Commit 87dfe848 authored by Kostya Shishkov's avatar Kostya Shishkov

Some optimization and fixes - mostly reworked MC and bitplanes.

Originally committed as revision 5543 to svn://svn.ffmpeg.org/ffmpeg/trunk
parent e8ba1cea
......@@ -203,30 +203,6 @@ enum CodingSet {
CS_HIGH_RATE_INTER
};
/** Bitplane struct
* We mainly need data and is_raw, so this struct could be avoided
* to save a level of indirection; feel free to modify
* @fixme For now, stride=width
* @warning Data are bits, either 1 or 0
*/
typedef struct BitPlane {
uint8_t *data; ///< Data buffer
int width; ///< Width of the buffer
int stride; ///< Stride of the buffer
int height; ///< Plane height
uint8_t is_raw; ///< Bit values must be read at MB level
} BitPlane;
/** Block data for DC/AC prediction
*/
typedef struct Block {
uint16_t dc;
int16_t hor_ac[7];
int16_t vert_ac[7];
int16_t dcstep, step;
} Block;
/** The VC1 Context
* @fixme Change size wherever another size is more efficient
* Many members are only used for Advanced Profile
......@@ -334,9 +310,11 @@ typedef struct VC1Context{
uint8_t *previous_line_cbpcy; ///< To use for predicted CBPCY
VLC *cbpcy_vlc; ///< CBPCY VLC table
int tt_index; ///< Index for Transform Type tables
BitPlane mv_type_mb_plane; ///< bitplane for mv_type == (4MV)
BitPlane skip_mb_plane; ///< bitplane for skipped MBs
BitPlane direct_mb_plane; ///< bitplane for "direct" MBs
uint8_t* mv_type_mb_plane; ///< bitplane for mv_type == (4MV)
uint8_t* skip_mb_plane; ///< bitplane for skipped MBs
// BitPlane direct_mb_plane; ///< bitplane for "direct" MBs
int mv_type_is_raw; ///< mv type mb plane is not coded
int skip_is_raw; ///< skip mb plane is not coded
/** Frame decoding info for S/M profiles only */
//@{
......@@ -359,8 +337,8 @@ typedef struct VC1Context{
int hrd_num_leaky_buckets;
uint8_t bit_rate_exponent;
uint8_t buffer_size_exponent;
BitPlane ac_pred_plane; ///< AC prediction flags bitplane
BitPlane over_flags_plane; ///< Overflags bitplane
// BitPlane ac_pred_plane; ///< AC prediction flags bitplane
// BitPlane over_flags_plane; ///< Overflags bitplane
uint8_t condover;
uint16_t *hrd_rate, *hrd_buffer;
uint8_t *hrd_fullness;
......@@ -436,11 +414,6 @@ static int vc1_init_common(VC1Context *v)
static int done = 0;
int i = 0;
/* Set the bit planes */
v->mv_type_mb_plane = (struct BitPlane) { NULL, 0, 0, 0 };
v->direct_mb_plane = (struct BitPlane) { NULL, 0, 0, 0 };
v->skip_mb_plane = (struct BitPlane) { NULL, 0, 0, 0 };
v->ac_pred_plane = v->over_flags_plane = (struct BitPlane) { NULL, 0, 0, 0 };
v->hrd_rate = v->hrd_buffer = NULL;
/* VLC tables */
......@@ -521,33 +494,6 @@ enum Imode {
};
/** @} */ //imode defines
/** Allocate the buffer from a bitplane, given its dimensions
* @param bp Bitplane which buffer is to allocate
* @param[in] width Width of the buffer
* @param[in] height Height of the buffer
* @return Status
* @todo TODO: Take into account stride
* @todo TODO: Allow use of external buffers ?
*/
static int alloc_bitplane(BitPlane *bp, int width, int height)
{
if (!bp || bp->width<0 || bp->height<0) return -1;
bp->data = (uint8_t*)av_malloc(width*height);
if (!bp->data) return -1;
bp->width = bp->stride = width;
bp->height = height;
return 0;
}
/** Free the bitplane's buffer
* @param bp Bitplane which buffer is to free
*/
static void free_bitplane(BitPlane *bp)
{
bp->width = bp->stride = bp->height = 0;
if (bp->data) av_freep(&bp->data);
}
/** Decode rows by checking if they are skipped
* @param plane Buffer to store decoded bits
* @param[in] width Width of this buffer
......@@ -595,53 +541,57 @@ static void decode_colskip(uint8_t* plane, int width, int height, int stride, Ge
* @fixme FIXME: Optimize
* @todo TODO: Decide if a struct is needed
*/
static int bitplane_decoding(BitPlane *bp, VC1Context *v)
static int bitplane_decoding(uint8_t* data, int *raw_flag, VC1Context *v)
{
GetBitContext *gb = &v->s.gb;
int imode, x, y, code, offset;
uint8_t invert, *planep = bp->data;
uint8_t invert, *planep = data;
int width, height, stride;
width = v->s.mb_width;
height = v->s.mb_height;
stride = v->s.mb_stride;
invert = get_bits(gb, 1);
imode = get_vlc2(gb, vc1_imode_vlc.table, VC1_IMODE_VLC_BITS, 1);
bp->is_raw = 0;
*raw_flag = 0;
switch (imode)
{
case IMODE_RAW:
//Data is actually read in the MB layer (same for all tests == "raw")
bp->is_raw = 1; //invert ignored
*raw_flag = 1; //invert ignored
return invert;
case IMODE_DIFF2:
case IMODE_NORM2:
if ((bp->height * bp->width) & 1)
if ((height * width) & 1)
{
*planep++ = get_bits(gb, 1);
offset = 1;
}
else offset = 0;
// decode bitplane as one long line
for (y = offset; y < bp->height * bp->width; y += 2) {
for (y = offset; y < height * width; y += 2) {
code = get_vlc2(gb, vc1_norm2_vlc.table, VC1_NORM2_VLC_BITS, 1);
*planep++ = code & 1;
offset++;
if(offset == bp->width) {
if(offset == width) {
offset = 0;
planep += bp->stride - bp->width;
planep += stride - width;
}
*planep++ = code >> 1;
offset++;
if(offset == bp->width) {
if(offset == width) {
offset = 0;
planep += bp->stride - bp->width;
planep += stride - width;
}
}
break;
case IMODE_DIFF6:
case IMODE_NORM6:
if(!(bp->height % 3) && (bp->width % 3)) { // use 2x3 decoding
for(y = 0; y < bp->height; y+= 3) {
for(x = bp->width & 1; x < bp->width; x += 2) {
if(!(height % 3) && (width % 3)) { // use 2x3 decoding
for(y = 0; y < height; y+= 3) {
for(x = width & 1; x < width; x += 2) {
code = get_vlc2(gb, vc1_norm6_vlc.table, VC1_NORM6_VLC_BITS, 2);
if(code < 0){
av_log(v->s.avctx, AV_LOG_DEBUG, "invalid NORM-6 VLC\n");
......@@ -649,17 +599,17 @@ static int bitplane_decoding(BitPlane *bp, VC1Context *v)
}
planep[x + 0] = (code >> 0) & 1;
planep[x + 1] = (code >> 1) & 1;
planep[x + 0 + bp->stride] = (code >> 2) & 1;
planep[x + 1 + bp->stride] = (code >> 3) & 1;
planep[x + 0 + bp->stride * 2] = (code >> 4) & 1;
planep[x + 1 + bp->stride * 2] = (code >> 5) & 1;
planep[x + 0 + stride] = (code >> 2) & 1;
planep[x + 1 + stride] = (code >> 3) & 1;
planep[x + 0 + stride * 2] = (code >> 4) & 1;
planep[x + 1 + stride * 2] = (code >> 5) & 1;
}
planep += bp->stride * 3;
planep += stride * 3;
}
if(bp->width & 1) decode_colskip(bp->data, 1, bp->height, bp->stride, &v->s.gb);
if(width & 1) decode_colskip(data, 1, height, stride, &v->s.gb);
} else { // 3x2
for(y = bp->height & 1; y < bp->height; y += 2) {
for(x = bp->width % 3; x < bp->width; x += 3) {
for(y = height & 1; y < height; y += 2) {
for(x = width % 3; x < width; x += 3) {
code = get_vlc2(gb, vc1_norm6_vlc.table, VC1_NORM6_VLC_BITS, 2);
if(code < 0){
av_log(v->s.avctx, AV_LOG_DEBUG, "invalid NORM-6 VLC\n");
......@@ -668,22 +618,22 @@ static int bitplane_decoding(BitPlane *bp, VC1Context *v)
planep[x + 0] = (code >> 0) & 1;
planep[x + 1] = (code >> 1) & 1;
planep[x + 2] = (code >> 2) & 1;
planep[x + 0 + bp->stride] = (code >> 3) & 1;
planep[x + 1 + bp->stride] = (code >> 4) & 1;
planep[x + 2 + bp->stride] = (code >> 5) & 1;
planep[x + 0 + stride] = (code >> 3) & 1;
planep[x + 1 + stride] = (code >> 4) & 1;
planep[x + 2 + stride] = (code >> 5) & 1;
}
planep += bp->stride * 2;
planep += stride * 2;
}
x = bp->width % 3;
if(x) decode_colskip(bp->data , x, bp->height , bp->stride, &v->s.gb);
if(bp->height & 1) decode_rowskip(bp->data+x, bp->width - x, bp->height & 1, bp->stride, &v->s.gb);
x = width % 3;
if(x) decode_colskip(data , x, height , stride, &v->s.gb);
if(height & 1) decode_rowskip(data+x, width - x, 1, stride, &v->s.gb);
}
break;
case IMODE_ROWSKIP:
decode_rowskip(bp->data, bp->width, bp->height, bp->stride, &v->s.gb);
decode_rowskip(data, width, height, stride, &v->s.gb);
break;
case IMODE_COLSKIP:
decode_colskip(bp->data, bp->width, bp->height, bp->stride, &v->s.gb);
decode_colskip(data, width, height, stride, &v->s.gb);
break;
default: break;
}
......@@ -691,28 +641,29 @@ static int bitplane_decoding(BitPlane *bp, VC1Context *v)
/* Applying diff operator */
if (imode == IMODE_DIFF2 || imode == IMODE_DIFF6)
{
planep = bp->data;
planep = data;
planep[0] ^= invert;
for (x=1; x<bp->width; x++)
for (x=1; x<width; x++)
planep[x] ^= planep[x-1];
for (y=1; y<bp->height; y++)
for (y=1; y<height; y++)
{
planep += bp->stride;
planep[0] ^= planep[-bp->stride];
for (x=1; x<bp->width; x++)
planep += stride;
planep[0] ^= planep[-stride];
for (x=1; x<width; x++)
{
if (planep[x-1] != planep[x-bp->stride]) planep[x] ^= invert;
else planep[x] ^= planep[x-1];
if (planep[x-1] != planep[x-stride]) planep[x] ^= invert;
else planep[x] ^= planep[x-1];
}
}
}
else if (invert)
{
planep = bp->data;
for (x=0; x<bp->width*bp->height; x++) planep[x] = !planep[x]; //FIXME stride
planep = data;
for (x=0; x<stride*height; x++) planep[x] = !planep[x]; //FIXME stride
}
return (imode<<1) + invert;
}
/** @} */ //Bitplane group
/***********************************************************************/
......@@ -886,6 +837,42 @@ static void vc1_overlap_block(MpegEncContext *s, DCTELEM block[64], int n, int d
}
static void vc1_v_overlap(uint8_t* src, int stride)
{
int i;
int a, b, c, d;
for(i = 0; i < 8; i++) {
a = src[-2*stride];
b = src[-stride];
c = src[0];
d = src[stride];
src[-2*stride] = (7*a + d) >> 3;
src[-stride] = (-a + 7*b + c + d) >> 3;
src[0] = (a + b + 7*c - d) >> 3;
src[stride] = (a + 7*d) >> 3;
src++;
}
}
static void vc1_h_overlap(uint8_t* src, int stride)
{
int i;
int a, b, c, d;
for(i = 0; i < 8; i++) {
a = src[-2];
b = src[-1];
c = src[0];
d = src[1];
src[-2] = (7*a + d) >> 3;
src[-1] = (-a + 7*b + c + d) >> 3;
src[0] = (a + b + 7*c - d) >> 3;
src[1] = (a + 7*d) >> 3;
src += stride;
}
}
/** Put block onto picture
* @todo move to DSPContext
*/
......@@ -910,6 +897,11 @@ static void vc1_put_block(VC1Context *v, DCTELEM block[6][64])
dsp->put_pixels_clamped(block[5], v->s.dest[2], vs);
}
/* clip motion vector as specified in 8.3.6.5 */
#define CLIP_RANGE(mv, src, lim, bs) \
if(mv < -bs) mv = -bs - src * bs; \
if(mv > lim) mv = lim - src * bs;
/** Do motion compensation over 1 macroblock
* Mostly adapted hpel_motion and qpel_motion from mpegvideo.c
*/
......@@ -918,89 +910,69 @@ static void vc1_mc_1mv(VC1Context *v)
MpegEncContext *s = &v->s;
DSPContext *dsp = &v->s.dsp;
uint8_t *srcY, *srcU, *srcV;
int dxy, mx, my, src_x, src_y;
int width = s->mb_width * 16, height = s->mb_height * 16;
int dxy, uvdxy, mx, my, uvmx, uvmy, src_x, src_y, uvsrc_x, uvsrc_y;
if(!v->s.last_picture.data[0])return;
mx = s->mv[0][0][0] >> s->mspel;
my = s->mv[0][0][1] >> s->mspel;
mx = s->mv[0][0][0];
my = s->mv[0][0][1];
uvmx = (mx + ((mx & 3) == 3)) >> 1;
uvmy = (my + ((my & 3) == 3)) >> 1;
srcY = s->last_picture.data[0];
srcU = s->last_picture.data[1];
srcV = s->last_picture.data[2];
if(s->mspel) { // hpel mc
if(v->fastuvmc) { // XXX: 8.3.5.4.5 specifies something different
uvmx = (uvmx + 1) >> 1;
uvmy = (uvmy + 1) >> 1;
}
src_x = s->mb_x * 16 + (mx >> 2);
src_y = s->mb_y * 16 + (my >> 2);
uvsrc_x = s->mb_x * 8 + (uvmx >> 2);
uvsrc_y = s->mb_y * 8 + (uvmy >> 2);
CLIP_RANGE( src_x, s->mb_x, s->mb_width * 16, 16);
CLIP_RANGE( src_y, s->mb_y, s->mb_height * 16, 16);
CLIP_RANGE(uvsrc_x, s->mb_x, s->mb_width * 8, 8);
CLIP_RANGE(uvsrc_y, s->mb_y, s->mb_height * 8, 8);
srcY += src_y * s->linesize + src_x;
srcU += uvsrc_y * s->uvlinesize + uvsrc_x;
srcV += uvsrc_y * s->uvlinesize + uvsrc_x;
if((unsigned)src_x > s->h_edge_pos - (mx&3) - 16
|| (unsigned)src_y > s->v_edge_pos - (my&3) - 16){
uint8_t *uvbuf= s->edge_emu_buffer + 18 * s->linesize;
ff_emulated_edge_mc(s->edge_emu_buffer, srcY, s->linesize, 16+1, 16+1,
src_x, src_y, s->h_edge_pos, s->v_edge_pos);
srcY = s->edge_emu_buffer;
ff_emulated_edge_mc(uvbuf , srcU, s->uvlinesize, 8+1, 8+1,
uvsrc_x, uvsrc_y, s->h_edge_pos >> 1, s->v_edge_pos >> 1);
ff_emulated_edge_mc(uvbuf + 16, srcV, s->uvlinesize, 8+1, 8+1,
uvsrc_x, uvsrc_y, s->h_edge_pos >> 1, s->v_edge_pos >> 1);
srcU = uvbuf;
srcV = uvbuf + 16;
}
if(!s->quarter_sample) { // hpel mc
mx >>= 1;
my >>= 1;
uvmx >>= 1;
uvmy >>= 1;
dxy = ((my & 1) << 1) | (mx & 1);
src_x = s->mb_x * 16 + (mx >> 1);
src_y = s->mb_y * 16 + (my >> 1);
/* src_x = clip(src_x, -16, width); //FIXME unneeded for emu?
if (src_x == width)
dxy &= ~1;
src_y = clip(src_y, -16, height);
if (src_y == height)
dxy &= ~2;*/
srcY += src_y * s->linesize + src_x;
srcU += (src_y >> 1) * s->uvlinesize + (src_x >> 1);
srcV += (src_y >> 1) * s->uvlinesize + (src_x >> 1);
if((unsigned)src_x > s->h_edge_pos - (mx&1) - 16
|| (unsigned)src_y > s->v_edge_pos - (my&1) - 16){
uint8_t *uvbuf= s->edge_emu_buffer + 18 * s->linesize;
ff_emulated_edge_mc(s->edge_emu_buffer, srcY, s->linesize, 16+1, 16+1,
src_x, src_y, s->h_edge_pos, s->v_edge_pos);
srcY = s->edge_emu_buffer;
ff_emulated_edge_mc(uvbuf, srcU, s->uvlinesize, 8+1, 8+1,
src_x >> 1, src_y >> 1, s->h_edge_pos >> 1, s->v_edge_pos >> 1);
ff_emulated_edge_mc(uvbuf + 16, srcV, s->uvlinesize, 8+1, 8+1,
src_x >> 1, src_y >> 1, s->h_edge_pos >> 1, s->v_edge_pos >> 1);
srcU = uvbuf;
srcV = uvbuf + 16;
}
uvdxy = 0;
dsp->put_no_rnd_pixels_tab[0][dxy](s->dest[0], srcY, s->linesize, 16);
dsp->put_no_rnd_pixels_tab[1][0](s->dest[1], srcU, s->uvlinesize, 8);
dsp->put_no_rnd_pixels_tab[1][0](s->dest[2], srcV, s->uvlinesize, 8);
} else {
int motion_x = mx, motion_y = my, uvdxy, uvsrc_x, uvsrc_y;
dxy = ((motion_y & 3) << 2) | (motion_x & 3);
src_x = s->mb_x * 16 + (mx >> 2);
src_y = s->mb_y * 16 + (my >> 2);
mx= motion_x/2;
my= motion_y/2;
mx= (mx>>1)|(mx&1);
my= (my>>1)|(my&1);
uvdxy= (mx&1) | ((my&1)<<1);
mx>>=1;
my>>=1;
uvsrc_x = s->mb_x * 8 + mx;
uvsrc_y = s->mb_y * 8 + my;
srcY = s->last_picture.data[0] + src_y * s->linesize + src_x;
srcU = s->last_picture.data[1] + uvsrc_y * s->uvlinesize + uvsrc_x;
srcV = s->last_picture.data[2] + uvsrc_y * s->uvlinesize + uvsrc_x;
if( (unsigned)src_x > s->h_edge_pos - (motion_x&3) - 16
|| (unsigned)src_y > s->v_edge_pos - (motion_y&3) - 16 ){
uint8_t *uvbuf= s->edge_emu_buffer + 18*s->linesize;
ff_emulated_edge_mc(s->edge_emu_buffer, srcY, s->linesize, 17, 17,
src_x, src_y, s->h_edge_pos, s->v_edge_pos);
srcY = s->edge_emu_buffer;
ff_emulated_edge_mc(uvbuf, srcU, s->uvlinesize, 9, 9,
uvsrc_x, uvsrc_y, s->h_edge_pos>>1, s->v_edge_pos>>1);
ff_emulated_edge_mc(uvbuf + 16, srcV, s->uvlinesize, 9, 9,
uvsrc_x, uvsrc_y, s->h_edge_pos>>1, s->v_edge_pos>>1);
srcU = uvbuf;
srcV = uvbuf + 16;
}
dxy = ((my & 3) << 2) | (mx & 3);
uvdxy = ((uvmy & 1) << 1) | (uvmx & 1);
dsp->put_no_rnd_qpel_pixels_tab[0][dxy](s->dest[0], srcY, s->linesize);
dsp->put_no_rnd_pixels_tab[1][uvdxy](s->dest[1], srcU, s->uvlinesize, 8);
dsp->put_no_rnd_pixels_tab[1][uvdxy](s->dest[2], srcV, s->uvlinesize, 8);
}
dsp->put_mspel_pixels_tab[uvdxy](s->dest[1], srcU, s->uvlinesize);
dsp->put_mspel_pixels_tab[uvdxy](s->dest[2], srcV, s->uvlinesize);
}
/**
......@@ -1215,9 +1187,9 @@ static int vc1_parse_frame_header(VC1Context *v, GetBitContext* gb)
v->lumshift = get_bits(gb, 6);
}
if(v->mv_mode == MV_PMODE_1MV_HPEL || v->mv_mode == MV_PMODE_1MV_HPEL_BILIN)
v->s.mspel = 1;
v->s.quarter_sample = 0;
else
v->s.mspel = 0;
v->s.quarter_sample = 1;
if(v->mv_mode != MV_PMODE_1MV && v->mv_mode != MV_PMODE_1MV_HPEL && v->mv_mode != MV_PMODE_1MV_HPEL_BILIN) {
av_log(v->s.avctx, AV_LOG_ERROR, "Only 1MV P-frames are supported by now\n");
......@@ -1227,12 +1199,15 @@ if(v->mv_mode != MV_PMODE_1MV && v->mv_mode != MV_PMODE_1MV_HPEL && v->mv_mode !
v->mv_mode2 == MV_PMODE_MIXED_MV)
|| v->mv_mode == MV_PMODE_MIXED_MV)
{
status = bitplane_decoding(&v->mv_type_mb_plane, v);
status = bitplane_decoding(v->mv_type_mb_plane, &v->mv_type_is_raw, v);
if (status < 0) return -1;
av_log(v->s.avctx, AV_LOG_DEBUG, "MB MV Type plane encoding: "
"Imode: %i, Invert: %i\n", status>>1, status&1);
} else {
v->mv_type_is_raw = 0;
memset(v->mv_type_mb_plane, 0, v->s.mb_stride * v->s.mb_height);
}
status = bitplane_decoding(&v->skip_mb_plane, v);
status = bitplane_decoding(v->skip_mb_plane, &v->skip_is_raw, v);
if (status < 0) return -1;
av_log(v->s.avctx, AV_LOG_DEBUG, "MB Skip plane encoding: "
"Imode: %i, Invert: %i\n", status>>1, status&1);
......@@ -1327,8 +1302,8 @@ if(v->mv_mode != MV_PMODE_1MV && v->mv_mode != MV_PMODE_1MV_HPEL && v->mv_mode !
if (!index) { _dmv_x = _dmv_y = 0; } \
else if (index == 35) \
{ \
_dmv_x = get_bits(gb, v->k_x - s->mspel); \
_dmv_y = get_bits(gb, v->k_y - s->mspel); \
_dmv_x = get_bits(gb, v->k_x - 1 + s->quarter_sample); \
_dmv_y = get_bits(gb, v->k_y - 1 + s->quarter_sample); \
} \
else if (index == 36) \
{ \
......@@ -1339,8 +1314,8 @@ if(v->mv_mode != MV_PMODE_1MV && v->mv_mode != MV_PMODE_1MV_HPEL && v->mv_mode !
else \
{ \
index1 = index%6; \
if (s->mspel && index1 == 5) val = 1; \
else val = 0; \
if (!s->quarter_sample && index1 == 5) val = 1; \
else val = 0; \
if(size_table[index1] - val > 0) \
val = get_bits(gb, size_table[index1] - val); \
else val = 0; \
......@@ -1348,8 +1323,8 @@ if(v->mv_mode != MV_PMODE_1MV && v->mv_mode != MV_PMODE_1MV_HPEL && v->mv_mode !
_dmv_x = (sign ^ ((val>>1) + offset_table[index1])) - sign; \
\
index1 = index/6; \
if (s->mspel && index1 == 5) val = 1; \
else val = 0; \
if (!s->quarter_sample && index1 == 5) val = 1; \
else val = 0; \
if(size_table[index1] - val > 0) \
val = get_bits(gb, size_table[index1] - val); \
else val = 0; \
......@@ -1368,8 +1343,8 @@ static inline void vc1_pred_mv(MpegEncContext *s, int dmv_x, int dmv_y, int mv1,
int mb_pos = s->mb_x + s->mb_y * s->mb_stride;
/* scale MV difference to be quad-pel */
dmv_x <<= s->mspel;
dmv_y <<= s->mspel;
dmv_x <<= 1 - s->quarter_sample;
dmv_y <<= 1 - s->quarter_sample;
wrap = s->b8_stride;
xy = s->block_index[0];
......@@ -1627,26 +1602,22 @@ static void vc1_decode_ac_coeff(VC1Context *v, int *last, int *skip, int *value,
level = -level;
} else {
escape = decode210(gb);
if (escape == 0) {
index = get_vlc2(gb, vc1_ac_coeff_table[codingset].table, AC_VLC_BITS, 3);
run = vc1_index_decode_table[codingset][index][0];
level = vc1_index_decode_table[codingset][index][1];
lst = index >= vc1_last_decode_table[codingset];
if(lst)
level += vc1_last_delta_level_table[codingset][run];
else
level += vc1_delta_level_table[codingset][run];
if(get_bits(gb, 1))
level = -level;
} else if (escape == 1) {
if (escape != 2) {
index = get_vlc2(gb, vc1_ac_coeff_table[codingset].table, AC_VLC_BITS, 3);
run = vc1_index_decode_table[codingset][index][0];
level = vc1_index_decode_table[codingset][index][1];
lst = index >= vc1_last_decode_table[codingset];
if(lst)
run += vc1_last_delta_run_table[codingset][level] + 1;
else
run += vc1_delta_run_table[codingset][level] + 1;
if(escape == 0) {
if(lst)
level += vc1_last_delta_level_table[codingset][run];
else
level += vc1_delta_level_table[codingset][run];
} else {
if(lst)
run += vc1_last_delta_run_table[codingset][level] + 1;
else
run += vc1_delta_run_table[codingset][level] + 1;
}
if(get_bits(gb, 1))
level = -level;
} else {
......@@ -1944,7 +1915,7 @@ static int vc1_decode_intra_block(VC1Context *v, DCTELEM block[64], int n, int c
}
/* apply AC prediction if needed */
if(s->ac_pred) {
if(s->ac_pred && (v->a_avail || v->c_avail)) {
/* scale predictors if needed*/
int mb_pos2, q1, q2;
......@@ -1960,7 +1931,7 @@ static int vc1_decode_intra_block(VC1Context *v, DCTELEM block[64], int n, int c
memset(ac_val + 8, 0, 8 * sizeof(ac_val[0]));
dc_pred_dir = 1;
}
if(!q1 && q1 && q2 && q1 != q2) {
if(q2 && q1 != q2) {
q1 = q1 * 2 - 1;
q2 = q2 * 2 - 1;
......@@ -2017,16 +1988,16 @@ not_coded:
memset(ac_val2, 0, 16 * 2);
if(dc_pred_dir) {//left
ac_val -= 16;
if(s->ac_pred)
if(s->ac_pred && (v->a_avail || v->c_avail))
memcpy(ac_val2, ac_val, 8 * 2);
} else {//top
ac_val -= 16 * s->block_wrap[n];
if(s->ac_pred)
if(s->ac_pred && (v->a_avail || v->c_avail))
memcpy(ac_val2 + 8, ac_val + 8, 8 * 2);
}
/* apply AC prediction if needed */
if(s->ac_pred) {
if(s->ac_pred && (v->a_avail || v->c_avail)) {
if(dc_pred_dir) { //left
for(k = 1; k < 8; k++) {
block[k << 3] = ac_val[k] * scale;
......@@ -2082,7 +2053,6 @@ static int vc1_decode_p_block(VC1Context *v, DCTELEM block[64], int n, int mquan
ttblk = TT_4X8;
subblkpat = 2 - (ttblk == TT_4X8_LEFT);
}
switch(ttblk) {
case TT_8X8:
i = 0;
......@@ -2101,7 +2071,7 @@ static int vc1_decode_p_block(VC1Context *v, DCTELEM block[64], int n, int mquan
for(j = 0; j < 4; j++) {
last = subblkpat & (1 << (3 - j));
i = 0;
off = (j & 1) * 4 + (j & 2) * 32;
off = (j & 1) * 4 + (j & 2) * 16;
while (!last) {
vc1_decode_ac_coeff(v, &last, &skip, &value, v->codingset2);
i += skip;
......@@ -2110,7 +2080,8 @@ static int vc1_decode_p_block(VC1Context *v, DCTELEM block[64], int n, int mquan
idx = vc1_simple_progressive_4x4_zz[i++];
block[idx + off] = value * scale;
}
vc1_inv_trans(block + off, 4, 4);
if(!(subblkpat & (1 << (3 - j))))
vc1_inv_trans(block + off, 4, 4);
}
break;
case TT_8X4:
......@@ -2126,7 +2097,8 @@ static int vc1_decode_p_block(VC1Context *v, DCTELEM block[64], int n, int mquan
idx = vc1_simple_progressive_8x4_zz[i++];
block[idx + off] = value * scale;
}
if(!(subblkpat & (1 << (1 - j)))) vc1_inv_trans(block + off, 8, 4);
if(!(subblkpat & (1 << (1 - j))))
vc1_inv_trans(block + off, 8, 4);
}
break;
case TT_4X8:
......@@ -2142,7 +2114,8 @@ static int vc1_decode_p_block(VC1Context *v, DCTELEM block[64], int n, int mquan
idx = vc1_simple_progressive_8x4_zz[i++];
block[idx + off] = value * scale;
}
vc1_inv_trans(block + off, 4, 8);
if(!(subblkpat & (1 << (1 - j))))
vc1_inv_trans(block + off, 4, 8);
}
break;
}
......@@ -2158,7 +2131,7 @@ static int vc1_decode_p_mb(VC1Context *v, DCTELEM block[6][64])
{
MpegEncContext *s = &v->s;
GetBitContext *gb = &s->gb;
int i, j, mb_offset = s->mb_x + s->mb_y*s->mb_width; /* XXX: mb_stride */
int i, j;
int mb_pos = s->mb_x + s->mb_y * s->mb_stride;
int cbp; /* cbp decoding stuff */
int hybrid_pred; /* Prediction types */
......@@ -2174,17 +2147,22 @@ static int vc1_decode_p_mb(VC1Context *v, DCTELEM block[6][64])
int val, sign; /* temp values */
int first_block = 1;
int dst_idx, off;
int skipped, fourmv;
mquant = v->pq; /* Loosy initialization */
if (v->mv_type_mb_plane.is_raw)
v->mv_type_mb_plane.data[mb_offset] = get_bits(gb, 1);
if (v->skip_mb_plane.is_raw)
v->skip_mb_plane.data[mb_offset] = get_bits(gb, 1);
s->current_picture.mbskip_table[mb_pos] = v->skip_mb_plane.data[mb_offset];
if (!v->mv_type_mb_plane.data[mb_offset]) /* 1MV mode */
if (v->mv_type_is_raw)
fourmv = get_bits1(gb);
else
fourmv = v->mv_type_mb_plane[mb_pos];
if (v->skip_is_raw)
skipped = get_bits1(gb);
else
skipped = v->skip_mb_plane[mb_pos];
if (!fourmv) /* 1MV mode */
{
if (!v->skip_mb_plane.data[mb_offset])
if (!skipped)
{
GET_MVDATA(dmv_x, dmv_y);
......@@ -2232,19 +2210,21 @@ static int vc1_decode_p_mb(VC1Context *v, DCTELEM block[6][64])
v->c_avail = 1;
vc1_decode_intra_block(v, block[i], i, val, mquant, (i&4)?v->codingset2:v->codingset);
vc1_inv_trans(s->block[i], 8, 8);
for(j = 0; j < 64; j++) s->block[i][j] += 128;
s->dsp.put_pixels_clamped(s->block[i], s->dest[dst_idx] + off, s->linesize >> ((i & 4) >> 2));
vc1_inv_trans(block[i], 8, 8);
for(j = 0; j < 64; j++) block[i][j] += 128;
s->dsp.put_pixels_clamped(block[i], s->dest[dst_idx] + off, s->linesize >> ((i & 4) >> 2));
/* TODO: proper loop filtering */
if(v->a_avail)
s->dsp.h263_v_loop_filter(s->dest[dst_idx] + off, s->linesize >> ((i & 4) >> 2), s->y_dc_scale);
if(v->c_avail)
s->dsp.h263_h_loop_filter(s->dest[dst_idx] + off, s->linesize >> ((i & 4) >> 2), s->y_dc_scale);
if(v->pq >= 9 && v->overlap) {
if(v->a_avail)
s->dsp.h263_v_loop_filter(s->dest[dst_idx] + off, s->linesize >> ((i & 4) >> 2), s->y_dc_scale);
if(v->c_avail)
s->dsp.h263_h_loop_filter(s->dest[dst_idx] + off, s->linesize >> ((i & 4) >> 2), s->y_dc_scale);
}
} else if(val) {
vc1_decode_p_block(v, block[i], i, mquant, ttmb, first_block);
if(!v->ttmbf && ttmb < 8) ttmb = -1;
first_block = 0;
s->dsp.add_pixels_clamped(s->block[i], s->dest[dst_idx] + off, (i&4)?s->uvlinesize:s->linesize);
s->dsp.add_pixels_clamped(block[i], s->dest[dst_idx] + off, (i&4)?s->uvlinesize:s->linesize);
}
}
}
......@@ -2259,7 +2239,7 @@ static int vc1_decode_p_mb(VC1Context *v, DCTELEM block[6][64])
} //1MV mode
else //4MV mode
{//FIXME: looks not conforming to standard and is not even theoretically complete
if (!v->skip_mb_plane.data[mb_offset] /* unskipped MB */)
if (!skipped /* unskipped MB */)
{
int blk_intra[4], blk_coded[4];
/* Get CBPCY */
......@@ -2495,6 +2475,8 @@ static int vc1_decode_init(AVCodecContext *avctx)
if (!avctx->extradata_size || !avctx->extradata) return -1;
avctx->pix_fmt = PIX_FMT_YUV420P;
v->s.avctx = avctx;
avctx->flags |= CODEC_FLAG_EMU_EDGE;
v->s.flags |= CODEC_FLAG_EMU_EDGE;
if(ff_h263_decode_init(avctx) < 0)
return -1;
......@@ -2503,7 +2485,6 @@ static int vc1_decode_init(AVCodecContext *avctx)
av_log(avctx, AV_LOG_INFO, "This decoder is not supposed to produce picture. Dont report this as a bug!\n");
av_log(avctx, AV_LOG_INFO, "If you see a picture, don't believe your eyes.\n");
avctx->flags |= CODEC_FLAG_EMU_EDGE;
avctx->coded_width = avctx->width;
avctx->coded_height = avctx->height;
if (avctx->codec_id == CODEC_ID_WMV3)
......@@ -2537,14 +2518,8 @@ static int vc1_decode_init(AVCodecContext *avctx)
s->mb_height = (avctx->coded_height+15)>>4;
/* Allocate mb bitplanes */
if (alloc_bitplane(&v->mv_type_mb_plane, s->mb_width, s->mb_height) < 0)
return -1;
if (alloc_bitplane(&v->mv_type_mb_plane, s->mb_width, s->mb_height) < 0)
return -1;
if (alloc_bitplane(&v->skip_mb_plane, s->mb_width, s->mb_height) < 0)
return -1;
if (alloc_bitplane(&v->direct_mb_plane, s->mb_width, s->mb_height) < 0)
return -1;
v->mv_type_mb_plane = av_malloc(s->mb_stride * s->mb_height);
v->skip_mb_plane = av_malloc(s->mb_stride * s->mb_height);
/* For predictors */
v->previous_line_cbpcy = (uint8_t *)av_malloc(s->mb_stride*4);
......@@ -2553,10 +2528,10 @@ static int vc1_decode_init(AVCodecContext *avctx)
/* Init coded blocks info */
if (v->profile == PROFILE_ADVANCED)
{
if (alloc_bitplane(&v->over_flags_plane, s->mb_width, s->mb_height) < 0)
return -1;
if (alloc_bitplane(&v->ac_pred_plane, s->mb_width, s->mb_height) < 0)
return -1;
// if (alloc_bitplane(&v->over_flags_plane, s->mb_width, s->mb_height) < 0)
// return -1;
// if (alloc_bitplane(&v->ac_pred_plane, s->mb_width, s->mb_height) < 0)
// return -1;
}
return 0;
......@@ -2670,9 +2645,8 @@ static int vc1_decode_end(AVCodecContext *avctx)
av_freep(&v->hrd_rate);
av_freep(&v->hrd_buffer);
MPV_common_end(&v->s);
free_bitplane(&v->mv_type_mb_plane);
free_bitplane(&v->skip_mb_plane);
free_bitplane(&v->direct_mb_plane);
av_freep(&v->mv_type_mb_plane);
av_freep(&v->skip_mb_plane);
return 0;
}
......
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