                                 const int16_t *filter_x, int x_step_q4,
                                 const int16_t *filter_y, int y_step_q4,
                                 int w, int h, int bd) {
  const InterpKernel *const filters_y = get_filter_base(filter_y);
  const int y0_q4 = get_filter_offset(filter_y, filters_y);
  (void)filter_x;
  (void)x_step_q4;

  highbd_convolve_vert(src, src_stride, dst, dst_stride, filters_y,
                       y0_q4, y_step_q4, w, h, bd);
}

void vp9_highbd_convolve8_avg_vert_c(const uint8_t *src, ptrdiff_t src_stride,
                                     uint8_t *dst, ptrdiff_t dst_stride,
                                     const int16_t *filter_x, int x_step_q4,
                                     const int16_t *filter_y, int y_step_q4,
                                     int w, int h, int bd) {
  const InterpKernel *const filters_y = get_filter_base(filter_y);
  const int y0_q4 = get_filter_offset(filter_y, filters_y);
  (void)filter_x;
  (void)x_step_q4;

  highbd_convolve_avg_vert(src, src_stride, dst, dst_stride, filters_y,
                           y0_q4, y_step_q4, w, h, bd);
}

void vp9_highbd_convolve8_c(const uint8_t *src, ptrdiff_t src_stride,
                            uint8_t *dst, ptrdiff_t dst_stride,
                            const int16_t *filter_x, int x_step_q4,
                            const int16_t *filter_y, int y_step_q4,
                            int w, int h, int bd) {
  const InterpKernel *const filters_x = get_filter_base(filter_x);
  const int x0_q4 = get_filter_offset(filter_x, filters_x);

  const InterpKernel *const filters_y = get_filter_base(filter_y);
  const int y0_q4 = get_filter_offset(filter_y, filters_y);

  highbd_convolve(src, src_stride, dst, dst_stride,
                  filters_x, x0_q4, x_step_q4,
                  filters_y, y0_q4, y_step_q4, w, h, bd);
}

void vp9_highbd_convolve8_avg_c(const uint8_t *src, ptrdiff_t src_stride,
                                uint8_t *dst, ptrdiff_t dst_stride,
                                const int16_t *filter_x, int x_step_q4,
                                const int16_t *filter_y, int y_step_q4,
                                int w, int h, int bd) {
  // Fixed size intermediate buffer places limits on parameters.
  DECLARE_ALIGNED(16, uint16_t, temp[64 * 64]);
  assert(w <= 64);
  assert(h <= 64);

  vp9_highbd_convolve8_c(src, src_stride, CONVERT_TO_BYTEPTR(temp), 64,
                         filter_x, x_step_q4, filter_y, y_step_q4, w, h, bd);
  vp9_highbd_convolve_avg_c(CONVERT_TO_BYTEPTR(temp), 64, dst, dst_stride,
                            NULL, 0, NULL, 0, w, h, bd);
}

void vp9_highbd_convolve_copy_c(const uint8_t *src8, ptrdiff_t src_stride,
                                uint8_t *dst8, ptrdiff_t dst_stride,
                                const int16_t *filter_x, int filter_x_stride,
                                const int16_t *filter_y, int filter_y_stride,
                                int w, int h, int bd) {
  int r;
  uint16_t *src = CONVERT_TO_SHORTPTR(src8);
  uint16_t *dst = CONVERT_TO_SHORTPTR(dst8);
  (void)filter_x;
  (void)filter_y;
  (void)filter_x_stride;
  (void)filter_y_stride;
  (void)bd;

  for (r = h; r > 0; --r) {
    memcpy(dst, src, w * sizeof(uint16_t));
    src += src_stride;
    dst += dst_stride;
  }
}

void vp9_highbd_convolve_avg_c(const uint8_t *src8, ptrdiff_t src_stride,
                               uint8_t *dst8, ptrdiff_t dst_stride,
                               const int16_t *filter_x, int filter_x_stride,
                               const int16_t *filter_y, int filter_y_stride,
                               int w, int h, int bd) {
  int x, y;
  uint16_t *src = CONVERT_TO_SHORTPTR(src8);
  uint16_t *dst = CONVERT_TO_SHORTPTR(dst8);
  (void)filter_x;
  (void)filter_y;
  (void)filter_x_stride;
  (void)filter_y_stride;
  (void)bd;

  for (y = 0; y < h; ++y) {
    for (x = 0; x < w; ++x) {
      dst[x] = ROUND_POWER_OF_TWO(dst[x] + src[x], 1);
    }
    src += src_stride;
    dst += dst_stride;
  }
}
#endif