new files for the CAVS decoder by (Stefan Gehrer <stefan gehrer gmx de)

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@5567 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

libavcodec/cavs.c

+/*
+ * Chinese AVS video (AVS1-P2, JiZhun profile) decoder.
+ * Copyright (c) 2006  Stefan Gehrer <stefan.gehrer@gmx.de>
+ *
+ * This library is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2 of the License, or (at your option) any later version.
+ *
+ * This library is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with this library; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+
+#include "avcodec.h"
+#include "bitstream.h"
+#include "golomb.h"
+#include "mpegvideo.h"
+#include "cavsdata.h"
+
+typedef struct {
+    MpegEncContext s;
+    Picture picture; //currently decoded frame
+    Picture DPB[2];  //reference frames
+    int dist[2];     //temporal distances from current frame to ref frames
+    int profile, level;
+    int aspect_ratio;
+    int mb_width, mb_height;
+    int pic_type;
+    int progressive;
+    int pic_structure;
+    int skip_mode_flag;
+    int loop_filter_disable;
+    int alpha_offset, beta_offset;
+    int ref_flag;
+    int mbx, mby;
+    int flags;
+    int stc;
+    uint8_t *cy, *cu, *cv;
+    int left_qp;
+    uint8_t *top_qp;
+
+    /* mv motion vector cache
+       0:    D3  B2  B3  C2
+       4:    A1  X0  X1   -
+       8:    A3  X2  X3   -
+
+       X are the vectors in the current macroblock (5,6,9,10)
+       A is the macroblock to the left (4,8)
+       B is the macroblock to the top (1,2)
+       C is the macroblock to the top-right (3)
+       D is the macroblock to the top-left (0)
+
+       the same is repeated for backward motion vectors */
+    vector_t mv[2*4*3];
+    vector_t *top_mv[2];
+    vector_t *col_mv;
+
+    /* luma pred mode cache
+       0:    --  B2  B3
+       3:    A1  X0  X1
+       6:    A3  X2  X3   */
+    int pred_mode_Y[3*3];
+    int *top_pred_Y;
+    int l_stride, c_stride;
+    int luma_scan[4];
+    int qp;
+    int qp_fixed;
+    int cbp;
+
+    /* intra prediction is done with un-deblocked samples
+     they are saved here before deblocking the MB  */
+    uint8_t *top_border_y, *top_border_u, *top_border_v;
+    uint8_t left_border_y[16], left_border_u[8], left_border_v[8];
+    uint8_t topleft_border_y, topleft_border_u, topleft_border_v;
+
+    void (*intra_pred_l[8])(uint8_t *d,uint8_t *top,uint8_t *left,int stride);
+    void (*intra_pred_c[7])(uint8_t *d,uint8_t *top,uint8_t *left,int stride);
+    uint8_t *col_type_base;
+    uint8_t *col_type;
+    int sym_factor;
+    int direct_den[2];
+    int scale_den[2];
+    int got_keyframe;
+} AVSContext;
+
+/*****************************************************************************
+ *
+ * in-loop deblocking filter
+ *
+ ****************************************************************************/
+
+static inline int get_bs_p(vector_t *mvP, vector_t *mvQ) {
+    if((mvP->ref == REF_INTRA) || (mvQ->ref == REF_INTRA))
+        return 2;
+    if(mvP->ref != mvQ->ref)
+        return 1;
+    if( (abs(mvP->x - mvQ->x) >= 4) ||  (abs(mvP->y - mvQ->y) >= 4) )
+        return 1;
+    return 0;
+}
+
+static inline int get_bs_b(vector_t *mvP, vector_t *mvQ) {
+    if((mvP->ref == REF_INTRA) || (mvQ->ref == REF_INTRA)) {
+        return 2;
+    } else {
+        vector_t *mvPbw = mvP + MV_BWD_OFFS;
+        vector_t *mvQbw = mvQ + MV_BWD_OFFS;
+        if( (abs(  mvP->x -   mvQ->x) >= 4) ||
+            (abs(  mvP->y -   mvQ->y) >= 4) ||
+            (abs(mvPbw->x - mvQbw->x) >= 4) ||
+            (abs(mvPbw->y - mvQbw->y) >= 4) )
+            return 1;
+    }
+    return 0;
+}
+
+/* boundary strength (bs) mapping:
+ *
+ * --4---5--
+ * 0   2   |
+ * | 6 | 7 |
+ * 1   3   |
+ * ---------
+ *
+ */
+
+#define SET_PARAMS                                            \
+    alpha = alpha_tab[clip(qp_avg + h->alpha_offset,0,63)];   \
+    beta  =  beta_tab[clip(qp_avg + h->beta_offset, 0,63)];   \
+    tc    =    tc_tab[clip(qp_avg + h->alpha_offset,0,63)];
+
+static void filter_mb(AVSContext *h, enum mb_t mb_type) {
+    uint8_t bs[8];
+    int qp_avg, alpha, beta, tc;
+    int i;
+
+    /* save un-deblocked lines */
+    h->topleft_border_y = h->top_border_y[h->mbx*16+15];
+    h->topleft_border_u = h->top_border_u[h->mbx*8+7];
+    h->topleft_border_v = h->top_border_v[h->mbx*8+7];
+    memcpy(&h->top_border_y[h->mbx*16], h->cy + 15* h->l_stride,16);
+    memcpy(&h->top_border_u[h->mbx* 8], h->cu +  7* h->c_stride,8);
+    memcpy(&h->top_border_v[h->mbx* 8], h->cv +  7* h->c_stride,8);
+    for(i=0;i<8;i++) {
+        h->left_border_y[i*2+0] = *(h->cy + 15 + (i*2+0)*h->l_stride);
+        h->left_border_y[i*2+1] = *(h->cy + 15 + (i*2+1)*h->l_stride);
+        h->left_border_u[i] = *(h->cu + 7 + i*h->c_stride);
+        h->left_border_v[i] = *(h->cv + 7 + i*h->c_stride);
+    }
+    if(!h->loop_filter_disable) {
+        /* clear bs */
+        *((uint64_t *)bs) = 0;
+        /* determine bs */
+        switch(mb_type) {
+        case I_8X8:
+            *((uint64_t *)bs) = 0x0202020202020202ULL;
+            break;
+        case P_8X8:
+        case P_8X16:
+            bs[2] = get_bs_p(&h->mv[MV_FWD_X0], &h->mv[MV_FWD_X1]);
+            bs[3] = get_bs_p(&h->mv[MV_FWD_X2], &h->mv[MV_FWD_X3]);
+        case P_16X8:
+            bs[6] = get_bs_p(&h->mv[MV_FWD_X0], &h->mv[MV_FWD_X2]);
+            bs[7] = get_bs_p(&h->mv[MV_FWD_X1], &h->mv[MV_FWD_X3]);
+        case P_16X16:
+        case P_SKIP:
+            bs[0] = get_bs_p(&h->mv[MV_FWD_A1], &h->mv[MV_FWD_X0]);
+            bs[1] = get_bs_p(&h->mv[MV_FWD_A3], &h->mv[MV_FWD_X2]);
+            bs[4] = get_bs_p(&h->mv[MV_FWD_B2], &h->mv[MV_FWD_X0]);
+            bs[5] = get_bs_p(&h->mv[MV_FWD_B3], &h->mv[MV_FWD_X1]);
+            break;
+        case B_SKIP:
+        case B_DIRECT:
+        case B_8X8:
+            bs[2] = get_bs_b(&h->mv[MV_FWD_X0], &h->mv[MV_FWD_X1]);
+            bs[3] = get_bs_b(&h->mv[MV_FWD_X2], &h->mv[MV_FWD_X3]);
+            bs[6] = get_bs_b(&h->mv[MV_FWD_X0], &h->mv[MV_FWD_X2]);
+            bs[7] = get_bs_b(&h->mv[MV_FWD_X1], &h->mv[MV_FWD_X3]);
+        case B_FWD_16X16:
+        case B_BWD_16X16:
+        case B_SYM_16X16:
+            bs[0] = get_bs_b(&h->mv[MV_FWD_A1], &h->mv[MV_FWD_X0]);
+            bs[1] = get_bs_b(&h->mv[MV_FWD_A3], &h->mv[MV_FWD_X2]);
+            bs[4] = get_bs_b(&h->mv[MV_FWD_B2], &h->mv[MV_FWD_X0]);
+            bs[5] = get_bs_b(&h->mv[MV_FWD_B3], &h->mv[MV_FWD_X1]);
+            break;
+        default:
+            if(mb_type & 1) { //16X8
+                bs[6] = bs[7] = get_bs_b(&h->mv[MV_FWD_X0], &h->mv[MV_FWD_X2]);
+            } else {          //8X16
+                bs[2] = bs[3] = get_bs_b(&h->mv[MV_FWD_X0], &h->mv[MV_FWD_X1]);
+            }
+            bs[0] = get_bs_b(&h->mv[MV_FWD_A1], &h->mv[MV_FWD_X0]);
+            bs[1] = get_bs_b(&h->mv[MV_FWD_A3], &h->mv[MV_FWD_X2]);
+            bs[4] = get_bs_b(&h->mv[MV_FWD_B2], &h->mv[MV_FWD_X0]);
+            bs[5] = get_bs_b(&h->mv[MV_FWD_B3], &h->mv[MV_FWD_X1]);
+        }
+        if( *((uint64_t *)bs) ) {
+            if(h->flags & A_AVAIL) {
+                qp_avg = (h->qp + h->left_qp + 1) >> 1;
+                SET_PARAMS;
+                h->s.dsp.cavs_filter_lv(h->cy,h->l_stride,alpha,beta,tc,bs[0],bs[1]);
+                h->s.dsp.cavs_filter_cv(h->cu,h->c_stride,alpha,beta,tc,bs[0],bs[1]);
+                h->s.dsp.cavs_filter_cv(h->cv,h->c_stride,alpha,beta,tc,bs[0],bs[1]);
+            }
+            qp_avg = h->qp;
+            SET_PARAMS;
+            h->s.dsp.cavs_filter_lv(h->cy + 8,h->l_stride,alpha,beta,tc,bs[2],bs[3]);
+            h->s.dsp.cavs_filter_lh(h->cy + 8*h->l_stride,h->l_stride,alpha,beta,tc,
+                           bs[6],bs[7]);
+
+            if(h->flags & B_AVAIL) {
+                qp_avg = (h->qp + h->top_qp[h->mbx] + 1) >> 1;
+                SET_PARAMS;
+                h->s.dsp.cavs_filter_lh(h->cy,h->l_stride,alpha,beta,tc,bs[4],bs[5]);
+                h->s.dsp.cavs_filter_ch(h->cu,h->c_stride,alpha,beta,tc,bs[4],bs[5]);
+                h->s.dsp.cavs_filter_ch(h->cv,h->c_stride,alpha,beta,tc,bs[4],bs[5]);
+            }
+        }
+    }
+    h->left_qp = h->qp;
+    h->top_qp[h->mbx] = h->qp;
+}
+
+#undef SET_PARAMS
+
+/*****************************************************************************
+ *
+ * spatial intra prediction
+ *
+ ****************************************************************************/
+
+static inline void load_intra_pred_luma(AVSContext *h, uint8_t *top,
+                                        uint8_t *left, int block) {
+    int i;
+
+    switch(block) {
+    case 0:
+        memcpy(&left[1],h->left_border_y,16);
+        left[0] = left[1];
+        left[17] = left[16];
+        memcpy(&top[1],&h->top_border_y[h->mbx*16],16);
+        top[17] = top[16];
+        top[0] = top[1];
+        if((h->flags & A_AVAIL) && (h->flags & B_AVAIL))
+            left[0] = top[0] = h->topleft_border_y;
+        break;
+    case 1:
+        for(i=0;i<8;i++)
+            left[i+1] = *(h->cy + 7 + i*h->l_stride);
+        memset(&left[9],left[8],9);
+        left[0] = left[1];
+        memcpy(&top[1],&h->top_border_y[h->mbx*16+8],8);
+        if(h->flags & C_AVAIL)
+            memcpy(&top[9],&h->top_border_y[(h->mbx + 1)*16],8);
+        else
+            memset(&top[9],top[8],9);
+        top[17] = top[16];
+        top[0] = top[1];
+        if(h->flags & B_AVAIL)
+            left[0] = top[0] = h->top_border_y[h->mbx*16+7];
+        break;
+    case 2:
+        memcpy(&left[1],&h->left_border_y[8],8);
+        memset(&left[9],left[8],9);
+        memcpy(&top[1],h->cy + 7*h->l_stride,16);
+        top[17] = top[16];
+        left[0] = h->left_border_y[7];
+        top[0] = top[1];
+        if(h->flags & A_AVAIL)
+            top[0] = left[0];
+        break;
+    case 3:
+        for(i=0;i<9;i++)
+            left[i] = *(h->cy + 7 + (i+7)*h->l_stride);
+        memset(&left[9],left[8],9);
+        memcpy(&top[0],h->cy + 7 + 7*h->l_stride,9);
+        memset(&top[9],top[8],9);
+        break;
+    }
+}
+
+static inline void load_intra_pred_chroma(uint8_t *stop, uint8_t *sleft,
+                                          uint8_t stopleft, uint8_t *dtop,
+                                          uint8_t *dleft, int stride, int flags) {
+    int i;
+
+    if(flags & A_AVAIL) {
+        for(i=0; i<8; i++)
+            dleft[i+1] = sleft[i];
+        dleft[0] = dleft[1];
+        dleft[9] = dleft[8];
+    }
+    if(flags & B_AVAIL) {
+        for(i=0; i<8; i++)
+            dtop[i+1] = stop[i];
+        dtop[0] = dtop[1];
+        dtop[9] = dtop[8];
+        if(flags & A_AVAIL)
+            dleft[0] = dtop[0] = stopleft;
+    }
+}
+
+static void intra_pred_vert(uint8_t *d,uint8_t *top,uint8_t *left,int stride) {
+    int y;
+    uint64_t a = *((uint64_t *)(&top[1]));
+    for(y=0;y<8;y++) {
+        *((uint64_t *)(d+y*stride)) = a;
+    }
+}
+
+static void intra_pred_horiz(uint8_t *d,uint8_t *top,uint8_t *left,int stride) {
+    int y;
+    uint64_t a;
+    for(y=0;y<8;y++) {
+        a = left[y+1] * 0x0101010101010101ULL;
+        *((uint64_t *)(d+y*stride)) = a;
+    }
+}
+
+static void intra_pred_dc_128(uint8_t *d,uint8_t *top,uint8_t *left,int stride) {
+    int y;
+    uint64_t a = 0x8080808080808080ULL;
+    for(y=0;y<8;y++)
+        *((uint64_t *)(d+y*stride)) = a;
+}
+
+static void intra_pred_plane(uint8_t *d,uint8_t *top,uint8_t *left,int stride) {
+    int x,y,ia;
+    int ih = 0;
+    int iv = 0;
+    uint8_t *cm = cropTbl + MAX_NEG_CROP;
+
+    for(x=0; x<4; x++) {
+        ih += (x+1)*(top[5+x]-top[3-x]);
+        iv += (x+1)*(left[5+x]-left[3-x]);
+    }
+    ia = (top[8]+left[8])<<4;
+    ih = (17*ih+16)>>5;
+    iv = (17*iv+16)>>5;
+    for(y=0; y<8; y++)
+        for(x=0; x<8; x++)
+            d[y*stride+x] = cm[(ia+(x-3)*ih+(y-3)*iv+16)>>5];
+}
+
+#define LOWPASS(ARRAY,INDEX)                                            \
+    (( ARRAY[(INDEX)-1] + 2*ARRAY[(INDEX)] + ARRAY[(INDEX)+1] + 2) >> 2)
+
+static void intra_pred_lp(uint8_t *d,uint8_t *top,uint8_t *left,int stride) {
+    int x,y;
+    for(y=0; y<8; y++)
+        for(x=0; x<8; x++)
+            d[y*stride+x] = (LOWPASS(top,x+1) + LOWPASS(left,y+1)) >> 1;
+}
+
+static void intra_pred_down_left(uint8_t *d,uint8_t *top,uint8_t *left,int stride) {
+    int x,y;
+    for(y=0; y<8; y++)
+        for(x=0; x<8; x++)
+            d[y*stride+x] = (LOWPASS(top,x+y+2) + LOWPASS(left,x+y+2)) >> 1;
+}
+
+static void intra_pred_down_right(uint8_t *d,uint8_t *top,uint8_t *left,int stride) {
+    int x,y;
+    for(y=0; y<8; y++)
+        for(x=0; x<8; x++)
+            if(x==y)
+                d[y*stride+x] = (left[1]+2*top[0]+top[1]+2)>>2;
+            else if(x>y)
+                d[y*stride+x] = LOWPASS(top,x-y);
+            else
+                d[y*stride+x] = LOWPASS(left,y-x);
+}
+
+static void intra_pred_lp_left(uint8_t *d,uint8_t *top,uint8_t *left,int stride) {
+    int x,y;
+    for(y=0; y<8; y++)
+        for(x=0; x<8; x++)
+            d[y*stride+x] = LOWPASS(left,y+1);
+}
+
+static void intra_pred_lp_top(uint8_t *d,uint8_t *top,uint8_t *left,int stride) {
+    int x,y;
+    for(y=0; y<8; y++)
+        for(x=0; x<8; x++)
+            d[y*stride+x] = LOWPASS(top,x+1);
+}
+
+#undef LOWPASS
+
+static inline void modify_pred(const int8_t *mod_table, int *mode) {
+    int newmode = mod_table[(int)*mode];
+    if(newmode < 0) {
+        av_log(NULL, AV_LOG_ERROR, "Illegal intra prediction mode\n");
+        *mode = 0;
+    } else {
+        *mode = newmode;
+    }
+}
+
+/*****************************************************************************
+ *
+ * motion compensation
+ *
+ ****************************************************************************/
+
+static inline void mc_dir_part(AVSContext *h,Picture *pic,int square,
+                        int chroma_height,int delta,int list,uint8_t *dest_y,
+                        uint8_t *dest_cb,uint8_t *dest_cr,int src_x_offset,
+                        int src_y_offset,qpel_mc_func *qpix_op,
+                        h264_chroma_mc_func chroma_op,vector_t *mv){
+    MpegEncContext * const s = &h->s;
+    const int mx= mv->x + src_x_offset*8;
+    const int my= mv->y + src_y_offset*8;
+    const int luma_xy= (mx&3) + ((my&3)<<2);
+    uint8_t * src_y = pic->data[0] + (mx>>2) + (my>>2)*h->l_stride;
+    uint8_t * src_cb= pic->data[1] + (mx>>3) + (my>>3)*h->c_stride;
+    uint8_t * src_cr= pic->data[2] + (mx>>3) + (my>>3)*h->c_stride;
+    int extra_width= 0; //(s->flags&CODEC_FLAG_EMU_EDGE) ? 0 : 16;
+    int extra_height= extra_width;
+    int emu=0;
+    const int full_mx= mx>>2;
+    const int full_my= my>>2;
+    const int pic_width  = 16*h->mb_width;
+    const int pic_height = 16*h->mb_height;
+
+    if(!pic->data[0])
+        return;
+    if(mx&7) extra_width -= 3;
+    if(my&7) extra_height -= 3;
+
+    if(   full_mx < 0-extra_width
+          || full_my < 0-extra_height
+          || full_mx + 16/*FIXME*/ > pic_width + extra_width
+          || full_my + 16/*FIXME*/ > pic_height + extra_height){
+        ff_emulated_edge_mc(s->edge_emu_buffer, src_y - 2 - 2*h->l_stride, h->l_stride,
+                            16+5, 16+5/*FIXME*/, full_mx-2, full_my-2, pic_width, pic_height);
+        src_y= s->edge_emu_buffer + 2 + 2*h->l_stride;
+        emu=1;
+    }
+
+    qpix_op[luma_xy](dest_y, src_y, h->l_stride); //FIXME try variable height perhaps?
+    if(!square){
+        qpix_op[luma_xy](dest_y + delta, src_y + delta, h->l_stride);
+    }
+
+    if(emu){
+        ff_emulated_edge_mc(s->edge_emu_buffer, src_cb, h->c_stride,
+                            9, 9/*FIXME*/, (mx>>3), (my>>3), pic_width>>1, pic_height>>1);
+        src_cb= s->edge_emu_buffer;
+    }
+    chroma_op(dest_cb, src_cb, h->c_stride, chroma_height, mx&7, my&7);
+
+    if(emu){
+        ff_emulated_edge_mc(s->edge_emu_buffer, src_cr, h->c_stride,
+                            9, 9/*FIXME*/, (mx>>3), (my>>3), pic_width>>1, pic_height>>1);
+        src_cr= s->edge_emu_buffer;
+    }
+    chroma_op(dest_cr, src_cr, h->c_stride, chroma_height, mx&7, my&7);
+}
+
+static inline void mc_part_std(AVSContext *h,int square,int chroma_height,int delta,
+                        uint8_t *dest_y,uint8_t *dest_cb,uint8_t *dest_cr,
+                        int x_offset, int y_offset,qpel_mc_func *qpix_put,
+                        h264_chroma_mc_func chroma_put,qpel_mc_func *qpix_avg,
+                        h264_chroma_mc_func chroma_avg, vector_t *mv){
+    qpel_mc_func *qpix_op=  qpix_put;
+    h264_chroma_mc_func chroma_op= chroma_put;
+
+    dest_y  += 2*x_offset + 2*y_offset*h->l_stride;
+    dest_cb +=   x_offset +   y_offset*h->c_stride;
+    dest_cr +=   x_offset +   y_offset*h->c_stride;
+    x_offset += 8*h->mbx;
+    y_offset += 8*h->mby;
+
+    if(mv->ref >= 0){
+        Picture *ref= &h->DPB[mv->ref];
+        mc_dir_part(h, ref, square, chroma_height, delta, 0,
+                    dest_y, dest_cb, dest_cr, x_offset, y_offset,
+                    qpix_op, chroma_op, mv);
+
+        qpix_op=  qpix_avg;
+        chroma_op= chroma_avg;
+    }
+
+    if((mv+MV_BWD_OFFS)->ref >= 0){
+        Picture *ref= &h->DPB[0];
+        mc_dir_part(h, ref, square, chroma_height, delta, 1,
+                    dest_y, dest_cb, dest_cr, x_offset, y_offset,
+                    qpix_op, chroma_op, mv+MV_BWD_OFFS);
+    }
+}
+
+static void inter_pred(AVSContext *h) {
+    /* always do 8x8 blocks TODO: are larger blocks worth it? */
+    mc_part_std(h, 1, 4, 0, h->cy, h->cu, h->cv, 0, 0,
+                h->s.dsp.put_cavs_qpel_pixels_tab[1],
+                h->s.dsp.put_h264_chroma_pixels_tab[1],
+                h->s.dsp.avg_cavs_qpel_pixels_tab[1],
+                h->s.dsp.avg_h264_chroma_pixels_tab[1],&h->mv[MV_FWD_X0]);
+    mc_part_std(h, 1, 4, 0, h->cy, h->cu, h->cv, 4, 0,
+                h->s.dsp.put_cavs_qpel_pixels_tab[1],
+                h->s.dsp.put_h264_chroma_pixels_tab[1],
+                h->s.dsp.avg_cavs_qpel_pixels_tab[1],
+                h->s.dsp.avg_h264_chroma_pixels_tab[1],&h->mv[MV_FWD_X1]);
+    mc_part_std(h, 1, 4, 0, h->cy, h->cu, h->cv, 0, 4,
+                h->s.dsp.put_cavs_qpel_pixels_tab[1],
+                h->s.dsp.put_h264_chroma_pixels_tab[1],
+                h->s.dsp.avg_cavs_qpel_pixels_tab[1],
+                h->s.dsp.avg_h264_chroma_pixels_tab[1],&h->mv[MV_FWD_X2]);
+    mc_part_std(h, 1, 4, 0, h->cy, h->cu, h->cv, 4, 4,
+                h->s.dsp.put_cavs_qpel_pixels_tab[1],
+                h->s.dsp.put_h264_chroma_pixels_tab[1],
+                h->s.dsp.avg_cavs_qpel_pixels_tab[1],
+                h->s.dsp.avg_h264_chroma_pixels_tab[1],&h->mv[MV_FWD_X3]);
+    /* set intra prediction modes to default values */
+    h->pred_mode_Y[3] =  h->pred_mode_Y[6] = INTRA_L_LP;
+    h->top_pred_Y[h->mbx*2+0] = h->top_pred_Y[h->mbx*2+1] = INTRA_L_LP;
+}
+
+/*****************************************************************************
+ *
+ * motion vector prediction
+ *
+ ****************************************************************************/
+
+static inline void veccpy(vector_t *dst, vector_t *src) {
+    *((uint64_t *)dst) = *((uint64_t *)src);
+}
+
+static inline void set_mvs(vector_t *mv, enum block_t size) {
+    switch(size) {
+    case BLK_16X16:
+        veccpy(mv+MV_STRIDE  ,mv);
+        veccpy(mv+MV_STRIDE+1,mv);
+    case BLK_16X8:
+        veccpy(mv          +1,mv);
+        break;
+    case BLK_8X16:
+        veccpy(mv+MV_STRIDE  ,mv);
+        break;
+    }
+}
+
+static inline void store_mvs(AVSContext *h) {
+    veccpy(&h->col_mv[(h->mby*h->mb_width + h->mbx)*4 + 0], &h->mv[MV_FWD_X0]);
+    veccpy(&h->col_mv[(h->mby*h->mb_width + h->mbx)*4 + 1], &h->mv[MV_FWD_X1]);
+    veccpy(&h->col_mv[(h->mby*h->mb_width + h->mbx)*4 + 2], &h->mv[MV_FWD_X2]);
+    veccpy(&h->col_mv[(h->mby*h->mb_width + h->mbx)*4 + 3], &h->mv[MV_FWD_X3]);
+}
+
+static inline void scale_mv(AVSContext *h, int *d_x, int *d_y, vector_t *src, int distp) {
+    int den = h->scale_den[src->ref];
+
+    *d_x = (src->x*distp*den + 256 + (src->x>>31)) >> 9;
+    *d_y = (src->y*distp*den + 256 + (src->y>>31)) >> 9;
+}
+
+static inline void mv_pred_median(AVSContext *h, vector_t *mvP, vector_t *mvA, vector_t *mvB, vector_t *mvC) {
+    int ax, ay, bx, by, cx, cy;
+    int len_ab, len_bc, len_ca, len_mid;
+
+    /* scale candidates according to their temporal span */
+    scale_mv(h, &ax, &ay, mvA, mvP->dist);
+    scale_mv(h, &bx, &by, mvB, mvP->dist);
+    scale_mv(h, &cx, &cy, mvC, mvP->dist);
+    /* find the geometrical median of the three candidates */
+    len_ab = abs(ax - bx) + abs(ay - by);
+    len_bc = abs(bx - cx) + abs(by - cy);
+    len_ca = abs(cx - ax) + abs(cy - ay);
+    len_mid = mid_pred(len_ab, len_bc, len_ca);
+    if(len_mid == len_ab) {
+        mvP->x = cx;
+        mvP->y = cy;
+    } else if(len_mid == len_bc) {
+        mvP->x = ax;
+        mvP->y = ay;
+    } else {
+        mvP->x = bx;
+        mvP->y = by;
+    }
+}
+
+static inline void mv_pred_direct(AVSContext *h, vector_t *pmv_fw,
+                                  vector_t *pmv_bw, vector_t *col_mv) {
+    int den = h->direct_den[col_mv->ref];
+    int m = col_mv->x >> 31;
+
+    pmv_fw->dist = h->dist[1];
+    pmv_bw->dist = h->dist[0];
+    pmv_fw->ref = 1;
+    pmv_bw->ref = 0;
+    /* scale the co-located motion vector according to its temporal span */
+    pmv_fw->x = (((den+(den*col_mv->x*pmv_fw->dist^m)-m-1)>>14)^m)-m;
+    pmv_bw->x = m-(((den+(den*col_mv->x*pmv_bw->dist^m)-m-1)>>14)^m);
+    m = col_mv->y >> 31;
+    pmv_fw->y = (((den+(den*col_mv->y*pmv_fw->dist^m)-m-1)>>14)^m)-m;
+    pmv_bw->y = m-(((den+(den*col_mv->y*pmv_bw->dist^m)-m-1)>>14)^m);
+}
+
+static inline void mv_pred_sym(AVSContext *h, vector_t *src, enum block_t size) {
+    vector_t *dst = src + MV_BWD_OFFS;
+
+    /* backward mv is the scaled and negated forward mv */
+    dst->x = -((src->x * h->sym_factor + 256) >> 9);
+    dst->y = -((src->y * h->sym_factor + 256) >> 9);
+    dst->ref = 0;
+    dst->dist = h->dist[0];
+    set_mvs(dst, size);
+}
+
+static void mv_pred(AVSContext *h, enum mv_loc_t nP, enum mv_loc_t nC,
+                    enum mv_pred_t mode, enum block_t size, int ref) {
+    vector_t *mvP = &h->mv[nP];
+    vector_t *mvA = &h->mv[nP-1];
+    vector_t *mvB = &h->mv[nP-4];
+    vector_t *mvC = &h->mv[nC];
+    int mvAref = mvA->ref;
+    int mvBref = mvB->ref;
+    int mvCref;
+
+    mvP->ref = ref;
+    mvP->dist = h->dist[mvP->ref];
+    if(mvC->ref == NOT_AVAIL)
+        mvC = &h->mv[nP-5]; // set to top-left (mvD)
+    mvCref = mvC->ref;
+    if(mode == MV_PRED_PSKIP) {
+        if((mvAref == NOT_AVAIL) || (mvBref == NOT_AVAIL) ||
+           ((mvA->x | mvA->y | mvA->ref) == 0)  ||
+           ((mvB->x | mvB->y | mvB->ref) == 0) ) {
+            mvP->x = mvP->y = 0;
+            set_mvs(mvP,size);
+            return;
+        }
+    }
+    /* if there is only one suitable candidate, take it */
+    if((mvAref >= 0) && (mvBref < 0) && (mvCref < 0)) {
+        mvP->x = mvA->x;
+        mvP->y = mvA->y;
+    } else if((mvAref < 0) && (mvBref >= 0) && (mvCref < 0)) {
+        mvP->x = mvB->x;
+        mvP->y = mvB->y;
+    } else if((mvAref < 0) && (mvBref < 0) && (mvCref >= 0)) {
+        mvP->x = mvC->x;
+        mvP->y = mvC->y;
+    } else {
+        switch(mode) {
+        case MV_PRED_LEFT:
+            if(mvAref == mvP->ref) {
+                mvP->x = mvA->x;
+                mvP->y = mvA->y;
+            } else
+                mv_pred_median(h, mvP, mvA, mvB, mvC);
+            break;
+        case MV_PRED_TOP:
+            if(mvBref == mvP->ref) {
+                mvP->x = mvB->x;
+                mvP->y = mvB->y;
+            } else
+                mv_pred_median(h, mvP, mvA, mvB, mvC);
+            break;
+        case MV_PRED_TOPRIGHT:
+            if(mvCref == mvP->ref) {
+                mvP->x = mvC->x;
+                mvP->y = mvC->y;
+            } else
+                mv_pred_median(h, mvP, mvA, mvB, mvC);
+            break;
+        default:
+            mv_pred_median(h, mvP, mvA, mvB, mvC);
+            break;
+        }
+    }
+    if(mode < MV_PRED_PSKIP) {
+        mvP->x += get_se_golomb(&h->s.gb);
+        mvP->y += get_se_golomb(&h->s.gb);
+    }
+    set_mvs(mvP,size);
+}
+
+/*****************************************************************************
+ *
+ * residual data decoding
+ *
+ ****************************************************************************/
+
+/* kth-order exponential golomb code */
+static inline int get_ue_code(GetBitContext *gb, int order) {
+    if(order)
+        return (get_ue_golomb(gb) << order) + get_bits(gb,order);
+    return get_ue_golomb(gb);
+}
+
+static int decode_residual_block(AVSContext *h, GetBitContext *gb,
+                                 const residual_vlc_t *r, int esc_golomb_order,
+                                 int qp, uint8_t *dst, int stride) {
+    int i,pos = -1;
+    int level_code, esc_code, level, run, mask;
+    int level_buf[64];
+    int run_buf[64];
+    int dqm = dequant_mul[qp];
+    int dqs = dequant_shift[qp];
+    int dqa = 1 << (dqs - 1);
+    const uint8_t *scantab = ff_zigzag_direct;
+    DCTELEM block[64];
+
+    memset(block,0,64*sizeof(DCTELEM));
+    for(i=0;i<65;i++) {
+        level_code = get_ue_code(gb,r->golomb_order);
+        if(level_code >= ESCAPE_CODE) {
+            run = (level_code - ESCAPE_CODE) >> 1;
+            esc_code = get_ue_code(gb,esc_golomb_order);
+            level = esc_code + (run > r->max_run ? 1 : r->level_add[run]);
+            while(level > r->inc_limit)
+                r++;
+            mask = -(level_code & 1);
+            level = (level^mask) - mask;
+        } else {
+            if(level_code < 0)
+                return -1;
+            level = r->rltab[level_code][0];
+            if(!level) //end of block signal
+                break;
+            run   = r->rltab[level_code][1];
+            r += r->rltab[level_code][2];
+        }
+        level_buf[i] = level;
+        run_buf[i] = run;
+    }
+    /* inverse scan and dequantization */
+    for(i=i-1;i>=0;i--) {
+        pos += 1 + run_buf[i];
+        if(pos > 63) {
+            av_log(h->s.avctx, AV_LOG_ERROR,
+                   "position out of block bounds at pic %d MB(%d,%d)\n",
+                   h->picture.poc, h->mbx, h->mby);
+            return -1;
+        }
+        block[scantab[pos]] = (level_buf[i]*dqm + dqa) >> dqs;
+    }
+    h->s.dsp.cavs_idct8_add(dst,block,stride);
+    return 0;
+}
+
+
+static inline void decode_residual_chroma(AVSContext *h) {
+    if(h->cbp & (1<<4))
+        decode_residual_block(h,&h->s.gb,chroma_2dvlc,0, chroma_qp[h->qp],
+                              h->cu,h->c_stride);
+    if(h->cbp & (1<<5))
+        decode_residual_block(h,&h->s.gb,chroma_2dvlc,0, chroma_qp[h->qp],
+                              h->cv,h->c_stride);
+}
+
+static inline void decode_residual_inter(AVSContext *h) {
+    int block;
+
+    /* get coded block pattern */
+    h->cbp = cbp_tab[get_ue_golomb(&h->s.gb)][1];
+    /* get quantizer */
+    if(h->cbp && !h->qp_fixed)
+        h->qp += get_se_golomb(&h->s.gb);
+    for(block=0;block<4;block++)
+        if(h->cbp & (1<<block))
+            decode_residual_block(h,&h->s.gb,inter_2dvlc,0,h->qp,
+                                  h->cy + h->luma_scan[block], h->l_stride);
+    decode_residual_chroma(h);
+}
+
+/*****************************************************************************
+ *
+ * macroblock level
+ *
+ ****************************************************************************/
+
+static inline void init_mb(AVSContext *h) {
+    int i;
+
+    /* copy predictors from top line (MB B and C) into cache */
+    for(i=0;i<3;i++) {
+        veccpy(&h->mv[MV_FWD_B2+i],&h->top_mv[0][h->mbx*2+i]);
+        veccpy(&h->mv[MV_BWD_B2+i],&h->top_mv[1][h->mbx*2+i]);
+    }
+    h->pred_mode_Y[1] = h->top_pred_Y[h->mbx*2+0];
+    h->pred_mode_Y[2] = h->top_pred_Y[h->mbx*2+1];
+    /* clear top predictors if MB B is not available */
+    if(!(h->flags & B_AVAIL)) {
+        veccpy(&h->mv[MV_FWD_B2],(vector_t *)&un_mv);
+        veccpy(&h->mv[MV_FWD_B3],(vector_t *)&un_mv);
+        veccpy(&h->mv[MV_BWD_B2],(vector_t *)&un_mv);
+        veccpy(&h->mv[MV_BWD_B3],(vector_t *)&un_mv);
+        h->pred_mode_Y[1] = h->pred_mode_Y[2] = NOT_AVAIL;
+        h->flags &= ~(C_AVAIL|D_AVAIL);
+    } else if(h->mbx) {
+        h->flags |= D_AVAIL;
+    }
+    if(h->mbx == h->mb_width-1) //MB C not available
+        h->flags &= ~C_AVAIL;
+    /* clear top-right predictors if MB C is not available */
+    if(!(h->flags & C_AVAIL)) {
+        veccpy(&h->mv[MV_FWD_C2],(vector_t *)&un_mv);
+        veccpy(&h->mv[MV_BWD_C2],(vector_t *)&un_mv);
+    }
+    /* clear top-left predictors if MB D is not available */
+    if(!(h->flags & D_AVAIL)) {
+        veccpy(&h->mv[MV_FWD_D3],(vector_t *)&un_mv);
+        veccpy(&h->mv[MV_BWD_D3],(vector_t *)&un_mv);
+    }
+    /* set pointer for co-located macroblock type */
+    h->col_type = &h->col_type_base[h->mby*h->mb_width + h->mbx];
+}
+
+static inline void check_for_slice(AVSContext *h);
+
+static inline int next_mb(AVSContext *h) {
+    int i;
+
+    h->flags |= A_AVAIL;
+    h->cy += 16;
+    h->cu += 8;
+    h->cv += 8;
+    /* copy mvs as predictors to the left */
+    for(i=0;i<=20;i+=4)
+        veccpy(&h->mv[i],&h->mv[i+2]);
+    /* copy bottom mvs from cache to top line */
+    veccpy(&h->top_mv[0][h->mbx*2+0],&h->mv[MV_FWD_X2]);
+    veccpy(&h->top_mv[0][h->mbx*2+1],&h->mv[MV_FWD_X3]);
+    veccpy(&h->top_mv[1][h->mbx*2+0],&h->mv[MV_BWD_X2]);
+    veccpy(&h->top_mv[1][h->mbx*2+1],&h->mv[MV_BWD_X3]);
+    /* next MB address */
+    h->mbx++;
+    if(h->mbx == h->mb_width) { //new mb line
+        h->flags = B_AVAIL|C_AVAIL;
+        /* clear left pred_modes */
+        h->pred_mode_Y[3] = h->pred_mode_Y[6] = NOT_AVAIL;
+        /* clear left mv predictors */
+        for(i=0;i<=20;i+=4)
+            veccpy(&h->mv[i],(vector_t *)&un_mv);
+        h->mbx = 0;
+        h->mby++;
+        /* re-calculate sample pointers */
+        h->cy = h->picture.data[0] + h->mby*16*h->l_stride;
+        h->cu = h->picture.data[1] + h->mby*8*h->c_stride;
+        h->cv = h->picture.data[2] + h->mby*8*h->c_stride;
+        if(h->mby == h->mb_height) { //frame end
+            return 0;
+        } else {
+            //check_for_slice(h);
+        }
+    }
+    return 1;
+}
+
+static void decode_mb_i(AVSContext *h, int is_i_pic) {
+    GetBitContext *gb = &h->s.gb;
+    int block, pred_mode_uv;
+    uint8_t top[18];
+    uint8_t left[18];
+    uint8_t *d;
+
+    /* get intra prediction modes from stream */
+    for(block=0;block<4;block++) {
+        int nA,nB,predpred;
+        int pos = scan3x3[block];
+
+        nA = h->pred_mode_Y[pos-1];
+        nB = h->pred_mode_Y[pos-3];
+        if((nA == NOT_AVAIL) || (nB == NOT_AVAIL))
+            predpred = 2;
+        else
+            predpred = FFMIN(nA,nB);
+        if(get_bits1(gb))
+            h->pred_mode_Y[pos] = predpred;
+        else {
+            h->pred_mode_Y[pos] = get_bits(gb,2);
+            if(h->pred_mode_Y[pos] >= predpred)
+                h->pred_mode_Y[pos]++;
+        }
+    }
+    pred_mode_uv = get_ue_golomb(gb);
+    if(pred_mode_uv > 6) {
+        av_log(h->s.avctx, AV_LOG_ERROR, "illegal intra chroma pred mode\n");
+        pred_mode_uv = 0;
+    }
+
+    /* save pred modes before they get modified */
+    h->pred_mode_Y[3] =  h->pred_mode_Y[5];
+    h->pred_mode_Y[6] =  h->pred_mode_Y[8];
+    h->top_pred_Y[h->mbx*2+0] = h->pred_mode_Y[7];
+    h->top_pred_Y[h->mbx*2+1] = h->pred_mode_Y[8];
+
+    /* modify pred modes according to availability of neighbour samples */
+    if(!(h->flags & A_AVAIL)) {
+        modify_pred(left_modifier_l, &h->pred_mode_Y[4] );
+        modify_pred(left_modifier_l, &h->pred_mode_Y[7] );
+        modify_pred(left_modifier_c, &pred_mode_uv );
+    }
+    if(!(h->flags & B_AVAIL)) {
+        modify_pred(top_modifier_l, &h->pred_mode_Y[4] );
+        modify_pred(top_modifier_l, &h->pred_mode_Y[5] );
+        modify_pred(top_modifier_c, &pred_mode_uv );
+    }
+
+    /* get coded block pattern */
+    if(is_i_pic)
+        h->cbp = cbp_tab[get_ue_golomb(gb)][0];
+    if(h->cbp && !h->qp_fixed)
+        h->qp += get_se_golomb(gb); //qp_delta
+
+    /* luma intra prediction interleaved with residual decode/transform/add */
+    for(block=0;block<4;block++) {
+        d = h->cy + h->luma_scan[block];
+        load_intra_pred_luma(h, top, left, block);
+        h->intra_pred_l[(int)h->pred_mode_Y[scan3x3[block]]]
+            (d, top, left, h->l_stride);
+        if(h->cbp & (1<<block))
+            decode_residual_block(h,gb,intra_2dvlc,1,h->qp,d,h->l_stride);
+    }
+
+    /* chroma intra prediction */
+    load_intra_pred_chroma(&h->top_border_u[h->mbx*8], h->left_border_u,
+                           h->topleft_border_u, top, left, h->c_stride, h->flags);
+    h->intra_pred_c[pred_mode_uv](h->cu, top, left, h->c_stride);
+    load_intra_pred_chroma(&h->top_border_v[h->mbx*8], h->left_border_v,
+                           h->topleft_border_v, top, left, h->c_stride, h->flags);
+    h->intra_pred_c[pred_mode_uv](h->cv, top, left, h->c_stride);
+
+    decode_residual_chroma(h);
+    filter_mb(h,I_8X8);
+
+    /* mark motion vectors as intra */
+    veccpy( &h->mv[MV_FWD_X0], (vector_t *)&intra_mv);
+    set_mvs(&h->mv[MV_FWD_X0], BLK_16X16);
+    veccpy( &h->mv[MV_BWD_X0], (vector_t *)&intra_mv);
+    set_mvs(&h->mv[MV_BWD_X0], BLK_16X16);
+    if(h->pic_type != FF_B_TYPE)
+        *h->col_type = I_8X8;
+}
+
+static void mb_skip_p(AVSContext *h) {
+    mv_pred(h, MV_FWD_X0, MV_FWD_C2, MV_PRED_PSKIP, BLK_16X16, 0);
+    inter_pred(h);
+    store_mvs(h);
+    filter_mb(h,P_SKIP);
+    *h->col_type = P_SKIP;
+}
+
+
+static void mb_skip_b(AVSContext *h) {
+    int i;
+
+    if(!(*h->col_type)) {
+        /* intra MB at co-location, do in-plane prediction */
+        mv_pred(h, MV_FWD_X0, MV_FWD_C2, MV_PRED_BSKIP, BLK_16X16, 1);
+        mv_pred(h, MV_BWD_X0, MV_BWD_C2, MV_PRED_BSKIP, BLK_16X16, 0);
+    } else {
+        /* direct prediction from co-located P MB, block-wise */
+        for(i=0;i<4;i++)
+            mv_pred_direct(h,&h->mv[mv_scan[i]],
+                           &h->mv[mv_scan[i]+MV_BWD_OFFS],
+                           &h->col_mv[(h->mby*h->mb_width + h->mbx)*4 + i]);
+    }
+}
+
+static void decode_mb_p(AVSContext *h, enum mb_t mb_type) {
+    GetBitContext *gb = &h->s.gb;
+    int ref[4];
+
+    switch(mb_type) {
+    case P_SKIP:
+        mb_skip_p(h);
+        return;
+    case P_16X16:
+        ref[0] = h->ref_flag ? 0 : get_bits1(gb);
+        mv_pred(h, MV_FWD_X0, MV_FWD_C2, MV_PRED_MEDIAN,   BLK_16X16,ref[0]);
+        break;
+    case P_16X8:
+        ref[0] = h->ref_flag ? 0 : get_bits1(gb);
+        ref[2] = h->ref_flag ? 0 : get_bits1(gb);
+        mv_pred(h, MV_FWD_X0, MV_FWD_C2, MV_PRED_TOP,      BLK_16X8, ref[0]);
+        mv_pred(h, MV_FWD_X2, MV_FWD_A1, MV_PRED_LEFT,     BLK_16X8, ref[2]);
+        break;
+    case P_8X16:
+        ref[0] = h->ref_flag ? 0 : get_bits1(gb);
+        ref[1] = h->ref_flag ? 0 : get_bits1(gb);
+        mv_pred(h, MV_FWD_X0, MV_FWD_B3, MV_PRED_LEFT,     BLK_8X16, ref[0]);
+        mv_pred(h, MV_FWD_X1, MV_FWD_C2, MV_PRED_TOPRIGHT, BLK_8X16, ref[1]);
+        break;
+    case P_8X8:
+        ref[0] = h->ref_flag ? 0 : get_bits1(gb);
+        ref[1] = h->ref_flag ? 0 : get_bits1(gb);
+        ref[2] = h->ref_flag ? 0 : get_bits1(gb);
+        ref[3] = h->ref_flag ? 0 : get_bits1(gb);
+        mv_pred(h, MV_FWD_X0, MV_FWD_B3, MV_PRED_MEDIAN,   BLK_8X8, ref[0]);
+        mv_pred(h, MV_FWD_X1, MV_FWD_C2, MV_PRED_MEDIAN,   BLK_8X8, ref[1]);
+        mv_pred(h, MV_FWD_X2, MV_FWD_X1, MV_PRED_MEDIAN,   BLK_8X8, ref[2]);
+        mv_pred(h, MV_FWD_X3, MV_FWD_X0, MV_PRED_MEDIAN,   BLK_8X8, ref[3]);
+    }
+    inter_pred(h);
+    store_mvs(h);
+    decode_residual_inter(h);
+    filter_mb(h,mb_type);
+    *h->col_type = mb_type;
+}
+
+static void decode_mb_b(AVSContext *h, enum mb_t mb_type) {
+    int block;
+    enum sub_mb_t sub_type[4];
+    int flags;
+
+    /* reset all MVs */
+    veccpy( &h->mv[MV_FWD_X0], (vector_t *)&dir_mv);
+    set_mvs(&h->mv[MV_FWD_X0], BLK_16X16);
+    veccpy( &h->mv[MV_BWD_X0], (vector_t *)&dir_mv);
+    set_mvs(&h->mv[MV_BWD_X0], BLK_16X16);
+    switch(mb_type) {
+    case B_SKIP:
+        mb_skip_b(h);
+        inter_pred(h);
+        filter_mb(h,B_SKIP);
+        return;
+    case B_DIRECT:
+        mb_skip_b(h);
+        break;
+    case B_FWD_16X16:
+        mv_pred(h, MV_FWD_X0, MV_FWD_C2, MV_PRED_MEDIAN, BLK_16X16, 1);
+        break;
+    case B_SYM_16X16:
+        mv_pred(h, MV_FWD_X0, MV_FWD_C2, MV_PRED_MEDIAN, BLK_16X16, 1);
+        mv_pred_sym(h, &h->mv[MV_FWD_X0], BLK_16X16);
+        break;
+    case B_BWD_16X16:
+        mv_pred(h, MV_BWD_X0, MV_BWD_C2, MV_PRED_MEDIAN, BLK_16X16, 0);
+        break;
+    case B_8X8:
+        for(block=0;block<4;block++)
+            sub_type[block] = get_bits(&h->s.gb,2);
+        for(block=0;block<4;block++) {
+            switch(sub_type[block]) {
+            case B_SUB_DIRECT:
+                if(!(*h->col_type)) {
+                    /* intra MB at co-location, do in-plane prediction */
+                    mv_pred(h, mv_scan[block], mv_scan[block]-3,
+                            MV_PRED_BSKIP, BLK_8X8, 1);
+                    mv_pred(h, mv_scan[block]+MV_BWD_OFFS,
+                            mv_scan[block]-3+MV_BWD_OFFS,
+                            MV_PRED_BSKIP, BLK_8X8, 0);
+                } else
+                    mv_pred_direct(h,&h->mv[mv_scan[block]],
+                                   &h->mv[mv_scan[block]+MV_BWD_OFFS],
+                                   &h->col_mv[(h->mby*h->mb_width + h->mbx)*4 + block]);
+                break;
+            case B_SUB_FWD:
+                mv_pred(h, mv_scan[block], mv_scan[block]-3,
+                        MV_PRED_MEDIAN, BLK_8X8, 1);
+                break;
+            case B_SUB_SYM:
+                mv_pred(h, mv_scan[block], mv_scan[block]-3,
+                        MV_PRED_MEDIAN, BLK_8X8, 1);
+                mv_pred_sym(h, &h->mv[mv_scan[block]], BLK_8X8);
+                break;
+            }
+        }
+        for(block=0;block<4;block++) {
+            if(sub_type[block] == B_SUB_BWD)
+                mv_pred(h, mv_scan[block]+MV_BWD_OFFS,
+                        mv_scan[block]+MV_BWD_OFFS-3,
+                        MV_PRED_MEDIAN, BLK_8X8, 0);
+        }
+        break;
+    default:
+        assert((mb_type > B_SYM_16X16) && (mb_type < B_8X8));
+        flags = b_partition_flags[(mb_type-1)>>1];
+        if(mb_type & 1) { /* 16x8 macroblock types */
+            if(flags & FWD0)
+                mv_pred(h, MV_FWD_X0, MV_FWD_C2, MV_PRED_TOP,  BLK_16X8, 1);
+            if(flags & SYM0) {
+                mv_pred(h, MV_FWD_X0, MV_FWD_C2, MV_PRED_TOP,  BLK_16X8, 1);
+                mv_pred_sym(h, &h->mv[MV_FWD_X0], BLK_16X8);
+            }
+            if(flags & FWD1)
+                mv_pred(h, MV_FWD_X2, MV_FWD_A1, MV_PRED_LEFT, BLK_16X8, 1);
+            if(flags & SYM1) {
+                mv_pred(h, MV_FWD_X2, MV_FWD_A1, MV_PRED_LEFT, BLK_16X8, 1);
+                mv_pred_sym(h, &h->mv[9], BLK_16X8);
+            }
+            if(flags & BWD0)
+                mv_pred(h, MV_BWD_X0, MV_BWD_C2, MV_PRED_TOP,  BLK_16X8, 0);
+            if(flags & BWD1)
+                mv_pred(h, MV_BWD_X2, MV_BWD_A1, MV_PRED_LEFT, BLK_16X8, 0);
+        } else {          /* 8x16 macroblock types */
+            if(flags & FWD0)
+                mv_pred(h, MV_FWD_X0, MV_FWD_B3, MV_PRED_LEFT, BLK_8X16, 1);
+            if(flags & SYM0) {
+                mv_pred(h, MV_FWD_X0, MV_FWD_B3, MV_PRED_LEFT, BLK_8X16, 1);
+                mv_pred_sym(h, &h->mv[MV_FWD_X0], BLK_8X16);
+            }
+            if(flags & FWD1)
+                mv_pred(h, MV_FWD_X1, MV_FWD_C2, MV_PRED_TOPRIGHT,BLK_8X16, 1);
+            if(flags & SYM1) {
+                mv_pred(h, MV_FWD_X1, MV_FWD_C2, MV_PRED_TOPRIGHT,BLK_8X16, 1);
+                mv_pred_sym(h, &h->mv[6], BLK_8X16);
+            }
+            if(flags & BWD0)
+                mv_pred(h, MV_BWD_X0, MV_BWD_B3, MV_PRED_LEFT, BLK_8X16, 0);
+            if(flags & BWD1)
+                mv_pred(h, MV_BWD_X1, MV_BWD_C2, MV_PRED_TOPRIGHT,BLK_8X16, 0);
+        }
+    }
+    inter_pred(h);
+    decode_residual_inter(h);
+    filter_mb(h,mb_type);
+}
+
+/*****************************************************************************
+ *
+ * slice level
+ *
+ ****************************************************************************/
+
+static inline int decode_slice_header(AVSContext *h, GetBitContext *gb) {
+    if(h->stc > 0xAF)
+        av_log(h->s.avctx, AV_LOG_ERROR, "unexpected start code 0x%02x\n", h->stc);
+    h->mby = h->stc;
+    if((h->mby == 0) && (!h->qp_fixed)){
+        h->qp_fixed = get_bits1(gb);
+        h->qp = get_bits(gb,6);
+    }
+    /* inter frame or second slice can have weighting params */
+    if((h->pic_type != FF_I_TYPE) || (!h->pic_structure && h->mby >= h->mb_width/2))
+        if(get_bits1(gb)) { //slice_weighting_flag
+            av_log(h->s.avctx, AV_LOG_ERROR,
+                   "weighted prediction not yet supported\n");
+        }
+    return 0;
+}
+
+static inline void check_for_slice(AVSContext *h) {
+    GetBitContext *gb = &h->s.gb;
+    int align;
+    align = (-get_bits_count(gb)) & 7;
+    if((show_bits_long(gb,24+align) & 0xFFFFFF) == 0x000001) {
+        get_bits_long(gb,24+align);
+        h->stc = get_bits(gb,8);
+        decode_slice_header(h,gb);
+    }
+}
+
+/*****************************************************************************
+ *
+ * frame level
+ *
+ ****************************************************************************/
+
+static void init_pic(AVSContext *h) {
+    int i;
+
+    /* clear some predictors */
+    for(i=0;i<=20;i+=4)
+        veccpy(&h->mv[i],(vector_t *)&un_mv);
+    veccpy(&h->mv[MV_BWD_X0], (vector_t *)&dir_mv);
+    set_mvs(&h->mv[MV_BWD_X0], BLK_16X16);
+    veccpy(&h->mv[MV_FWD_X0], (vector_t *)&dir_mv);
+    set_mvs(&h->mv[MV_FWD_X0], BLK_16X16);
+    h->pred_mode_Y[3] = h->pred_mode_Y[6] = NOT_AVAIL;
+    h->cy = h->picture.data[0];
+    h->cu = h->picture.data[1];
+    h->cv = h->picture.data[2];
+    h->l_stride = h->picture.linesize[0];
+    h->c_stride = h->picture.linesize[1];
+    h->luma_scan[2] = 8*h->l_stride;
+    h->luma_scan[3] = 8*h->l_stride+8;
+    h->mbx = h->mby = 0;
+    h->flags = 0;
+}
+
+static int decode_pic(AVSContext *h) {
+    MpegEncContext *s = &h->s;
+    int i,skip_count;
+    enum mb_t mb_type;
+
+    if (!s->context_initialized) {
+        if (MPV_common_init(s) < 0)
+            return -1;
+    }
+    get_bits(&s->gb,16);//bbv_dwlay
+    if(h->stc == PIC_PB_START_CODE) {
+        h->pic_type = get_bits(&s->gb,2) + FF_I_TYPE;
+        /* make sure we have the reference frames we need */
+        if(!h->DPB[0].data[0] ||
+          (!h->DPB[1].data[0] && h->pic_type == FF_B_TYPE))
+            return -1;
+    } else {
+        h->pic_type = FF_I_TYPE;
+        if(get_bits1(&s->gb))
+            get_bits(&s->gb,16);//time_code
+    }
+    /* release last B frame */
+    if(h->picture.data[0])
+        s->avctx->release_buffer(s->avctx, (AVFrame *)&h->picture);
+
+    s->avctx->get_buffer(s->avctx, (AVFrame *)&h->picture);
+    init_pic(h);
+    h->picture.poc = get_bits(&s->gb,8)*2;
+
+    /* get temporal distances and MV scaling factors */
+    if(h->pic_type != FF_B_TYPE) {
+        h->dist[0] = (h->picture.poc - h->DPB[0].poc  + 512) % 512;
+    } else {
+        h->dist[0] = (h->DPB[0].poc  - h->picture.poc + 512) % 512;
+    }
+    h->dist[1] = (h->picture.poc - h->DPB[1].poc  + 512) % 512;
+    h->scale_den[0] = h->dist[0] ? 512/h->dist[0] : 0;
+    h->scale_den[1] = h->dist[1] ? 512/h->dist[1] : 0;
+    if(h->pic_type == FF_B_TYPE) {
+        h->sym_factor = h->dist[0]*h->scale_den[1];
+    } else {
+        h->direct_den[0] = h->dist[0] ? 16384/h->dist[0] : 0;
+        h->direct_den[1] = h->dist[1] ? 16384/h->dist[1] : 0;
+    }
+
+    if(s->low_delay)
+        get_ue_golomb(&s->gb); //bbv_check_times
+    h->progressive             = get_bits1(&s->gb);
+    if(h->progressive)
+        h->pic_structure = 1;
+    else if(!(h->pic_structure = get_bits1(&s->gb) && (h->stc == PIC_PB_START_CODE)) )
+        get_bits1(&s->gb);     //advanced_pred_mode_disable
+    skip_bits1(&s->gb);        //top_field_first
+    skip_bits1(&s->gb);        //repeat_first_field
+    h->qp_fixed                = get_bits1(&s->gb);
+    h->qp                      = get_bits(&s->gb,6);
+    if(h->pic_type == FF_I_TYPE) {
+        if(!h->progressive && !h->pic_structure)
+            skip_bits1(&s->gb);//what is this?
+        skip_bits(&s->gb,4);   //reserved bits
+    } else {
+        if(!(h->pic_type == FF_B_TYPE && h->pic_structure == 1))
+            h->ref_flag        = get_bits1(&s->gb);
+        skip_bits(&s->gb,4);   //reserved bits
+        h->skip_mode_flag      = get_bits1(&s->gb);
+    }
+    h->loop_filter_disable     = get_bits1(&s->gb);
+    if(!h->loop_filter_disable && get_bits1(&s->gb)) {
+        h->alpha_offset        = get_se_golomb(&s->gb);
+        h->beta_offset         = get_se_golomb(&s->gb);
+    } else {
+        h->alpha_offset = h->beta_offset  = 0;
+    }
+    check_for_slice(h);
+    if(h->pic_type == FF_I_TYPE) {
+        do {
+            init_mb(h);
+            decode_mb_i(h,1);
+        } while(next_mb(h));
+    } else if(h->pic_type == FF_P_TYPE) {
+        do {
+            if(h->skip_mode_flag) {
+                skip_count = get_ue_golomb(&s->gb);
+                for(i=0;i<skip_count;i++) {
+                    init_mb(h);
+                    mb_skip_p(h);
+                    if(!next_mb(h))
+                        goto done;
+                }
+                mb_type = get_ue_golomb(&s->gb) + P_16X16;
+            } else {
+                mb_type = get_ue_golomb(&s->gb) + P_SKIP;
+            }
+            init_mb(h);
+            if(mb_type > P_8X8) {
+                h->cbp = cbp_tab[mb_type - P_8X8 - 1][0];
+                decode_mb_i(h,0);
+            } else {
+                decode_mb_p(h,mb_type);
+            }
+        } while(next_mb(h));
+    } else { //FF_B_TYPE
+        do {
+            if(h->skip_mode_flag) {
+                skip_count = get_ue_golomb(&s->gb);
+                for(i=0;i<skip_count;i++) {
+                    init_mb(h);
+                    mb_skip_b(h);
+                    inter_pred(h);
+                    filter_mb(h,B_SKIP);
+                    if(!next_mb(h))
+                        goto done;
+                }
+                mb_type = get_ue_golomb(&s->gb) + B_DIRECT;
+            } else {
+                mb_type = get_ue_golomb(&s->gb) + B_SKIP;
+            }
+            init_mb(h);
+            if(mb_type > B_8X8) {
+                h->cbp = cbp_tab[mb_type - B_8X8 - 1][0];
+                decode_mb_i(h,0);
+            } else {
+                decode_mb_b(h,mb_type);
+            }
+        } while(next_mb(h));
+    }
+ done:
+    if(h->pic_type != FF_B_TYPE) {
+        if(h->DPB[1].data[0])
+            s->avctx->release_buffer(s->avctx, (AVFrame *)&h->DPB[1]);
+        memcpy(&h->DPB[1], &h->DPB[0], sizeof(Picture));
+        memcpy(&h->DPB[0], &h->picture, sizeof(Picture));
+        memset(&h->picture,0,sizeof(Picture));
+    }
+    return 0;
+}
+
+/*****************************************************************************
+ *
+ * headers and interface
+ *
+ ****************************************************************************/
+
+static void init_top_lines(AVSContext *h) {
+    /* alloc top line of predictors */
+    h->top_qp       = av_malloc( h->mb_width);
+    h->top_mv[0]    = av_malloc((h->mb_width*2+1)*sizeof(vector_t));
+    h->top_mv[1]    = av_malloc((h->mb_width*2+1)*sizeof(vector_t));
+    h->top_pred_Y   = av_malloc( h->mb_width*2*sizeof(int));
+    h->top_border_y = av_malloc((h->mb_width+1)*16);
+    h->top_border_u = av_malloc((h->mb_width+1)*8);
+    h->top_border_v = av_malloc((h->mb_width+1)*8);
+
+    /* alloc space for co-located MVs and types */
+    h->col_mv       = av_malloc( h->mb_width*h->mb_height*4*sizeof(vector_t));
+    h->col_type_base = av_malloc(h->mb_width*h->mb_height);
+}
+
+static int decode_seq_header(AVSContext *h) {
+    MpegEncContext *s = &h->s;
+    extern const AVRational frame_rate_tab[];
+    int frame_rate_code;
+
+    h->profile =         get_bits(&s->gb,8);
+    h->level =           get_bits(&s->gb,8);
+    skip_bits1(&s->gb); //progressive sequence
+    s->width =           get_bits(&s->gb,14);
+    s->height =          get_bits(&s->gb,14);
+    skip_bits(&s->gb,2); //chroma format
+    skip_bits(&s->gb,3); //sample_precision
+    h->aspect_ratio =    get_bits(&s->gb,4);
+    frame_rate_code =    get_bits(&s->gb,4);
+    skip_bits(&s->gb,18);//bit_rate_lower
+    skip_bits1(&s->gb);  //marker_bit
+    skip_bits(&s->gb,12);//bit_rate_upper
+    s->low_delay =       get_bits1(&s->gb);
+    h->mb_width  = (s->width  + 15) >> 4;
+    h->mb_height = (s->height + 15) >> 4;
+    h->s.avctx->time_base.den = frame_rate_tab[frame_rate_code].num;
+    h->s.avctx->time_base.num = frame_rate_tab[frame_rate_code].den;
+    h->s.avctx->width  = s->width;
+    h->s.avctx->height = s->height;
+    if(!h->top_qp)
+        init_top_lines(h);
+    return 0;
+}
+
+/**
+ * finds the end of the current frame in the bitstream.
+ * @return the position of the first byte of the next frame, or -1
+ */
+int ff_cavs_find_frame_end(ParseContext *pc, const uint8_t *buf, int buf_size) {
+    int pic_found, i;
+    uint32_t state;
+
+    pic_found= pc->frame_start_found;
+    state= pc->state;
+
+    i=0;
+    if(!pic_found){
+        for(i=0; i<buf_size; i++){
+            state= (state<<8) | buf[i];
+            if(state == PIC_I_START_CODE || state == PIC_PB_START_CODE){
+                i++;
+                pic_found=1;
+                break;
+            }
+        }
+    }
+
+    if(pic_found){
+        /* EOF considered as end of frame */
+        if (buf_size == 0)
+            return 0;
+        for(; i<buf_size; i++){
+            state= (state<<8) | buf[i];
+            if((state&0xFFFFFF00) == 0x100){
+                if(state < SLICE_MIN_START_CODE || state > SLICE_MAX_START_CODE){
+                    pc->frame_start_found=0;
+                    pc->state=-1;
+                    return i-3;
+                }
+            }
+        }
+    }
+    pc->frame_start_found= pic_found;
+    pc->state= state;
+    return END_NOT_FOUND;
+}
+
+void ff_cavs_flush(AVCodecContext * avctx) {
+    AVSContext *h = (AVSContext *)avctx->priv_data;
+    h->got_keyframe = 0;
+}
+
+static int cavs_decode_frame(AVCodecContext * avctx,void *data, int *data_size,
+                             uint8_t * buf, int buf_size) {
+    AVSContext *h = avctx->priv_data;
+    MpegEncContext *s = &h->s;
+    int input_size;
+    const uint8_t *buf_end;
+    const uint8_t *buf_ptr;
+    AVFrame *picture = data;
+    uint32_t stc;
+
+    s->avctx = avctx;
+
+    if (buf_size == 0) {
+        if(!s->low_delay && h->DPB[0].data[0]) {
+            *data_size = sizeof(AVPicture);
+            *picture = *(AVFrame *) &h->DPB[0];
+        }
+        return 0;
+    }
+
+    buf_ptr = buf;
+    buf_end = buf + buf_size;
+    for(;;) {
+        buf_ptr = ff_find_start_code(buf_ptr,buf_end, &stc);
+        if(stc & 0xFFFFFE00)
+            return FFMAX(0, buf_ptr - buf - s->parse_context.last_index);
+        input_size = (buf_end - buf_ptr)*8;
+        switch(stc) {
+        case SEQ_START_CODE:
+            init_get_bits(&s->gb, buf_ptr, input_size);
+            decode_seq_header(h);
+            break;
+        case PIC_I_START_CODE:
+            if(!h->got_keyframe) {
+                if(h->DPB[0].data[0])
+                    avctx->release_buffer(avctx, (AVFrame *)&h->DPB[0]);
+                if(h->DPB[1].data[0])
+                    avctx->release_buffer(avctx, (AVFrame *)&h->DPB[1]);
+                h->got_keyframe = 1;
+            }
+        case PIC_PB_START_CODE:
+            *data_size = 0;
+            if(!h->got_keyframe)
+                break;
+            init_get_bits(&s->gb, buf_ptr, input_size);
+            h->stc = stc;
+            if(decode_pic(h))
+                break;
+            *data_size = sizeof(AVPicture);
+            if(h->pic_type != FF_B_TYPE) {
+                if(h->DPB[1].data[0]) {
+                    *picture = *(AVFrame *) &h->DPB[1];
+                } else {
+                    *data_size = 0;
+                }
+            } else
+                *picture = *(AVFrame *) &h->picture;
+            break;
+        case EXT_START_CODE:
+            //mpeg_decode_extension(avctx,buf_ptr, input_size);
+            break;
+        case USER_START_CODE:
+            //mpeg_decode_user_data(avctx,buf_ptr, input_size);
+            break;
+        default:
+            if (stc >= SLICE_MIN_START_CODE &&
+                stc <= SLICE_MAX_START_CODE) {
+                init_get_bits(&s->gb, buf_ptr, input_size);
+                decode_slice_header(h, &s->gb);
+            }
+            break;
+        }
+    }
+}
+
+static int cavs_decode_init(AVCodecContext * avctx) {
+    AVSContext *h = (AVSContext *)avctx->priv_data;
+    MpegEncContext * const s = &h->s;
+
+    MPV_decode_defaults(s);
+    s->avctx = avctx;
+
+    avctx->pix_fmt= PIX_FMT_YUV420P;
+
+    h->luma_scan[0] = 0;
+    h->luma_scan[1] = 8;
+    h->intra_pred_l[      INTRA_L_VERT] = intra_pred_vert;
+    h->intra_pred_l[     INTRA_L_HORIZ] = intra_pred_horiz;
+    h->intra_pred_l[        INTRA_L_LP] = intra_pred_lp;
+    h->intra_pred_l[ INTRA_L_DOWN_LEFT] = intra_pred_down_left;
+    h->intra_pred_l[INTRA_L_DOWN_RIGHT] = intra_pred_down_right;
+    h->intra_pred_l[   INTRA_L_LP_LEFT] = intra_pred_lp_left;
+    h->intra_pred_l[    INTRA_L_LP_TOP] = intra_pred_lp_top;
+    h->intra_pred_l[    INTRA_L_DC_128] = intra_pred_dc_128;
+    h->intra_pred_c[        INTRA_C_LP] = intra_pred_lp;
+    h->intra_pred_c[     INTRA_C_HORIZ] = intra_pred_horiz;
+    h->intra_pred_c[      INTRA_C_VERT] = intra_pred_vert;
+    h->intra_pred_c[     INTRA_C_PLANE] = intra_pred_plane;
+    h->intra_pred_c[   INTRA_C_LP_LEFT] = intra_pred_lp_left;
+    h->intra_pred_c[    INTRA_C_LP_TOP] = intra_pred_lp_top;
+    h->intra_pred_c[    INTRA_C_DC_128] = intra_pred_dc_128;
+    veccpy(&h->mv[ 7], (vector_t *)&un_mv);
+    veccpy(&h->mv[19], (vector_t *)&un_mv);
+    return 0;
+}
+
+static int cavs_decode_end(AVCodecContext * avctx) {
+    AVSContext *h = (AVSContext *)avctx->priv_data;
+
+    av_free(h->top_qp);
+    av_free(h->top_mv[0]);
+    av_free(h->top_mv[1]);
+    av_free(h->top_pred_Y);
+    av_free(h->top_border_y);
+    av_free(h->top_border_u);
+    av_free(h->top_border_v);
+    av_free(h->col_mv);
+    av_free(h->col_type_base);
+    return 0;
+}
+
+AVCodec cavs_decoder = {
+    "cavs",
+    CODEC_TYPE_VIDEO,
+    CODEC_ID_CAVS,
+    sizeof(AVSContext),
+    cavs_decode_init,
+    NULL,
+    cavs_decode_end,
+    cavs_decode_frame,
+    CODEC_CAP_TRUNCATED | CODEC_CAP_DELAY, //FIXME is this correct ?
+    .flush= ff_cavs_flush,
+};

libavcodec/cavsdata.h

+/*
+ * Chinese AVS video (AVS1-P2, JiZhun profile) decoder.
+ * Copyright (c) 2006  Stefan Gehrer <stefan.gehrer@gmx.de>
+ *
+ * This library is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2 of the License, or (at your option) any later version.
+ *
+ * This library is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with this library; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+
+#define SLICE_MIN_START_CODE    0x00000101
+#define SLICE_MAX_START_CODE    0x000001af
+#define EXT_START_CODE          0x000001b5
+#define USER_START_CODE         0x000001b2
+#define SEQ_START_CODE          0x000001b0
+#define PIC_I_START_CODE        0x000001b3
+#define PIC_PB_START_CODE       0x000001b6
+
+#define A_AVAIL                          1
+#define B_AVAIL                          2
+#define C_AVAIL                          4
+#define D_AVAIL                          8
+#define NOT_AVAIL                       -1
+#define REF_INTRA                       -2
+#define REF_DIR                         -3
+
+#define ESCAPE_CODE                     59
+
+#define FWD0                          0x01
+#define FWD1                          0x02
+#define BWD0                          0x04
+#define BWD1                          0x08
+#define SYM0                          0x10
+#define SYM1                          0x20
+
+#define MV_BWD_OFFS                     12
+#define MV_STRIDE                        4
+
+enum mb_t {
+  I_8X8 = 0,
+  P_SKIP,
+  P_16X16,
+  P_16X8,
+  P_8X16,
+  P_8X8,
+  B_SKIP,
+  B_DIRECT,
+  B_FWD_16X16,
+  B_BWD_16X16,
+  B_SYM_16X16,
+  B_8X8 = 29
+};
+
+enum sub_mb_t {
+  B_SUB_DIRECT,
+  B_SUB_FWD,
+  B_SUB_BWD,
+  B_SUB_SYM
+};
+
+enum intra_luma_t {
+  INTRA_L_VERT,
+  INTRA_L_HORIZ,
+  INTRA_L_LP,
+  INTRA_L_DOWN_LEFT,
+  INTRA_L_DOWN_RIGHT,
+  INTRA_L_LP_LEFT,
+  INTRA_L_LP_TOP,
+  INTRA_L_DC_128
+};
+
+enum intra_chroma_t {
+  INTRA_C_LP,
+  INTRA_C_HORIZ,
+  INTRA_C_VERT,
+  INTRA_C_PLANE,
+  INTRA_C_LP_LEFT,
+  INTRA_C_LP_TOP,
+  INTRA_C_DC_128,
+};
+
+enum mv_pred_t {
+  MV_PRED_MEDIAN,
+  MV_PRED_LEFT,
+  MV_PRED_TOP,
+  MV_PRED_TOPRIGHT,
+  MV_PRED_PSKIP,
+  MV_PRED_BSKIP
+};
+
+enum block_t {
+  BLK_16X16,
+  BLK_16X8,
+  BLK_8X16,
+  BLK_8X8
+};
+
+enum mv_loc_t {
+  MV_FWD_D3 = 0,
+  MV_FWD_B2,
+  MV_FWD_B3,
+  MV_FWD_C2,
+  MV_FWD_A1,
+  MV_FWD_X0,
+  MV_FWD_X1,
+  MV_FWD_A3 = 8,
+  MV_FWD_X2,
+  MV_FWD_X3,
+  MV_BWD_D3 = MV_BWD_OFFS,
+  MV_BWD_B2,
+  MV_BWD_B3,
+  MV_BWD_C2,
+  MV_BWD_A1,
+  MV_BWD_X0,
+  MV_BWD_X1,
+  MV_BWD_A3 = MV_BWD_OFFS+8,
+  MV_BWD_X2,
+  MV_BWD_X3
+};
+
+static const uint8_t b_partition_flags[14] = {
+  0,0,0,0,0,
+  FWD0|FWD1,
+  BWD0|BWD1,
+  FWD0|BWD1,
+  BWD0|FWD1,
+  FWD0|SYM1,
+  BWD0|SYM1,
+  SYM0|FWD1,
+  SYM0|BWD1,
+  SYM0|SYM1
+};
+
+static const uint8_t scan3x3[4] = {4,5,7,8};
+
+static const uint8_t mv_scan[4] = {
+    MV_FWD_X0,MV_FWD_X1,
+    MV_FWD_X2,MV_FWD_X3
+};
+
+static const uint8_t cbp_tab[64][2] = {
+  {63, 0},{15,15},{31,63},{47,31},{ 0,16},{14,32},{13,47},{11,13},
+  { 7,14},{ 5,11},{10,12},{ 8, 5},{12,10},{61, 7},{ 4,48},{55, 3},
+  { 1, 2},{ 2, 8},{59, 4},{ 3, 1},{62,61},{ 9,55},{ 6,59},{29,62},
+  {45,29},{51,27},{23,23},{39,19},{27,30},{46,28},{53, 9},{30, 6},
+  {43,60},{37,21},{60,44},{16,26},{21,51},{28,35},{19,18},{35,20},
+  {42,24},{26,53},{44,17},{32,37},{58,39},{24,45},{20,58},{17,43},
+  {18,42},{48,46},{22,36},{33,33},{25,34},{49,40},{40,52},{36,49},
+  {34,50},{50,56},{52,25},{54,22},{41,54},{56,57},{38,41},{57,38}
+};
+
+static const uint8_t chroma_qp[64] = {
+  0,  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,
+  32,33,34,35,36,37,38,39,40,41,42,42,43,43,44,44,
+  45,45,46,46,47,47,48,48,48,49,49,49,50,50,50,51
+};
+
+static const uint8_t dequant_shift[64] = {
+  14,14,14,14,14,14,14,14,
+  13,13,13,13,13,13,13,13,
+  13,12,12,12,12,12,12,12,
+  11,11,11,11,11,11,11,11,
+  11,10,10,10,10,10,10,10,
+  10, 9, 9, 9, 9, 9, 9, 9,
+  9, 8, 8, 8, 8, 8, 8, 8,
+  7, 7, 7, 7, 7, 7, 7, 7
+};
+
+static const uint16_t dequant_mul[64] = {
+  32768,36061,38968,42495,46341,50535,55437,60424,
+  32932,35734,38968,42495,46177,50535,55109,59933,
+  65535,35734,38968,42577,46341,50617,55027,60097,
+  32809,35734,38968,42454,46382,50576,55109,60056,
+  65535,35734,38968,42495,46320,50515,55109,60076,
+  65535,35744,38968,42495,46341,50535,55099,60087,
+  65535,35734,38973,42500,46341,50535,55109,60097,
+  32771,35734,38965,42497,46341,50535,55109,60099
+};
+
+typedef struct {
+    int16_t x;
+    int16_t y;
+    int16_t dist;
+    int16_t ref;
+} vector_t;
+
+// marks block as unavailable, i.e. out of picture
+//  or not yet decoded
+static const vector_t un_mv    = {0,0,1,NOT_AVAIL};
+
+//marks block as "no prediction from this direction"
+// e.g. forward motion vector in BWD partition
+static const vector_t dir_mv   = {0,0,1,REF_DIR};
+
+//marks block as using intra prediction
+static const vector_t intra_mv = {0,0,1,REF_INTRA};
+
+typedef struct residual_vlc_t {
+  int8_t rltab[59][3];
+  int8_t level_add[26];
+  int8_t golomb_order;
+  int inc_limit;
+  int8_t max_run;
+} residual_vlc_t;
+
+static const residual_vlc_t intra_2dvlc[7] = {
+  {
+    { //level / run / table_inc
+      {  1, 0, 1},{ -1, 0, 1},{  1, 1, 1},{ -1, 1, 1},{  1, 2, 1},{ -1, 2, 1},
+      {  1, 3, 1},{ -1, 3, 1},{  1, 4, 1},{ -1, 4, 1},{  1, 5, 1},{ -1, 5, 1},
+      {  1, 6, 1},{ -1, 6, 1},{  1, 7, 1},{ -1, 7, 1},{  1, 8, 1},{ -1, 8, 1},
+      {  1, 9, 1},{ -1, 9, 1},{  1,10, 1},{ -1,10, 1},{  2, 0, 2},{ -2, 0, 2},
+      {  1,11, 1},{ -1,11, 1},{  1,12, 1},{ -1,12, 1},{  1,13, 1},{ -1,13, 1},
+      {  1,14, 1},{ -1,14, 1},{  2, 1, 2},{ -2, 1, 2},{  1,15, 1},{ -1,15, 1},
+      {  1,16, 1},{ -1,16, 1},{  3, 0, 3},{ -3, 0, 3},{  1,17, 1},{ -1,17, 1},
+      {  1,18, 1},{ -1,18, 1},{  2, 2, 2},{ -2, 2, 2},{  1,19, 1},{ -1,19, 1},
+      {  1,20, 1},{ -1,20, 1},{  2, 3, 2},{ -2, 3, 2},{  1,21, 1},{ -1,21, 1},
+      {  2, 4, 2},{ -2, 4, 2},{  1,22, 1},{ -1,22, 1},{  0, 0,-1}
+    },
+    //level_add
+    { 4, 3, 3, 3, 3, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
+      2, 2, 2, 2, 2, 2, 2,-1,-1,-1},
+    2, //golomb_order
+    0, //inc_limit
+    22, //max_run
+  },{
+    { //level / run
+      {  1, 0, 0},{ -1, 0, 0},{  1, 1, 0},{ -1, 1, 0},{  2, 0, 1},{ -2, 0, 1},
+      {  1, 2, 0},{ -1, 2, 0},{  0, 0, 0},{  1, 3, 0},{ -1, 3, 0},{  1, 4, 0},
+      { -1, 4, 0},{  1, 5, 0},{ -1, 5, 0},{  3, 0, 2},{ -3, 0, 2},{  2, 1, 1},
+      { -2, 1, 1},{  1, 6, 0},{ -1, 6, 0},{  1, 7, 0},{ -1, 7, 0},{  1, 8, 0},
+      { -1, 8, 0},{  2, 2, 1},{ -2, 2, 1},{  4, 0, 2},{ -4, 0, 2},{  1, 9, 0},
+      { -1, 9, 0},{  1,10, 0},{ -1,10, 0},{  2, 3, 1},{ -2, 3, 1},{  3, 1, 2},
+      { -3, 1, 2},{  1,11, 0},{ -1,11, 0},{  2, 4, 1},{ -2, 4, 1},{  5, 0, 3},
+      { -5, 0, 3},{  1,12, 0},{ -1,12, 0},{  2, 5, 1},{ -2, 5, 1},{  1,13, 0},
+      { -1,13, 0},{  2, 6, 1},{ -2, 6, 1},{  2, 7, 1},{ -2, 7, 1},{  3, 2, 2},
+      { -3, 2, 2},{  6, 0, 3},{ -6, 0, 3},{  1,14, 0},{ -1,14, 0}
+    },
+    //level_add
+    { 7, 4, 4, 3, 3, 3, 3, 3, 2, 2, 2, 2, 2, 2, 2,-1,
+      -1,-1,-1,-1,-1,-1,-1,-1,-1,-1},
+    2, //golomb_order
+    1, //inc_limit
+    14, //max_run
+  },{
+    { //level / run
+      {  1, 0, 0},{ -1, 0, 0},{  2, 0, 0},{ -2, 0, 0},{  1, 1, 0},{ -1, 1, 0},
+      {  3, 0, 1},{ -3, 0, 1},{  0, 0, 0},{  1, 2, 0},{ -1, 2, 0},{  2, 1, 0},
+      { -2, 1, 0},{  4, 0, 1},{ -4, 0, 1},{  1, 3, 0},{ -1, 3, 0},{  5, 0, 2},
+      { -5, 0, 2},{  1, 4, 0},{ -1, 4, 0},{  3, 1, 1},{ -3, 1, 1},{  2, 2, 0},
+      { -2, 2, 0},{  1, 5, 0},{ -1, 5, 0},{  6, 0, 2},{ -6, 0, 2},{  2, 3, 0},
+      { -2, 3, 0},{  1, 6, 0},{ -1, 6, 0},{  4, 1, 1},{ -4, 1, 1},{  7, 0, 2},
+      { -7, 0, 2},{  3, 2, 1},{ -3, 2, 1},{  2, 4, 0},{ -2, 4, 0},{  1, 7, 0},
+      { -1, 7, 0},{  2, 5, 0},{ -2, 5, 0},{  8, 0, 3},{ -8, 0, 3},{  1, 8, 0},
+      { -1, 8, 0},{  5, 1, 2},{ -5, 1, 2},{  3, 3, 1},{ -3, 3, 1},{  2, 6, 0},
+      { -2, 6, 0},{  9, 0, 3},{ -9, 0, 3},{  1, 9, 0},{ -1, 9, 0}
+    },
+    //level_add
+    {10, 6, 4, 4, 3, 3, 3, 2, 2, 2,-1,-1,-1,-1,-1,-1,
+     -1,-1,-1,-1,-1,-1,-1,-1,-1,-1},
+    2, //golomb_order
+    2, //inc_limit
+    9, //max_run
+  },{
+    { //level / run
+      {  1, 0, 0},{ -1, 0, 0},{  2, 0, 0},{ -2, 0, 0},{  3, 0, 0},{ -3, 0, 0},
+      {  1, 1, 0},{ -1, 1, 0},{  0, 0, 0},{  4, 0, 0},{ -4, 0, 0},{  5, 0, 1},
+      { -5, 0, 1},{  2, 1, 0},{ -2, 1, 0},{  1, 2, 0},{ -1, 2, 0},{  6, 0, 1},
+      { -6, 0, 1},{  3, 1, 0},{ -3, 1, 0},{  7, 0, 1},{ -7, 0, 1},{  1, 3, 0},
+      { -1, 3, 0},{  8, 0, 2},{ -8, 0, 2},{  2, 2, 0},{ -2, 2, 0},{  4, 1, 0},
+      { -4, 1, 0},{  1, 4, 0},{ -1, 4, 0},{  9, 0, 2},{ -9, 0, 2},{  5, 1, 1},
+      { -5, 1, 1},{  2, 3, 0},{ -2, 3, 0},{ 10, 0, 2},{-10, 0, 2},{  3, 2, 0},
+      { -3, 2, 0},{  1, 5, 0},{ -1, 5, 0},{ 11, 0, 3},{-11, 0, 3},{  6, 1, 1},
+      { -6, 1, 1},{  1, 6, 0},{ -1, 6, 0},{  2, 4, 0},{ -2, 4, 0},{  3, 3, 0},
+      { -3, 3, 0},{ 12, 0, 3},{-12, 0, 3},{  4, 2, 0},{ -4, 2, 0}
+    },
+    //level_add
+    {13, 7, 5, 4, 3, 2, 2,-1,-1,-1 -1,-1,-1,-1,-1,-1,
+     -1,-1,-1,-1,-1,-1,-1,-1,-1,-1},
+    2, //golomb_order
+    4, //inc_limit
+    6, //max_run
+  },{
+    { //level / run
+      {  1, 0, 0},{ -1, 0, 0},{  2, 0, 0},{ -2, 0, 0},{  3, 0, 0},{ -3, 0, 0},
+      {  0, 0, 0},{  4, 0, 0},{ -4, 0, 0},{  5, 0, 0},{ -5, 0, 0},{  6, 0, 0},
+      { -6, 0, 0},{  1, 1, 0},{ -1, 1, 0},{  7, 0, 0},{ -7, 0, 0},{  8, 0, 1},
+      { -8, 0, 1},{  2, 1, 0},{ -2, 1, 0},{  9, 0, 1},{ -9, 0, 1},{ 10, 0, 1},
+      {-10, 0, 1},{  1, 2, 0},{ -1, 2, 0},{  3, 1, 0},{ -3, 1, 0},{ 11, 0, 2},
+      {-11, 0, 2},{  4, 1, 0},{ -4, 1, 0},{ 12, 0, 2},{-12, 0, 2},{ 13, 0, 2},
+      {-13, 0, 2},{  5, 1, 0},{ -5, 1, 0},{  1, 3, 0},{ -1, 3, 0},{  2, 2, 0},
+      { -2, 2, 0},{ 14, 0, 2},{-14, 0, 2},{  6, 1, 0},{ -6, 1, 0},{ 15, 0, 2},
+      {-15, 0, 2},{ 16, 0, 2},{-16, 0, 2},{  3, 2, 0},{ -3, 2, 0},{  1, 4, 0},
+      { -1, 4, 0},{  7, 1, 0},{ -7, 1, 0},{ 17, 0, 2},{-17, 0, 2},
+    },
+    //level_add
+    {18, 8, 4, 2, 2,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,
+     -1,-1,-1,-1,-1,-1,-1,-1,-1,-1},
+    2, //golomb_order
+    7, //inc_limit
+    4, //max_run
+  },{
+    { //level / run
+      {  0, 0, 0},{  1, 0, 0},{ -1, 0, 0},{  2, 0, 0},{ -2, 0, 0},{  3, 0, 0},
+      { -3, 0, 0},{  4, 0, 0},{ -4, 0, 0},{  5, 0, 0},{ -5, 0, 0},{  6, 0, 0},
+      { -6, 0, 0},{  7, 0, 0},{ -7, 0, 0},{  8, 0, 0},{ -8, 0, 0},{  9, 0, 0},
+      { -9, 0, 0},{ 10, 0, 0},{-10, 0, 0},{  1, 1, 0},{ -1, 1, 0},{ 11, 0, 1},
+      {-11, 0, 1},{ 12, 0, 1},{-12, 0, 1},{ 13, 0, 1},{-13, 0, 1},{  2, 1, 0},
+      { -2, 1, 0},{ 14, 0, 1},{-14, 0, 1},{ 15, 0, 1},{-15, 0, 1},{  3, 1, 0},
+      { -3, 1, 0},{ 16, 0, 1},{-16, 0, 1},{  1, 2, 0},{ -1, 2, 0},{ 17, 0, 1},
+      {-17, 0, 1},{  4, 1, 0},{ -4, 1, 0},{ 18, 0, 1},{-18, 0, 1},{  5, 1, 0},
+      { -5, 1, 0},{ 19, 0, 1},{-19, 0, 1},{ 20, 0, 1},{-20, 0, 1},{  6, 1, 0},
+      { -6, 1, 0},{ 21, 0, 1},{-21, 0, 1},{  2, 2, 0},{ -2, 2, 0},
+    },
+    //level_add
+    {22, 7, 3,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,
+     -1,-1,-1,-1,-1,-1,-1,-1,-1,-1},
+    2, //golomb_order
+    10, //inc_limit
+    2, //max_run
+  },{
+    { //level / run
+      {  0, 0, 0},{  1, 0, 0},{ -1, 0, 0},{  2, 0, 0},{ -2, 0, 0},{  3, 0, 0},
+      { -3, 0, 0},{  4, 0, 0},{ -4, 0, 0},{  5, 0, 0},{ -5, 0, 0},{  6, 0, 0},
+      { -6, 0, 0},{  7, 0, 0},{ -7, 0, 0},{  8, 0, 0},{ -8, 0, 0},{  9, 0, 0},
+      { -9, 0, 0},{ 10, 0, 0},{-10, 0, 0},{ 11, 0, 0},{-11, 0, 0},{ 12, 0, 0},
+      {-12, 0, 0},{ 13, 0, 0},{-13, 0, 0},{ 14, 0, 0},{-14, 0, 0},{ 15, 0, 0},
+      {-15, 0, 0},{ 16, 0, 0},{-16, 0, 0},{  1, 1, 0},{ -1, 1, 0},{ 17, 0, 0},
+      {-17, 0, 0},{ 18, 0, 0},{-18, 0, 0},{ 19, 0, 0},{-19, 0, 0},{ 20, 0, 0},
+      {-20, 0, 0},{ 21, 0, 0},{-21, 0, 0},{  2, 1, 0},{ -2, 1, 0},{ 22, 0, 0},
+      {-22, 0, 0},{ 23, 0, 0},{-23, 0, 0},{ 24, 0, 0},{-24, 0, 0},{ 25, 0, 0},
+      {-25, 0, 0},{  3, 1, 0},{ -3, 1, 0},{ 26, 0, 0},{-26, 0, 0}
+    },
+    //level_add
+    {27, 4,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,
+     -1,-1,-1,-1,-1,-1,-1,-1,-1,-1},
+    2, //golomb_order
+    INT_MAX, //inc_limit
+    1, //max_run
+  }
+};
+
+static const residual_vlc_t inter_2dvlc[7] = {
+  {
+    { //level / run
+      {  1, 0, 1},{ -1, 0, 1},{  1, 1, 1},{ -1, 1, 1},{  1, 2, 1},{ -1, 2, 1},
+      {  1, 3, 1},{ -1, 3, 1},{  1, 4, 1},{ -1, 4, 1},{  1, 5, 1},{ -1, 5, 1},
+      {  1, 6, 1},{ -1, 6, 1},{  1, 7, 1},{ -1, 7, 1},{  1, 8, 1},{ -1, 8, 1},
+      {  1, 9, 1},{ -1, 9, 1},{  1,10, 1},{ -1,10, 1},{  1,11, 1},{ -1,11, 1},
+      {  1,12, 1},{ -1,12, 1},{  2, 0, 2},{ -2, 0, 2},{  1,13, 1},{ -1,13, 1},
+      {  1,14, 1},{ -1,14, 1},{  1,15, 1},{ -1,15, 1},{  1,16, 1},{ -1,16, 1},
+      {  1,17, 1},{ -1,17, 1},{  1,18, 1},{ -1,18, 1},{  3, 0, 3},{ -3, 0, 3},
+      {  1,19, 1},{ -1,19, 1},{  1,20, 1},{ -1,20, 1},{  2, 1, 2},{ -2, 1, 2},
+      {  1,21, 1},{ -1,21, 1},{  1,22, 1},{ -1,22, 1},{  1,23, 1},{ -1,23, 1},
+      {  1,24, 1},{ -1,24, 1},{  1,25, 1},{ -1,25, 1},{  0, 0,-1}
+    },
+    //level_add
+    { 4, 3, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
+      2, 2, 2, 2, 2, 2, 2, 2, 2, 2},
+    3, //golomb_order
+    0, //inc_limit
+    25 //max_run
+  },{
+    { //level / run
+      {  1, 0, 0},{ -1, 0, 0},{  0, 0, 0},{  1, 1, 0},{ -1, 1, 0},{  1, 2, 0},
+      { -1, 2, 0},{  1, 3, 0},{ -1, 3, 0},{  1, 4, 0},{ -1, 4, 0},{  1, 5, 0},
+      { -1, 5, 0},{  2, 0, 1},{ -2, 0, 1},{  1, 6, 0},{ -1, 6, 0},{  1, 7, 0},
+      { -1, 7, 0},{  1, 8, 0},{ -1, 8, 0},{  1, 9, 0},{ -1, 9, 0},{  2, 1, 1},
+      { -2, 1, 1},{  1,10, 0},{ -1,10, 0},{  1,11, 0},{ -1,11, 0},{  3, 0, 2},
+      { -3, 0, 2},{  1,12, 0},{ -1,12, 0},{  1,13, 0},{ -1,13, 0},{  2, 2, 1},
+      { -2, 2, 1},{  1,14, 0},{ -1,14, 0},{  2, 3, 1},{ -2, 3, 1},{  1,15, 0},
+      { -1,15, 0},{  2, 4, 1},{ -2, 4, 1},{  1,16, 0},{ -1,16, 0},{  4, 0, 3},
+      { -4, 0, 3},{  2, 5, 1},{ -2, 5, 1},{  1,17, 0},{ -1,17, 0},{  1,18, 0},
+      { -1,18, 0},{  2, 6, 1},{ -2, 6, 1},{  3, 1, 2},{ -3, 1, 2},
+    },
+    //level_add
+    { 5, 4, 3, 3, 3, 3, 3, 2, 2, 2, 2, 2, 2, 2, 2, 2,
+      2, 2, 2,-1,-1,-1,-1,-1,-1,-1},
+    2, //golomb_order
+    1, //inc_limit
+    18 //max_run
+  },{
+    { //level / run
+      {  1, 0, 0},{ -1, 0, 0},{  0, 0, 0},{  1, 1, 0},{ -1, 1, 0},{  2, 0, 0},
+      { -2, 0, 0},{  1, 2, 0},{ -1, 2, 0},{  1, 3, 0},{ -1, 3, 0},{  3, 0, 1},
+      { -3, 0, 1},{  2, 1, 0},{ -2, 1, 0},{  1, 4, 0},{ -1, 4, 0},{  1, 5, 0},
+      { -1, 5, 0},{  1, 6, 0},{ -1, 6, 0},{  2, 2, 0},{ -2, 2, 0},{  4, 0, 2},
+      { -4, 0, 2},{  1, 7, 0},{ -1, 7, 0},{  3, 1, 1},{ -3, 1, 1},{  2, 3, 0},
+      { -2, 3, 0},{  1, 8, 0},{ -1, 8, 0},{  1, 9, 0},{ -1, 9, 0},{  5, 0, 2},
+      { -5, 0, 2},{  2, 4, 0},{ -2, 4, 0},{  1,10, 0},{ -1,10, 0},{  2, 5, 0},
+      { -2, 5, 0},{  1,11, 0},{ -1,11, 0},{  3, 2, 1},{ -3, 2, 1},{  6, 0, 2},
+      { -6, 0, 2},{  4, 1, 2},{ -4, 1, 2},{  1,12, 0},{ -1,12, 0},{  2, 6, 0},
+      { -2, 6, 0},{  3, 3, 1},{ -3, 3, 1},{  1,13, 0},{ -1,13, 0},
+    },
+    //level_add
+    { 7, 5, 4, 4, 3, 3, 3, 2, 2, 2, 2, 2, 2, 2,-1,-1,
+      -1,-1,-1,-1,-1,-1,-1,-1,-1,-1},
+    2, //golomb_order
+    2, //inc_limit
+    13 //max_run
+  },{
+    { //level / run
+      {  1, 0, 0},{ -1, 0, 0},{  0, 0, 0},{  2, 0, 0},{ -2, 0, 0},{  1, 1, 0},
+      { -1, 1, 0},{  3, 0, 0},{ -3, 0, 0},{  1, 2, 0},{ -1, 2, 0},{  2, 1, 0},
+      { -2, 1, 0},{  4, 0, 1},{ -4, 0, 1},{  1, 3, 0},{ -1, 3, 0},{  5, 0, 1},
+      { -5, 0, 1},{  1, 4, 0},{ -1, 4, 0},{  3, 1, 0},{ -3, 1, 0},{  2, 2, 0},
+      { -2, 2, 0},{  1, 5, 0},{ -1, 5, 0},{  6, 0, 1},{ -6, 0, 1},{  2, 3, 0},
+      { -2, 3, 0},{  1, 6, 0},{ -1, 6, 0},{  4, 1, 1},{ -4, 1, 1},{  7, 0, 2},
+      { -7, 0, 2},{  3, 2, 0},{ -3, 2, 0},{  1, 7, 0},{ -1, 7, 0},{  2, 4, 0},
+      { -2, 4, 0},{  8, 0, 2},{ -8, 0, 2},{  1, 8, 0},{ -1, 8, 0},{  3, 3, 0},
+      { -3, 3, 0},{  2, 5, 0},{ -2, 5, 0},{  5, 1, 1},{ -5, 1, 1},{  1, 9, 0},
+      { -1, 9, 0},{  9, 0, 2},{ -9, 0, 2},{  4, 2, 1},{ -4, 2, 1},
+    },
+    //level_add
+    {10, 6, 5, 4, 3, 3, 2, 2, 2, 2,-1,-1,-1,-1,-1,-1,
+     -1,-1,-1,-1,-1,-1,-1,-1,-1,-1},
+    2, //golomb_order
+    3, //inc_limit
+    9 //max_run
+  },{
+    { //level / run
+      {  1, 0, 0},{ -1, 0, 0},{  0, 0, 0},{  2, 0, 0},{ -2, 0, 0},{  3, 0, 0},
+      { -3, 0, 0},{  1, 1, 0},{ -1, 1, 0},{  4, 0, 0},{ -4, 0, 0},{  5, 0, 0},
+      { -5, 0, 0},{  2, 1, 0},{ -2, 1, 0},{  1, 2, 0},{ -1, 2, 0},{  6, 0, 0},
+      { -6, 0, 0},{  3, 1, 0},{ -3, 1, 0},{  7, 0, 1},{ -7, 0, 1},{  1, 3, 0},
+      { -1, 3, 0},{  8, 0, 1},{ -8, 0, 1},{  2, 2, 0},{ -2, 2, 0},{  4, 1, 0},
+      { -4, 1, 0},{  1, 4, 0},{ -1, 4, 0},{  9, 0, 1},{ -9, 0, 1},{  5, 1, 0},
+      { -5, 1, 0},{  2, 3, 0},{ -2, 3, 0},{  1, 5, 0},{ -1, 5, 0},{ 10, 0, 2},
+      {-10, 0, 2},{  3, 2, 0},{ -3, 2, 0},{ 11, 0, 2},{-11, 0, 2},{  1, 6, 0},
+      { -1, 6, 0},{  6, 1, 0},{ -6, 1, 0},{  3, 3, 0},{ -3, 3, 0},{  2, 4, 0},
+      { -2, 4, 0},{ 12, 0, 2},{-12, 0, 2},{  4, 2, 0},{ -4, 2, 0},
+    },
+    //level_add
+    {13, 7, 5, 4, 3, 2, 2,-1,-1,-1,-1,-1,-1,-1,-1,-1,
+     -1,-1,-1,-1,-1,-1,-1,-1,-1,-1},
+    2, //golomb_order
+    6, //inc_limit
+    6 //max_run
+  },{
+    { //level / run
+      {  0, 0, 0},{  1, 0, 0},{ -1, 0, 0},{  2, 0, 0},{ -2, 0, 0},{  3, 0, 0},
+      { -3, 0, 0},{  4, 0, 0},{ -4, 0, 0},{  5, 0, 0},{ -5, 0, 0},{  1, 1, 0},
+      { -1, 1, 0},{  6, 0, 0},{ -6, 0, 0},{  7, 0, 0},{ -7, 0, 0},{  8, 0, 0},
+      { -8, 0, 0},{  2, 1, 0},{ -2, 1, 0},{  9, 0, 0},{ -9, 0, 0},{  1, 2, 0},
+      { -1, 2, 0},{ 10, 0, 1},{-10, 0, 1},{  3, 1, 0},{ -3, 1, 0},{ 11, 0, 1},
+      {-11, 0, 1},{  4, 1, 0},{ -4, 1, 0},{ 12, 0, 1},{-12, 0, 1},{  1, 3, 0},
+      { -1, 3, 0},{  2, 2, 0},{ -2, 2, 0},{ 13, 0, 1},{-13, 0, 1},{  5, 1, 0},
+      { -5, 1, 0},{ 14, 0, 1},{-14, 0, 1},{  6, 1, 0},{ -6, 1, 0},{  1, 4, 0},
+      { -1, 4, 0},{ 15, 0, 1},{-15, 0, 1},{  3, 2, 0},{ -3, 2, 0},{ 16, 0, 1},
+      {-16, 0, 1},{  2, 3, 0},{ -2, 3, 0},{  7, 1, 0},{ -7, 1, 0},
+    },
+    //level_add
+    {17, 8, 4, 3, 2,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,
+     -1,-1,-1,-1,-1,-1,-1,-1,-1,-1},
+    2, //golomb_order
+    9, //inc_limit
+    4 //max_run
+  },{
+    { //level / run
+      {  0, 0, 0},{  1, 0, 0},{ -1, 0, 0},{  2, 0, 0},{ -2, 0, 0},{  3, 0, 0},
+      { -3, 0, 0},{  4, 0, 0},{ -4, 0, 0},{  5, 0, 0},{ -5, 0, 0},{  6, 0, 0},
+      { -6, 0, 0},{  7, 0, 0},{ -7, 0, 0},{  1, 1, 0},{ -1, 1, 0},{  8, 0, 0},
+      { -8, 0, 0},{  9, 0, 0},{ -9, 0, 0},{ 10, 0, 0},{-10, 0, 0},{ 11, 0, 0},
+      {-11, 0, 0},{ 12, 0, 0},{-12, 0, 0},{  2, 1, 0},{ -2, 1, 0},{ 13, 0, 0},
+      {-13, 0, 0},{  1, 2, 0},{ -1, 2, 0},{ 14, 0, 0},{-14, 0, 0},{ 15, 0, 0},
+      {-15, 0, 0},{  3, 1, 0},{ -3, 1, 0},{ 16, 0, 0},{-16, 0, 0},{ 17, 0, 0},
+      {-17, 0, 0},{ 18, 0, 0},{-18, 0, 0},{  4, 1, 0},{ -4, 1, 0},{ 19, 0, 0},
+      {-19, 0, 0},{ 20, 0, 0},{-20, 0, 0},{  2, 2, 0},{ -2, 2, 0},{  1, 3, 0},
+      { -1, 3, 0},{  5, 1, 0},{ -5, 1, 0},{ 21, 0, 0},{-21, 0, 0},
+    },
+    //level_add
+    {22, 6, 3, 2,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,
+     -1,-1,-1,-1,-1,-1,-1,-1,-1,-1},
+    2, //golomb_order
+    INT_MAX, //inc_limit
+    3 //max_run
+  }
+};
+
+static const residual_vlc_t chroma_2dvlc[5] = {
+  {
+    { //level / run
+      {  1, 0, 1},{ -1, 0, 1},{  1, 1, 1},{ -1, 1, 1},{  1, 2, 1},{ -1, 2, 1},
+      {  1, 3, 1},{ -1, 3, 1},{  1, 4, 1},{ -1, 4, 1},{  1, 5, 1},{ -1, 5, 1},
+      {  1, 6, 1},{ -1, 6, 1},{  2, 0, 2},{ -2, 0, 2},{  1, 7, 1},{ -1, 7, 1},
+      {  1, 8, 1},{ -1, 8, 1},{  1, 9, 1},{ -1, 9, 1},{  1,10, 1},{ -1,10, 1},
+      {  1,11, 1},{ -1,11, 1},{  1,12, 1},{ -1,12, 1},{  1,13, 1},{ -1,13, 1},
+      {  1,14, 1},{ -1,14, 1},{  3, 0, 3},{ -3, 0, 3},{  1,15, 1},{ -1,15, 1},
+      {  1,16, 1},{ -1,16, 1},{  1,17, 1},{ -1,17, 1},{  1,18, 1},{ -1,18, 1},
+      {  1,19, 1},{ -1,19, 1},{  1,20, 1},{ -1,20, 1},{  1,21, 1},{ -1,21, 1},
+      {  2, 1, 2},{ -2, 1, 2},{  1,22, 1},{ -1,22, 1},{  1,23, 1},{ -1,23, 1},
+      {  1,24, 1},{ -1,24, 1},{  4, 0, 3},{ -4, 0, 3},{  0, 0,-1}
+    },
+    //level_add
+    { 5, 3, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
+      2, 2, 2, 2, 2, 2, 2, 2, 2,-1},
+    2, //golomb_order
+    0, //inc_limit
+    24, //max_run
+  },{
+    { //level / run
+      {  0, 0, 0},{  1, 0, 0},{ -1, 0, 0},{  1, 1, 0},{ -1, 1, 0},{  2, 0, 1},
+      { -2, 0, 1},{  1, 2, 0},{ -1, 2, 0},{  1, 3, 0},{ -1, 3, 0},{  1, 4, 0},
+      { -1, 4, 0},{  1, 5, 0},{ -1, 5, 0},{  3, 0, 2},{ -3, 0, 2},{  1, 6, 0},
+      { -1, 6, 0},{  1, 7, 0},{ -1, 7, 0},{  2, 1, 1},{ -2, 1, 1},{  1, 8, 0},
+      { -1, 8, 0},{  1, 9, 0},{ -1, 9, 0},{  1,10, 0},{ -1,10, 0},{  4, 0, 2},
+      { -4, 0, 2},{  1,11, 0},{ -1,11, 0},{  1,12, 0},{ -1,12, 0},{  1,13, 0},
+      { -1,13, 0},{  2, 2, 1},{ -2, 2, 1},{  1,14, 0},{ -1,14, 0},{  2, 3, 1},
+      { -2, 3, 1},{  5, 0, 3},{ -5, 0, 3},{  3, 1, 2},{ -3, 1, 2},{  1,15, 0},
+      { -1,15, 0},{  1,16, 0},{ -1,16, 0},{  1,17, 0},{ -1,17, 0},{  2, 4, 1},
+      { -2, 4, 1},{  1,18, 0},{ -1,18, 0},{  1,19, 0},{ -1,19, 0},
+    },
+    //level_add
+    { 6, 4, 3, 3, 3, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
+      2, 2, 2, 2,-1,-1,-1,-1,-1,-1},
+    0, //golomb_order
+    1, //inc_limit
+    19, //max_run
+  },{
+    { //level / run
+      {  1, 0, 0},{ -1, 0, 0},{  0, 0, 0},{  2, 0, 0},{ -2, 0, 0},{  1, 1, 0},
+      { -1, 1, 0},{  3, 0, 1},{ -3, 0, 1},{  1, 2, 0},{ -1, 2, 0},{  4, 0, 1},
+      { -4, 0, 1},{  2, 1, 0},{ -2, 1, 0},{  1, 3, 0},{ -1, 3, 0},{  5, 0, 2},
+      { -5, 0, 2},{  1, 4, 0},{ -1, 4, 0},{  3, 1, 1},{ -3, 1, 1},{  2, 2, 0},
+      { -2, 2, 0},{  1, 5, 0},{ -1, 5, 0},{  6, 0, 2},{ -6, 0, 2},{  1, 6, 0},
+      { -1, 6, 0},{  2, 3, 0},{ -2, 3, 0},{  7, 0, 2},{ -7, 0, 2},{  1, 7, 0},
+      { -1, 7, 0},{  4, 1, 1},{ -4, 1, 1},{  1, 8, 0},{ -1, 8, 0},{  3, 2, 1},
+      { -3, 2, 1},{  2, 4, 0},{ -2, 4, 0},{  2, 5, 0},{ -2, 5, 0},{  8, 0, 2},
+      { -8, 0, 2},{  1, 9, 0},{ -1, 9, 0},{  1,10, 0},{ -1,10, 0},{  9, 0, 2},
+      { -9, 0, 2},{  5, 1, 2},{ -5, 1, 2},{  3, 3, 1},{ -3, 3, 1},
+    },
+    //level_add
+    {10, 6, 4, 4, 3, 3, 2, 2, 2, 2, 2,-1,-1,-1,-1,-1,
+     -1,-1,-1,-1,-1,-1,-1,-1,-1,-1},
+    1, //golomb_order
+    2, //inc_limit
+    10, //max_run
+  },{
+    { //level / run
+      {  0, 0, 0},{  1, 0, 0},{ -1, 0, 0},{  2, 0, 0},{ -2, 0, 0},{  3, 0, 0},
+      { -3, 0, 0},{  4, 0, 0},{ -4, 0, 0},{  1, 1, 0},{ -1, 1, 0},{  5, 0, 1},
+      { -5, 0, 1},{  2, 1, 0},{ -2, 1, 0},{  6, 0, 1},{ -6, 0, 1},{  1, 2, 0},
+      { -1, 2, 0},{  7, 0, 1},{ -7, 0, 1},{  3, 1, 0},{ -3, 1, 0},{  8, 0, 1},
+      { -8, 0, 1},{  1, 3, 0},{ -1, 3, 0},{  2, 2, 0},{ -2, 2, 0},{  9, 0, 1},
+      { -9, 0, 1},{  4, 1, 0},{ -4, 1, 0},{  1, 4, 0},{ -1, 4, 0},{ 10, 0, 1},
+      {-10, 0, 1},{  3, 2, 0},{ -3, 2, 0},{  5, 1, 1},{ -5, 1, 1},{  2, 3, 0},
+      { -2, 3, 0},{ 11, 0, 1},{-11, 0, 1},{  1, 5, 0},{ -1, 5, 0},{ 12, 0, 1},
+      {-12, 0, 1},{  1, 6, 0},{ -1, 6, 0},{  6, 1, 1},{ -6, 1, 1},{ 13, 0, 1},
+      {-13, 0, 1},{  2, 4, 0},{ -2, 4, 0},{  1, 7, 0},{ -1, 7, 0},
+    },
+    //level_add
+    {14, 7, 4, 3, 3, 2, 2, 2,-1,-1,-1,-1,-1,-1,-1,-1,
+     -1,-1,-1,-1,-1,-1,-1,-1,-1,-1},
+    1, //golomb_order
+    4, //inc_limit
+    7, //max_run
+  },{
+    { //level / run
+      {  0, 0, 0},{  1, 0, 0},{ -1, 0, 0},{  2, 0, 0},{ -2, 0, 0},{  3, 0, 0},
+      { -3, 0, 0},{  4, 0, 0},{ -4, 0, 0},{  5, 0, 0},{ -5, 0, 0},{  6, 0, 0},
+      { -6, 0, 0},{  7, 0, 0},{ -7, 0, 0},{  8, 0, 0},{ -8, 0, 0},{  1, 1, 0},
+      { -1, 1, 0},{  9, 0, 0},{ -9, 0, 0},{ 10, 0, 0},{-10, 0, 0},{ 11, 0, 0},
+      {-11, 0, 0},{  2, 1, 0},{ -2, 1, 0},{ 12, 0, 0},{-12, 0, 0},{ 13, 0, 0},
+      {-13, 0, 0},{  3, 1, 0},{ -3, 1, 0},{ 14, 0, 0},{-14, 0, 0},{  1, 2, 0},
+      { -1, 2, 0},{ 15, 0, 0},{-15, 0, 0},{  4, 1, 0},{ -4, 1, 0},{ 16, 0, 0},
+      {-16, 0, 0},{ 17, 0, 0},{-17, 0, 0},{  5, 1, 0},{ -5, 1, 0},{  1, 3, 0},
+      { -1, 3, 0},{  2, 2, 0},{ -2, 2, 0},{ 18, 0, 0},{-18, 0, 0},{  6, 1, 0},
+      { -6, 1, 0},{ 19, 0, 0},{-19, 0, 0},{  1, 4, 0},{ -1, 4, 0},
+    },
+    //level_add
+    {20, 7, 3, 2, 2,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,
+     -1,-1,-1,-1,-1,-1,-1,-1,-1,-1},
+    0, //golomb_order
+    INT_MAX, //inc_limit
+    4, //max_run
+  }
+};
+
+static const uint8_t alpha_tab[64] = {
+   0,  0,  0,  0,  0,  0,  1,  1,  1,  1,  1,  2,  2,  2,  3,  3,
+   4,  4,  5,  5,  6,  7,  8,  9, 10, 11, 12, 13, 15, 16, 18, 20,
+  22, 24, 26, 28, 30, 33, 33, 35, 35, 36, 37, 37, 39, 39, 42, 44,
+  46, 48, 50, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64
+};
+
+static const uint8_t beta_tab[64] = {
+   0,  0,  0,  0,  0,  0,  1,  1,  1,  1,  1,  1,  1,  2,  2,  2,
+   2,  2,  3,  3,  3,  3,  4,  4,  4,  4,  5,  5,  5,  5,  6,  6,
+   6,  7,  7,  7,  8,  8,  8,  9,  9, 10, 10, 11, 11, 12, 13, 14,
+  15, 16, 17, 18, 19, 20, 21, 22, 23, 23, 24, 24, 25, 25, 26, 27
+};
+
+static const uint8_t tc_tab[64] = {
+  0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+  1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 2,
+  2, 2, 2, 2, 2, 2, 3, 3, 3, 3, 3, 3, 3, 4, 4, 4,
+  5, 5, 5, 6, 6, 6, 7, 7, 7, 7, 8, 8, 8, 9, 9, 9
+};
+
+static const int_fast8_t left_modifier_l[8] = { 0,-1, 6,-1,-1, 7, 6, 7};
+static const int_fast8_t top_modifier_l[8]  = {-1, 1, 5,-1,-1, 5, 7, 7};
+static const int_fast8_t left_modifier_c[7] = { 5,-1, 2,-1, 6, 5, 6};
+static const int_fast8_t top_modifier_c[7]  = { 4, 1,-1,-1, 4, 6, 6};

libavcodec/cavsdsp.c

+/*
+ * Chinese AVS video (AVS1-P2, JiZhun profile) decoder.
+ *
+ * DSP functions
+ *
+ * Copyright (c) 2006  Stefan Gehrer <stefan.gehrer@gmx.de>
+ *
+ * This library is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2 of the License, or (at your option) any later version.
+ *
+ * This library is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with this library; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+
+#include <stdio.h>
+#include "dsputil.h"
+#include "cavsdsp.h"
+
+/*****************************************************************************
+ *
+ * in-loop deblocking filter
+ *
+ ****************************************************************************/
+
+#define P2 p0_p[-3*stride]
+#define P1 p0_p[-2*stride]
+#define P0 p0_p[-1*stride]
+#define Q0 p0_p[ 0*stride]
+#define Q1 p0_p[ 1*stride]
+#define Q2 p0_p[ 2*stride]
+
+static inline void loop_filter_l2(uint8_t *p0_p,int stride,int alpha, int beta) {
+    int p0 = P0;
+    int q0 = Q0;
+
+    if(abs(p0-q0)<alpha && abs(P1-p0)<beta && abs(Q1-q0)<beta) {
+        int s = p0 + q0 + 2;
+        alpha = (alpha>>2) + 2;
+        if(abs(P2-p0) < beta && abs(p0-q0) < alpha) {
+            P0 = (P1 + p0 + s) >> 2;
+            P1 = (2*P1 + s) >> 2;
+        } else
+            P0 = (2*P1 + s) >> 2;
+        if(abs(Q2-q0) < beta && abs(q0-p0) < alpha) {
+            Q0 = (Q1 + q0 + s) >> 2;
+            Q1 = (2*Q1 + s) >> 2;
+        } else
+            Q0 = (2*Q1 + s) >> 2;
+    }
+}
+
+static inline void loop_filter_l1(uint8_t *p0_p, int stride, int alpha, int beta, int tc) {
+    int p0 = P0;
+    int q0 = Q0;
+
+    if(abs(p0-q0)<alpha && abs(P1-p0)<beta && abs(Q1-q0)<beta) {
+        int delta = clip(((q0-p0)*3+P1-Q1+4)>>3,-tc, tc);
+        P0 = clip_uint8(p0+delta);
+        Q0 = clip_uint8(q0-delta);
+        if(abs(P2-p0)<beta) {
+            delta = clip(((P0-P1)*3+P2-Q0+4)>>3, -tc, tc);
+            P1 = clip_uint8(P1+delta);
+        }
+        if(abs(Q2-q0)<beta) {
+            delta = clip(((Q1-Q0)*3+P0-Q2+4)>>3, -tc, tc);
+            Q1 = clip_uint8(Q1-delta);
+        }
+    }
+}
+
+static inline void loop_filter_c2(uint8_t *p0_p,int stride,int alpha, int beta) {
+    int p0 = P0;
+    int q0 = Q0;
+
+    if(abs(p0-q0)<alpha && abs(P1-p0)<beta && abs(Q1-q0)<beta) {
+        int s = p0 + q0 + 2;
+        alpha = (alpha>>2) + 2;
+        if(abs(P2-p0) < beta && abs(p0-q0) < alpha) {
+            P0 = (P1 + p0 + s) >> 2;
+        } else
+            P0 = (2*P1 + s) >> 2;
+        if(abs(Q2-q0) < beta && abs(q0-p0) < alpha) {
+            Q0 = (Q1 + q0 + s) >> 2;
+        } else
+            Q0 = (2*Q1 + s) >> 2;
+    }
+}
+
+static inline void loop_filter_c1(uint8_t *p0_p,int stride,int alpha, int beta,
+                                  int tc) {
+    if(abs(P0-Q0)<alpha && abs(P1-P0)<beta && abs(Q1-Q0)<beta) {
+        int delta = clip(((Q0-P0)*3+P1-Q1+4)>>3, -tc, tc);
+        P0 = clip_uint8(P0+delta);
+        Q0 = clip_uint8(Q0-delta);
+    }
+}
+
+#undef P0
+#undef P1
+#undef P2
+#undef Q0
+#undef Q1
+#undef Q2
+
+void cavs_filter_lv_c(uint8_t *d, int stride, int alpha, int beta, int tc,
+                           int bs1, int bs2) {
+    int i;
+    if(bs1==2)
+        for(i=0;i<16;i++)
+            loop_filter_l2(d + i*stride,1,alpha,beta);
+    else {
+        if(bs1)
+            for(i=0;i<8;i++)
+                loop_filter_l1(d + i*stride,1,alpha,beta,tc);
+        if (bs2)
+            for(i=8;i<16;i++)
+                loop_filter_l1(d + i*stride,1,alpha,beta,tc);
+    }
+}
+
+void cavs_filter_lh_c(uint8_t *d, int stride, int alpha, int beta, int tc,
+                           int bs1, int bs2) {
+    int i;
+    if(bs1==2)
+        for(i=0;i<16;i++)
+            loop_filter_l2(d + i,stride,alpha,beta);
+    else {
+        if(bs1)
+            for(i=0;i<8;i++)
+                loop_filter_l1(d + i,stride,alpha,beta,tc);
+        if (bs2)
+            for(i=8;i<16;i++)
+                loop_filter_l1(d + i,stride,alpha,beta,tc);
+    }
+}
+
+void cavs_filter_cv_c(uint8_t *d, int stride, int alpha, int beta, int tc,
+                           int bs1, int bs2) {
+    int i;
+    if(bs1==2)
+        for(i=0;i<8;i++)
+            loop_filter_c2(d + i*stride,1,alpha,beta);
+    else {
+        if(bs1)
+            for(i=0;i<4;i++)
+                loop_filter_c1(d + i*stride,1,alpha,beta,tc);
+        if (bs2)
+            for(i=4;i<8;i++)
+                loop_filter_c1(d + i*stride,1,alpha,beta,tc);
+    }
+}
+
+void cavs_filter_ch_c(uint8_t *d, int stride, int alpha, int beta, int tc,
+                           int bs1, int bs2) {
+    int i;
+    if(bs1==2)
+        for(i=0;i<8;i++)
+            loop_filter_c2(d + i,stride,alpha,beta);
+    else {
+        if(bs1)
+            for(i=0;i<4;i++)
+                loop_filter_c1(d + i,stride,alpha,beta,tc);
+        if (bs2)
+            for(i=4;i<8;i++)
+                loop_filter_c1(d + i,stride,alpha,beta,tc);
+    }
+}
+
+/*****************************************************************************
+ *
+ * inverse transform
+ *
+ ****************************************************************************/
+
+void cavs_idct8_add_c(uint8_t *dst, DCTELEM *block, int stride) {
+    int i;
+    DCTELEM (*src)[8] = (DCTELEM(*)[8])block;
+    uint8_t *cm = cropTbl + MAX_NEG_CROP;
+
+    for( i = 0; i < 8; i++ ) {
+        const int a0 =  3*src[i][1] - (src[i][7]<<1);
+        const int a1 =  3*src[i][3] + (src[i][5]<<1);
+        const int a2 =  (src[i][3]<<1) - 3*src[i][5];
+        const int a3 =  (src[i][1]<<1) + 3*src[i][7];
+
+        const int b4 = ((a0 + a1 + a3)<<1) + a1;
+        const int b5 = ((a0 - a1 + a2)<<1) + a0;
+        const int b6 = ((a3 - a2 - a1)<<1) + a3;
+        const int b7 = ((a0 - a2 - a3)<<1) - a2;
+
+        const int a7 = (src[i][2]<<2) - 10*src[i][6];
+        const int a6 = (src[i][6]<<2) + 10*src[i][2];
+        const int a5 = (src[i][0] - src[i][4]) << 3;
+        const int a4 = (src[i][0] + src[i][4]) << 3;
+
+        const int b0 = a4 + a6;
+        const int b1 = a5 + a7;
+        const int b2 = a5 - a7;
+        const int b3 = a4 - a6;
+
+        src[i][0] = (b0 + b4 + 4) >> 3;
+        src[i][1] = (b1 + b5 + 4) >> 3;
+        src[i][2] = (b2 + b6 + 4) >> 3;
+        src[i][3] = (b3 + b7 + 4) >> 3;
+        src[i][4] = (b3 - b7 + 4) >> 3;
+        src[i][5] = (b2 - b6 + 4) >> 3;
+        src[i][6] = (b1 - b5 + 4) >> 3;
+        src[i][7] = (b0 - b4 + 4) >> 3;
+    }
+    for( i = 0; i < 8; i++ ) {
+        const int a0 =  3*src[1][i] - (src[7][i]<<1);
+        const int a1 =  3*src[3][i] + (src[5][i]<<1);
+        const int a2 =  (src[3][i]<<1) - 3*src[5][i];
+        const int a3 =  (src[1][i]<<1) + 3*src[7][i];
+
+        const int b4 = ((a0 + a1 + a3)<<1) + a1;
+        const int b5 = ((a0 - a1 + a2)<<1) + a0;
+        const int b6 = ((a3 - a2 - a1)<<1) + a3;
+        const int b7 = ((a0 - a2 - a3)<<1) - a2;
+
+        const int a7 = (src[2][i]<<2) - 10*src[6][i];
+        const int a6 = (src[6][i]<<2) + 10*src[2][i];
+        const int a5 = (src[0][i] - src[4][i]) << 3;
+        const int a4 = (src[0][i] + src[4][i]) << 3;
+
+        const int b0 = a4 + a6;
+        const int b1 = a5 + a7;
+        const int b2 = a5 - a7;
+        const int b3 = a4 - a6;
+
+        dst[i + 0*stride] = cm[ dst[i + 0*stride] + ((b0 + b4 + 64) >> 7)];
+        dst[i + 1*stride] = cm[ dst[i + 1*stride] + ((b1 + b5 + 64) >> 7)];
+        dst[i + 2*stride] = cm[ dst[i + 2*stride] + ((b2 + b6 + 64) >> 7)];
+        dst[i + 3*stride] = cm[ dst[i + 3*stride] + ((b3 + b7 + 64) >> 7)];
+        dst[i + 4*stride] = cm[ dst[i + 4*stride] + ((b3 - b7 + 64) >> 7)];
+        dst[i + 5*stride] = cm[ dst[i + 5*stride] + ((b2 - b6 + 64) >> 7)];
+        dst[i + 6*stride] = cm[ dst[i + 6*stride] + ((b1 - b5 + 64) >> 7)];
+        dst[i + 7*stride] = cm[ dst[i + 7*stride] + ((b0 - b4 + 64) >> 7)];
+    }
+}
+
+/*****************************************************************************
+ *
+ * motion compensation
+ *
+ ****************************************************************************/
+
+#define CAVS_SUBPIX(OPNAME, OP, NAME, A, B, C, D, E, F) \
+static void OPNAME ## cavs_filt8_h_ ## NAME(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\
+    const int h=8;\
+    uint8_t *cm = cropTbl + MAX_NEG_CROP;\
+    int i;\
+    for(i=0; i<h; i++)\
+    {\
+        OP(dst[0], A*src[-2] + B*src[-1] + C*src[0] + D*src[1] + E*src[2] + F*src[3]);\
+        OP(dst[1], A*src[-1] + B*src[ 0] + C*src[1] + D*src[2] + E*src[3] + F*src[4]);\
+        OP(dst[2], A*src[ 0] + B*src[ 1] + C*src[2] + D*src[3] + E*src[4] + F*src[5]);\
+        OP(dst[3], A*src[ 1] + B*src[ 2] + C*src[3] + D*src[4] + E*src[5] + F*src[6]);\
+        OP(dst[4], A*src[ 2] + B*src[ 3] + C*src[4] + D*src[5] + E*src[6] + F*src[7]);\
+        OP(dst[5], A*src[ 3] + B*src[ 4] + C*src[5] + D*src[6] + E*src[7] + F*src[8]);\
+        OP(dst[6], A*src[ 4] + B*src[ 5] + C*src[6] + D*src[7] + E*src[8] + F*src[9]);\
+        OP(dst[7], A*src[ 5] + B*src[ 6] + C*src[7] + D*src[8] + E*src[9] + F*src[10]);\
+        dst+=dstStride;\
+        src+=srcStride;\
+    }\
+}\
+\
+static void OPNAME ## cavs_filt8_v_  ## NAME(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\
+    const int w=8;\
+    uint8_t *cm = cropTbl + MAX_NEG_CROP;\
+    int i;\
+    for(i=0; i<w; i++)\
+    {\
+        const int srcB= src[-2*srcStride];\
+        const int srcA= src[-1*srcStride];\
+        const int src0= src[0 *srcStride];\
+        const int src1= src[1 *srcStride];\
+        const int src2= src[2 *srcStride];\
+        const int src3= src[3 *srcStride];\
+        const int src4= src[4 *srcStride];\
+        const int src5= src[5 *srcStride];\
+        const int src6= src[6 *srcStride];\
+        const int src7= src[7 *srcStride];\
+        const int src8= src[8 *srcStride];\
+        const int src9= src[9 *srcStride];\
+        const int src10= src[10 *srcStride];\
+        OP(dst[0*dstStride], A*srcB + B*srcA + C*src0 + D*src1 + E*src2 + F*src3);\
+        OP(dst[1*dstStride], A*srcA + B*src0 + C*src1 + D*src2 + E*src3 + F*src4);\
+        OP(dst[2*dstStride], A*src0 + B*src1 + C*src2 + D*src3 + E*src4 + F*src5);\
+        OP(dst[3*dstStride], A*src1 + B*src2 + C*src3 + D*src4 + E*src5 + F*src6);\
+        OP(dst[4*dstStride], A*src2 + B*src3 + C*src4 + D*src5 + E*src6 + F*src7);\
+        OP(dst[5*dstStride], A*src3 + B*src4 + C*src5 + D*src6 + E*src7 + F*src8);\
+        OP(dst[6*dstStride], A*src4 + B*src5 + C*src6 + D*src7 + E*src8 + F*src9);\
+        OP(dst[7*dstStride], A*src5 + B*src6 + C*src7 + D*src8 + E*src9 + F*src10);\
+        dst++;\
+        src++;\
+    }\
+}\
+\
+static void OPNAME ## cavs_filt16_v_ ## NAME(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\
+    OPNAME ## cavs_filt8_v_ ## NAME(dst  , src  , dstStride, srcStride);\
+    OPNAME ## cavs_filt8_v_ ## NAME(dst+8, src+8, dstStride, srcStride);\
+    src += 8*srcStride;\
+    dst += 8*dstStride;\
+    OPNAME ## cavs_filt8_v_ ## NAME(dst  , src  , dstStride, srcStride);\
+    OPNAME ## cavs_filt8_v_ ## NAME(dst+8, src+8, dstStride, srcStride);\
+}\
+\
+static void OPNAME ## cavs_filt16_h_ ## NAME(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\
+    OPNAME ## cavs_filt8_h_ ## NAME(dst  , src  , dstStride, srcStride);\
+    OPNAME ## cavs_filt8_h_ ## NAME(dst+8, src+8, dstStride, srcStride);\
+    src += 8*srcStride;\
+    dst += 8*dstStride;\
+    OPNAME ## cavs_filt8_h_ ## NAME(dst  , src  , dstStride, srcStride);\
+    OPNAME ## cavs_filt8_h_ ## NAME(dst+8, src+8, dstStride, srcStride);\
+}\
+
+#define CAVS_SUBPIX_HV(OPNAME, OP, NAME, AH, BH, CH, DH, EH, FH, AV, BV, CV, DV, EV, FV, FULL) \
+static void OPNAME ## cavs_filt8_hv_ ## NAME(uint8_t *dst, uint8_t *src1, uint8_t *src2, int dstStride, int srcStride){\
+    int16_t temp[8*(8+5)];\
+    int16_t *tmp = temp;\
+    const int h=8;\
+    const int w=8;\
+    uint8_t *cm = cropTbl + MAX_NEG_CROP;\
+    int i;\
+    src1 -= 2*srcStride;\
+    for(i=0; i<h+5; i++)\
+    {\
+        tmp[0]= AH*src1[-2] + BH*src1[-1] + CH*src1[0] + DH*src1[1] + EH*src1[2] + FH*src1[3];\
+        tmp[1]= AH*src1[-1] + BH*src1[ 0] + CH*src1[1] + DH*src1[2] + EH*src1[3] + FH*src1[4];\
+        tmp[2]= AH*src1[ 0] + BH*src1[ 1] + CH*src1[2] + DH*src1[3] + EH*src1[4] + FH*src1[5];\
+        tmp[3]= AH*src1[ 1] + BH*src1[ 2] + CH*src1[3] + DH*src1[4] + EH*src1[5] + FH*src1[6];\
+        tmp[4]= AH*src1[ 2] + BH*src1[ 3] + CH*src1[4] + DH*src1[5] + EH*src1[6] + FH*src1[7];\
+        tmp[5]= AH*src1[ 3] + BH*src1[ 4] + CH*src1[5] + DH*src1[6] + EH*src1[7] + FH*src1[8];\
+        tmp[6]= AH*src1[ 4] + BH*src1[ 5] + CH*src1[6] + DH*src1[7] + EH*src1[8] + FH*src1[9];\
+        tmp[7]= AH*src1[ 5] + BH*src1[ 6] + CH*src1[7] + DH*src1[8] + EH*src1[9] + FH*src1[10];\
+        tmp+=8;\
+        src1+=srcStride;\
+    }\
+    if(FULL) {\
+      tmp = temp+8*2;                           \
+      for(i=0; i<w; i++)                        \
+        {                                       \
+          const int tmpB= tmp[-2*8];    \
+          const int tmpA= tmp[-1*8];    \
+          const int tmp0= tmp[0 *8];    \
+          const int tmp1= tmp[1 *8];    \
+          const int tmp2= tmp[2 *8];    \
+          const int tmp3= tmp[3 *8];    \
+          const int tmp4= tmp[4 *8];    \
+          const int tmp5= tmp[5 *8];    \
+          const int tmp6= tmp[6 *8];    \
+          const int tmp7= tmp[7 *8];    \
+          const int tmp8= tmp[8 *8];    \
+          const int tmp9= tmp[9 *8];    \
+          const int tmp10=tmp[10*8];                            \
+          OP(dst[0*dstStride], AV*tmpB + BV*tmpA + CV*tmp0 + DV*tmp1 + EV*tmp2 + FV*tmp3 + 64*src2[0*srcStride]); \
+          OP(dst[1*dstStride], AV*tmpA + BV*tmp0 + CV*tmp1 + DV*tmp2 + EV*tmp3 + FV*tmp4 + 64*src2[1*srcStride]); \
+          OP(dst[2*dstStride], AV*tmp0 + BV*tmp1 + CV*tmp2 + DV*tmp3 + EV*tmp4 + FV*tmp5 + 64*src2[2*srcStride]); \
+          OP(dst[3*dstStride], AV*tmp1 + BV*tmp2 + CV*tmp3 + DV*tmp4 + EV*tmp5 + FV*tmp6 + 64*src2[3*srcStride]); \
+          OP(dst[4*dstStride], AV*tmp2 + BV*tmp3 + CV*tmp4 + DV*tmp5 + EV*tmp6 + FV*tmp7 + 64*src2[4*srcStride]); \
+          OP(dst[5*dstStride], AV*tmp3 + BV*tmp4 + CV*tmp5 + DV*tmp6 + EV*tmp7 + FV*tmp8 + 64*src2[5*srcStride]); \
+          OP(dst[6*dstStride], AV*tmp4 + BV*tmp5 + CV*tmp6 + DV*tmp7 + EV*tmp8 + FV*tmp9 + 64*src2[6*srcStride]); \
+          OP(dst[7*dstStride], AV*tmp5 + BV*tmp6 + CV*tmp7 + DV*tmp8 + EV*tmp9 + FV*tmp10 + 64*src2[7*srcStride]); \
+          dst++;                                                        \
+          tmp++;                                                        \
+          src2++;                                                       \
+        }                                                               \
+    } else {\
+      tmp = temp+8*2;                           \
+      for(i=0; i<w; i++)                        \
+        {                                       \
+          const int tmpB= tmp[-2*8];    \
+          const int tmpA= tmp[-1*8];    \
+          const int tmp0= tmp[0 *8];    \
+          const int tmp1= tmp[1 *8];    \
+          const int tmp2= tmp[2 *8];    \
+          const int tmp3= tmp[3 *8];    \
+          const int tmp4= tmp[4 *8];    \
+          const int tmp5= tmp[5 *8];    \
+          const int tmp6= tmp[6 *8];    \
+          const int tmp7= tmp[7 *8];    \
+          const int tmp8= tmp[8 *8];    \
+          const int tmp9= tmp[9 *8];    \
+          const int tmp10=tmp[10*8];                            \
+          OP(dst[0*dstStride], AV*tmpB + BV*tmpA + CV*tmp0 + DV*tmp1 + EV*tmp2 + FV*tmp3); \
+          OP(dst[1*dstStride], AV*tmpA + BV*tmp0 + CV*tmp1 + DV*tmp2 + EV*tmp3 + FV*tmp4); \
+          OP(dst[2*dstStride], AV*tmp0 + BV*tmp1 + CV*tmp2 + DV*tmp3 + EV*tmp4 + FV*tmp5); \
+          OP(dst[3*dstStride], AV*tmp1 + BV*tmp2 + CV*tmp3 + DV*tmp4 + EV*tmp5 + FV*tmp6); \
+          OP(dst[4*dstStride], AV*tmp2 + BV*tmp3 + CV*tmp4 + DV*tmp5 + EV*tmp6 + FV*tmp7); \
+          OP(dst[5*dstStride], AV*tmp3 + BV*tmp4 + CV*tmp5 + DV*tmp6 + EV*tmp7 + FV*tmp8); \
+          OP(dst[6*dstStride], AV*tmp4 + BV*tmp5 + CV*tmp6 + DV*tmp7 + EV*tmp8 + FV*tmp9); \
+          OP(dst[7*dstStride], AV*tmp5 + BV*tmp6 + CV*tmp7 + DV*tmp8 + EV*tmp9 + FV*tmp10); \
+          dst++;                                                        \
+          tmp++;                                                        \
+        }                                                               \
+    }\
+}\
+\
+static void OPNAME ## cavs_filt16_hv_ ## NAME(uint8_t *dst, uint8_t *src1, uint8_t *src2, int dstStride, int srcStride){ \
+    OPNAME ## cavs_filt8_hv_ ## NAME(dst  , src1,   src2  , dstStride, srcStride); \
+    OPNAME ## cavs_filt8_hv_ ## NAME(dst+8, src1+8, src2+8, dstStride, srcStride); \
+    src1 += 8*srcStride;\
+    src2 += 8*srcStride;\
+    dst += 8*dstStride;\
+    OPNAME ## cavs_filt8_hv_ ## NAME(dst  , src1,   src2  , dstStride, srcStride); \
+    OPNAME ## cavs_filt8_hv_ ## NAME(dst+8, src1+8, src2+8, dstStride, srcStride); \
+}\
+
+#define CAVS_MC(OPNAME, SIZE) \
+void OPNAME ## cavs_qpel ## SIZE ## _mc00_c (uint8_t *dst, uint8_t *src, int stride){\
+    OPNAME ## pixels ## SIZE ## _c(dst, src, stride, SIZE);\
+}\
+void OPNAME ## cavs_qpel ## SIZE ## _mc10_c(uint8_t *dst, uint8_t *src, int stride){\
+    OPNAME ## cavs_filt ## SIZE ## _h_qpel_l(dst, src, stride, stride);\
+}\
+\
+void OPNAME ## cavs_qpel ## SIZE ## _mc20_c(uint8_t *dst, uint8_t *src, int stride){\
+    OPNAME ## cavs_filt ## SIZE ## _h_hpel(dst, src, stride, stride);\
+}\
+\
+void OPNAME ## cavs_qpel ## SIZE ## _mc30_c(uint8_t *dst, uint8_t *src, int stride){\
+    OPNAME ## cavs_filt ## SIZE ## _h_qpel_r(dst, src, stride, stride);\
+}\
+\
+void OPNAME ## cavs_qpel ## SIZE ## _mc01_c(uint8_t *dst, uint8_t *src, int stride){\
+    OPNAME ## cavs_filt ## SIZE ## _v_qpel_l(dst, src, stride, stride);\
+}\
+\
+void OPNAME ## cavs_qpel ## SIZE ## _mc02_c(uint8_t *dst, uint8_t *src, int stride){\
+    OPNAME ## cavs_filt ## SIZE ## _v_hpel(dst, src, stride, stride);\
+}\
+\
+void OPNAME ## cavs_qpel ## SIZE ## _mc03_c(uint8_t *dst, uint8_t *src, int stride){\
+    OPNAME ## cavs_filt ## SIZE ## _v_qpel_r(dst, src, stride, stride);\
+}\
+\
+void OPNAME ## cavs_qpel ## SIZE ## _mc22_c(uint8_t *dst, uint8_t *src, int stride){\
+  OPNAME ## cavs_filt ## SIZE ## _hv_jj(dst, src, NULL, stride, stride); \
+}\
+\
+void OPNAME ## cavs_qpel ## SIZE ## _mc11_c(uint8_t *dst, uint8_t *src, int stride){\
+  OPNAME ## cavs_filt ## SIZE ## _hv_egpr(dst, src, src, stride, stride); \
+}\
+\
+void OPNAME ## cavs_qpel ## SIZE ## _mc13_c(uint8_t *dst, uint8_t *src, int stride){\
+  OPNAME ## cavs_filt ## SIZE ## _hv_egpr(dst, src, src+stride, stride, stride); \
+}\
+\
+void OPNAME ## cavs_qpel ## SIZE ## _mc31_c(uint8_t *dst, uint8_t *src, int stride){\
+  OPNAME ## cavs_filt ## SIZE ## _hv_egpr(dst, src, src+1, stride, stride); \
+}\
+\
+void OPNAME ## cavs_qpel ## SIZE ## _mc33_c(uint8_t *dst, uint8_t *src, int stride){\
+  OPNAME ## cavs_filt ## SIZE ## _hv_egpr(dst, src, src+stride+1,stride, stride); \
+}\
+\
+void OPNAME ## cavs_qpel ## SIZE ## _mc21_c(uint8_t *dst, uint8_t *src, int stride){\
+  OPNAME ## cavs_filt ## SIZE ## _hv_ff(dst, src, src+stride+1,stride, stride); \
+}\
+\
+void OPNAME ## cavs_qpel ## SIZE ## _mc12_c(uint8_t *dst, uint8_t *src, int stride){\
+  OPNAME ## cavs_filt ## SIZE ## _hv_ii(dst, src, src+stride+1,stride, stride); \
+}\
+\
+void OPNAME ## cavs_qpel ## SIZE ## _mc32_c(uint8_t *dst, uint8_t *src, int stride){\
+  OPNAME ## cavs_filt ## SIZE ## _hv_kk(dst, src, src+stride+1,stride, stride); \
+}\
+\
+void OPNAME ## cavs_qpel ## SIZE ## _mc23_c(uint8_t *dst, uint8_t *src, int stride){\
+  OPNAME ## cavs_filt ## SIZE ## _hv_qq(dst, src, src+stride+1,stride, stride); \
+}\
+
+#define op_put1(a, b)  a = cm[((b)+4)>>3]
+#define op_put2(a, b)  a = cm[((b)+64)>>7]
+#define op_put3(a, b)  a = cm[((b)+32)>>6]
+#define op_put4(a, b)  a = cm[((b)+512)>>10]
+#define op_avg1(a, b)  a = ((a)+cm[((b)+4)>>3]   +1)>>1
+#define op_avg2(a, b)  a = ((a)+cm[((b)+64)>>7]  +1)>>1
+#define op_avg3(a, b)  a = ((a)+cm[((b)+32)>>6]  +1)>>1
+#define op_avg4(a, b)  a = ((a)+cm[((b)+512)>>10]+1)>>1
+CAVS_SUBPIX(put_   , op_put1, hpel,    0, -1,  5,  5, -1,  0)
+CAVS_SUBPIX(put_   , op_put2, qpel_l, -1, -2, 96, 42, -7,  0)
+CAVS_SUBPIX(put_   , op_put2, qpel_r,  0, -7, 42, 96, -2, -1)
+CAVS_SUBPIX_HV(put_, op_put3, jj,      0, -1,  5,  5, -1,  0,  0, -1,  5,  5, -1, 0, 0)
+CAVS_SUBPIX_HV(put_, op_put4, ff,      0, -1,  5,  5, -1,  0, -1, -2, 96, 42, -7, 0, 0)
+CAVS_SUBPIX_HV(put_, op_put4, ii,     -1, -2, 96, 42, -7,  0,  0, -1,  5,  5, -1, 0, 0)
+CAVS_SUBPIX_HV(put_, op_put4, kk,      0, -7, 42, 96, -2, -1,  0, -1,  5,  5, -1, 0, 0)
+CAVS_SUBPIX_HV(put_, op_put4, qq,      0, -1,  5,  5, -1,  0,  0, -7, 42, 96, -2,-1, 0)
+CAVS_SUBPIX_HV(put_, op_put2, egpr,    0, -1,  5,  5, -1,  0,  0, -1,  5,  5, -1, 0, 1)
+CAVS_SUBPIX(avg_   , op_avg1, hpel,    0, -1,  5,  5, -1,  0)
+CAVS_SUBPIX(avg_   , op_avg2, qpel_l, -1, -2, 96, 42, -7,  0)
+CAVS_SUBPIX(avg_   , op_avg2, qpel_r,  0, -7, 42, 96, -2, -1)
+CAVS_SUBPIX_HV(avg_, op_avg3, jj,      0, -1,  5,  5, -1,  0,  0, -1,  5,  5, -1, 0, 0)
+CAVS_SUBPIX_HV(avg_, op_avg4, ff,      0, -1,  5,  5, -1,  0, -1, -2, 96, 42, -7, 0, 0)
+CAVS_SUBPIX_HV(avg_, op_avg4, ii,     -1, -2, 96, 42, -7,  0,  0, -1,  5,  5, -1, 0, 0)
+CAVS_SUBPIX_HV(avg_, op_avg4, kk,      0, -7, 42, 96, -2, -1,  0, -1,  5,  5, -1, 0, 0)
+CAVS_SUBPIX_HV(avg_, op_avg4, qq,      0, -1,  5,  5, -1,  0,  0, -7, 42, 96, -2,-1, 0)
+CAVS_SUBPIX_HV(avg_, op_avg2, egpr,    0, -1,  5,  5, -1,  0,  0, -1,  5,  5, -1, 0, 1)
+CAVS_MC(put_, 8)
+CAVS_MC(put_, 16)
+CAVS_MC(avg_, 8)
+CAVS_MC(avg_, 16)

libavcodec/cavsdsp.h

+/*
+ * Chinese AVS video (AVS1-P2, JiZhun profile) decoder.
+ * Copyright (c) 2006  Stefan Gehrer <stefan.gehrer@gmx.de>
+ *
+ * DSP function prototypes
+ *
+ * This library is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2 of the License, or (at your option) any later version.
+ *
+ * This library is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with this library; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+
+void put_cavs_qpel16_mc00_c(uint8_t *dst, uint8_t *src, int stride);
+void put_cavs_qpel16_mc01_c(uint8_t *dst, uint8_t *src, int stride);
+void put_cavs_qpel16_mc02_c(uint8_t *dst, uint8_t *src, int stride);
+void put_cavs_qpel16_mc03_c(uint8_t *dst, uint8_t *src, int stride);
+void put_cavs_qpel16_mc10_c(uint8_t *dst, uint8_t *src, int stride);
+void put_cavs_qpel16_mc11_c(uint8_t *dst, uint8_t *src, int stride);
+void put_cavs_qpel16_mc12_c(uint8_t *dst, uint8_t *src, int stride);
+void put_cavs_qpel16_mc13_c(uint8_t *dst, uint8_t *src, int stride);
+void put_cavs_qpel16_mc20_c(uint8_t *dst, uint8_t *src, int stride);
+void put_cavs_qpel16_mc21_c(uint8_t *dst, uint8_t *src, int stride);
+void put_cavs_qpel16_mc22_c(uint8_t *dst, uint8_t *src, int stride);
+void put_cavs_qpel16_mc23_c(uint8_t *dst, uint8_t *src, int stride);
+void put_cavs_qpel16_mc30_c(uint8_t *dst, uint8_t *src, int stride);
+void put_cavs_qpel16_mc31_c(uint8_t *dst, uint8_t *src, int stride);
+void put_cavs_qpel16_mc32_c(uint8_t *dst, uint8_t *src, int stride);
+void put_cavs_qpel16_mc33_c(uint8_t *dst, uint8_t *src, int stride);
+void put_cavs_qpel8_mc00_c(uint8_t *dst, uint8_t *src, int stride);
+void put_cavs_qpel8_mc01_c(uint8_t *dst, uint8_t *src, int stride);
+void put_cavs_qpel8_mc02_c(uint8_t *dst, uint8_t *src, int stride);
+void put_cavs_qpel8_mc03_c(uint8_t *dst, uint8_t *src, int stride);
+void put_cavs_qpel8_mc10_c(uint8_t *dst, uint8_t *src, int stride);
+void put_cavs_qpel8_mc11_c(uint8_t *dst, uint8_t *src, int stride);
+void put_cavs_qpel8_mc12_c(uint8_t *dst, uint8_t *src, int stride);
+void put_cavs_qpel8_mc13_c(uint8_t *dst, uint8_t *src, int stride);
+void put_cavs_qpel8_mc20_c(uint8_t *dst, uint8_t *src, int stride);
+void put_cavs_qpel8_mc21_c(uint8_t *dst, uint8_t *src, int stride);
+void put_cavs_qpel8_mc22_c(uint8_t *dst, uint8_t *src, int stride);
+void put_cavs_qpel8_mc23_c(uint8_t *dst, uint8_t *src, int stride);
+void put_cavs_qpel8_mc30_c(uint8_t *dst, uint8_t *src, int stride);
+void put_cavs_qpel8_mc31_c(uint8_t *dst, uint8_t *src, int stride);
+void put_cavs_qpel8_mc32_c(uint8_t *dst, uint8_t *src, int stride);
+void put_cavs_qpel8_mc33_c(uint8_t *dst, uint8_t *src, int stride);
+void avg_cavs_qpel16_mc00_c(uint8_t *dst, uint8_t *src, int stride);
+void avg_cavs_qpel16_mc01_c(uint8_t *dst, uint8_t *src, int stride);
+void avg_cavs_qpel16_mc02_c(uint8_t *dst, uint8_t *src, int stride);
+void avg_cavs_qpel16_mc03_c(uint8_t *dst, uint8_t *src, int stride);
+void avg_cavs_qpel16_mc10_c(uint8_t *dst, uint8_t *src, int stride);
+void avg_cavs_qpel16_mc11_c(uint8_t *dst, uint8_t *src, int stride);
+void avg_cavs_qpel16_mc12_c(uint8_t *dst, uint8_t *src, int stride);
+void avg_cavs_qpel16_mc13_c(uint8_t *dst, uint8_t *src, int stride);
+void avg_cavs_qpel16_mc20_c(uint8_t *dst, uint8_t *src, int stride);
+void avg_cavs_qpel16_mc21_c(uint8_t *dst, uint8_t *src, int stride);
+void avg_cavs_qpel16_mc22_c(uint8_t *dst, uint8_t *src, int stride);
+void avg_cavs_qpel16_mc23_c(uint8_t *dst, uint8_t *src, int stride);
+void avg_cavs_qpel16_mc30_c(uint8_t *dst, uint8_t *src, int stride);
+void avg_cavs_qpel16_mc31_c(uint8_t *dst, uint8_t *src, int stride);
+void avg_cavs_qpel16_mc32_c(uint8_t *dst, uint8_t *src, int stride);
+void avg_cavs_qpel16_mc33_c(uint8_t *dst, uint8_t *src, int stride);
+void avg_cavs_qpel8_mc00_c(uint8_t *dst, uint8_t *src, int stride);
+void avg_cavs_qpel8_mc01_c(uint8_t *dst, uint8_t *src, int stride);
+void avg_cavs_qpel8_mc02_c(uint8_t *dst, uint8_t *src, int stride);
+void avg_cavs_qpel8_mc03_c(uint8_t *dst, uint8_t *src, int stride);
+void avg_cavs_qpel8_mc10_c(uint8_t *dst, uint8_t *src, int stride);
+void avg_cavs_qpel8_mc11_c(uint8_t *dst, uint8_t *src, int stride);
+void avg_cavs_qpel8_mc12_c(uint8_t *dst, uint8_t *src, int stride);
+void avg_cavs_qpel8_mc13_c(uint8_t *dst, uint8_t *src, int stride);
+void avg_cavs_qpel8_mc20_c(uint8_t *dst, uint8_t *src, int stride);
+void avg_cavs_qpel8_mc21_c(uint8_t *dst, uint8_t *src, int stride);
+void avg_cavs_qpel8_mc22_c(uint8_t *dst, uint8_t *src, int stride);
+void avg_cavs_qpel8_mc23_c(uint8_t *dst, uint8_t *src, int stride);
+void avg_cavs_qpel8_mc30_c(uint8_t *dst, uint8_t *src, int stride);
+void avg_cavs_qpel8_mc31_c(uint8_t *dst, uint8_t *src, int stride);
+void avg_cavs_qpel8_mc32_c(uint8_t *dst, uint8_t *src, int stride);
+void avg_cavs_qpel8_mc33_c(uint8_t *dst, uint8_t *src, int stride);
+void cavs_filter_lv_c(uint8_t *d, int stride, int alpha, int beta, int tc, int bs1, int bs2);
+void cavs_filter_lh_c(uint8_t *d, int stride, int alpha, int beta, int tc, int bs1, int bs2);
+void cavs_filter_cv_c(uint8_t *d, int stride, int alpha, int beta, int tc, int bs1, int bs2);
+void cavs_filter_ch_c(uint8_t *d, int stride, int alpha, int beta, int tc, int bs1, int bs2);
+void cavs_idct8_add_c(uint8_t *dst, DCTELEM *block, int stride);
+
+void put_pixels8_c(uint8_t *block, const uint8_t *pixels, int line_size, int h);
+void put_pixels16_c(uint8_t *block, const uint8_t *pixels, int line_size, int h);
+void avg_pixels8_c(uint8_t *block, const uint8_t *pixels, int line_size, int h);
+void avg_pixels16_c(uint8_t *block, const uint8_t *pixels, int line_size, int h);