From 91029be79005b17311fcf75f3f9f977f4d20cfd7 Mon Sep 17 00:00:00 2001 From: Michael Niedermayer Date: Fri, 19 Apr 2002 03:25:20 +0000 Subject: [PATCH] much better ME for b frames (a bit slow though) fixed MC rounding for b frames fixed hq mode with b-frames Originally committed as revision 406 to svn://svn.ffmpeg.org/ffmpeg/trunk --- libavcodec/common.h | 10 ++ libavcodec/h263.c | 56 ++++++--- libavcodec/motion_est.c | 269 ++++++++++++++++++++++++++++++++++++---- libavcodec/mpegvideo.c | 194 ++++++++++++++++++++++------- libavcodec/mpegvideo.h | 11 +- 5 files changed, 446 insertions(+), 94 deletions(-) diff --git a/libavcodec/common.h b/libavcodec/common.h index 96b545d0dc..21cdd541b4 100644 --- a/libavcodec/common.h +++ b/libavcodec/common.h @@ -881,6 +881,16 @@ static inline int mid_pred(int a, int b, int c) return a + b + c - vmin - vmax; } +static inline int clip(int a, int amin, int amax) +{ + if (a < amin) + return amin; + else if (a > amax) + return amax; + else + return a; +} + /* memory */ void *av_mallocz(int size); diff --git a/libavcodec/h263.c b/libavcodec/h263.c index bcb9ea2508..42ae076ce4 100644 --- a/libavcodec/h263.c +++ b/libavcodec/h263.c @@ -18,7 +18,7 @@ * along with this program; if not, write to the Free Software * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. * - * ac prediction encoding by Michael Niedermayer + * ac prediction encoding & b-frame support by Michael Niedermayer */ #include "common.h" #include "dsputil.h" @@ -282,7 +282,7 @@ void mpeg4_encode_mb(MpegEncContext * s, s->mv[0][0][1]= s->mv[1][0][0]= s->mv[1][0][1]= 0; -// s->mv_dir= MV_DIR_FORWARD; //doesnt matter + s->mv_dir= MV_DIR_FORWARD; //doesnt matter return; } @@ -334,7 +334,8 @@ void mpeg4_encode_mb(MpegEncContext * s, s->last_mv[0][0][0]= motion_x; s->last_mv[0][0][1]= motion_y; break; - default: + default: + printf("unknown mb type\n"); return; } bits= get_bit_count(&s->pb); @@ -959,6 +960,31 @@ static void put_string(PutBitContext * pbc, char *s) put_bits(pbc, 8, 0); } +/* must be called before writing the header */ +void ff_set_mpeg4_time(MpegEncContext * s, int picture_number){ + int time_div, time_mod; + + if(s->pict_type==I_TYPE){ //we will encode a vol header + s->time_increment_resolution= s->frame_rate/ff_gcd(s->frame_rate, FRAME_RATE_BASE); + if(s->time_increment_resolution>=256*256) s->time_increment_resolution= 256*128; + + s->time_increment_bits = av_log2(s->time_increment_resolution - 1) + 1; + } + + s->time= picture_number*(int64_t)FRAME_RATE_BASE*s->time_increment_resolution/s->frame_rate; + time_div= s->time/s->time_increment_resolution; + time_mod= s->time%s->time_increment_resolution; + + if(s->pict_type==B_TYPE){ + s->bp_time= s->last_non_b_time - s->time; + }else{ + s->last_time_base= s->time_base; + s->time_base= time_div; + s->pp_time= s->time - s->last_non_b_time; + s->last_non_b_time= s->time; + } +} + static void mpeg4_encode_vol_header(MpegEncContext * s) { int vo_ver_id=1; //must be 2 if we want GMC or q-pel @@ -983,11 +1009,7 @@ static void mpeg4_encode_vol_header(MpegEncContext * s) put_bits(&s->pb, 2, RECT_SHAPE); /* vol shape= rectangle */ put_bits(&s->pb, 1, 1); /* marker bit */ - s->time_increment_resolution= s->frame_rate/ff_gcd(s->frame_rate, FRAME_RATE_BASE); - if(s->time_increment_resolution>=256*256) s->time_increment_resolution= 256*128; - put_bits(&s->pb, 16, s->time_increment_resolution); - s->time_increment_bits = av_log2(s->time_increment_resolution - 1) + 1; if (s->time_increment_bits < 1) s->time_increment_bits = 1; put_bits(&s->pb, 1, 1); /* marker bit */ @@ -1034,9 +1056,6 @@ void mpeg4_encode_picture_header(MpegEncContext * s, int picture_number) if(s->pict_type==I_TYPE) mpeg4_encode_vol_header(s); - s->time= s->picture_number*(int64_t)FRAME_RATE_BASE*s->time_increment_resolution/s->frame_rate; - time_div= s->time/s->time_increment_resolution; - time_mod= s->time%s->time_increment_resolution; //printf("num:%d rate:%d base:%d\n", s->picture_number, s->frame_rate, FRAME_RATE_BASE); if(get_bit_count(&s->pb)!=0) mpeg4_stuffing(&s->pb); @@ -1044,15 +1063,8 @@ void mpeg4_encode_picture_header(MpegEncContext * s, int picture_number) put_bits(&s->pb, 16, 0x1B6); /* vop header */ put_bits(&s->pb, 2, s->pict_type - 1); /* pict type: I = 0 , P = 1 */ - if(s->pict_type==B_TYPE){ - s->bp_time= s->last_non_b_time - s->time; - }else{ - s->last_time_base= s->time_base; - s->time_base= time_div; - s->pp_time= s->time - s->last_non_b_time; - s->last_non_b_time= s->time; - } - + time_div= s->time/s->time_increment_resolution; + time_mod= s->time%s->time_increment_resolution; time_incr= time_div - s->last_time_base; while(time_incr--) put_bits(&s->pb, 1, 1); @@ -1770,6 +1782,7 @@ int h263_decode_mb(MpegEncContext *s, s->last_mv[0][0][1]= s->last_mv[1][0][0]= s->last_mv[1][0][1]= 0; +// printf("\n"); } /* if we skipped it in the future P Frame than skip it now too */ @@ -1789,6 +1802,7 @@ int h263_decode_mb(MpegEncContext *s, //FIXME is this correct? /* s->last_mv[0][0][0]= s->last_mv[0][0][1]=0;*/ +// printf("S"); return 0; } @@ -1837,6 +1851,7 @@ int h263_decode_mb(MpegEncContext *s, s->mv[0][0][1] = s->mv[1][0][0] = s->mv[1][0][1] = 1000;*/ +// printf("D"); break; case 1: s->mv_dir = MV_DIR_FORWARD | MV_DIR_BACKWARD; @@ -1849,6 +1864,7 @@ int h263_decode_mb(MpegEncContext *s, my = h263_decode_motion(s, s->last_mv[1][0][1], s->b_code); s->last_mv[1][0][0]= s->mv[1][0][0] = mx; s->last_mv[1][0][1]= s->mv[1][0][1] = my; +// printf("I"); break; case 2: s->mv_dir = MV_DIR_BACKWARD; @@ -1856,6 +1872,7 @@ int h263_decode_mb(MpegEncContext *s, my = h263_decode_motion(s, s->last_mv[1][0][1], s->b_code); s->last_mv[1][0][0]= s->mv[1][0][0] = mx; s->last_mv[1][0][1]= s->mv[1][0][1] = my; +// printf("B"); break; case 3: s->mv_dir = MV_DIR_FORWARD; @@ -1863,6 +1880,7 @@ int h263_decode_mb(MpegEncContext *s, my = h263_decode_motion(s, s->last_mv[0][0][1], s->f_code); s->last_mv[0][0][0]= s->mv[0][0][0] = mx; s->last_mv[0][0][1]= s->mv[0][0][1] = my; +// printf("F"); break; default: return -1; } diff --git a/libavcodec/motion_est.c b/libavcodec/motion_est.c index 7fc0cec18c..1ba246ae5c 100644 --- a/libavcodec/motion_est.c +++ b/libavcodec/motion_est.c @@ -29,7 +29,7 @@ #define MAX(a,b) ((a) > (b) ? (a) : (b)) #define INTER_BIAS 257 -static void halfpel_motion_search(MpegEncContext * s, +static int halfpel_motion_search(MpegEncContext * s, int *mx_ptr, int *my_ptr, int dmin, int xmin, int ymin, int xmax, int ymax, int pred_x, int pred_y, uint8_t *ref_picture); @@ -673,7 +673,7 @@ static int epzs_motion_search4(MpegEncContext * s, int block, /* The idea would be to make half pel ME after Inter/Intra decision to save time. */ -static inline void halfpel_motion_search(MpegEncContext * s, +static inline int halfpel_motion_search(MpegEncContext * s, int *mx_ptr, int *my_ptr, int dmin, int xmin, int ymin, int xmax, int ymax, int pred_x, int pred_y, uint8_t *ref_picture) @@ -702,7 +702,7 @@ static inline void halfpel_motion_search(MpegEncContext * s, if(dmin < Z_THRESHOLD && mx==0 && my==0){ *mx_ptr = 0; *my_ptr = 0; - return; + return dmin; } pen_x= pred_x + mx; @@ -727,6 +727,7 @@ static inline void halfpel_motion_search(MpegEncContext * s, *mx_ptr = mx; *my_ptr = my; + return dminh; } static inline void halfpel_motion_search4(MpegEncContext * s, @@ -1044,17 +1045,15 @@ void ff_estimate_p_frame_motion(MpegEncContext * s, set_p_mv_tables(s, mx, my); } -void ff_estimate_motion_b(MpegEncContext * s, +int ff_estimate_motion_b(MpegEncContext * s, int mb_x, int mb_y, int16_t (*mv_table)[2], uint8_t *ref_picture, int f_code) { - UINT8 *pix, *ppix; - int sum, varc, vard, mx, my, range, dmin, xx, yy; + int mx, my, range, dmin; int xmin, ymin, xmax, ymax; int rel_xmin, rel_ymin, rel_xmax, rel_ymax; int pred_x=0, pred_y=0; int P[6][2]; const int shift= 1+s->quarter_sample; - int mb_type=0; const int mot_stride = s->mb_width + 2; const int mot_xy = (mb_y + 1)*mot_stride + mb_x + 1; @@ -1124,18 +1123,210 @@ void ff_estimate_motion_b(MpegEncContext * s, /* At this point (mx,my) are full-pell and the absolute displacement */ // ppix = ref_picture + (my * s->linesize) + mx; - halfpel_motion_search(s, &mx, &my, dmin, xmin, ymin, xmax, ymax, pred_x, pred_y, ref_picture); + dmin= halfpel_motion_search(s, &mx, &my, dmin, xmin, ymin, xmax, ymax, pred_x, pred_y, ref_picture); // s->mb_type[mb_y*s->mb_width + mb_x]= mb_type; mv_table[mot_xy][0]= mx; mv_table[mot_xy][1]= my; + return dmin; } -int ff_decide_type(MpegEncContext * s, - int mb_x, int mb_y) +static inline int check_bidir_mv(MpegEncContext * s, + int mb_x, int mb_y, + int motion_fx, int motion_fy, + int motion_bx, int motion_by, + int pred_fx, int pred_fy, + int pred_bx, int pred_by) { + //FIXME optimize? + UINT16 *mv_penalty= s->mv_penalty[s->f_code] + MAX_MV; // f_code of the prev frame + uint8_t *dest_y = s->me_scratchpad; + uint8_t *ptr; + int dxy; + int src_x, src_y; + int fbmin; + + fbmin = (mv_penalty[motion_fx-pred_fx] + mv_penalty[motion_fy-pred_fy])*s->qscale; + + dxy = ((motion_fy & 1) << 1) | (motion_fx & 1); + src_x = mb_x * 16 + (motion_fx >> 1); + src_y = mb_y * 16 + (motion_fy >> 1); + + ptr = s->last_picture[0] + (src_y * s->linesize) + src_x; + put_pixels_tab[dxy](dest_y , ptr , s->linesize, 16); + put_pixels_tab[dxy](dest_y + 8, ptr + 8, s->linesize, 16); + + fbmin += (mv_penalty[motion_bx-pred_bx] + mv_penalty[motion_by-pred_by])*s->qscale; + + dxy = ((motion_by & 1) << 1) | (motion_bx & 1); + src_x = mb_x * 16 + (motion_bx >> 1); + src_y = mb_y * 16 + (motion_by >> 1); + + ptr = s->next_picture[0] + (src_y * s->linesize) + src_x; + avg_pixels_tab[dxy](dest_y , ptr , s->linesize, 16); + avg_pixels_tab[dxy](dest_y + 8, ptr + 8, s->linesize, 16); + + fbmin += pix_abs16x16(s->new_picture[0] + mb_x*16 + mb_y*16*s->linesize, dest_y, s->linesize); + return fbmin; +} +/* refine the bidir vectors in hq mode and return the score in both lq & hq mode*/ +static inline int bidir_refine(MpegEncContext * s, + int mb_x, int mb_y) +{ + const int mot_stride = s->mb_width + 2; + const int xy = (mb_y + 1)*mot_stride + mb_x + 1; + int fbmin; + int pred_fx= s->b_bidir_forw_mv_table[xy-1][0]; + int pred_fy= s->b_bidir_forw_mv_table[xy-1][1]; + int pred_bx= s->b_bidir_back_mv_table[xy-1][0]; + int pred_by= s->b_bidir_back_mv_table[xy-1][1]; + int motion_fx= s->b_bidir_forw_mv_table[xy][0]= s->b_forw_mv_table[xy][0]; + int motion_fy= s->b_bidir_forw_mv_table[xy][1]= s->b_forw_mv_table[xy][1]; + int motion_bx= s->b_bidir_back_mv_table[xy][0]= s->b_back_mv_table[xy][0]; + int motion_by= s->b_bidir_back_mv_table[xy][1]= s->b_back_mv_table[xy][1]; + + //FIXME do refinement and add flag + + fbmin= check_bidir_mv(s, mb_x, mb_y, + motion_fx, motion_fy, + motion_bx, motion_by, + pred_fx, pred_fy, + pred_bx, pred_by); + + return fbmin; +} + +static inline int direct_search(MpegEncContext * s, + int mb_x, int mb_y) +{ + int P[6][2]; + const int mot_stride = s->mb_width + 2; + const int mot_xy = (mb_y + 1)*mot_stride + mb_x + 1; + int dmin, dmin2; + int motion_fx, motion_fy, motion_bx, motion_by, motion_bx0, motion_by0; + int motion_dx, motion_dy; + const int motion_px= s->p_mv_table[mot_xy][0]; + const int motion_py= s->p_mv_table[mot_xy][1]; + const int time_pp= s->pp_time; + const int time_bp= s->bp_time; + const int time_pb= time_pp - time_bp; + int bx, by; + int mx, my, mx2, my2; + uint8_t *ref_picture= s->me_scratchpad - (mb_x + 1 + (mb_y + 1)*s->linesize)*16; + int16_t (*mv_table)[2]= s->b_direct_mv_table; + uint16_t *mv_penalty= s->mv_penalty[s->f_code] + MAX_MV; // f_code of the prev frame + + /* thanks to iso-mpeg the rounding is different for the zero vector, so we need to handle that ... */ + motion_fx= (motion_px*time_pb)/time_pp; + motion_fy= (motion_py*time_pb)/time_pp; + motion_bx0= (-motion_px*time_bp)/time_pp; + motion_by0= (-motion_py*time_bp)/time_pp; + motion_dx= motion_dy=0; + dmin2= check_bidir_mv(s, mb_x, mb_y, + motion_fx, motion_fy, + motion_bx0, motion_by0, + motion_fx, motion_fy, + motion_bx0, motion_by0) - s->qscale; + + motion_bx= motion_fx - motion_px; + motion_by= motion_fy - motion_py; + for(by=-1; by<2; by++){ + for(bx=-1; bx<2; bx++){ + uint8_t *dest_y = s->me_scratchpad + (by+1)*s->linesize*16 + (bx+1)*16; + uint8_t *ptr; + int dxy; + int src_x, src_y; + const int width= s->width; + const int height= s->height; + + dxy = ((motion_fy & 1) << 1) | (motion_fx & 1); + src_x = (mb_x + bx) * 16 + (motion_fx >> 1); + src_y = (mb_y + by) * 16 + (motion_fy >> 1); + src_x = clip(src_x, -16, width); + if (src_x == width) dxy &= ~1; + src_y = clip(src_y, -16, height); + if (src_y == height) dxy &= ~2; + + ptr = s->last_picture[0] + (src_y * s->linesize) + src_x; + put_pixels_tab[dxy](dest_y , ptr , s->linesize, 16); + put_pixels_tab[dxy](dest_y + 8, ptr + 8, s->linesize, 16); + + dxy = ((motion_by & 1) << 1) | (motion_bx & 1); + src_x = (mb_x + bx) * 16 + (motion_bx >> 1); + src_y = (mb_y + by) * 16 + (motion_by >> 1); + src_x = clip(src_x, -16, width); + if (src_x == width) dxy &= ~1; + src_y = clip(src_y, -16, height); + if (src_y == height) dxy &= ~2; + + avg_pixels_tab[dxy](dest_y , ptr , s->linesize, 16); + avg_pixels_tab[dxy](dest_y + 8, ptr + 8, s->linesize, 16); + } + } + + P[0][0] = mv_table[mot_xy ][0]; + P[0][1] = mv_table[mot_xy ][1]; + P[1][0] = mv_table[mot_xy - 1][0]; + P[1][1] = mv_table[mot_xy - 1][1]; + + /* special case for first line */ + if ((mb_y == 0 || s->first_slice_line || s->first_gob_line)) { + P[4][0] = P[1][0]; + P[4][1] = P[1][1]; + } else { + P[2][0] = mv_table[mot_xy - mot_stride ][0]; + P[2][1] = mv_table[mot_xy - mot_stride ][1]; + P[3][0] = mv_table[mot_xy - mot_stride + 1 ][0]; + P[3][1] = mv_table[mot_xy - mot_stride + 1 ][1]; + + P[4][0]= mid_pred(P[1][0], P[2][0], P[3][0]); + P[4][1]= mid_pred(P[1][1], P[2][1], P[3][1]); + } + dmin = epzs_motion_search(s, &mx, &my, P, 0, 0, -16, -16, 15, 15, ref_picture); + if(mx==0 && my==0) dmin=99999999; // not representable, due to rounding stuff + if(dmin2qscale; + + if(dmin2b_direct_mv_table[mot_xy][0]= mx; + s->b_direct_mv_table[mot_xy][1]= my; + s->b_direct_forw_mv_table[mot_xy][0]= motion_fx + mx; + s->b_direct_forw_mv_table[mot_xy][1]= motion_fy + my; + s->b_direct_back_mv_table[mot_xy][0]= motion_bx + mx; + s->b_direct_back_mv_table[mot_xy][1]= motion_by + my; + return dmin; } void ff_estimate_b_frame_motion(MpegEncContext * s, @@ -1143,16 +1334,41 @@ void ff_estimate_b_frame_motion(MpegEncContext * s, { const int mot_stride = s->mb_width + 2; const int xy = (mb_y + 1)*mot_stride + mb_x + 1; + const int quant= s->qscale; + int fmin, bmin, dmin, fbmin; + int type=0; + int motion_fx, motion_fy, motion_bx, motion_by; + + dmin= direct_search(s, mb_x, mb_y); - ff_estimate_motion_b(s, mb_x, mb_y, s->b_forw_mv_table, s->last_picture[0], s->f_code); - ff_estimate_motion_b(s, mb_x, mb_y, s->b_back_mv_table, s->next_picture[0], s->b_code); + fmin= ff_estimate_motion_b(s, mb_x, mb_y, s->b_forw_mv_table, s->last_picture[0], s->f_code); + bmin= ff_estimate_motion_b(s, mb_x, mb_y, s->b_back_mv_table, s->next_picture[0], s->b_code) - quant; //printf(" %d %d ", s->b_forw_mv_table[xy][0], s->b_forw_mv_table[xy][1]); - s->b_bidir_forw_mv_table[xy][0]= s->b_forw_mv_table[xy][0]; - s->b_bidir_forw_mv_table[xy][1]= s->b_forw_mv_table[xy][1]; - s->b_bidir_back_mv_table[xy][0]= s->b_back_mv_table[xy][0]; - s->b_bidir_back_mv_table[xy][1]= s->b_back_mv_table[xy][1]; - - s->mb_type[mb_y*s->mb_width + mb_x]= MB_TYPE_FORWARD; //FIXME + + fbmin= bidir_refine(s, mb_x, mb_y); + + if(s->flags&CODEC_FLAG_HQ){ + type= MB_TYPE_FORWARD | MB_TYPE_BACKWARD | MB_TYPE_BIDIR | MB_TYPE_DIRECT; + }else{ + int score= dmin; + type=MB_TYPE_DIRECT; + + if(fminmc_mb_var += score; + } + + s->mb_type[mb_y*s->mb_width + mb_x]= type; } /* find best f_code for ME which do unlimited searches */ @@ -1184,8 +1400,12 @@ int ff_get_best_fcode(MpegEncContext * s, int16_t (*mv_table)[2], int type) } for(i=MAX_FCODE; i>1; i--){ + int threshold; loose+= mv_num[i]; - if(loose > s->mb_num/20) break; //FIXME this is pretty ineffective + + if(s->pict_type==B_TYPE) threshold= 0; + else threshold= s->mb_num/20; //FIXME + if(loose > threshold) break; } // printf("fcode: %d type: %d\n", i, s->pict_type); return i; @@ -1275,11 +1495,12 @@ void ff_fix_long_b_mvs(MpegEncContext * s, int16_t (*mv_table)[2], int f_code, i || fcode_tab[mv_table[xy][0] + MAX_MV] == 0 || fcode_tab[mv_table[xy][1] + MAX_MV] > f_code || fcode_tab[mv_table[xy][1] + MAX_MV] == 0 ){ - s->mb_type[i] &= ~type; - if(s->mb_type[i]==0) s->mb_type[i]= MB_TYPE_FORWARD; //FIXME - mv_table[xy][0] = 0; - mv_table[xy][1] = 0; - //this is certainly bad FIXME + if(s->mb_type[i]&(~type)) s->mb_type[i] &= ~type; + else{ + mv_table[xy][0] = 0; + mv_table[xy][1] = 0; + //this is certainly bad FIXME + } } } xy++; diff --git a/libavcodec/mpegvideo.c b/libavcodec/mpegvideo.c index af1d1c3396..d1c7e16372 100644 --- a/libavcodec/mpegvideo.c +++ b/libavcodec/mpegvideo.c @@ -225,6 +225,12 @@ int MPV_common_init(MpegEncContext *s) goto fail; } + s->me_scratchpad = av_mallocz( s->linesize*16*3*sizeof(uint8_t)); + if (s->me_scratchpad == NULL) { + perror("malloc"); + goto fail; + } + if(s->max_b_frames){ for(j=0; jmbskip_table) goto fail; - s->block= s->intra_block; + s->block= s->blocks[0]; s->context_initialized = 1; return 0; @@ -333,6 +339,7 @@ void MPV_common_end(MpegEncContext *s) CHECK_FREE(s->ac_val[0]); CHECK_FREE(s->coded_block); CHECK_FREE(s->mbintra_table); + CHECK_FREE(s->me_scratchpad); CHECK_FREE(s->mbskip_table); for(i=0;i<3;i++) { @@ -761,16 +768,6 @@ int MPV_encode_picture(AVCodecContext *avctx, return pbBufPtr(&s->pb) - s->pb.buf; } -static inline int clip(int a, int amin, int amax) -{ - if (a < amin) - return amin; - else if (a > amax) - return amax; - else - return a; -} - static inline void gmc1_motion(MpegEncContext *s, UINT8 *dest_y, UINT8 *dest_cb, UINT8 *dest_cr, int dest_offset, @@ -1225,7 +1222,7 @@ void MPV_decode_mb(MpegEncContext *s, DCTELEM block[6][64]) if (!s->mb_intra) { /* motion handling */ if((s->flags&CODEC_FLAG_HQ) || (!s->encoding)){ - if (!s->no_rounding){ + if ((!s->no_rounding) || s->pict_type==B_TYPE){ op_pix = put_pixels_tab; op_qpix= qpel_mc_rnd_tab; }else{ @@ -1235,7 +1232,7 @@ void MPV_decode_mb(MpegEncContext *s, DCTELEM block[6][64]) if (s->mv_dir & MV_DIR_FORWARD) { MPV_motion(s, dest_y, dest_cb, dest_cr, 0, s->last_picture, op_pix, op_qpix); - if (!s->no_rounding) + if ((!s->no_rounding) || s->pict_type==B_TYPE) op_pix = avg_pixels_tab; else op_pix = avg_no_rnd_pixels_tab; @@ -1312,7 +1309,7 @@ static void encode_mb(MpegEncContext *s, int motion_x, int motion_y) dest_cb = s->current_picture[1] + (mb_y * 8 * (s->linesize >> 1)) + mb_x * 8; dest_cr = s->current_picture[2] + (mb_y * 8 * (s->linesize >> 1)) + mb_x * 8; - if (!s->no_rounding){ + if ((!s->no_rounding) || s->pict_type==B_TYPE){ op_pix = put_pixels_tab; op_qpix= qpel_mc_rnd_tab; }else{ @@ -1322,7 +1319,7 @@ static void encode_mb(MpegEncContext *s, int motion_x, int motion_y) if (s->mv_dir & MV_DIR_FORWARD) { MPV_motion(s, dest_y, dest_cb, dest_cr, 0, s->last_picture, op_pix, op_qpix); - if (!s->no_rounding) + if ((!s->no_rounding) || s->pict_type==B_TYPE) op_pix = avg_pixels_tab; else op_pix = avg_no_rnd_pixels_tab; @@ -1429,6 +1426,8 @@ static void copy_context_before_encode(MpegEncContext *d, MpegEncContext *s, int d->skip_count= s->skip_count; d->misc_bits= s->misc_bits; d->last_bits= s->last_bits; + + d->mb_skiped= s->mb_skiped; } static void copy_context_after_encode(MpegEncContext *d, MpegEncContext *s, int type){ @@ -1453,6 +1452,7 @@ static void copy_context_after_encode(MpegEncContext *d, MpegEncContext *s, int d->last_bits= s->last_bits; d->mb_intra= s->mb_intra; + d->mb_skiped= s->mb_skiped; d->mv_type= s->mv_type; d->mv_dir= s->mv_dir; d->pb= s->pb; @@ -1468,7 +1468,7 @@ static void encode_picture(MpegEncContext *s, int picture_number) int i; int bits; MpegEncContext best_s, backup_s; - UINT8 bit_buf[4][3000]; //FIXME check that this is ALLWAYS large enogh for a MB + UINT8 bit_buf[7][3000]; //FIXME check that this is ALLWAYS large enogh for a MB s->picture_number = picture_number; @@ -1483,7 +1483,11 @@ static void encode_picture(MpegEncContext *s, int picture_number) /* Reset the average MB variance */ s->avg_mb_var = 0; s->mc_mb_var = 0; - + + /* we need to initialize some time vars before we can encode b-frames */ + if (s->h263_pred && !s->h263_msmpeg4) + ff_set_mpeg4_time(s, s->picture_number); + /* Estimate motion for every MB */ if(s->pict_type != I_TYPE){ // int16_t (*tmp)[2]= s->p_mv_table; @@ -1535,9 +1539,11 @@ static void encode_picture(MpegEncContext *s, int picture_number) if(s->pict_type==B_TYPE){ s->f_code= ff_get_best_fcode(s, s->b_forw_mv_table, MB_TYPE_FORWARD); s->b_code= ff_get_best_fcode(s, s->b_back_mv_table, MB_TYPE_BACKWARD); - //FIXME if BIDIR != for&back - ff_fix_long_b_mvs(s, s->b_forw_mv_table, s->f_code, MB_TYPE_FORWARD |MB_TYPE_BIDIR); - ff_fix_long_b_mvs(s, s->b_back_mv_table, s->b_code, MB_TYPE_BACKWARD|MB_TYPE_BIDIR); + + ff_fix_long_b_mvs(s, s->b_forw_mv_table, s->f_code, MB_TYPE_FORWARD); + ff_fix_long_b_mvs(s, s->b_back_mv_table, s->b_code, MB_TYPE_BACKWARD); + ff_fix_long_b_mvs(s, s->b_bidir_forw_mv_table, s->f_code, MB_TYPE_BIDIR); + ff_fix_long_b_mvs(s, s->b_bidir_back_mv_table, s->b_code, MB_TYPE_BIDIR); } //printf("f_code %d ///\n", s->f_code); @@ -1632,7 +1638,8 @@ static void encode_picture(MpegEncContext *s, int picture_number) s->block_index[4]= s->block_wrap[4]*(mb_y + 1) + s->block_wrap[0]*(s->mb_height*2 + 2); s->block_index[5]= s->block_wrap[4]*(mb_y + 1 + s->mb_height + 2) + s->block_wrap[0]*(s->mb_height*2 + 2); for(mb_x=0; mb_x < s->mb_width; mb_x++) { - /*const */int mb_type= s->mb_type[mb_y * s->mb_width + mb_x]; + const int mb_type= s->mb_type[mb_y * s->mb_width + mb_x]; + const int xy= (mb_y+1) * (s->mb_width+2) + mb_x + 1; PutBitContext pb; int d; int dmin=10000000; @@ -1647,19 +1654,19 @@ static void encode_picture(MpegEncContext *s, int picture_number) s->block_index[4]++; s->block_index[5]++; if(mb_type & (mb_type-1)){ // more than 1 MB type possible + int next_block=0; pb= s->pb; - s->mv_dir = MV_DIR_FORWARD; copy_context_before_encode(&backup_s, s, -1); if(mb_type&MB_TYPE_INTER){ - int xy= (mb_y+1) * (s->mb_width+2) + mb_x + 1; + s->mv_dir = MV_DIR_FORWARD; s->mv_type = MV_TYPE_16X16; s->mb_intra= 0; s->mv[0][0][0] = s->p_mv_table[xy][0]; s->mv[0][0][1] = s->p_mv_table[xy][1]; init_put_bits(&s->pb, bit_buf[1], 3000, NULL, NULL); - s->block= s->inter_block; + s->block= s->blocks[next_block]; encode_mb(s, s->mv[0][0][0], s->mv[0][0][1]); d= get_bit_count(&s->pb); @@ -1668,10 +1675,12 @@ static void encode_picture(MpegEncContext *s, int picture_number) dmin=d; copy_context_after_encode(&best_s, s, MB_TYPE_INTER); best=1; + next_block^=1; } } if(mb_type&MB_TYPE_INTER4V){ copy_context_before_encode(s, &backup_s, MB_TYPE_INTER4V); + s->mv_dir = MV_DIR_FORWARD; s->mv_type = MV_TYPE_8X8; s->mb_intra= 0; for(i=0; i<4; i++){ @@ -1679,25 +1688,111 @@ static void encode_picture(MpegEncContext *s, int picture_number) s->mv[0][i][1] = s->motion_val[s->block_index[i]][1]; } init_put_bits(&s->pb, bit_buf[2], 3000, NULL, NULL); - s->block= s->inter4v_block; + s->block= s->blocks[next_block]; encode_mb(s, 0, 0); d= get_bit_count(&s->pb); - if(dpb); dmin=d; copy_context_after_encode(&best_s, s, MB_TYPE_INTER4V); best=2; + next_block^=1; + } + } + if(mb_type&MB_TYPE_FORWARD){ + copy_context_before_encode(s, &backup_s, MB_TYPE_FORWARD); + s->mv_dir = MV_DIR_FORWARD; + s->mv_type = MV_TYPE_16X16; + s->mb_intra= 0; + s->mv[0][0][0] = s->b_forw_mv_table[xy][0]; + s->mv[0][0][1] = s->b_forw_mv_table[xy][1]; + init_put_bits(&s->pb, bit_buf[3], 3000, NULL, NULL); + s->block= s->blocks[next_block]; + + encode_mb(s, s->mv[0][0][0], s->mv[0][0][1]); + d= get_bit_count(&s->pb); + if(dpb); + dmin=d; + copy_context_after_encode(&best_s, s, MB_TYPE_FORWARD); + best=3; + next_block^=1; + } + } + if(mb_type&MB_TYPE_BACKWARD){ + copy_context_before_encode(s, &backup_s, MB_TYPE_BACKWARD); + s->mv_dir = MV_DIR_BACKWARD; + s->mv_type = MV_TYPE_16X16; + s->mb_intra= 0; + s->mv[1][0][0] = s->b_back_mv_table[xy][0]; + s->mv[1][0][1] = s->b_back_mv_table[xy][1]; + init_put_bits(&s->pb, bit_buf[4], 3000, NULL, NULL); + s->block= s->blocks[next_block]; + + encode_mb(s, s->mv[1][0][0], s->mv[1][0][1]); + d= get_bit_count(&s->pb); + if(dpb); + dmin=d; + copy_context_after_encode(&best_s, s, MB_TYPE_BACKWARD); + best=4; + next_block^=1; + } + } + if(mb_type&MB_TYPE_BIDIR){ + copy_context_before_encode(s, &backup_s, MB_TYPE_BIDIR); + s->mv_dir = MV_DIR_FORWARD | MV_DIR_BACKWARD; + s->mv_type = MV_TYPE_16X16; + s->mb_intra= 0; + s->mv[0][0][0] = s->b_bidir_forw_mv_table[xy][0]; + s->mv[0][0][1] = s->b_bidir_forw_mv_table[xy][1]; + s->mv[1][0][0] = s->b_bidir_back_mv_table[xy][0]; + s->mv[1][0][1] = s->b_bidir_back_mv_table[xy][1]; + init_put_bits(&s->pb, bit_buf[5], 3000, NULL, NULL); + s->block= s->blocks[next_block]; + + encode_mb(s, 0, 0); + d= get_bit_count(&s->pb); + if(dpb); + dmin=d; + copy_context_after_encode(&best_s, s, MB_TYPE_BIDIR); + best=5; + next_block^=1; + } + } + if(mb_type&MB_TYPE_DIRECT){ + copy_context_before_encode(s, &backup_s, MB_TYPE_DIRECT); + s->mv_dir = MV_DIR_FORWARD | MV_DIR_BACKWARD | MV_DIRECT; + s->mv_type = MV_TYPE_16X16; //FIXME + s->mb_intra= 0; + s->mv[0][0][0] = s->b_direct_forw_mv_table[xy][0]; + s->mv[0][0][1] = s->b_direct_forw_mv_table[xy][1]; + s->mv[1][0][0] = s->b_direct_back_mv_table[xy][0]; + s->mv[1][0][1] = s->b_direct_back_mv_table[xy][1]; + init_put_bits(&s->pb, bit_buf[6], 3000, NULL, NULL); + s->block= s->blocks[next_block]; + + encode_mb(s, s->b_direct_mv_table[xy][0], s->b_direct_mv_table[xy][1]); + d= get_bit_count(&s->pb); + if(dpb); + dmin=d; + copy_context_after_encode(&best_s, s, MB_TYPE_DIRECT); + best=6; + next_block^=1; } } if(mb_type&MB_TYPE_INTRA){ copy_context_before_encode(s, &backup_s, MB_TYPE_INTRA); + s->mv_dir = MV_DIR_FORWARD; s->mv_type = MV_TYPE_16X16; s->mb_intra= 1; s->mv[0][0][0] = 0; s->mv[0][0][1] = 0; init_put_bits(&s->pb, bit_buf[0], 3000, NULL, NULL); - s->block= s->intra_block; + s->block= s->blocks[next_block]; encode_mb(s, 0, 0); d= get_bit_count(&s->pb); @@ -1706,6 +1801,7 @@ static void encode_picture(MpegEncContext *s, int picture_number) dmin=d; copy_context_after_encode(&best_s, s, MB_TYPE_INTRA); best=0; + next_block^=1; } /* force cleaning of ac/dc pred stuff if needed ... */ if(s->h263_pred || s->h263_aic) @@ -1718,30 +1814,30 @@ static void encode_picture(MpegEncContext *s, int picture_number) int motion_x, motion_y; s->mv_type=MV_TYPE_16X16; // only one MB-Type possible - //FIXME convert to swicth() - if(mb_type&MB_TYPE_INTRA){ + switch(mb_type){ + case MB_TYPE_INTRA: s->mv_dir = MV_DIR_FORWARD; s->mb_intra= 1; motion_x= s->mv[0][0][0] = 0; motion_y= s->mv[0][0][1] = 0; - }else if(mb_type&MB_TYPE_INTER){ - int xy= (mb_y+1) * (s->mb_width+2) + mb_x + 1; + break; + case MB_TYPE_INTER: s->mv_dir = MV_DIR_FORWARD; s->mb_intra= 0; motion_x= s->mv[0][0][0] = s->p_mv_table[xy][0]; motion_y= s->mv[0][0][1] = s->p_mv_table[xy][1]; - }else if(mb_type&MB_TYPE_DIRECT){ - int xy= (mb_y+1) * (s->mb_width+2) + mb_x + 1; + break; + case MB_TYPE_DIRECT: s->mv_dir = MV_DIR_FORWARD | MV_DIR_BACKWARD | MV_DIRECT; s->mb_intra= 0; - motion_x=0; - motion_y=0; - s->mv[0][0][0] = 0; - s->mv[0][0][1] = 0; - s->mv[1][0][0] = 0; - s->mv[1][0][1] = 0; - }else if(mb_type&MB_TYPE_BIDIR){ - int xy= (mb_y+1) * (s->mb_width+2) + mb_x + 1; + motion_x=s->b_direct_mv_table[xy][0]; + motion_y=s->b_direct_mv_table[xy][1]; + s->mv[0][0][0] = s->b_direct_forw_mv_table[xy][0]; + s->mv[0][0][1] = s->b_direct_forw_mv_table[xy][1]; + s->mv[1][0][0] = s->b_direct_back_mv_table[xy][0]; + s->mv[1][0][1] = s->b_direct_back_mv_table[xy][1]; + break; + case MB_TYPE_BIDIR: s->mv_dir = MV_DIR_FORWARD | MV_DIR_BACKWARD; s->mb_intra= 0; motion_x=0; @@ -1750,25 +1846,31 @@ static void encode_picture(MpegEncContext *s, int picture_number) s->mv[0][0][1] = s->b_bidir_forw_mv_table[xy][1]; s->mv[1][0][0] = s->b_bidir_back_mv_table[xy][0]; s->mv[1][0][1] = s->b_bidir_back_mv_table[xy][1]; - }else if(mb_type&MB_TYPE_BACKWARD){ - int xy= (mb_y+1) * (s->mb_width+2) + mb_x + 1; + break; + case MB_TYPE_BACKWARD: s->mv_dir = MV_DIR_BACKWARD; s->mb_intra= 0; motion_x= s->mv[1][0][0] = s->b_back_mv_table[xy][0]; motion_y= s->mv[1][0][1] = s->b_back_mv_table[xy][1]; - }else if(mb_type&MB_TYPE_FORWARD){ - int xy= (mb_y+1) * (s->mb_width+2) + mb_x + 1; + break; + case MB_TYPE_FORWARD: s->mv_dir = MV_DIR_FORWARD; s->mb_intra= 0; motion_x= s->mv[0][0][0] = s->b_forw_mv_table[xy][0]; motion_y= s->mv[0][0][1] = s->b_forw_mv_table[xy][1]; // printf(" %d %d ", motion_x, motion_y); - }else{ + break; + default: motion_x=motion_y=0; //gcc warning fix printf("illegal MB type\n"); } encode_mb(s, motion_x, motion_y); } + /* clean the MV table in IPS frames for direct mode in B frames */ + if(s->mb_intra /* && I,P,S_TYPE */){ + s->p_mv_table[xy][0]=0; + s->p_mv_table[xy][1]=0; + } MPV_decode_mb(s, s->block); } diff --git a/libavcodec/mpegvideo.h b/libavcodec/mpegvideo.h index 841d40fb3c..24943e5b19 100644 --- a/libavcodec/mpegvideo.h +++ b/libavcodec/mpegvideo.h @@ -141,7 +141,8 @@ typedef struct MpegEncContext { INT16 (*b_direct_forw_mv_table)[2];/* MV table (1MV per MB) direct mode b-frame encoding */ INT16 (*b_direct_back_mv_table)[2];/* MV table (1MV per MB) direct mode b-frame encoding */ INT16 (*b_direct_mv_table)[2]; /* MV table (1MV per MB) direct mode b-frame encoding */ - int me_method; /* ME algorithm */ + int me_method; /* ME algorithm */ + uint8_t *me_scratchpad; /* data area for the me algo, so that the ME doesnt need to malloc/free */ int mv_dir; #define MV_DIR_BACKWARD 1 #define MV_DIR_FORWARD 2 @@ -164,7 +165,8 @@ typedef struct MpegEncContext { UINT8 *fcode_tab; /* smallest fcode needed for each MV */ int has_b_frames; - int no_rounding; /* apply no rounding to motion compensation (MPEG4, msmpeg4, ...) */ + int no_rounding; /* apply no rounding to motion compensation (MPEG4, msmpeg4, ...) + for b-frames rounding mode is allways 0 */ /* macroblock layer */ int mb_x, mb_y; @@ -335,9 +337,7 @@ typedef struct MpegEncContext { UINT32 mb_line_avgsize; DCTELEM (*block)[64]; /* points to one of the following blocks */ - DCTELEM intra_block[6][64] __align8; - DCTELEM inter_block[6][64] __align8; - DCTELEM inter4v_block[6][64] __align8; + DCTELEM blocks[2][6][64] __align8; // for HQ mode we need to keep the best block void (*dct_unquantize_mpeg1)(struct MpegEncContext *s, DCTELEM *block, int n, int qscale); void (*dct_unquantize_mpeg2)(struct MpegEncContext *s, @@ -421,6 +421,7 @@ INT16 *h263_pred_motion(MpegEncContext * s, int block, int *px, int *py); void mpeg4_pred_ac(MpegEncContext * s, INT16 *block, int n, int dir); +void ff_set_mpeg4_time(MpegEncContext * s, int picture_number); void mpeg4_encode_picture_header(MpegEncContext *s, int picture_number); void h263_encode_init(MpegEncContext *s);