Commit e6deea4e authored by John Koleszar's avatar John Koleszar

detokenize: use consistent structure for all block sizes

Restructure the code to avoid the majority of per-block-size
switches, code duplication, etc. All block types (mb/sb32/sb64)
can be handled by the same code.

Change-Id: I4022718d66e31a15a7074e43f3b98cd0a5124ea7
parent 05a79f2f
......@@ -215,6 +215,14 @@ typedef enum {
BLOCK_SIZE_SB64X64 = 2,
} BLOCK_SIZE_TYPE;
typedef enum {
BLOCK_4X4_LG2 = 0,
BLOCK_8X8_LG2 = 2,
BLOCK_16X16_LG2 = 4,
BLOCK_32X32_LG2 = 6,
BLOCK_64X64_LG2 = 8
} BLOCK_SIZE_LG2;
typedef struct {
MB_PREDICTION_MODE mode, uv_mode;
#if CONFIG_COMP_INTERINTRA_PRED
......@@ -299,6 +307,9 @@ struct mb_plane {
DECLARE_ALIGNED(16, int16_t, qcoeff[64 * 64]);
DECLARE_ALIGNED(16, int16_t, dqcoeff[64 * 64]);
DECLARE_ALIGNED(16, uint16_t, eobs[256]);
PLANE_TYPE plane_type;
int subsampling_x;
int subsampling_y;
};
#define BLOCK_OFFSET(x, i, n) ((x) + (i) * (n))
......
......@@ -70,7 +70,7 @@ static void setup_macroblock(MACROBLOCKD *mb, BLOCKSET bs) {
}
void vp9_setup_block_dptrs(MACROBLOCKD *mb) {
int r, c;
int r, c, i;
BLOCKD *blockd = mb->block;
for (r = 0; r < 4; r++) {
......@@ -99,6 +99,12 @@ void vp9_setup_block_dptrs(MACROBLOCKD *mb) {
blockd[to].predictor = &mb->predictor[from];
}
}
for (i = 0; i < MAX_MB_PLANE; i++) {
mb->plane[i].plane_type = i ? PLANE_TYPE_UV : PLANE_TYPE_Y_WITH_DC;
mb->plane[i].subsampling_x = !!i;
mb->plane[i].subsampling_y = !!i;
}
}
void vp9_build_block_doffsets(MACROBLOCKD *mb) {
......
......@@ -380,182 +380,122 @@ static int get_eob(MACROBLOCKD* const xd, int segment_id, int eob_max) {
return vp9_get_segdata(xd, segment_id, SEG_LVL_SKIP) ? 0 : eob_max;
}
static INLINE int decode_sb(VP9D_COMP* const pbi,
MACROBLOCKD* const xd,
BOOL_DECODER* const bc,
int offset, int count, int inc,
int eob_max, TX_SIZE tx_size) {
const int segment_id = xd->mode_info_context->mbmi.segment_id;
const int seg_eob = get_eob(xd, segment_id, eob_max);
/* TODO(jkoleszar): Probably best to remove instances that require this,
* as the data likely becomes per-plane and stored in the per-plane structures.
* This is a stub to work with the existing code.
*/
static INLINE int block_idx_4x4(MACROBLOCKD* const xd, int block_size_b,
int plane, int i) {
const int luma_blocks = 1 << block_size_b;
assert(xd->plane[0].subsampling_x == 0);
assert(xd->plane[0].subsampling_y == 0);
assert(xd->plane[1].subsampling_x == 1);
assert(xd->plane[1].subsampling_y == 1);
assert(xd->plane[2].subsampling_x == 1);
assert(xd->plane[2].subsampling_y == 1);
return plane == 0 ? i :
plane == 1 ? luma_blocks + i :
luma_blocks * 5 / 4 + i;
}
static INLINE int decode_block_plane(VP9D_COMP* const pbi,
MACROBLOCKD* const xd,
BOOL_DECODER* const bc,
BLOCK_SIZE_LG2 block_size,
int segment_id,
int plane,
int is_split) {
// block and transform sizes, in number of 4x4 blocks log 2 ("*_b")
// 4x4=0, 8x8=2, 16x16=4, 32x32=6, 64x64=8
const TX_SIZE tx_size = xd->mode_info_context->mbmi.txfm_size;
const BLOCK_SIZE_LG2 block_size_b = block_size;
const BLOCK_SIZE_LG2 txfrm_size_b = tx_size * 2;
// subsampled size of the block
const int ss_sum = xd->plane[plane].subsampling_x +
xd->plane[plane].subsampling_y;
const BLOCK_SIZE_LG2 ss_block_size = block_size_b - ss_sum;
// size of the transform to use. scale the transform down if it's larger
// than the size of the subsampled data, or forced externally by the mb mode.
const int ss_max = MAX(xd->plane[plane].subsampling_x,
xd->plane[plane].subsampling_y);
const BLOCK_SIZE_LG2 ss_txfrm_size = txfrm_size_b > ss_block_size || is_split
? txfrm_size_b - ss_max * 2
: txfrm_size_b;
const TX_SIZE ss_tx_size = ss_txfrm_size / 2;
// TODO(jkoleszar): 1 may not be correct here with larger chroma planes.
const int inc = is_split ? 1 : (1 << ss_txfrm_size);
// find the maximum eob for this transform size, adjusted by segment
const int seg_eob = get_eob(xd, segment_id, 16 << ss_txfrm_size);
int i, eobtotal = 0;
assert(count == offset * 3 / 2);
assert(txfrm_size_b <= block_size_b);
assert(ss_txfrm_size <= ss_block_size);
// luma blocks
for (i = 0; i < offset; i += inc) {
const int c = decode_coefs(pbi, xd, bc, i, PLANE_TYPE_Y_WITH_DC, seg_eob,
BLOCK_OFFSET(xd->plane[0].qcoeff, i, 16),
tx_size);
xd->plane[0].eobs[i] = c;
eobtotal += c;
}
// step through the block by the size of the transform in use.
for (i = 0; i < (1 << ss_block_size); i += inc) {
const int block_idx = block_idx_4x4(xd, block_size_b, plane, i);
// chroma blocks
for (i = offset; i < offset * 5 / 4; i += inc) {
const int b = i - offset;
const int c = decode_coefs(pbi, xd, bc, i, PLANE_TYPE_UV, seg_eob,
BLOCK_OFFSET(xd->plane[1].qcoeff, b, 16),
tx_size);
xd->plane[1].eobs[b] = c;
const int c = decode_coefs(pbi, xd, bc, block_idx,
xd->plane[plane].plane_type, seg_eob,
BLOCK_OFFSET(xd->plane[plane].qcoeff, i, 16),
ss_tx_size);
xd->plane[plane].eobs[i] = c;
eobtotal += c;
}
for (i = offset * 5 / 4; i < count; i += inc) {
const int b = i - offset * 5 / 4;
const int c = decode_coefs(pbi, xd, bc, i, PLANE_TYPE_UV, seg_eob,
BLOCK_OFFSET(xd->plane[2].qcoeff, b, 16),
tx_size);
xd->plane[2].eobs[b] = c;
eobtotal += c;
}
return eobtotal;
}
int vp9_decode_sb_tokens(VP9D_COMP* const pbi,
MACROBLOCKD* const xd,
BOOL_DECODER* const bc) {
switch (xd->mode_info_context->mbmi.txfm_size) {
case TX_32X32: {
// 32x32 luma block
const int segment_id = xd->mode_info_context->mbmi.segment_id;
int eobtotal = 0, seg_eob;
int c = decode_coefs(pbi, xd, bc, 0, PLANE_TYPE_Y_WITH_DC,
get_eob(xd, segment_id, 1024),
xd->plane[0].qcoeff, TX_32X32);
xd->plane[0].eobs[0] = c;
eobtotal += c;
// 16x16 chroma blocks
seg_eob = get_eob(xd, segment_id, 256);
c = decode_coefs(pbi, xd, bc, 64, PLANE_TYPE_UV, seg_eob,
xd->plane[1].qcoeff, TX_16X16);
xd->plane[1].eobs[0] = c;
eobtotal += c;
c = decode_coefs(pbi, xd, bc, 80, PLANE_TYPE_UV, seg_eob,
xd->plane[2].qcoeff, TX_16X16);
xd->plane[2].eobs[0] = c;
eobtotal += c;
return eobtotal;
}
case TX_16X16:
return decode_sb(pbi, xd, bc, 64, 96, 16, 16 * 16, TX_16X16);
case TX_8X8:
return decode_sb(pbi, xd, bc, 64, 96, 4, 8 * 8, TX_8X8);
case TX_4X4:
return decode_sb(pbi, xd, bc, 64, 96, 1, 4 * 4, TX_4X4);
default:
assert(0);
return 0;
static INLINE int decode_blocks_helper(VP9D_COMP* const pbi,
MACROBLOCKD* const xd,
BOOL_DECODER* const bc,
int block_size,
int is_split_chroma) {
const int segment_id = xd->mode_info_context->mbmi.segment_id;
int plane, eobtotal = 0;
for (plane = 0; plane < MAX_MB_PLANE; plane++) {
const int is_split = is_split_chroma &&
xd->plane[plane].plane_type == PLANE_TYPE_UV;
eobtotal += decode_block_plane(pbi, xd, bc, block_size, segment_id,
plane, is_split);
}
return eobtotal;
}
static INLINE int decode_blocks(VP9D_COMP* const pbi,
MACROBLOCKD* const xd,
BOOL_DECODER* const bc,
int block_size) {
const MB_PREDICTION_MODE mode = xd->mode_info_context->mbmi.mode;
const TX_SIZE tx_size = xd->mode_info_context->mbmi.txfm_size;
return decode_blocks_helper(pbi, xd, bc, block_size,
tx_size == TX_8X8 && (mode == I8X8_PRED || mode == SPLITMV));
}
int vp9_decode_sb64_tokens(VP9D_COMP* const pbi,
MACROBLOCKD* const xd,
BOOL_DECODER* const bc) {
switch (xd->mode_info_context->mbmi.txfm_size) {
case TX_32X32:
return decode_sb(pbi, xd, bc, 256, 384, 64, 32 * 32, TX_32X32);
case TX_16X16:
return decode_sb(pbi, xd, bc, 256, 384, 16, 16 * 16, TX_16X16);
case TX_8X8:
return decode_sb(pbi, xd, bc, 256, 384, 4, 8 * 8, TX_8X8);
case TX_4X4:
return decode_sb(pbi, xd, bc, 256, 384, 1, 4 * 4, TX_4X4);
default:
assert(0);
return 0;
}
return decode_blocks(pbi, xd, bc, BLOCK_64X64_LG2);
}
static int vp9_decode_mb_tokens_16x16(VP9D_COMP* const pbi,
MACROBLOCKD* const xd,
BOOL_DECODER* const bc) {
const int segment_id = xd->mode_info_context->mbmi.segment_id;
int eobtotal = 0, seg_eob;
// Luma block
int c = decode_coefs(pbi, xd, bc, 0, PLANE_TYPE_Y_WITH_DC,
get_eob(xd, segment_id, 256),
xd->plane[0].qcoeff, TX_16X16);
xd->plane[0].eobs[0] = c;
eobtotal += c;
// 8x8 chroma blocks
seg_eob = get_eob(xd, segment_id, 64);
c = decode_coefs(pbi, xd, bc, 16, PLANE_TYPE_UV,
seg_eob, xd->plane[1].qcoeff, TX_8X8);
xd->plane[1].eobs[0] = c;
eobtotal += c;
c = decode_coefs(pbi, xd, bc, 20, PLANE_TYPE_UV,
seg_eob, xd->plane[2].qcoeff, TX_8X8);
xd->plane[2].eobs[0] = c;
eobtotal += c;
return eobtotal;
int vp9_decode_sb_tokens(VP9D_COMP* const pbi,
MACROBLOCKD* const xd,
BOOL_DECODER* const bc) {
return decode_blocks(pbi, xd, bc, BLOCK_32X32_LG2);
}
static int vp9_decode_mb_tokens_8x8(VP9D_COMP* const pbi,
MACROBLOCKD* const xd,
BOOL_DECODER* const bc) {
int i, eobtotal = 0;
const int segment_id = xd->mode_info_context->mbmi.segment_id;
// luma blocks
int seg_eob = get_eob(xd, segment_id, 64);
for (i = 0; i < 16; i += 4) {
const int c = decode_coefs(pbi, xd, bc, i, PLANE_TYPE_Y_WITH_DC, seg_eob,
BLOCK_OFFSET(xd->plane[0].qcoeff, i, 16),
TX_8X8);
xd->plane[0].eobs[i] = c;
eobtotal += c;
}
// chroma blocks
if (xd->mode_info_context->mbmi.mode == I8X8_PRED ||
xd->mode_info_context->mbmi.mode == SPLITMV) {
// use 4x4 transform for U, V components in I8X8/splitmv prediction mode
seg_eob = get_eob(xd, segment_id, 16);
for (i = 16; i < 20; i++) {
const int c = decode_coefs(pbi, xd, bc, i, PLANE_TYPE_UV, seg_eob,
BLOCK_OFFSET(xd->plane[1].qcoeff, i - 16, 16),
TX_4X4);
xd->plane[1].eobs[i - 16] = c;
eobtotal += c;
}
for (i = 20; i < 24; i++) {
const int c = decode_coefs(pbi, xd, bc, i, PLANE_TYPE_UV, seg_eob,
BLOCK_OFFSET(xd->plane[2].qcoeff, i - 20, 16),
TX_4X4);
xd->plane[2].eobs[i - 20] = c;
eobtotal += c;
}
} else {
int c;
c = decode_coefs(pbi, xd, bc, 16, PLANE_TYPE_UV, seg_eob,
xd->plane[1].qcoeff, TX_8X8);
xd->plane[1].eobs[0] = c;
eobtotal += c;
c = decode_coefs(pbi, xd, bc, 20, PLANE_TYPE_UV, seg_eob,
xd->plane[2].qcoeff, TX_8X8);
xd->plane[2].eobs[0] = c;
eobtotal += c;
}
return eobtotal;
int vp9_decode_mb_tokens(VP9D_COMP* const pbi,
MACROBLOCKD* const xd,
BOOL_DECODER* const bc) {
return decode_blocks(pbi, xd, bc, BLOCK_16X16_LG2);
}
#if CONFIG_NEWBINTRAMODES
static int decode_coefs_4x4(VP9D_COMP *dx, MACROBLOCKD *xd,
BOOL_DECODER* const bc,
PLANE_TYPE type, int i, int seg_eob) {
......@@ -588,39 +528,6 @@ int vp9_decode_mb_tokens_4x4_uv(VP9D_COMP* const dx,
return decode_mb_tokens_4x4_uv(dx, xd, bc, seg_eob);
}
static int vp9_decode_mb_tokens_4x4(VP9D_COMP* const dx,
MACROBLOCKD* const xd,
BOOL_DECODER* const bc) {
int i, eobtotal = 0;
const int segment_id = xd->mode_info_context->mbmi.segment_id;
const int seg_eob = get_eob(xd, segment_id, 16);
// luma blocks
for (i = 0; i < 16; ++i)
eobtotal += decode_coefs_4x4(dx, xd, bc, PLANE_TYPE_Y_WITH_DC, i, seg_eob);
// chroma blocks
eobtotal += decode_mb_tokens_4x4_uv(dx, xd, bc, seg_eob);
return eobtotal;
}
int vp9_decode_mb_tokens(VP9D_COMP* const dx,
MACROBLOCKD* const xd,
BOOL_DECODER* const bc) {
const TX_SIZE tx_size = xd->mode_info_context->mbmi.txfm_size;
switch (tx_size) {
case TX_16X16:
return vp9_decode_mb_tokens_16x16(dx, xd, bc);
case TX_8X8:
return vp9_decode_mb_tokens_8x8(dx, xd, bc);
default:
assert(tx_size == TX_4X4);
return vp9_decode_mb_tokens_4x4(dx, xd, bc);
}
}
#if CONFIG_NEWBINTRAMODES
int vp9_decode_coefs_4x4(VP9D_COMP *dx, MACROBLOCKD *xd,
BOOL_DECODER* const bc,
PLANE_TYPE type, int i) {
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment