huffyuv
authorMichael Niedermayer <michaelni@gmx.at>
Thu, 14 Nov 2002 19:20:04 +0000 (19:20 +0000)
committerMichael Niedermayer <michaelni@gmx.at>
Thu, 14 Nov 2002 19:20:04 +0000 (19:20 +0000)
Originally committed as revision 1211 to svn://svn.ffmpeg.org/ffmpeg/trunk

libavcodec/Makefile
libavcodec/allcodecs.c
libavcodec/avcodec.h
libavcodec/dsputil.c
libavcodec/dsputil.h
libavcodec/huffyuv.c [new file with mode: 0644]
libavcodec/i386/dsputil_mmx.c

index a79111d..4652aeb 100644 (file)
@@ -16,7 +16,7 @@ OBJS= common.o utils.o mem.o allcodecs.o \
       motion_est.o imgconvert.o imgresample.o msmpeg4.o \
       mpeg12.o h263dec.o svq1.o rv10.o mpegaudiodec.o pcm.o simple_idct.o \
       ratecontrol.o adpcm.o eval.o dv.o error_resilience.o \
-      wmadec.o fft.o mdct.o mace.o
+      wmadec.o fft.o mdct.o mace.o huffyuv.o
 ASM_OBJS=
 
 # currently using liba52 for ac3 decoding
index f07eff9..384326e 100644 (file)
@@ -53,6 +53,7 @@ void avcodec_register_all(void)
     register_avcodec(&msmpeg4v3_encoder);
     register_avcodec(&wmv1_encoder);
     register_avcodec(&wmv2_encoder);
+    register_avcodec(&huffyuv_encoder);
 #endif /* CONFIG_ENCODERS */
     register_avcodec(&rawvideo_codec);
 
@@ -78,6 +79,7 @@ void avcodec_register_all(void)
     register_avcodec(&wmav2_decoder);
     register_avcodec(&mace3_decoder);
     register_avcodec(&mace6_decoder);
+    register_avcodec(&huffyuv_decoder);
 #ifdef CONFIG_AC3
     register_avcodec(&ac3_decoder);
 #endif
index 3e7bb46..8ca71a0 100644 (file)
@@ -5,8 +5,8 @@
 
 #define LIBAVCODEC_VERSION_INT 0x000406
 #define LIBAVCODEC_VERSION     "0.4.6"
-#define LIBAVCODEC_BUILD       4638
-#define LIBAVCODEC_BUILD_STR   "4638"
+#define LIBAVCODEC_BUILD       4639
+#define LIBAVCODEC_BUILD_STR   "4639"
 
 enum CodecID {
     CODEC_ID_NONE, 
@@ -34,6 +34,7 @@ enum CodecID {
     CODEC_ID_WMAV2,
     CODEC_ID_MACE3,
     CODEC_ID_MACE6,
+    CODEC_ID_HUFFYUV,
 
     /* various pcm "codecs" */
     CODEC_ID_PCM_S16LE,
@@ -770,6 +771,23 @@ typedef struct AVCodecContext {
      * CPU features (i.e. MMX, SSE. ...)
      */
      unsigned dsp_mask;
+
+    /**
+     * bits per sample/pixel from the demuxer (needed for huffyuv)
+     * encoding; set by lavc
+     * decoding: set by user
+     */
+     int bits_per_sample;
+    
+    /**
+     * prediction method (needed for huffyuv)
+     * encoding; set by user
+     * decoding: unused
+     */
+     int prediction_method;
+#define FF_PRED_LEFT   0
+#define FF_PRED_PLANE  1
+#define FF_PRED_MEDIAN 2
 } AVCodecContext;
 
 typedef struct AVCodec {
@@ -810,6 +828,7 @@ extern AVCodec msmpeg4v2_encoder;
 extern AVCodec msmpeg4v3_encoder;
 extern AVCodec wmv1_encoder;
 extern AVCodec wmv2_encoder;
+extern AVCodec huffyuv_encoder;
 
 extern AVCodec h263_decoder;
 extern AVCodec mpeg4_decoder;
@@ -831,6 +850,7 @@ extern AVCodec mp2_decoder;
 extern AVCodec mp3_decoder;
 extern AVCodec mace3_decoder;
 extern AVCodec mace6_decoder;
+extern AVCodec huffyuv_decoder;
 
 /* pcm codecs */
 #define PCM_CODEC(id, name) \
index affdfc5..5952d92 100644 (file)
@@ -1342,6 +1342,38 @@ static void clear_blocks_c(DCTELEM *blocks)
     memset(blocks, 0, sizeof(DCTELEM)*6*64);
 }
 
+static void add_bytes_c(uint8_t *dst, uint8_t *src, int w){
+    int i;
+    for(i=0; i+7<w; i++){
+        dst[i+0] += src[i+0];
+        dst[i+1] += src[i+1];
+        dst[i+2] += src[i+2];
+        dst[i+3] += src[i+3];
+        dst[i+4] += src[i+4];
+        dst[i+5] += src[i+5];
+        dst[i+6] += src[i+6];
+        dst[i+7] += src[i+7];
+    }
+    for(; i<w; i++)
+        dst[i+0] += src[i+0];
+}
+
+static void diff_bytes_c(uint8_t *dst, uint8_t *src1, uint8_t *src2, int w){
+    int i;
+    for(i=0; i+7<w; i++){
+        dst[i+0] = src1[i+0]-src2[i+0];
+        dst[i+1] = src1[i+1]-src2[i+1];
+        dst[i+2] = src1[i+2]-src2[i+2];
+        dst[i+3] = src1[i+3]-src2[i+3];
+        dst[i+4] = src1[i+4]-src2[i+4];
+        dst[i+5] = src1[i+5]-src2[i+5];
+        dst[i+6] = src1[i+6]-src2[i+6];
+        dst[i+7] = src1[i+7]-src2[i+7];
+    }
+    for(; i<w; i++)
+        dst[i+0] = src1[i+0]-src2[i+0];
+}
+
 void dsputil_init(DSPContext* c, unsigned mask)
 {
     static int init_done = 0;
@@ -1431,6 +1463,9 @@ void dsputil_init(DSPContext* c, unsigned mask)
     /* dspfunc(avg_no_rnd_qpel, 1, 8); */
 #undef dspfunc
 
+    c->add_bytes= add_bytes_c;
+    c->diff_bytes= diff_bytes_c;
+
 #ifdef HAVE_MMX
     dsputil_init_mmx(c, mask);
     if (ff_bit_exact)
index 2ff6fb9..4f036b0 100644 (file)
@@ -117,6 +117,11 @@ typedef struct DSPContext {
     op_pixels_abs_func pix_abs8x8_x2;
     op_pixels_abs_func pix_abs8x8_y2;
     op_pixels_abs_func pix_abs8x8_xy2;
+    
+    /* huffyuv specific */
+    //FIXME note: alignment isnt guranteed currently but could be if needed
+    void (*add_bytes)(uint8_t *dst/*align 16*/, uint8_t *src/*align 16*/, int w);
+    void (*diff_bytes)(uint8_t *dst/*align 16*/, uint8_t *src1/*align 16*/, uint8_t *src2/*align 16*/,int w);
 } DSPContext;
 
 void dsputil_init(DSPContext* p, unsigned mask);
diff --git a/libavcodec/huffyuv.c b/libavcodec/huffyuv.c
new file mode 100644 (file)
index 0000000..ad8b814
--- /dev/null
@@ -0,0 +1,978 @@
+/*
+ * huffyuv codec for libavcodec
+ *
+ * Copyright (c) 2002 Michael Niedermayer <michaelni@gmx.at>
+ *
+ * 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
+ *
+ * see http://www.pcisys.net/~melanson/codecs/huffyuv.txt for a description of
+ * the algorithm used 
+ */
+
+#include "common.h"
+#include "avcodec.h"
+#include "dsputil.h"
+
+#ifndef MAX_INT64
+#define MAX_INT64 9223372036854775807LL
+#endif
+
+#define VLC_BITS 11
+  
+typedef enum Predictor{
+    LEFT= 0,
+    PLANE,
+    MEDIAN,
+} Predictor;
+typedef struct HYuvContext{
+    AVCodecContext *avctx;
+    Predictor predictor;
+    GetBitContext gb;
+    PutBitContext pb;
+    int interlaced;
+    int decorrelate;
+    int bitstream_bpp;
+    int version;
+    int yuy2;                               //use yuy2 instead of 422P
+    int bgr32;                              //use bgr32 instead of bgr24
+    int width, height;
+    int flags;
+    int picture_number;
+    int linesize[3];
+    uint8_t __align8 temp[3][2500];
+    uint64_t stats[3][256];
+    uint8_t len[3][256];
+    uint32_t bits[3][256];
+    VLC vlc[3];
+    uint8_t __align8 *picture[3];
+    uint8_t __align8 bitstream_buffer[1024*1024*3]; //FIXME dynamic alloc or some other solution
+    DSPContext dsp; 
+}HYuvContext;
+
+static inline void bswap_buf(uint32_t *dst, uint32_t *src, int w){
+    int i;
+    
+    for(i=0; i+8<=w; i+=8){
+        dst[i+0]= bswap_32(src[i+0]);
+        dst[i+1]= bswap_32(src[i+1]);
+        dst[i+2]= bswap_32(src[i+2]);
+        dst[i+3]= bswap_32(src[i+3]);
+        dst[i+4]= bswap_32(src[i+4]);
+        dst[i+5]= bswap_32(src[i+5]);
+        dst[i+6]= bswap_32(src[i+6]);
+        dst[i+7]= bswap_32(src[i+7]);
+    }
+    for(;i<w; i++){
+        dst[i+0]= bswap_32(src[i+0]);
+    }
+}
+
+static inline int add_left_prediction(uint8_t *dst, uint8_t *src, int w, int acc){
+    int i;
+
+    for(i=0; i<w-1; i++){
+        acc+= src[i];
+        dst[i]= acc;
+        i++;
+        acc+= src[i];
+        dst[i]= acc;
+    }
+
+    for(; i<w; i++){
+        acc+= src[i];
+        dst[i]= acc;
+    }
+
+    return acc;
+}
+
+static inline void add_median_prediction(uint8_t *dst, uint8_t *src1, uint8_t *diff, int w, int *left, int *left_top){
+    int i;
+    uint8_t l, lt;
+
+    l= *left;
+    lt= *left_top;
+
+    for(i=0; i<w; i++){
+        l= mid_pred(l, src1[i], (l + src1[i] - lt)&0xFF) + diff[i];
+        lt= src1[i];
+        dst[i]= l;
+    }    
+
+    *left= l;
+    *left_top= lt;
+}
+//FIXME optimize
+static inline void sub_median_prediction(uint8_t *dst, uint8_t *src1, uint8_t *src2, int w, int *left, int *left_top){
+    int i;
+    uint8_t l, lt;
+
+    l= *left;
+    lt= *left_top;
+
+    for(i=0; i<w; i++){
+        const int pred= mid_pred(l, src1[i], (l + src1[i] - lt)&0xFF);
+        lt= src1[i];
+        l= src2[i];
+        dst[i]= l - pred;
+    }    
+
+    *left= l;
+    *left_top= lt;
+}
+
+
+static inline void add_left_prediction_bgr32(uint8_t *dst, uint8_t *src, int w, int *red, int *green, int *blue){
+    int i;
+    int r,g,b;
+    r= *red;
+    g= *green;
+    b= *blue;
+
+    for(i=0; i<w; i++){
+        b+= src[4*i+0];
+        g+= src[4*i+1];
+        r+= src[4*i+2];
+        
+        dst[4*i+0]= b;
+        dst[4*i+1]= g;
+        dst[4*i+2]= r;
+    }
+
+    *red= r;
+    *green= g;
+    *blue= b;
+}
+
+//FIXME optimize
+static inline int sub_left_prediction(uint8_t *dst, uint8_t *src, int w, int left){
+    int i;
+    for(i=0; i<w; i++){
+        const int temp= src[i];
+        dst[i]= temp - left;
+        left= temp;
+    }
+    return left;
+}
+
+static void read_len_table(uint8_t *dst, GetBitContext *gb){
+    int i, val, repeat;
+  
+    for(i=0; i<256;){
+        repeat= get_bits(gb, 3);
+        val   = get_bits(gb, 5);
+        if(repeat==0)
+            repeat= get_bits(gb, 8);
+//printf("%d %d\n", val, repeat);
+        while (repeat--)
+            dst[i++] = val;
+    }
+}
+
+static int generate_bits_table(uint32_t *dst, uint8_t *len_table){
+    int len, index;
+    uint32_t bits=0;
+
+    for(len=32; len>0; len--){
+        int bit= 1<<(32-len);
+        for(index=0; index<256; index++){
+            if(len_table[index]==len){
+                if(bits & (bit-1)){
+                    fprintf(stderr, "Error generating huffman table\n");
+                    return -1;
+                }
+                dst[index]= bits>>(32-len);
+                bits+= bit;
+            }
+        }
+    }
+    return 0;
+}
+
+static void generate_len_table(uint8_t *dst, uint64_t *stats, int size){
+    uint64_t counts[2*size];
+    int up[2*size];
+    int offset, i, next;
+    
+    for(offset=1; ; offset<<=1){
+        for(i=0; i<size; i++){
+            counts[i]= stats[i] + offset - 1;
+        }
+        
+        for(next=size; next<size*2; next++){
+            uint64_t min1, min2;
+            int min1_i, min2_i;
+            
+            min1=min2= INT64_MAX;
+            min1_i= min2_i=-1;
+            
+            for(i=0; i<next; i++){
+                if(min2 > counts[i]){
+                    if(min1 > counts[i]){
+                        min2= min1;
+                        min2_i= min1_i;
+                        min1= counts[i];
+                        min1_i= i;
+                    }else{
+                        min2= counts[i];
+                        min2_i= i;
+                    }
+                }
+            }
+            
+            if(min2==INT64_MAX) break;
+            
+            counts[next]= min1 + min2;
+            counts[min1_i]=
+            counts[min2_i]= MAX_INT64;
+            up[min1_i]=
+            up[min2_i]= next;
+            up[next]= -1;
+        }
+        
+        for(i=0; i<size; i++){
+            int len;
+            int index=i;
+            
+            for(len=0; up[index] != -1; len++)
+                index= up[index];
+                
+            if(len > 32) break;
+            
+            dst[i]= len;
+        }
+        if(i==size) break;
+    }
+}
+
+static int read_huffman_tables(HYuvContext *s, uint8_t *src, int length){
+    GetBitContext gb;
+    int i;
+    
+    init_get_bits(&gb, src, length);
+    
+    for(i=0; i<3; i++){
+        read_len_table(s->len[i], &gb);
+        
+        if(generate_bits_table(s->bits[i], s->len[i])<0){
+            return -1;
+        }
+#if 0
+for(j=0; j<256; j++){
+printf("%6X, %2d,  %3d\n", s->bits[i][j], s->len[i][j], j);
+}
+#endif
+        init_vlc(&s->vlc[i], VLC_BITS, 256, s->len[i], 1, 1, s->bits[i], 4, 4);
+    }
+    
+    return 0;
+}
+
+static int read_old_huffman_tables(HYuvContext *s){
+#if 0    
+    GetBitContext gb;
+    int i;
+
+    init_get_bits(&gb, classic_shift_luma, sizeof(classic_shift_luma));
+    read_len_table(s->len[0], &gb);
+    init_get_bits(&gb, classic_shift_chroma, sizeof(classic_shift_chroma));
+    read_len_table(s->len[1], &gb);
+    
+    for(i=0; i<256; i++) s->bits[0][i] = classic_add_luma  [i];
+    for(i=0; i<256; i++) s->bits[1][i] = classic_add_chroma[i];
+
+    if(s->bitstream_bpp >= 24){
+        memcpy(s->bits[1], s->bits[0], 256*sizeof(uint32_t));
+        memcpy(s->len[1] , s->len [0], 256*sizeof(uint8_t));
+    }
+    memcpy(s->bits[2], s->bits[1], 256*sizeof(uint32_t));
+    memcpy(s->len[2] , s->len [1], 256*sizeof(uint8_t));
+    
+    for(i=0; i<3; i++)
+        init_vlc(&s->vlc[i], VLC_BITS, 256, s->len[i], 1, 1, s->bits[i], 4, 4);
+    
+    return 0;
+#else
+    fprintf(stderr, "v1 huffyuv is not supported \n");
+    return -1;
+#endif
+}
+
+static int decode_init(AVCodecContext *avctx)
+{
+    HYuvContext *s = avctx->priv_data;
+    int width, height, y_size, c_size, stride;
+
+    s->avctx= avctx;
+    s->flags= avctx->flags;
+        
+    dsputil_init(&s->dsp, avctx->dsp_mask);
+    
+    width= s->width= avctx->width;
+    height= s->height= avctx->height;
+s->bgr32=1;
+    assert(width && height);
+//if(avctx->extradata)
+//  printf("extradata:%X, extradata_size:%d\n", *(uint32_t*)avctx->extradata, avctx->extradata_size);
+    if(avctx->extradata_size){
+        if(avctx->bits_per_sample&7)
+            s->version=1; // do such files exist at all?
+        else
+            s->version=2;
+    }else
+        s->version=0;
+    
+    if(s->version==2){
+        int method;
+
+        method= ((uint8_t*)avctx->extradata)[0];
+        s->decorrelate= method&64 ? 1 : 0;
+        s->predictor= method&63;
+        s->bitstream_bpp= ((uint8_t*)avctx->extradata)[1];
+        if(s->bitstream_bpp==0) 
+            s->bitstream_bpp= avctx->bits_per_sample&~7;
+            
+        if(read_huffman_tables(s, ((uint8_t*)avctx->extradata)+4, avctx->extradata_size) < 0)
+            return -1;
+    }else{
+        switch(avctx->bits_per_sample&7){
+        case 1:
+            s->predictor= LEFT;
+            s->decorrelate= 0;
+            break;
+        case 2:
+            s->predictor= LEFT;
+            s->decorrelate= 1;
+            break;
+        case 3:
+            s->predictor= PLANE;
+            s->decorrelate= avctx->bits_per_sample >= 24;
+            break;
+        case 4:
+            s->predictor= MEDIAN;
+            s->decorrelate= 0;
+            break;
+        default:
+            s->predictor= LEFT; //OLD
+            s->decorrelate= 0;
+            break;
+        }
+        s->bitstream_bpp= avctx->bits_per_sample & ~7;
+        
+        if(read_old_huffman_tables(s) < 0)
+            return -1;
+    }
+    
+    s->interlaced= height > 288;
+    
+    c_size= 0;
+    switch(s->bitstream_bpp){
+    case 12:
+        avctx->pix_fmt = PIX_FMT_YUV420P;
+        stride= (width+15)&~15;
+        c_size= height*stride/4;
+        break;
+    case 16:
+        if(s->yuy2){
+            avctx->pix_fmt = PIX_FMT_YUV422;
+            stride= (width*2+15)&~15;
+        }else{
+            avctx->pix_fmt = PIX_FMT_YUV422P;
+            stride= (width+15)&~15;
+            c_size= height*stride/2;
+        }
+        break;
+    case 24:
+    case 32:
+        if(s->bgr32){
+            avctx->pix_fmt = PIX_FMT_BGRA32;
+            stride= (width*4+15)&~15;
+        }else{
+            avctx->pix_fmt = PIX_FMT_BGR24;
+            stride= (width*3+15)&~15;
+        }
+        break;
+    default:
+        assert(0);
+        stride=0; //gcc fix
+    }
+    
+    y_size= height*stride;
+    
+    s->linesize[0]= stride;
+    s->picture[0]= av_mallocz(y_size);
+    if(c_size){
+        s->picture[1]= av_mallocz(c_size);
+        s->picture[2]= av_mallocz(c_size);
+        s->linesize[1]= s->linesize[2]= stride/2;
+        
+        memset(s->picture[1], 128, c_size);
+        memset(s->picture[2], 128, c_size);
+    }
+    
+//    printf("pred:%d bpp:%d hbpp:%d il:%d\n", s->predictor, s->bitstream_bpp, avctx->bits_per_sample, s->interlaced);
+    
+    return 0;
+}
+
+static void store_table(HYuvContext *s, uint8_t *len){
+    int i;
+    int index= s->avctx->extradata_size;
+
+    for(i=0; i<256;){
+        int cur=i;
+        int val= len[i];
+        int repeat;
+        
+        for(; i<256 && len[i]==val; i++);
+        
+        repeat= i - cur;
+        
+        if(repeat>7){
+            ((uint8_t*)s->avctx->extradata)[index++]= val;
+            ((uint8_t*)s->avctx->extradata)[index++]= repeat;
+        }else{
+            ((uint8_t*)s->avctx->extradata)[index++]= val | (repeat<<5);
+        }
+    }
+    
+    s->avctx->extradata_size= index;
+}
+
+static int encode_init(AVCodecContext *avctx)
+{
+    HYuvContext *s = avctx->priv_data;
+    int i, j, width, height;
+
+    s->avctx= avctx;
+    s->flags= avctx->flags;
+        
+    dsputil_init(&s->dsp, avctx->dsp_mask);
+    
+    width= s->width= avctx->width;
+    height= s->height= avctx->height;
+    
+    assert(width && height);
+    
+    avctx->extradata= av_mallocz(1024*10);
+    avctx->stats_out= av_mallocz(1024*10);
+    s->version=2;
+    
+    switch(avctx->pix_fmt){
+    case PIX_FMT_YUV422P:
+        s->bitstream_bpp= 16;
+        break;
+    default:
+        fprintf(stderr, "format not supported\n");
+        return -1;
+    }
+    avctx->bits_per_sample= s->bitstream_bpp;
+    s->decorrelate= s->bitstream_bpp >= 24;
+    s->predictor= avctx->prediction_method;
+    
+    ((uint8_t*)avctx->extradata)[0]= s->predictor;
+    ((uint8_t*)avctx->extradata)[1]= s->bitstream_bpp;
+    ((uint8_t*)avctx->extradata)[2]=
+    ((uint8_t*)avctx->extradata)[3]= 0;
+    s->avctx->extradata_size= 4;
+    
+    if(avctx->stats_in){
+        char *p= avctx->stats_in;
+    
+        for(i=0; i<3; i++)
+            for(j=0; j<256; j++)
+                s->stats[i][j]= 1;
+
+        for(;;){
+            for(i=0; i<3; i++){
+                char *next;
+
+                for(j=0; j<256; j++){
+                    s->stats[i][j]+= strtol(p, &next, 0);
+                    if(next==p) return -1;
+                    p=next;
+                }        
+            }
+            if(p[0]==0 || p[1]==0 || p[2]==0) break;
+        }
+    }else{
+        for(i=0; i<3; i++)
+            for(j=0; j<256; j++){
+                int d= FFMIN(j, 256-j);
+                
+                s->stats[i][j]= 100000000/(d+1);
+            }
+    }
+    
+    for(i=0; i<3; i++){
+        generate_len_table(s->len[i], s->stats[i], 256);
+
+        if(generate_bits_table(s->bits[i], s->len[i])<0){
+            return -1;
+        }
+        
+        store_table(s, s->len[i]);
+    }
+
+    for(i=0; i<3; i++)
+        for(j=0; j<256; j++)
+            s->stats[i][j]= 0;
+    
+    s->interlaced= height > 288;
+    
+//    printf("pred:%d bpp:%d hbpp:%d il:%d\n", s->predictor, s->bitstream_bpp, avctx->bits_per_sample, s->interlaced);
+    
+    s->picture_number=0;
+    
+    return 0;
+}
+
+static void decode_422_bitstream(HYuvContext *s, int count){
+    int i;
+    
+    count/=2;
+    
+    for(i=0; i<count; i++){
+        s->temp[0][2*i  ]= get_vlc2(&s->gb, s->vlc[0].table, VLC_BITS, 3); 
+        s->temp[1][  i  ]= get_vlc2(&s->gb, s->vlc[1].table, VLC_BITS, 3); 
+        s->temp[0][2*i+1]= get_vlc2(&s->gb, s->vlc[0].table, VLC_BITS, 3); 
+        s->temp[2][  i  ]= get_vlc2(&s->gb, s->vlc[2].table, VLC_BITS, 3); 
+    }
+}
+
+static void encode_422_bitstream(HYuvContext *s, int count){
+    int i;
+    
+    count/=2;
+    if(s->flags&CODEC_FLAG_PASS1){
+        for(i=0; i<count; i++){
+            s->stats[0][ s->temp[0][2*i  ] ]++;
+            s->stats[1][ s->temp[1][  i  ] ]++;
+            s->stats[0][ s->temp[0][2*i+1] ]++;
+            s->stats[2][ s->temp[2][  i  ] ]++;
+        }
+    }else{
+        for(i=0; i<count; i++){
+            put_bits(&s->pb, s->len[0][ s->temp[0][2*i  ] ], s->bits[0][ s->temp[0][2*i  ] ]);
+            put_bits(&s->pb, s->len[1][ s->temp[1][  i  ] ], s->bits[1][ s->temp[1][  i  ] ]);
+            put_bits(&s->pb, s->len[0][ s->temp[0][2*i+1] ], s->bits[0][ s->temp[0][2*i+1] ]);
+            put_bits(&s->pb, s->len[2][ s->temp[2][  i  ] ], s->bits[2][ s->temp[2][  i  ] ]);
+        }
+    }
+}
+
+static void decode_bgr_bitstream(HYuvContext *s, int count){
+    int i;
+    
+    if(s->decorrelate){
+        if(s->bitstream_bpp==24){
+            for(i=0; i<count; i++){
+                s->temp[0][4*i+1]= get_vlc2(&s->gb, s->vlc[1].table, VLC_BITS, 3); 
+                s->temp[0][4*i  ]= get_vlc2(&s->gb, s->vlc[0].table, VLC_BITS, 3) + s->temp[0][4*i+1];
+                s->temp[0][4*i+2]= get_vlc2(&s->gb, s->vlc[2].table, VLC_BITS, 3) + s->temp[0][4*i+1];
+            }
+        }else{
+            for(i=0; i<count; i++){
+                s->temp[0][4*i+1]= get_vlc2(&s->gb, s->vlc[1].table, VLC_BITS, 3); 
+                s->temp[0][4*i  ]= get_vlc2(&s->gb, s->vlc[0].table, VLC_BITS, 3) + s->temp[0][4*i+1];
+                s->temp[0][4*i+2]= get_vlc2(&s->gb, s->vlc[2].table, VLC_BITS, 3) + s->temp[0][4*i+1]; 
+                                   get_vlc2(&s->gb, s->vlc[2].table, VLC_BITS, 3); //?!
+            }
+        }
+    }else{
+        if(s->bitstream_bpp==24){
+            for(i=0; i<count; i++){
+                s->temp[0][4*i  ]= get_vlc2(&s->gb, s->vlc[0].table, VLC_BITS, 3);
+                s->temp[0][4*i+1]= get_vlc2(&s->gb, s->vlc[1].table, VLC_BITS, 3); 
+                s->temp[0][4*i+2]= get_vlc2(&s->gb, s->vlc[2].table, VLC_BITS, 3); 
+            }
+        }else{
+            for(i=0; i<count; i++){
+                s->temp[0][4*i  ]= get_vlc2(&s->gb, s->vlc[0].table, VLC_BITS, 3);
+                s->temp[0][4*i+1]= get_vlc2(&s->gb, s->vlc[1].table, VLC_BITS, 3); 
+                s->temp[0][4*i+2]= get_vlc2(&s->gb, s->vlc[2].table, VLC_BITS, 3); 
+                                   get_vlc2(&s->gb, s->vlc[2].table, VLC_BITS, 3); //?!
+            }
+        }
+    }
+}
+
+static int decode_frame(AVCodecContext *avctx, void *data, int *data_size, uint8_t *buf, int buf_size){
+    HYuvContext *s = avctx->priv_data;
+    const int width= s->width;
+    const int width2= s->width>>1;
+    const int height= s->height;
+    const int fake_ystride= s->interlaced ? s->linesize[0]*2  : s->linesize[0];
+    const int fake_ustride= s->interlaced ? s->linesize[1]*2  : s->linesize[1];
+    const int fake_vstride= s->interlaced ? s->linesize[2]*2  : s->linesize[2];
+    int i;
+
+    AVPicture *picture = data;
+
+    *data_size = 0;
+
+    /* no supplementary picture */
+    if (buf_size == 0)
+        return 0;
+
+    bswap_buf((uint32_t*)s->bitstream_buffer, (uint32_t*)buf, buf_size/4);
+    
+    init_get_bits(&s->gb, s->bitstream_buffer, buf_size);
+    
+    if(s->bitstream_bpp<24){
+        int y;
+        int lefty, leftu, leftv;
+        int lefttopy, lefttopu, lefttopv;
+        
+        if(s->yuy2){
+            s->picture[0][3]= get_bits(&s->gb, 8);
+            s->picture[0][2]= get_bits(&s->gb, 8);
+            s->picture[0][1]= get_bits(&s->gb, 8);
+            s->picture[0][0]= get_bits(&s->gb, 8);
+            
+            fprintf(stderr, "YUY2 output isnt implemenetd yet\n");
+            return -1;
+        }else{
+        
+            leftv= s->picture[2][0]= get_bits(&s->gb, 8);
+            lefty= s->picture[0][1]= get_bits(&s->gb, 8);
+            leftu= s->picture[1][0]= get_bits(&s->gb, 8);
+                   s->picture[0][0]= get_bits(&s->gb, 8);
+        
+            switch(s->predictor){
+            case LEFT:
+            case PLANE:
+                decode_422_bitstream(s, width-2);
+                lefty= add_left_prediction(s->picture[0] + 2, s->temp[0], width-2, lefty);
+                if(!(s->flags&CODEC_FLAG_GRAY)){
+                    leftu= add_left_prediction(s->picture[1] + 1, s->temp[1], width2-1, leftu);
+                    leftv= add_left_prediction(s->picture[2] + 1, s->temp[2], width2-1, leftv);
+                }
+
+                for(y=1; y<s->height; y++){
+                    uint8_t *ydst, *udst, *vdst;
+                    decode_422_bitstream(s, width);
+                    
+                    ydst= s->picture[0] + s->linesize[0]*y;
+                    udst= s->picture[1] + s->linesize[1]*y;
+                    vdst= s->picture[2] + s->linesize[2]*y;
+
+                    lefty= add_left_prediction(ydst, s->temp[0], width, lefty);
+                    if(!(s->flags&CODEC_FLAG_GRAY)){
+                        leftu= add_left_prediction(udst, s->temp[1], width2, leftu);
+                        leftv= add_left_prediction(vdst, s->temp[2], width2, leftv);
+                    }
+                    if(s->predictor == PLANE){
+                        if(y>s->interlaced){
+                            s->dsp.add_bytes(ydst, ydst - fake_ystride, width);
+                            if(!(s->flags&CODEC_FLAG_GRAY)){
+                                s->dsp.add_bytes(udst, udst - fake_ustride, width2);
+                                s->dsp.add_bytes(vdst, vdst - fake_vstride, width2);
+                            }
+                        }
+                    }
+                }
+                break;
+            case MEDIAN:
+                /* first line except first 2 pixels is left predicted */
+                decode_422_bitstream(s, width-2);
+                lefty= add_left_prediction(s->picture[0] + 2, s->temp[0], width-2, lefty);
+                if(!(s->flags&CODEC_FLAG_GRAY)){
+                    leftu= add_left_prediction(s->picture[1] + 1, s->temp[1], width2-1, leftu);
+                    leftv= add_left_prediction(s->picture[2] + 1, s->temp[2], width2-1, leftv);
+                }
+                
+                y=1;
+                
+                /* second line is left predicted for interlaced case */
+                if(s->interlaced){
+                    decode_422_bitstream(s, width);
+                    lefty= add_left_prediction(s->picture[0] + s->linesize[0], s->temp[0], width, lefty);
+                    if(!(s->flags&CODEC_FLAG_GRAY)){
+                        leftu= add_left_prediction(s->picture[1] + s->linesize[2], s->temp[1], width2, leftu);
+                        leftv= add_left_prediction(s->picture[2] + s->linesize[1], s->temp[2], width2, leftv);
+                    }
+                    y++;
+                }
+
+                /* next 4 pixels are left predicted too */
+                decode_422_bitstream(s, 4);
+                lefty= add_left_prediction(s->picture[0] + fake_ystride, s->temp[0], 4, lefty);
+                if(!(s->flags&CODEC_FLAG_GRAY)){
+                    leftu= add_left_prediction(s->picture[1] + fake_ustride, s->temp[1], 2, leftu);
+                    leftv= add_left_prediction(s->picture[2] + fake_vstride, s->temp[2], 2, leftv);
+                }
+
+                /* next line except the first 4 pixels is median predicted */
+                lefttopy= s->picture[0][3];
+                decode_422_bitstream(s, width-4);
+                add_median_prediction(s->picture[0] + fake_ystride+4, s->picture[0]+4, s->temp[0], width-4, &lefty, &lefttopy);
+                if(!(s->flags&CODEC_FLAG_GRAY)){
+                    lefttopu= s->picture[1][1];
+                    lefttopv= s->picture[2][1];
+                    add_median_prediction(s->picture[1] + fake_ustride+2, s->picture[1]+2, s->temp[1], width2-2, &leftu, &lefttopu);
+                    add_median_prediction(s->picture[2] + fake_vstride+2, s->picture[2]+2, s->temp[2], width2-2, &leftv, &lefttopv);
+                }
+                y++;
+
+                for(; y<height; y++){
+                    uint8_t *ydst, *udst, *vdst;
+                    decode_422_bitstream(s, width);
+                    
+                    ydst= s->picture[0] + s->linesize[0]*y;
+                    udst= s->picture[1] + s->linesize[1]*y;
+                    vdst= s->picture[2] + s->linesize[2]*y;
+
+                    add_median_prediction(ydst, ydst - fake_ystride, s->temp[0], width, &lefty, &lefttopy);
+                    if(!(s->flags&CODEC_FLAG_GRAY)){
+                        add_median_prediction(udst, udst - fake_ustride, s->temp[1], width2, &leftu, &lefttopu);
+                        add_median_prediction(vdst, vdst - fake_vstride, s->temp[2], width2, &leftv, &lefttopv);
+                    }
+                }
+                break;
+            }
+        }
+    }else{
+        int y;
+        int leftr, leftg, leftb;
+        const int last_line= (height-1)*s->linesize[0];
+        
+        if(s->bitstream_bpp==32){
+                   s->picture[0][last_line+3]= get_bits(&s->gb, 8);
+            leftr= s->picture[0][last_line+2]= get_bits(&s->gb, 8);
+            leftg= s->picture[0][last_line+1]= get_bits(&s->gb, 8);
+            leftb= s->picture[0][last_line+0]= get_bits(&s->gb, 8);
+        }else{
+            leftr= s->picture[0][last_line+2]= get_bits(&s->gb, 8);
+            leftg= s->picture[0][last_line+1]= get_bits(&s->gb, 8);
+            leftb= s->picture[0][last_line+0]= get_bits(&s->gb, 8);
+            skip_bits(&s->gb, 8);
+        }
+        
+        if(s->bgr32){
+            switch(s->predictor){
+            case LEFT:
+            case PLANE:
+                decode_bgr_bitstream(s, width-1);
+                add_left_prediction_bgr32(s->picture[0] + last_line+4, s->temp[0], width-1, &leftr, &leftg, &leftb);
+
+                for(y=s->height-2; y>=0; y--){ //yes its stored upside down
+                    decode_bgr_bitstream(s, width);
+                    
+                    add_left_prediction_bgr32(s->picture[0] + s->linesize[0]*y, s->temp[0], width, &leftr, &leftg, &leftb);
+                    if(s->predictor == PLANE){
+                        if((y&s->interlaced)==0){
+                            s->dsp.add_bytes(s->picture[0] + s->linesize[0]*y, 
+                                             s->picture[0] + s->linesize[0]*y + fake_ystride, fake_ystride);
+                        }
+                    }
+                }
+                break;
+            default:
+                fprintf(stderr, "prediction type not supported!\n");
+            }
+        }else{
+
+            fprintf(stderr, "BGR24 output isnt implemenetd yet\n");
+            return -1;
+        }
+    }
+    emms_c();
+    
+    for(i=0;i<3;i++) {
+        picture->data[i] = s->picture[i];
+        picture->linesize[i]= s->linesize[i];
+    }
+
+    *data_size = sizeof(AVPicture);
+    
+    return (get_bits_count(&s->gb)+7)>>3;
+}
+
+static int decode_end(AVCodecContext *avctx)
+{
+    HYuvContext *s = avctx->priv_data;
+    int i;
+    
+    for(i=0; i<3; i++){
+        av_freep(&s->picture[i]);
+        free_vlc(&s->vlc[i]);
+    }
+
+    return 0;
+}
+
+static int encode_frame(AVCodecContext *avctx, unsigned char *buf, int buf_size, void *data){
+    HYuvContext *s = avctx->priv_data;
+    AVPicture *pict = data;
+    const int width= s->width;
+    const int width2= s->width>>1;
+    const int height= s->height;
+    const int fake_ystride= s->interlaced ? pict->linesize[0]*2  : pict->linesize[0];
+    const int fake_ustride= s->interlaced ? pict->linesize[1]*2  : pict->linesize[1];
+    const int fake_vstride= s->interlaced ? pict->linesize[2]*2  : pict->linesize[2];
+    int i, size;
+
+    init_put_bits(&s->pb, buf, buf_size, NULL, NULL);
+    
+    for(i=0; i<3; i++){
+        s->picture[i]= pict->data[i];
+        s->linesize[i]= pict->linesize[i];
+    }
+    
+    if(avctx->pix_fmt == PIX_FMT_YUV422P){
+        int lefty, leftu, leftv, y;
+
+        put_bits(&s->pb, 8, leftv= s->picture[2][0]);
+        put_bits(&s->pb, 8, lefty= s->picture[0][1]);
+        put_bits(&s->pb, 8, leftu= s->picture[1][0]);
+        put_bits(&s->pb, 8,        s->picture[0][0]);
+        
+        lefty= sub_left_prediction(s->temp[0], s->picture[0]+2, width-2 , lefty);
+        leftu= sub_left_prediction(s->temp[1], s->picture[1]+1, width2-1, leftu);
+        leftv= sub_left_prediction(s->temp[2], s->picture[2]+1, width2-1, leftv);
+        
+        encode_422_bitstream(s, width-2);
+        
+        if(s->predictor==MEDIAN){
+            int lefttopy, lefttopu, lefttopv;
+            y=1;
+            if(s->interlaced){
+                lefty= sub_left_prediction(s->temp[0], s->picture[0]+s->linesize[0], width , lefty);
+                leftu= sub_left_prediction(s->temp[1], s->picture[1]+s->linesize[1], width2, leftu);
+                leftv= sub_left_prediction(s->temp[2], s->picture[2]+s->linesize[2], width2, leftv);
+        
+                encode_422_bitstream(s, width);
+                y++;
+            }
+            
+            lefty= sub_left_prediction(s->temp[0], s->picture[0]+fake_ystride, 4, lefty);
+            leftu= sub_left_prediction(s->temp[1], s->picture[1]+fake_ystride, 2, leftu);
+            leftv= sub_left_prediction(s->temp[2], s->picture[2]+fake_ystride, 2, leftv);
+        
+            encode_422_bitstream(s, 4);
+                        
+            lefttopy= s->picture[0][3];
+            lefttopu= s->picture[1][1];
+            lefttopv= s->picture[2][1];
+            sub_median_prediction(s->temp[0], s->picture[0]+4, s->picture[0] + fake_ystride+4, width-4 , &lefty, &lefttopy);
+            sub_median_prediction(s->temp[1], s->picture[1]+2, s->picture[1] + fake_ustride+2, width2-2, &leftu, &lefttopu);
+            sub_median_prediction(s->temp[2], s->picture[2]+2, s->picture[2] + fake_vstride+2, width2-2, &leftv, &lefttopv);
+            encode_422_bitstream(s, width-4);
+            y++;
+
+            for(; y<height; y++){
+                uint8_t *ydst, *udst, *vdst;
+                    
+                ydst= s->picture[0] + s->linesize[0]*y;
+                udst= s->picture[1] + s->linesize[1]*y;
+                vdst= s->picture[2] + s->linesize[2]*y;
+
+                sub_median_prediction(s->temp[0], ydst - fake_ystride, ydst, width , &lefty, &lefttopy);
+                sub_median_prediction(s->temp[1], udst - fake_ustride, udst, width2, &leftu, &lefttopu);
+                sub_median_prediction(s->temp[2], vdst - fake_vstride, vdst, width2, &leftv, &lefttopv);
+
+                encode_422_bitstream(s, width);
+            }
+        }else{
+            for(y=1; y<height; y++){
+                uint8_t *ydst, *udst, *vdst;
+                    
+                ydst= s->picture[0] + s->linesize[0]*y;
+                udst= s->picture[1] + s->linesize[1]*y;
+                vdst= s->picture[2] + s->linesize[2]*y;
+
+                if(s->predictor == PLANE && s->interlaced < y){
+                    s->dsp.diff_bytes(s->temp[0], ydst, ydst - fake_ystride, width);
+                    s->dsp.diff_bytes(s->temp[1], udst, udst - fake_ustride, width2);
+                    s->dsp.diff_bytes(s->temp[2], vdst, vdst - fake_vstride, width2);
+
+                    lefty= sub_left_prediction(s->temp[0], s->temp[0], width , lefty);
+                    leftu= sub_left_prediction(s->temp[1], s->temp[1], width2, leftu);
+                    leftv= sub_left_prediction(s->temp[2], s->temp[2], width2, leftv);
+                }else{
+                    lefty= sub_left_prediction(s->temp[0], ydst, width , lefty);
+                    leftu= sub_left_prediction(s->temp[1], udst, width2, leftu);
+                    leftv= sub_left_prediction(s->temp[2], vdst, width2, leftv);
+                }
+
+                encode_422_bitstream(s, width);
+            }
+        }        
+    }else{
+        fprintf(stderr, "Format not supported!\n");
+    }
+    emms_c();
+    
+    size= (get_bit_count(&s->pb)+31)/32;
+    
+    if((s->flags&CODEC_FLAG_PASS1) && (s->picture_number&31)==0){
+        int j;
+        char *p= avctx->stats_out;
+        for(i=0; i<3; i++){
+            for(j=0; j<256; j++){
+                sprintf(p, "%Ld ", s->stats[i][j]);
+                p+= strlen(p);
+                s->stats[i][j]= 0;
+            }
+            sprintf(p, "\n");
+            p++;
+        }
+    }else{
+        bswap_buf((uint32_t*)buf, (uint32_t*)buf, size);
+    }
+    
+    avctx->key_frame= 1;
+    avctx->pict_type= I_TYPE;
+    
+    s->picture_number++;
+    
+    return size*4;
+}
+
+static int encode_end(AVCodecContext *avctx)
+{
+//    HYuvContext *s = avctx->priv_data;
+
+    av_freep(&avctx->extradata);
+    av_freep(&avctx->stats_out);
+    
+    return 0;
+}
+
+AVCodec huffyuv_decoder = {
+    "huffyuv",
+    CODEC_TYPE_VIDEO,
+    CODEC_ID_HUFFYUV,
+    sizeof(HYuvContext),
+    decode_init,
+    NULL,
+    decode_end,
+    decode_frame,
+    0,
+    NULL
+};
+
+AVCodec huffyuv_encoder = {
+    "huffyuv",
+    CODEC_TYPE_VIDEO,
+    CODEC_ID_HUFFYUV,
+    sizeof(HYuvContext),
+    encode_init,
+    encode_frame,
+    encode_end,
+};
index 94b88b7..12a3601 100644 (file)
@@ -453,6 +453,51 @@ static int pix_sum16_mmx(UINT8 * pix, int line_size){
         return sum;
 }
 
+static void add_bytes_mmx(uint8_t *dst, uint8_t *src, int w){
+    int i=0;
+    asm volatile(
+        "1:                            \n\t"
+        "movq  (%1, %0), %%mm0         \n\t"
+        "movq  (%2, %0), %%mm1         \n\t"
+        "paddb %%mm0, %%mm1            \n\t"
+        "movq %%mm1, (%2, %0)          \n\t"
+        "movq 8(%1, %0), %%mm0         \n\t"
+        "movq 8(%2, %0), %%mm1         \n\t"
+        "paddb %%mm0, %%mm1            \n\t"
+        "movq %%mm1, 8(%2, %0)         \n\t"
+        "addl $16, %0                  \n\t"
+        "cmpl %3, %0                   \n\t"
+        " jb 1b                                \n\t"
+        : "+r" (i)
+        : "r"(src), "r"(dst), "r"(w-15)
+    );
+    for(; i<w; i++)
+        dst[i+0] += src[i+0];
+}
+
+static void diff_bytes_mmx(uint8_t *dst, uint8_t *src1, uint8_t *src2, int w){
+    int i=0;
+    asm volatile(
+        "1:                            \n\t"
+        "movq  (%2, %0), %%mm0         \n\t"
+        "movq  (%1, %0), %%mm1         \n\t"
+        "psubb %%mm0, %%mm1            \n\t"
+        "movq %%mm1, (%3, %0)          \n\t"
+        "movq 8(%2, %0), %%mm0         \n\t"
+        "movq 8(%1, %0), %%mm1         \n\t"
+        "psubb %%mm0, %%mm1            \n\t"
+        "movq %%mm1, 8(%3, %0)         \n\t"
+        "addl $16, %0                  \n\t"
+        "cmpl %4, %0                   \n\t"
+        " jb 1b                                \n\t"
+        : "+r" (i)
+        : "r"(src1), "r"(src2), "r"(dst), "r"(w-15)
+    );
+    for(; i<w; i++)
+        dst[i+0] = src1[i+0]-src2[i+0];
+}
+
+
 #if 0
 static void just_return() { return; }
 #endif
@@ -531,6 +576,9 @@ void dsputil_init_mmx(DSPContext* c, unsigned mask)
         c->avg_no_rnd_pixels_tab[1][1] = avg_no_rnd_pixels8_x2_mmx;
         c->avg_no_rnd_pixels_tab[1][2] = avg_no_rnd_pixels8_y2_mmx;
         c->avg_no_rnd_pixels_tab[1][3] = avg_no_rnd_pixels8_xy2_mmx;
+        
+        c->add_bytes= add_bytes_mmx;
+        c->diff_bytes= diff_bytes_mmx;
 
         if (mm_flags & MM_MMXEXT) {
             c->pix_abs16x16     = pix_abs16x16_mmx2;