4xm.c

Trolltech公司发布的图形界面操作系统。可在qt-embedded-2.3.10平台上编译为嵌入式图形界面操作系统。

/*
 * 4XM codec
 * Copyright (c) 2003 Michael Niedermayer
 *
 * 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., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
 */
 
/**
 * @file 4xm.c
 * 4XM codec.
 */
 
#include "avcodec.h"
#include "dsputil.h"
#include "mpegvideo.h"

//#undef NDEBUG
//#include <assert.h>

#define BLOCK_TYPE_VLC_BITS 5
#define ACDC_VLC_BITS 9

#define CFRAME_BUFFER_COUNT 100

static const uint8_t block_type_tab[4][8][2]={
  {   //{8,4,2}x{8,4,2}
    { 0,1}, { 2,2}, { 6,3}, {14,4}, {30,5}, {31,5}, { 0,0}
  },{ //{8,4}x1
    { 0,1}, { 0,0}, { 2,2}, { 6,3}, {14,4}, {15,4}, { 0,0}
  },{ //1x{8,4}
    { 0,1}, { 2,2}, { 0,0}, { 6,3}, {14,4}, {15,4}, { 0,0}
  },{ //1x2, 2x1
    { 0,1}, { 0,0}, { 0,0}, { 2,2}, { 6,3}, {14,4}, {15,4}
  }
};

static const uint8_t size2index[4][4]={
  {-1, 3, 1, 1},
  { 3, 0, 0, 0},
  { 2, 0, 0, 0},
  { 2, 0, 0, 0},
};

static const int8_t mv[256][2]={
{  0,  0},{  0, -1},{ -1,  0},{  1,  0},{  0,  1},{ -1, -1},{  1, -1},{ -1,  1},
{  1,  1},{  0, -2},{ -2,  0},{  2,  0},{  0,  2},{ -1, -2},{  1, -2},{ -2, -1},
{  2, -1},{ -2,  1},{  2,  1},{ -1,  2},{  1,  2},{ -2, -2},{  2, -2},{ -2,  2},
{  2,  2},{  0, -3},{ -3,  0},{  3,  0},{  0,  3},{ -1, -3},{  1, -3},{ -3, -1},
{  3, -1},{ -3,  1},{  3,  1},{ -1,  3},{  1,  3},{ -2, -3},{  2, -3},{ -3, -2},
{  3, -2},{ -3,  2},{  3,  2},{ -2,  3},{  2,  3},{  0, -4},{ -4,  0},{  4,  0},
{  0,  4},{ -1, -4},{  1, -4},{ -4, -1},{  4, -1},{  4,  1},{ -1,  4},{  1,  4},
{ -3, -3},{ -3,  3},{  3,  3},{ -2, -4},{ -4, -2},{  4, -2},{ -4,  2},{ -2,  4},
{  2,  4},{ -3, -4},{  3, -4},{  4, -3},{ -5,  0},{ -4,  3},{ -3,  4},{  3,  4},
{ -1, -5},{ -5, -1},{ -5,  1},{ -1,  5},{ -2, -5},{  2, -5},{  5, -2},{  5,  2},
{ -4, -4},{ -4,  4},{ -3, -5},{ -5, -3},{ -5,  3},{  3,  5},{ -6,  0},{  0,  6},
{ -6, -1},{ -6,  1},{  1,  6},{  2, -6},{ -6,  2},{  2,  6},{ -5, -4},{  5,  4},
{  4,  5},{ -6, -3},{  6,  3},{ -7,  0},{ -1, -7},{  5, -5},{ -7,  1},{ -1,  7},
{  4, -6},{  6,  4},{ -2, -7},{ -7,  2},{ -3, -7},{  7, -3},{  3,  7},{  6, -5},
{  0, -8},{ -1, -8},{ -7, -4},{ -8,  1},{  4,  7},{  2, -8},{ -2,  8},{  6,  6},
{ -8,  3},{  5, -7},{ -5,  7},{  8, -4},{  0, -9},{ -9, -1},{  1,  9},{  7, -6},
{ -7,  6},{ -5, -8},{ -5,  8},{ -9,  3},{  9, -4},{  7, -7},{  8, -6},{  6,  8},
{ 10,  1},{-10,  2},{  9, -5},{ 10, -3},{ -8, -7},{-10, -4},{  6, -9},{-11,  0},
{ 11,  1},{-11, -2},{ -2, 11},{  7, -9},{ -7,  9},{ 10,  6},{ -4, 11},{  8, -9},
{  8,  9},{  5, 11},{  7,-10},{ 12, -3},{ 11,  6},{ -9, -9},{  8, 10},{  5, 12},
{-11,  7},{ 13,  2},{  6,-12},{ 10,  9},{-11,  8},{ -7, 12},{  0, 14},{ 14, -2},
{ -9, 11},{ -6, 13},{-14, -4},{ -5,-14},{  5, 14},{-15, -1},{-14, -6},{  3,-15},
{ 11,-11},{ -7, 14},{ -5, 15},{  8,-14},{ 15,  6},{  3, 16},{  7,-15},{-16,  5},
{  0, 17},{-16, -6},{-10, 14},{-16,  7},{ 12, 13},{-16,  8},{-17,  6},{-18,  3},
{ -7, 17},{ 15, 11},{ 16, 10},{  2,-19},{  3,-19},{-11,-16},{-18,  8},{-19, -6},
{  2,-20},{-17,-11},{-10,-18},{  8, 19},{-21, -1},{-20,  7},{ -4, 21},{ 21,  5},
{ 15, 16},{  2,-22},{-10,-20},{-22,  5},{ 20,-11},{ -7,-22},{-12, 20},{ 23, -5},
{ 13,-20},{ 24, -2},{-15, 19},{-11, 22},{ 16, 19},{ 23,-10},{-18,-18},{ -9,-24},
{ 24,-10},{ -3, 26},{-23, 13},{-18,-20},{ 17, 21},{ -4, 27},{ 27,  6},{  1,-28},
{-11, 26},{-17,-23},{  7, 28},{ 11,-27},{ 29,  5},{-23,-19},{-28,-11},{-21, 22},
{-30,  7},{-17, 26},{-27, 16},{ 13, 29},{ 19,-26},{ 10,-31},{-14,-30},{ 20,-27},
{-29, 18},{-16,-31},{-28,-22},{ 21,-30},{-25, 28},{ 26,-29},{ 25,-32},{-32,-32}
};

// this is simply the scaled down elementwise product of the standard jpeg quantizer table and the AAN premul table
static const uint8_t dequant_table[64]={
 16, 15, 13, 19, 24, 31, 28, 17,
 17, 23, 25, 31, 36, 63, 45, 21,
 18, 24, 27, 37, 52, 59, 49, 20,
 16, 28, 34, 40, 60, 80, 51, 20,
 18, 31, 48, 66, 68, 86, 56, 21,
 19, 38, 56, 59, 64, 64, 48, 20,
 27, 48, 55, 55, 56, 51, 35, 15,
 20, 35, 34, 32, 31, 22, 15,  8,
};

static VLC block_type_vlc[4];


typedef struct CFrameBuffer{
    int allocated_size;
    int size;
    int id;
    uint8_t *data;
}CFrameBuffer;

typedef struct FourXContext{
    AVCodecContext *avctx;
    DSPContext dsp;
    AVFrame current_picture, last_picture;
    GetBitContext pre_gb;          ///< ac/dc prefix
    GetBitContext gb;
    uint8_t *bytestream;
    uint16_t *wordstream;
    int mv[256];
    VLC pre_vlc;
    int last_dc;
    DCTELEM __align8 block[6][64];
    uint8_t *bitstream_buffer;
    int bitstream_buffer_size;
    CFrameBuffer cfrm[CFRAME_BUFFER_COUNT];
} FourXContext;


#define FIX_1_082392200  70936
#define FIX_1_414213562  92682
#define FIX_1_847759065 121095
#define FIX_2_613125930 171254

#define MULTIPLY(var,const)  (((var)*(const)) >> 16)

static void idct(DCTELEM block[64]){
    int tmp0, tmp1, tmp2, tmp3, tmp4, tmp5, tmp6, tmp7;
    int tmp10, tmp11, tmp12, tmp13;
    int z5, z10, z11, z12, z13;
    int i;
    int temp[64];
    
    for(i=0; i<8; i++){
        tmp10 = block[8*0 + i] + block[8*4 + i];
        tmp11 = block[8*0 + i] - block[8*4 + i];

        tmp13 =          block[8*2 + i] + block[8*6 + i];
        tmp12 = MULTIPLY(block[8*2 + i] - block[8*6 + i], FIX_1_414213562) - tmp13;

        tmp0 = tmp10 + tmp13;
        tmp3 = tmp10 - tmp13;
        tmp1 = tmp11 + tmp12;
        tmp2 = tmp11 - tmp12;
        
        z13 = block[8*5 + i] + block[8*3 + i];
        z10 = block[8*5 + i] - block[8*3 + i];
        z11 = block[8*1 + i] + block[8*7 + i];
        z12 = block[8*1 + i] - block[8*7 + i];

        tmp7  =          z11 + z13;
        tmp11 = MULTIPLY(z11 - z13, FIX_1_414213562);

        z5    = MULTIPLY(z10 + z12, FIX_1_847759065);
        tmp10 = MULTIPLY(z12, FIX_1_082392200) - z5;
        tmp12 = MULTIPLY(z10, - FIX_2_613125930) + z5;

        tmp6 = tmp12 - tmp7;
        tmp5 = tmp11 - tmp6;
        tmp4 = tmp10 + tmp5;

        temp[8*0 + i] = tmp0 + tmp7;
        temp[8*7 + i] = tmp0 - tmp7;
        temp[8*1 + i] = tmp1 + tmp6;
        temp[8*6 + i] = tmp1 - tmp6;
        temp[8*2 + i] = tmp2 + tmp5;
        temp[8*5 + i] = tmp2 - tmp5;
        temp[8*4 + i] = tmp3 + tmp4;
        temp[8*3 + i] = tmp3 - tmp4;
    }
  
    for(i=0; i<8*8; i+=8){
        tmp10 = temp[0 + i] + temp[4 + i];
        tmp11 = temp[0 + i] - temp[4 + i];

        tmp13 = temp[2 + i] + temp[6 + i];
        tmp12 = MULTIPLY(temp[2 + i] - temp[6 + i], FIX_1_414213562) - tmp13;

        tmp0 = tmp10 + tmp13;
        tmp3 = tmp10 - tmp13;
        tmp1 = tmp11 + tmp12;
        tmp2 = tmp11 - tmp12;

        z13 = temp[5 + i] + temp[3 + i];
        z10 = temp[5 + i] - temp[3 + i];
        z11 = temp[1 + i] + temp[7 + i];
        z12 = temp[1 + i] - temp[7 + i];

        tmp7 = z11 + z13;
        tmp11 = MULTIPLY(z11 - z13, FIX_1_414213562);

        z5 = MULTIPLY(z10 + z12, FIX_1_847759065);
        tmp10 = MULTIPLY(z12, FIX_1_082392200) - z5;
        tmp12 = MULTIPLY(z10, - FIX_2_613125930) + z5;

        tmp6 = tmp12 - tmp7;
        tmp5 = tmp11 - tmp6;
        tmp4 = tmp10 + tmp5;

        block[0 + i] = (tmp0 + tmp7)>>6;
        block[7 + i] = (tmp0 - tmp7)>>6;
        block[1 + i] = (tmp1 + tmp6)>>6;
        block[6 + i] = (tmp1 - tmp6)>>6;
        block[2 + i] = (tmp2 + tmp5)>>6;
        block[5 + i] = (tmp2 - tmp5)>>6;
        block[4 + i] = (tmp3 + tmp4)>>6;
        block[3 + i] = (tmp3 - tmp4)>>6;
    }
}

static void init_vlcs(FourXContext *f){
    static int done = 0;
    int i;

    if (!done) {
        done = 1;

        for(i=0; i<4; i++){
            init_vlc(&block_type_vlc[i], BLOCK_TYPE_VLC_BITS, 7, 
                     &block_type_tab[i][0][1], 2, 1,
                     &block_type_tab[i][0][0], 2, 1);
        }
    }
}

static void init_mv(FourXContext *f){
    int i;

    for(i=0; i<256; i++){
        f->mv[i] = mv[i][0] + mv[i][1]*f->current_picture.linesize[0]/2;
    }
}

static inline void mcdc(uint16_t *dst, uint16_t *src, int log2w, int h, int stride, int scale, int dc){
   int i;
   dc*= 0x10001;

   switch(log2w){
   case 0:
        for(i=0; i<h; i++){
            dst[0] = scale*src[0] + dc;
            if(scale) src += stride;
            dst += stride;
        }
        break;
    case 1:
        for(i=0; i<h; i++){
            ((uint32_t*)dst)[0] = scale*((uint32_t*)src)[0] + dc;
            if(scale) src += stride;
            dst += stride;
        }
        break;
    case 2:
        for(i=0; i<h; i++){
            ((uint32_t*)dst)[0] = scale*((uint32_t*)src)[0] + dc;
            ((uint32_t*)dst)[1] = scale*((uint32_t*)src)[1] + dc;
            if(scale) src += stride;
            dst += stride;
        }
        break;
    case 3:
        for(i=0; i<h; i++){
            ((uint32_t*)dst)[0] = scale*((uint32_t*)src)[0] + dc;
            ((uint32_t*)dst)[1] = scale*((uint32_t*)src)[1] + dc;
            ((uint32_t*)dst)[2] = scale*((uint32_t*)src)[2] + dc;
            ((uint32_t*)dst)[3] = scale*((uint32_t*)src)[3] + dc;
            if(scale) src += stride;
            dst += stride;
        }
        break;
    default: assert(0);
    }
}

static void decode_p_block(FourXContext *f, uint16_t *dst, uint16_t *src, int log2w, int log2h, int stride){
    const int index= size2index[log2h][log2w];
    const int h= 1<<log2h;
    int code= get_vlc2(&f->gb, block_type_vlc[index].table, BLOCK_TYPE_VLC_BITS, 1);
    
    assert(code>=0 && code<=6);

    if(code == 0){
        src += f->mv[ *f->bytestream++ ];
        mcdc(dst, src, log2w, h, stride, 1, 0);
    }else if(code == 1){
        log2h--;
        decode_p_block(f, dst                  , src                  , log2w, log2h, stride);
        decode_p_block(f, dst + (stride<<log2h), src + (stride<<log2h), log2w, log2h, stride);
    }else if(code == 2){
        log2w--;
        decode_p_block(f, dst             , src             , log2w, log2h, stride);
        decode_p_block(f, dst + (1<<log2w), src + (1<<log2w), log2w, log2h, stride);
    }else if(code == 4){
        src += f->mv[ *f->bytestream++ ];
        mcdc(dst, src, log2w, h, stride, 1, le2me_16(*f->wordstream++));
    }else if(code == 5){
        mcdc(dst, src, log2w, h, stride, 0, le2me_16(*f->wordstream++));
    }else if(code == 6){
        if(log2w){
            dst[0] = le2me_16(*f->wordstream++);
            dst[1] = le2me_16(*f->wordstream++);
        }else{
            dst[0     ] = le2me_16(*f->wordstream++);
            dst[stride] = le2me_16(*f->wordstream++);
        }
    }
}

static int get32(void *p){
    return le2me_32(*(uint32_t*)p);
}

static int decode_p_frame(FourXContext *f, uint8_t *buf, int length){
    int x, y;
    const int width= f->avctx->width;
    const int height= f->avctx->height;
    uint16_t *src= (uint16_t*)f->last_picture.data[0];
    uint16_t *dst= (uint16_t*)f->current_picture.data[0];
    const int stride= f->current_picture.linesize[0]>>1;
    const int bitstream_size= get32(buf+8);
    const int bytestream_size= get32(buf+16);
    const int wordstream_size= get32(buf+12);
    
    if(bitstream_size+ bytestream_size+ wordstream_size + 20 != length)
        av_log(f->avctx, AV_LOG_ERROR, "lengths %d %d %d %d\n", bitstream_size, bytestream_size, wordstream_size, 
        bitstream_size+ bytestream_size+ wordstream_size - length);
    
    f->bitstream_buffer= av_fast_realloc(f->bitstream_buffer, &f->bitstream_buffer_size, bitstream_size + FF_INPUT_BUFFER_PADDING_SIZE);
    f->dsp.bswap_buf((uint32_t*)f->bitstream_buffer, (uint32_t*)(buf + 20), bitstream_size/4);
    init_get_bits(&f->gb, f->bitstream_buffer, 8*bitstream_size);

    f->wordstream= (uint16_t*)(buf + 20 + bitstream_size);
    f->bytestream= buf + 20 + bitstream_size + wordstream_size;
    
    init_mv(f);
    
    for(y=0; y<height; y+=8){
        for(x=0; x<width; x+=8){
            decode_p_block(f, dst + x, src + x, 3, 3, stride);
        }
        src += 8*stride; 
        dst += 8*stride; 
    }
    
    if(bitstream_size != (get_bits_count(&f->gb)+31)/32*4)
        av_log(f->avctx, AV_LOG_ERROR, " %d %d %d bytes left\n", 
            bitstream_size - (get_bits_count(&f->gb)+31)/32*4, 
            bytestream_size - (f->bytestream - (buf + 20 + bitstream_size + wordstream_size)),
            wordstream_size - (((uint8_t*)f->wordstream) - (buf + 20 + bitstream_size))
        );
    
    return 0;
}

/**
 * decode block and dequantize.
 * Note this is allmost identical to mjpeg
 */
static int decode_i_block(FourXContext *f, DCTELEM *block){
    int code, i, j, level, val;

    /* DC coef */
    val = get_vlc2(&f->pre_gb, f->pre_vlc.table, ACDC_VLC_BITS, 3);
    if (val>>4){
        av_log(f->avctx, AV_LOG_ERROR, "error dc run != 0\n");
    }

    if(val)