aboutsummaryrefslogblamecommitdiffstats
path: root/contrib/libs/libwebp/enc/iterator_enc.c
blob: 2de6545c24b348b87f0f5ed699b16aa08282cbe8 (plain) (tree)
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









































































































































































































































































































































































































































































                                                                                
// Copyright 2011 Google Inc. All Rights Reserved.
//
// Use of this source code is governed by a BSD-style license
// that can be found in the COPYING 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.
// -----------------------------------------------------------------------------
//
// VP8Iterator: block iterator
//
// Author: Skal (pascal.massimino@gmail.com)

#include <string.h>

#include "./vp8i_enc.h"

//------------------------------------------------------------------------------
// VP8Iterator
//------------------------------------------------------------------------------

static void InitLeft(VP8EncIterator* const it) {
  it->y_left_[-1] = it->u_left_[-1] = it->v_left_[-1] =
      (it->y_ > 0) ? 129 : 127;
  memset(it->y_left_, 129, 16);
  memset(it->u_left_, 129, 8);
  memset(it->v_left_, 129, 8);
  it->left_nz_[8] = 0;
  if (it->top_derr_ != NULL) {
    memset(&it->left_derr_, 0, sizeof(it->left_derr_));
  }
}

static void InitTop(VP8EncIterator* const it) {
  const VP8Encoder* const enc = it->enc_;
  const size_t top_size = enc->mb_w_ * 16;
  memset(enc->y_top_, 127, 2 * top_size);
  memset(enc->nz_, 0, enc->mb_w_ * sizeof(*enc->nz_));
  if (enc->top_derr_ != NULL) {
    memset(enc->top_derr_, 0, enc->mb_w_ * sizeof(*enc->top_derr_));
  }
}

void VP8IteratorSetRow(VP8EncIterator* const it, int y) {
  VP8Encoder* const enc = it->enc_;
  it->x_ = 0;
  it->y_ = y;
  it->bw_ = &enc->parts_[y & (enc->num_parts_ - 1)];
  it->preds_ = enc->preds_ + y * 4 * enc->preds_w_;
  it->nz_ = enc->nz_;
  it->mb_ = enc->mb_info_ + y * enc->mb_w_;
  it->y_top_ = enc->y_top_;
  it->uv_top_ = enc->uv_top_;
  InitLeft(it);
}

void VP8IteratorReset(VP8EncIterator* const it) {
  VP8Encoder* const enc = it->enc_;
  VP8IteratorSetRow(it, 0);
  VP8IteratorSetCountDown(it, enc->mb_w_ * enc->mb_h_);  // default
  InitTop(it);
  memset(it->bit_count_, 0, sizeof(it->bit_count_));
  it->do_trellis_ = 0;
}

void VP8IteratorSetCountDown(VP8EncIterator* const it, int count_down) {
  it->count_down_ = it->count_down0_ = count_down;
}

int VP8IteratorIsDone(const VP8EncIterator* const it) {
  return (it->count_down_ <= 0);
}

void VP8IteratorInit(VP8Encoder* const enc, VP8EncIterator* const it) {
  it->enc_ = enc;
  it->yuv_in_   = (uint8_t*)WEBP_ALIGN(it->yuv_mem_);
  it->yuv_out_  = it->yuv_in_ + YUV_SIZE_ENC;
  it->yuv_out2_ = it->yuv_out_ + YUV_SIZE_ENC;
  it->yuv_p_    = it->yuv_out2_ + YUV_SIZE_ENC;
  it->lf_stats_ = enc->lf_stats_;
  it->percent0_ = enc->percent_;
  it->y_left_ = (uint8_t*)WEBP_ALIGN(it->yuv_left_mem_ + 1);
  it->u_left_ = it->y_left_ + 16 + 16;
  it->v_left_ = it->u_left_ + 16;
  it->top_derr_ = enc->top_derr_;
  VP8IteratorReset(it);
}

int VP8IteratorProgress(const VP8EncIterator* const it, int delta) {
  VP8Encoder* const enc = it->enc_;
  if (delta && enc->pic_->progress_hook != NULL) {
    const int done = it->count_down0_ - it->count_down_;
    const int percent = (it->count_down0_ <= 0)
                      ? it->percent0_
                      : it->percent0_ + delta * done / it->count_down0_;
    return WebPReportProgress(enc->pic_, percent, &enc->percent_);
  }
  return 1;
}

//------------------------------------------------------------------------------
// Import the source samples into the cache. Takes care of replicating
// boundary pixels if necessary.

static WEBP_INLINE int MinSize(int a, int b) { return (a < b) ? a : b; }

static void ImportBlock(const uint8_t* src, int src_stride,
                        uint8_t* dst, int w, int h, int size) {
  int i;
  for (i = 0; i < h; ++i) {
    memcpy(dst, src, w);
    if (w < size) {
      memset(dst + w, dst[w - 1], size - w);
    }
    dst += BPS;
    src += src_stride;
  }
  for (i = h; i < size; ++i) {
    memcpy(dst, dst - BPS, size);
    dst += BPS;
  }
}

static void ImportLine(const uint8_t* src, int src_stride,
                       uint8_t* dst, int len, int total_len) {
  int i;
  for (i = 0; i < len; ++i, src += src_stride) dst[i] = *src;
  for (; i < total_len; ++i) dst[i] = dst[len - 1];
}

void VP8IteratorImport(VP8EncIterator* const it, uint8_t* const tmp_32) {
  const VP8Encoder* const enc = it->enc_;
  const int x = it->x_, y = it->y_;
  const WebPPicture* const pic = enc->pic_;
  const uint8_t* const ysrc = pic->y + (y * pic->y_stride  + x) * 16;
  const uint8_t* const usrc = pic->u + (y * pic->uv_stride + x) * 8;
  const uint8_t* const vsrc = pic->v + (y * pic->uv_stride + x) * 8;
  const int w = MinSize(pic->width - x * 16, 16);
  const int h = MinSize(pic->height - y * 16, 16);
  const int uv_w = (w + 1) >> 1;
  const int uv_h = (h + 1) >> 1;

  ImportBlock(ysrc, pic->y_stride,  it->yuv_in_ + Y_OFF_ENC, w, h, 16);
  ImportBlock(usrc, pic->uv_stride, it->yuv_in_ + U_OFF_ENC, uv_w, uv_h, 8);
  ImportBlock(vsrc, pic->uv_stride, it->yuv_in_ + V_OFF_ENC, uv_w, uv_h, 8);

  if (tmp_32 == NULL) return;

  // Import source (uncompressed) samples into boundary.
  if (x == 0) {
    InitLeft(it);
  } else {
    if (y == 0) {
      it->y_left_[-1] = it->u_left_[-1] = it->v_left_[-1] = 127;
    } else {
      it->y_left_[-1] = ysrc[- 1 - pic->y_stride];
      it->u_left_[-1] = usrc[- 1 - pic->uv_stride];
      it->v_left_[-1] = vsrc[- 1 - pic->uv_stride];
    }
    ImportLine(ysrc - 1, pic->y_stride,  it->y_left_, h,   16);
    ImportLine(usrc - 1, pic->uv_stride, it->u_left_, uv_h, 8);
    ImportLine(vsrc - 1, pic->uv_stride, it->v_left_, uv_h, 8);
  }

  it->y_top_  = tmp_32 + 0;
  it->uv_top_ = tmp_32 + 16;
  if (y == 0) {
    memset(tmp_32, 127, 32 * sizeof(*tmp_32));
  } else {
    ImportLine(ysrc - pic->y_stride,  1, tmp_32,          w,   16);
    ImportLine(usrc - pic->uv_stride, 1, tmp_32 + 16,     uv_w, 8);
    ImportLine(vsrc - pic->uv_stride, 1, tmp_32 + 16 + 8, uv_w, 8);
  }
}

//------------------------------------------------------------------------------
// Copy back the compressed samples into user space if requested.

static void ExportBlock(const uint8_t* src, uint8_t* dst, int dst_stride,
                        int w, int h) {
  while (h-- > 0) {
    memcpy(dst, src, w);
    dst += dst_stride;
    src += BPS;
  }
}

void VP8IteratorExport(const VP8EncIterator* const it) {
  const VP8Encoder* const enc = it->enc_;
  if (enc->config_->show_compressed) {
    const int x = it->x_, y = it->y_;
    const uint8_t* const ysrc = it->yuv_out_ + Y_OFF_ENC;
    const uint8_t* const usrc = it->yuv_out_ + U_OFF_ENC;
    const uint8_t* const vsrc = it->yuv_out_ + V_OFF_ENC;
    const WebPPicture* const pic = enc->pic_;
    uint8_t* const ydst = pic->y + (y * pic->y_stride + x) * 16;
    uint8_t* const udst = pic->u + (y * pic->uv_stride + x) * 8;
    uint8_t* const vdst = pic->v + (y * pic->uv_stride + x) * 8;
    int w = (pic->width - x * 16);
    int h = (pic->height - y * 16);

    if (w > 16) w = 16;
    if (h > 16) h = 16;

    // Luma plane
    ExportBlock(ysrc, ydst, pic->y_stride, w, h);

    {   // U/V planes
      const int uv_w = (w + 1) >> 1;
      const int uv_h = (h + 1) >> 1;
      ExportBlock(usrc, udst, pic->uv_stride, uv_w, uv_h);
      ExportBlock(vsrc, vdst, pic->uv_stride, uv_w, uv_h);
    }
  }
}

//------------------------------------------------------------------------------
// Non-zero contexts setup/teardown

// Nz bits:
//  0  1  2  3  Y
//  4  5  6  7
//  8  9 10 11
// 12 13 14 15
// 16 17        U
// 18 19
// 20 21        V
// 22 23
// 24           DC-intra16

// Convert packed context to byte array
#define BIT(nz, n) (!!((nz) & (1 << (n))))

void VP8IteratorNzToBytes(VP8EncIterator* const it) {
  const int tnz = it->nz_[0], lnz = it->nz_[-1];
  int* const top_nz = it->top_nz_;
  int* const left_nz = it->left_nz_;

  // Top-Y
  top_nz[0] = BIT(tnz, 12);
  top_nz[1] = BIT(tnz, 13);
  top_nz[2] = BIT(tnz, 14);
  top_nz[3] = BIT(tnz, 15);
  // Top-U
  top_nz[4] = BIT(tnz, 18);
  top_nz[5] = BIT(tnz, 19);
  // Top-V
  top_nz[6] = BIT(tnz, 22);
  top_nz[7] = BIT(tnz, 23);
  // DC
  top_nz[8] = BIT(tnz, 24);

  // left-Y
  left_nz[0] = BIT(lnz,  3);
  left_nz[1] = BIT(lnz,  7);
  left_nz[2] = BIT(lnz, 11);
  left_nz[3] = BIT(lnz, 15);
  // left-U
  left_nz[4] = BIT(lnz, 17);
  left_nz[5] = BIT(lnz, 19);
  // left-V
  left_nz[6] = BIT(lnz, 21);
  left_nz[7] = BIT(lnz, 23);
  // left-DC is special, iterated separately
}

void VP8IteratorBytesToNz(VP8EncIterator* const it) {
  uint32_t nz = 0;
  const int* const top_nz = it->top_nz_;
  const int* const left_nz = it->left_nz_;
  // top
  nz |= (top_nz[0] << 12) | (top_nz[1] << 13);
  nz |= (top_nz[2] << 14) | (top_nz[3] << 15);
  nz |= (top_nz[4] << 18) | (top_nz[5] << 19);
  nz |= (top_nz[6] << 22) | (top_nz[7] << 23);
  nz |= (top_nz[8] << 24);  // we propagate the _top_ bit, esp. for intra4
  // left
  nz |= (left_nz[0] << 3) | (left_nz[1] << 7);
  nz |= (left_nz[2] << 11);
  nz |= (left_nz[4] << 17) | (left_nz[6] << 21);

  *it->nz_ = nz;
}

#undef BIT

//------------------------------------------------------------------------------
// Advance to the next position, doing the bookkeeping.

void VP8IteratorSaveBoundary(VP8EncIterator* const it) {
  VP8Encoder* const enc = it->enc_;
  const int x = it->x_, y = it->y_;
  const uint8_t* const ysrc = it->yuv_out_ + Y_OFF_ENC;
  const uint8_t* const uvsrc = it->yuv_out_ + U_OFF_ENC;
  if (x < enc->mb_w_ - 1) {   // left
    int i;
    for (i = 0; i < 16; ++i) {
      it->y_left_[i] = ysrc[15 + i * BPS];
    }
    for (i = 0; i < 8; ++i) {
      it->u_left_[i] = uvsrc[7 + i * BPS];
      it->v_left_[i] = uvsrc[15 + i * BPS];
    }
    // top-left (before 'top'!)
    it->y_left_[-1] = it->y_top_[15];
    it->u_left_[-1] = it->uv_top_[0 + 7];
    it->v_left_[-1] = it->uv_top_[8 + 7];
  }
  if (y < enc->mb_h_ - 1) {  // top
    memcpy(it->y_top_, ysrc + 15 * BPS, 16);
    memcpy(it->uv_top_, uvsrc + 7 * BPS, 8 + 8);
  }
}

int VP8IteratorNext(VP8EncIterator* const it) {
  if (++it->x_ == it->enc_->mb_w_) {
    VP8IteratorSetRow(it, ++it->y_);
  } else {
    it->preds_ += 4;
    it->mb_ += 1;
    it->nz_ += 1;
    it->y_top_ += 16;
    it->uv_top_ += 16;
  }
  return (0 < --it->count_down_);
}

//------------------------------------------------------------------------------
// Helper function to set mode properties

void VP8SetIntra16Mode(const VP8EncIterator* const it, int mode) {
  uint8_t* preds = it->preds_;
  int y;
  for (y = 0; y < 4; ++y) {
    memset(preds, mode, 4);
    preds += it->enc_->preds_w_;
  }
  it->mb_->type_ = 1;
}

void VP8SetIntra4Mode(const VP8EncIterator* const it, const uint8_t* modes) {
  uint8_t* preds = it->preds_;
  int y;
  for (y = 4; y > 0; --y) {
    memcpy(preds, modes, 4 * sizeof(*modes));
    preds += it->enc_->preds_w_;
    modes += 4;
  }
  it->mb_->type_ = 0;
}

void VP8SetIntraUVMode(const VP8EncIterator* const it, int mode) {
  it->mb_->uv_mode_ = mode;
}

void VP8SetSkip(const VP8EncIterator* const it, int skip) {
  it->mb_->skip_ = skip;
}

void VP8SetSegment(const VP8EncIterator* const it, int segment) {
  it->mb_->segment_ = segment;
}

//------------------------------------------------------------------------------
// Intra4x4 sub-blocks iteration
//
//  We store and update the boundary samples into an array of 37 pixels. They
//  are updated as we iterate and reconstructs each intra4x4 blocks in turn.
//  The position of the samples has the following snake pattern:
//
// 16|17 18 19 20|21 22 23 24|25 26 27 28|29 30 31 32|33 34 35 36  <- Top-right
// --+-----------+-----------+-----------+-----------+
// 15|         19|         23|         27|         31|
// 14|         18|         22|         26|         30|
// 13|         17|         21|         25|         29|
// 12|13 14 15 16|17 18 19 20|21 22 23 24|25 26 27 28|
// --+-----------+-----------+-----------+-----------+
// 11|         15|         19|         23|         27|
// 10|         14|         18|         22|         26|
//  9|         13|         17|         21|         25|
//  8| 9 10 11 12|13 14 15 16|17 18 19 20|21 22 23 24|
// --+-----------+-----------+-----------+-----------+
//  7|         11|         15|         19|         23|
//  6|         10|         14|         18|         22|
//  5|          9|         13|         17|         21|
//  4| 5  6  7  8| 9 10 11 12|13 14 15 16|17 18 19 20|
// --+-----------+-----------+-----------+-----------+
//  3|          7|         11|         15|         19|
//  2|          6|         10|         14|         18|
//  1|          5|          9|         13|         17|
//  0| 1  2  3  4| 5  6  7  8| 9 10 11 12|13 14 15 16|
// --+-----------+-----------+-----------+-----------+

// Array to record the position of the top sample to pass to the prediction
// functions in dsp.c.
static const uint8_t VP8TopLeftI4[16] = {
  17, 21, 25, 29,
  13, 17, 21, 25,
  9,  13, 17, 21,
  5,   9, 13, 17
};

void VP8IteratorStartI4(VP8EncIterator* const it) {
  const VP8Encoder* const enc = it->enc_;
  int i;

  it->i4_ = 0;    // first 4x4 sub-block
  it->i4_top_ = it->i4_boundary_ + VP8TopLeftI4[0];

  // Import the boundary samples
  for (i = 0; i < 17; ++i) {    // left
    it->i4_boundary_[i] = it->y_left_[15 - i];
  }
  for (i = 0; i < 16; ++i) {    // top
    it->i4_boundary_[17 + i] = it->y_top_[i];
  }
  // top-right samples have a special case on the far right of the picture
  if (it->x_ < enc->mb_w_ - 1) {
    for (i = 16; i < 16 + 4; ++i) {
      it->i4_boundary_[17 + i] = it->y_top_[i];
    }
  } else {    // else, replicate the last valid pixel four times
    for (i = 16; i < 16 + 4; ++i) {
      it->i4_boundary_[17 + i] = it->i4_boundary_[17 + 15];
    }
  }
  VP8IteratorNzToBytes(it);  // import the non-zero context
}

int VP8IteratorRotateI4(VP8EncIterator* const it,
                        const uint8_t* const yuv_out) {
  const uint8_t* const blk = yuv_out + VP8Scan[it->i4_];
  uint8_t* const top = it->i4_top_;
  int i;

  // Update the cache with 7 fresh samples
  for (i = 0; i <= 3; ++i) {
    top[-4 + i] = blk[i + 3 * BPS];   // store future top samples
  }
  if ((it->i4_ & 3) != 3) {  // if not on the right sub-blocks #3, #7, #11, #15
    for (i = 0; i <= 2; ++i) {        // store future left samples
      top[i] = blk[3 + (2 - i) * BPS];
    }
  } else {  // else replicate top-right samples, as says the specs.
    for (i = 0; i <= 3; ++i) {
      top[i] = top[i + 4];
    }
  }
  // move pointers to next sub-block
  ++it->i4_;
  if (it->i4_ == 16) {    // we're done
    return 0;
  }

  it->i4_top_ = it->i4_boundary_ + VP8TopLeftI4[it->i4_];
  return 1;
}

//------------------------------------------------------------------------------