av1/decoder/decodetxb.c

/*
 * Copyright (c) 2017, Alliance for Open Media. All rights reserved
 *
 * This source code is subject to the terms of the BSD 2 Clause License and
 * the Alliance for Open Media Patent License 1.0. If the BSD 2 Clause License
 * was not distributed with this source code in the LICENSE file, you can
 * obtain it at www.aomedia.org/license/software. If the Alliance for Open
 * Media Patent License 1.0 was not distributed with this source code in the
 * PATENTS file, you can obtain it at www.aomedia.org/license/patent.
 */

#include "av1/decoder/decodetxb.h"

#include "aom_ports/mem.h"
#include "av1/common/idct.h"
#include "av1/common/scan.h"
#include "av1/common/txb_common.h"
#include "av1/decoder/decodemv.h"

#define ACCT_STR __func__

static int read_golomb(MACROBLOCKD *xd, aom_reader *r) {
  int x = 1;
  int length = 0;
  int i = 0;

  while (!i) {
    i = aom_read_bit(r, ACCT_STR);
    ++length;
    if (length > 20) {
      aom_internal_error(xd->error_info, AOM_CODEC_CORRUPT_FRAME,
                         "Invalid length in read_golomb");
      break;
    }
  }

  for (i = 0; i < length - 1; ++i) {
    x <<= 1;
    x += aom_read_bit(r, ACCT_STR);
  }

  return x - 1;
}

static INLINE int rec_eob_pos(const int eob_token, const int extra) {
  int eob = av1_eob_group_start[eob_token];
  if (eob > 2) {
    eob += extra;
  }
  return eob;
}

static INLINE int get_dqv(const int16_t *dequant, int coeff_idx,
                          const qm_val_t *iqmatrix) {
  int dqv = dequant[!!coeff_idx];
  if (iqmatrix != NULL)
    dqv =
        ((iqmatrix[coeff_idx] * dqv) + (1 << (AOM_QM_BITS - 1))) >> AOM_QM_BITS;
  return dqv;
}

static INLINE void read_coeffs_reverse_2d(aom_reader *r, TX_SIZE tx_size,
                                          int start_si, int end_si,
                                          const int16_t *scan, int bwl,
                                          uint8_t *levels,
                                          base_cdf_arr base_cdf,
                                          br_cdf_arr br_cdf) {
  for (int c = end_si; c >= start_si; --c) {
    const int pos = scan[c];
    const int coeff_ctx = get_lower_levels_ctx_2d(levels, pos, bwl, tx_size);
    const int nsymbs = 4;
    int level = aom_read_symbol(r, base_cdf[coeff_ctx], nsymbs, ACCT_STR);
    if (level > NUM_BASE_LEVELS) {
      const int br_ctx = get_br_ctx_2d(levels, pos, bwl);
      aom_cdf_prob *cdf = br_cdf[br_ctx];
      for (int idx = 0; idx < COEFF_BASE_RANGE; idx += BR_CDF_SIZE - 1) {
        const int k = aom_read_symbol(r, cdf, BR_CDF_SIZE, ACCT_STR);
        level += k;
        if (k < BR_CDF_SIZE - 1) break;
      }
    }
    levels[get_padded_idx(pos, bwl)] = level;
  }
}

static INLINE void read_coeffs_reverse(aom_reader *r, TX_SIZE tx_size,
                                       TX_CLASS tx_class, int start_si,
                                       int end_si, const int16_t *scan, int bwl,
                                       uint8_t *levels, base_cdf_arr base_cdf,
                                       br_cdf_arr br_cdf) {
  for (int c = end_si; c >= start_si; --c) {
    const int pos = scan[c];
    const int coeff_ctx =
        get_lower_levels_ctx(levels, pos, bwl, tx_size, tx_class);
    const int nsymbs = 4;
    int level = aom_read_symbol(r, base_cdf[coeff_ctx], nsymbs, ACCT_STR);
    if (level > NUM_BASE_LEVELS) {
      const int br_ctx = get_br_ctx(levels, pos, bwl, tx_class);
      aom_cdf_prob *cdf = br_cdf[br_ctx];
      for (int idx = 0; idx < COEFF_BASE_RANGE; idx += BR_CDF_SIZE - 1) {
        const int k = aom_read_symbol(r, cdf, BR_CDF_SIZE, ACCT_STR);
        level += k;
        if (k < BR_CDF_SIZE - 1) break;
      }
    }
    levels[get_padded_idx(pos, bwl)] = level;
  }
}

uint8_t av1_read_coeffs_txb(const AV1_COMMON *const cm, DecoderCodingBlock *dcb,
                            aom_reader *const r, const int blk_row,
                            const int blk_col, const int plane,
                            const TXB_CTX *const txb_ctx,
                            const TX_SIZE tx_size) {
  MACROBLOCKD *const xd = &dcb->xd;
  FRAME_CONTEXT *const ec_ctx = xd->tile_ctx;
  const int32_t max_value = (1 << (7 + xd->bd)) - 1;
  const int32_t min_value = -(1 << (7 + xd->bd));
  const TX_SIZE txs_ctx = get_txsize_entropy_ctx(tx_size);
  const PLANE_TYPE plane_type = get_plane_type(plane);
  MB_MODE_INFO *const mbmi = xd->mi[0];
  struct macroblockd_plane *const pd = &xd->plane[plane];
  const int16_t *const dequant = pd->seg_dequant_QTX[mbmi->segment_id];
  tran_low_t *const tcoeffs = dcb->dqcoeff_block[plane] + dcb->cb_offset[plane];
  const int shift = av1_get_tx_scale(tx_size);
  const int bwl = get_txb_bwl(tx_size);
  const int width = get_txb_wide(tx_size);
  const int height = get_txb_high(tx_size);
  int cul_level = 0;
  int dc_val = 0;
  uint8_t levels_buf[TX_PAD_2D];
  uint8_t *const levels = set_levels(levels_buf, width);
  const int all_zero = aom_read_symbol(
      r, ec_ctx->txb_skip_cdf[txs_ctx][txb_ctx->txb_skip_ctx], 2, ACCT_STR);
  eob_info *eob_data = dcb->eob_data[plane] + dcb->txb_offset[plane];
  uint16_t *const eob = &(eob_data->eob);
  uint16_t *const max_scan_line = &(eob_data->max_scan_line);
  *max_scan_line = 0;
  *eob = 0;

#if CONFIG_INSPECTION
  if (plane == 0) {
    const int txk_type_idx =
        av1_get_txk_type_index(mbmi->bsize, blk_row, blk_col);
    mbmi->tx_skip[txk_type_idx] = all_zero;
  }
#endif

  if (all_zero) {
    *max_scan_line = 0;
    if (plane == 0) {
      xd->tx_type_map[blk_row * xd->tx_type_map_stride + blk_col] = DCT_DCT;
    }
    return 0;
  }

  if (plane == AOM_PLANE_Y) {
    // only y plane's tx_type is transmitted
    av1_read_tx_type(cm, xd, blk_row, blk_col, tx_size, r);
  }
  const TX_TYPE tx_type =
      av1_get_tx_type(xd, plane_type, blk_row, blk_col, tx_size,
                      cm->features.reduced_tx_set_used);
  const TX_CLASS tx_class = tx_type_to_class[tx_type];
  const qm_val_t *iqmatrix =
      av1_get_iqmatrix(&cm->quant_params, xd, plane, tx_size, tx_type);
  const SCAN_ORDER *const scan_order = get_scan(tx_size, tx_type);
  const int16_t *const scan = scan_order->scan;
  int eob_extra = 0;
  int eob_pt = 1;

  const int eob_multi_size = txsize_log2_minus4[tx_size];
  const int eob_multi_ctx = (tx_class == TX_CLASS_2D) ? 0 : 1;
  switch (eob_multi_size) {
    case 0:
      eob_pt =
          aom_read_symbol(r, ec_ctx->eob_flag_cdf16[plane_type][eob_multi_ctx],
                          5, ACCT_STR) +
          1;
      break;
    case 1:
      eob_pt =
          aom_read_symbol(r, ec_ctx->eob_flag_cdf32[plane_type][eob_multi_ctx],
                          6, ACCT_STR) +
          1;
      break;
    case 2:
      eob_pt =
          aom_read_symbol(r, ec_ctx->eob_flag_cdf64[plane_type][eob_multi_ctx],
                          7, ACCT_STR) +
          1;
      break;
    case 3:
      eob_pt =
          aom_read_symbol(r, ec_ctx->eob_flag_cdf128[plane_type][eob_multi_ctx],
                          8, ACCT_STR) +
          1;
      break;
    case 4:
      eob_pt =
          aom_read_symbol(r, ec_ctx->eob_flag_cdf256[plane_type][eob_multi_ctx],
                          9, ACCT_STR) +
          1;
      break;
    case 5:
      eob_pt =
          aom_read_symbol(r, ec_ctx->eob_flag_cdf512[plane_type][eob_multi_ctx],
                          10, ACCT_STR) +
          1;
      break;
    case 6:
    default:
      eob_pt = aom_read_symbol(
                   r, ec_ctx->eob_flag_cdf1024[plane_type][eob_multi_ctx], 11,
                   ACCT_STR) +
               1;
      break;
  }

  const int eob_offset_bits = av1_eob_offset_bits[eob_pt];
  if (eob_offset_bits > 0) {
    const int eob_ctx = eob_pt - 3;
    int bit = aom_read_symbol(
        r, ec_ctx->eob_extra_cdf[txs_ctx][plane_type][eob_ctx], 2, ACCT_STR);
    if (bit) {
      eob_extra += (1 << (eob_offset_bits - 1));
    }

    for (int i = 1; i < eob_offset_bits; i++) {
      bit = aom_read_bit(r, ACCT_STR);
      if (bit) {
        eob_extra += (1 << (eob_offset_bits - 1 - i));
      }
    }
  }
  *eob = rec_eob_pos(eob_pt, eob_extra);

  if (*eob > 1) {
    memset(levels_buf, 0,
           sizeof(*levels_buf) *
               ((width + TX_PAD_HOR) * (height + TX_PAD_VER) + TX_PAD_END));
  }

  {
    // Read the non-zero coefficient with scan index eob-1
    // TODO(angiebird): Put this into a function
    const int c = *eob - 1;
    const int pos = scan[c];
    const int coeff_ctx = get_lower_levels_ctx_eob(bwl, height, c);
    const int nsymbs = 3;
    aom_cdf_prob *cdf =
        ec_ctx->coeff_base_eob_cdf[txs_ctx][plane_type][coeff_ctx];
    int level = aom_read_symbol(r, cdf, nsymbs, ACCT_STR) + 1;
    if (level > NUM_BASE_LEVELS) {
      const int br_ctx = get_br_ctx_eob(pos, bwl, tx_class);
      cdf = ec_ctx->coeff_br_cdf[AOMMIN(txs_ctx, TX_32X32)][plane_type][br_ctx];
      for (int idx = 0; idx < COEFF_BASE_RANGE; idx += BR_CDF_SIZE - 1) {
        const int k = aom_read_symbol(r, cdf, BR_CDF_SIZE, ACCT_STR);
        level += k;
        if (k < BR_CDF_SIZE - 1) break;
      }
    }
    levels[get_padded_idx(pos, bwl)] = level;
  }
  if (*eob > 1) {
    base_cdf_arr base_cdf = ec_ctx->coeff_base_cdf[txs_ctx][plane_type];
    br_cdf_arr br_cdf =
        ec_ctx->coeff_br_cdf[AOMMIN(txs_ctx, TX_32X32)][plane_type];
    if (tx_class == TX_CLASS_2D) {
      read_coeffs_reverse_2d(r, tx_size, 1, *eob - 1 - 1, scan, bwl, levels,
                             base_cdf, br_cdf);
      read_coeffs_reverse(r, tx_size, tx_class, 0, 0, scan, bwl, levels,
                          base_cdf, br_cdf);
    } else {
      read_coeffs_reverse(r, tx_size, tx_class, 0, *eob - 1 - 1, scan, bwl,
                          levels, base_cdf, br_cdf);
    }
  }

  for (int c = 0; c < *eob; ++c) {
    const int pos = scan[c];
    uint8_t sign;
    tran_low_t level = levels[get_padded_idx(pos, bwl)];
    if (level) {
      *max_scan_line = AOMMAX(*max_scan_line, pos);
      if (c == 0) {
        const int dc_sign_ctx = txb_ctx->dc_sign_ctx;
        sign = aom_read_symbol(r, ec_ctx->dc_sign_cdf[plane_type][dc_sign_ctx],
                               2, ACCT_STR);
      } else {
        sign = aom_read_bit(r, ACCT_STR);
      }
      if (level >= MAX_BASE_BR_RANGE) {
        level += read_golomb(xd, r);
      }

      if (c == 0) dc_val = sign ? -level : level;

      // Bitmasking to clamp level to valid range:
      //   The valid range for 8/10/12 bit vdieo is at most 14/16/18 bit
      level &= 0xfffff;
      cul_level += level;
      tran_low_t dq_coeff;
      // Bitmasking to clamp dq_coeff to valid range:
      //   The valid range for 8/10/12 bit video is at most 17/19/21 bit
      dq_coeff = (tran_low_t)(
          (int64_t)level * get_dqv(dequant, scan[c], iqmatrix) & 0xffffff);
      dq_coeff = dq_coeff >> shift;
      if (sign) {
        dq_coeff = -dq_coeff;
      }
      tcoeffs[pos] = clamp(dq_coeff, min_value, max_value);
    }
  }

  cul_level = AOMMIN(COEFF_CONTEXT_MASK, cul_level);

  // DC value
  set_dc_sign(&cul_level, dc_val);

  return cul_level;
}

void av1_read_coeffs_txb_facade(const AV1_COMMON *const cm,
                                DecoderCodingBlock *dcb, aom_reader *const r,
                                const int plane, const int row, const int col,
                                const TX_SIZE tx_size) {
#if TXCOEFF_TIMER
  struct aom_usec_timer timer;
  aom_usec_timer_start(&timer);
#endif
  MACROBLOCKD *const xd = &dcb->xd;
  MB_MODE_INFO *const mbmi = xd->mi[0];
  struct macroblockd_plane *const pd = &xd->plane[plane];

  const BLOCK_SIZE bsize = mbmi->bsize;
  assert(bsize < BLOCK_SIZES_ALL);
  const BLOCK_SIZE plane_bsize =
      get_plane_block_size(bsize, pd->subsampling_x, pd->subsampling_y);

  TXB_CTX txb_ctx;
  get_txb_ctx(plane_bsize, tx_size, plane, pd->above_entropy_context + col,
              pd->left_entropy_context + row, &txb_ctx);
  const uint8_t cul_level =
      av1_read_coeffs_txb(cm, dcb, r, row, col, plane, &txb_ctx, tx_size);
  av1_set_entropy_contexts(xd, pd, plane, plane_bsize, tx_size, cul_level, col,
                           row);

  if (is_inter_block(mbmi)) {
    const PLANE_TYPE plane_type = get_plane_type(plane);
    // tx_type will be read out in av1_read_coeffs_txb_facade
    const TX_TYPE tx_type = av1_get_tx_type(xd, plane_type, row, col, tx_size,
                                            cm->features.reduced_tx_set_used);

    if (plane == 0) {
      const int txw = tx_size_wide_unit[tx_size];
      const int txh = tx_size_high_unit[tx_size];
      // The 16x16 unit is due to the constraint from tx_64x64 which sets the
      // maximum tx size for chroma as 32x32. Coupled with 4x1 transform block
      // size, the constraint takes effect in 32x16 / 16x32 size too. To solve
      // the intricacy, cover all the 16x16 units inside a 64 level transform.
      if (txw == tx_size_wide_unit[TX_64X64] ||
          txh == tx_size_high_unit[TX_64X64]) {
        const int tx_unit = tx_size_wide_unit[TX_16X16];
        const int stride = xd->tx_type_map_stride;
        for (int idy = 0; idy < txh; idy += tx_unit) {
          for (int idx = 0; idx < txw; idx += tx_unit) {
            xd->tx_type_map[(row + idy) * stride + col + idx] = tx_type;
          }
        }
      }
    }
  }

#if TXCOEFF_TIMER
  aom_usec_timer_mark(&timer);
  const int64_t elapsed_time = aom_usec_timer_elapsed(&timer);
  cm->txcoeff_timer += elapsed_time;
  ++cm->txb_count;
#endif
}