DXR is a code search and navigation tool aimed at making sense of large projects. It supports full-text and regex searches as well as structural queries.

Mercurial (d38398e5144e)

VCS Links

Line Code
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
/*
 *  Copyright (c) 2014 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 <stdlib.h>

#include "./vpx_dsp_rtcd.h"
#include "vpx_ports/mem.h"

unsigned int vpx_avg_8x8_c(const uint8_t *s, int p) {
  int i, j;
  int sum = 0;
  for (i = 0; i < 8; ++i, s += p)
    for (j = 0; j < 8; sum += s[j], ++j) {
    }

  return (sum + 32) >> 6;
}

unsigned int vpx_avg_4x4_c(const uint8_t *s, int p) {
  int i, j;
  int sum = 0;
  for (i = 0; i < 4; ++i, s += p)
    for (j = 0; j < 4; sum += s[j], ++j) {
    }

  return (sum + 8) >> 4;
}

// src_diff: first pass, 9 bit, dynamic range [-255, 255]
//           second pass, 12 bit, dynamic range [-2040, 2040]
static void hadamard_col8(const int16_t *src_diff, int src_stride,
                          int16_t *coeff) {
  int16_t b0 = src_diff[0 * src_stride] + src_diff[1 * src_stride];
  int16_t b1 = src_diff[0 * src_stride] - src_diff[1 * src_stride];
  int16_t b2 = src_diff[2 * src_stride] + src_diff[3 * src_stride];
  int16_t b3 = src_diff[2 * src_stride] - src_diff[3 * src_stride];
  int16_t b4 = src_diff[4 * src_stride] + src_diff[5 * src_stride];
  int16_t b5 = src_diff[4 * src_stride] - src_diff[5 * src_stride];
  int16_t b6 = src_diff[6 * src_stride] + src_diff[7 * src_stride];
  int16_t b7 = src_diff[6 * src_stride] - src_diff[7 * src_stride];

  int16_t c0 = b0 + b2;
  int16_t c1 = b1 + b3;
  int16_t c2 = b0 - b2;
  int16_t c3 = b1 - b3;
  int16_t c4 = b4 + b6;
  int16_t c5 = b5 + b7;
  int16_t c6 = b4 - b6;
  int16_t c7 = b5 - b7;

  coeff[0] = c0 + c4;
  coeff[7] = c1 + c5;
  coeff[3] = c2 + c6;
  coeff[4] = c3 + c7;
  coeff[2] = c0 - c4;
  coeff[6] = c1 - c5;
  coeff[1] = c2 - c6;
  coeff[5] = c3 - c7;
}

// The order of the output coeff of the hadamard is not important. For
// optimization purposes the final transpose may be skipped.
void vpx_hadamard_8x8_c(const int16_t *src_diff, int src_stride,
                        int16_t *coeff) {
  int idx;
  int16_t buffer[64];
  int16_t *tmp_buf = &buffer[0];
  for (idx = 0; idx < 8; ++idx) {
    hadamard_col8(src_diff, src_stride, tmp_buf);  // src_diff: 9 bit
                                                   // dynamic range [-255, 255]
    tmp_buf += 8;
    ++src_diff;
  }

  tmp_buf = &buffer[0];
  for (idx = 0; idx < 8; ++idx) {
    hadamard_col8(tmp_buf, 8, coeff);  // tmp_buf: 12 bit
                                       // dynamic range [-2040, 2040]
    coeff += 8;                        // coeff: 15 bit
                                       // dynamic range [-16320, 16320]
    ++tmp_buf;
  }
}

// In place 16x16 2D Hadamard transform
void vpx_hadamard_16x16_c(const int16_t *src_diff, int src_stride,
                          int16_t *coeff) {
  int idx;
  for (idx = 0; idx < 4; ++idx) {
    // src_diff: 9 bit, dynamic range [-255, 255]
    const int16_t *src_ptr =
        src_diff + (idx >> 1) * 8 * src_stride + (idx & 0x01) * 8;
    vpx_hadamard_8x8_c(src_ptr, src_stride, coeff + idx * 64);
  }

  // coeff: 15 bit, dynamic range [-16320, 16320]
  for (idx = 0; idx < 64; ++idx) {
    int16_t a0 = coeff[0];
    int16_t a1 = coeff[64];
    int16_t a2 = coeff[128];
    int16_t a3 = coeff[192];

    int16_t b0 = (a0 + a1) >> 1;  // (a0 + a1): 16 bit, [-32640, 32640]
    int16_t b1 = (a0 - a1) >> 1;  // b0-b3: 15 bit, dynamic range
    int16_t b2 = (a2 + a3) >> 1;  // [-16320, 16320]
    int16_t b3 = (a2 - a3) >> 1;

    coeff[0] = b0 + b2;  // 16 bit, [-32640, 32640]
    coeff[64] = b1 + b3;
    coeff[128] = b0 - b2;
    coeff[192] = b1 - b3;

    ++coeff;
  }
}

// coeff: 16 bits, dynamic range [-32640, 32640].
// length: value range {16, 64, 256, 1024}.
int vpx_satd_c(const int16_t *coeff, int length) {
  int i;
  int satd = 0;
  for (i = 0; i < length; ++i) satd += abs(coeff[i]);

  // satd: 26 bits, dynamic range [-32640 * 1024, 32640 * 1024]
  return satd;
}

// Integer projection onto row vectors.
// height: value range {16, 32, 64}.
void vpx_int_pro_row_c(int16_t hbuf[16], const uint8_t *ref,
                       const int ref_stride, const int height) {
  int idx;
  const int norm_factor = height >> 1;
  for (idx = 0; idx < 16; ++idx) {
    int i;
    hbuf[idx] = 0;
    // hbuf[idx]: 14 bit, dynamic range [0, 16320].
    for (i = 0; i < height; ++i) hbuf[idx] += ref[i * ref_stride];
    // hbuf[idx]: 9 bit, dynamic range [0, 510].
    hbuf[idx] /= norm_factor;
    ++ref;
  }
}

// width: value range {16, 32, 64}.
int16_t vpx_int_pro_col_c(const uint8_t *ref, const int width) {
  int idx;
  int16_t sum = 0;
  // sum: 14 bit, dynamic range [0, 16320]
  for (idx = 0; idx < width; ++idx) sum += ref[idx];
  return sum;
}

// ref: [0 - 510]
// src: [0 - 510]
// bwl: {2, 3, 4}
int vpx_vector_var_c(const int16_t *ref, const int16_t *src, const int bwl) {
  int i;
  int width = 4 << bwl;
  int sse = 0, mean = 0, var;

  for (i = 0; i < width; ++i) {
    int diff = ref[i] - src[i];  // diff: dynamic range [-510, 510], 10 bits.
    mean += diff;                // mean: dynamic range 16 bits.
    sse += diff * diff;          // sse:  dynamic range 26 bits.
  }

  // (mean * mean): dynamic range 31 bits.
  var = sse - ((mean * mean) >> (bwl + 2));
  return var;
}

void vpx_minmax_8x8_c(const uint8_t *s, int p, const uint8_t *d, int dp,
                      int *min, int *max) {
  int i, j;
  *min = 255;
  *max = 0;
  for (i = 0; i < 8; ++i, s += p, d += dp) {
    for (j = 0; j < 8; ++j) {
      int diff = abs(s[j] - d[j]);
      *min = diff < *min ? diff : *min;
      *max = diff > *max ? diff : *max;
    }
  }
}

#if CONFIG_VP9_HIGHBITDEPTH
unsigned int vpx_highbd_avg_8x8_c(const uint8_t *s8, int p) {
  int i, j;
  int sum = 0;
  const uint16_t *s = CONVERT_TO_SHORTPTR(s8);
  for (i = 0; i < 8; ++i, s += p)
    for (j = 0; j < 8; sum += s[j], ++j) {
    }

  return (sum + 32) >> 6;
}

unsigned int vpx_highbd_avg_4x4_c(const uint8_t *s8, int p) {
  int i, j;
  int sum = 0;
  const uint16_t *s = CONVERT_TO_SHORTPTR(s8);
  for (i = 0; i < 4; ++i, s += p)
    for (j = 0; j < 4; sum += s[j], ++j) {
    }

  return (sum + 8) >> 4;
}

void vpx_highbd_minmax_8x8_c(const uint8_t *s8, int p, const uint8_t *d8,
                             int dp, int *min, int *max) {
  int i, j;
  const uint16_t *s = CONVERT_TO_SHORTPTR(s8);
  const uint16_t *d = CONVERT_TO_SHORTPTR(d8);
  *min = 255;
  *max = 0;
  for (i = 0; i < 8; ++i, s += p, d += dp) {
    for (j = 0; j < 8; ++j) {
      int diff = abs(s[j] - d[j]);
      *min = diff < *min ? diff : *min;
      *max = diff > *max ? diff : *max;
    }
  }
}
#endif  // CONFIG_VP9_HIGHBITDEPTH