]> git.sesse.net Git - ffmpeg/blobdiff - libavcodec/ffv1.c
Merge remote-tracking branch 'qatar/master'
[ffmpeg] / libavcodec / ffv1.c
index e2ca6fd9d97210e1e7ff87bee3aa4f4f41479f9e..0d63dd8bf77ccc0d679782eb4fd6756d44881189 100644 (file)
@@ -3,20 +3,20 @@
  *
  * Copyright (c) 2003 Michael Niedermayer <michaelni@gmx.at>
  *
- * This file is part of Libav.
+ * This file is part of FFmpeg.
  *
- * Libav is free software; you can redistribute it and/or
+ * FFmpeg 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 2.1 of the License, or (at your option) any later version.
  *
- * Libav is distributed in the hope that it will be useful,
+ * FFmpeg 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 Libav; if not, write to the Free Software
+ * License along with FFmpeg; if not, write to the Free Software
  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
  */
 
  */
 
 #include "avcodec.h"
+#include "internal.h"
 #include "get_bits.h"
 #include "put_bits.h"
 #include "dsputil.h"
 #include "rangecoder.h"
 #include "golomb.h"
 #include "mathops.h"
+#include "libavutil/pixdesc.h"
 #include "libavutil/avassert.h"
 
 #define MAX_PLANES 4
@@ -165,6 +167,8 @@ typedef struct FFV1Context{
     int version;
     int width, height;
     int chroma_h_shift, chroma_v_shift;
+    int chroma_planes;
+    int transparency;
     int flags;
     int picture_number;
     AVFrame picture;
@@ -180,6 +184,7 @@ typedef struct FFV1Context{
     int colorspace;
     int16_t *sample_buffer;
     int gob_count;
+    int packed_at_lsb;
 
     int quant_table_count;
 
@@ -193,6 +198,7 @@ typedef struct FFV1Context{
     int slice_height;
     int slice_x;
     int slice_y;
+    int bits_per_raw_sample;
 }FFV1Context;
 
 static av_always_inline int fold(int diff, int bits){
@@ -537,16 +543,22 @@ static void encode_plane(FFV1Context *s, uint8_t *src, int w, int h, int stride,
         sample[0][-1]= sample[1][0  ];
         sample[1][ w]= sample[1][w-1];
 //{START_TIMER
-        if(s->avctx->bits_per_raw_sample<=8){
+        if(s->bits_per_raw_sample<=8){
             for(x=0; x<w; x++){
                 sample[0][x]= src[x + stride*y];
             }
             encode_line(s, w, sample, plane_index, 8);
         }else{
-            for(x=0; x<w; x++){
-                sample[0][x]= ((uint16_t*)(src + stride*y))[x] >> (16 - s->avctx->bits_per_raw_sample);
+            if(s->packed_at_lsb){
+                for(x=0; x<w; x++){
+                    sample[0][x]= ((uint16_t*)(src + stride*y))[x];
+                }
+            }else{
+                for(x=0; x<w; x++){
+                    sample[0][x]= ((uint16_t*)(src + stride*y))[x] >> (16 - s->bits_per_raw_sample);
+                }
             }
-            encode_line(s, w, sample, plane_index, s->avctx->bits_per_raw_sample);
+            encode_line(s, w, sample, plane_index, s->bits_per_raw_sample);
         }
 //STOP_TIMER("encode line")}
     }
@@ -555,21 +567,22 @@ static void encode_plane(FFV1Context *s, uint8_t *src, int w, int h, int stride,
 static void encode_rgb_frame(FFV1Context *s, uint32_t *src, int w, int h, int stride){
     int x, y, p, i;
     const int ring_size= s->avctx->context_model ? 3 : 2;
-    int16_t *sample[3][3];
+    int16_t *sample[4][3];
     s->run_index=0;
 
-    memset(s->sample_buffer, 0, ring_size*3*(w+6)*sizeof(*s->sample_buffer));
+    memset(s->sample_buffer, 0, ring_size*4*(w+6)*sizeof(*s->sample_buffer));
 
     for(y=0; y<h; y++){
         for(i=0; i<ring_size; i++)
-            for(p=0; p<3; p++)
+            for(p=0; p<4; p++)
                 sample[p][i]= s->sample_buffer + p*ring_size*(w+6) + ((h+i-y)%ring_size)*(w+6) + 3;
 
         for(x=0; x<w; x++){
-            int v= src[x + stride*y];
+            unsigned v= src[x + stride*y];
             int b= v&0xFF;
             int g= (v>>8)&0xFF;
             int r= (v>>16)&0xFF;
+            int a=  v>>24;
 
             b -= g;
             r -= g;
@@ -582,11 +595,12 @@ static void encode_rgb_frame(FFV1Context *s, uint32_t *src, int w, int h, int st
             sample[0][0][x]= g;
             sample[1][0][x]= b;
             sample[2][0][x]= r;
+            sample[3][0][x]= a;
         }
-        for(p=0; p<3; p++){
+        for(p=0; p<3 + s->transparency; p++){
             sample[p][0][-1]= sample[p][1][0  ];
             sample[p][1][ w]= sample[p][1][w-1];
-            encode_line(s, w, sample[p], FFMIN(p, 1), 9);
+            encode_line(s, w, sample[p], (p+1)/2, 9);
         }
     }
 }
@@ -629,11 +643,11 @@ static void write_header(FFV1Context *f){
         }
         put_symbol(c, state, f->colorspace, 0); //YUV cs type
         if(f->version>0)
-            put_symbol(c, state, f->avctx->bits_per_raw_sample, 0);
-        put_rac(c, state, 1); //chroma planes
-            put_symbol(c, state, f->chroma_h_shift, 0);
-            put_symbol(c, state, f->chroma_v_shift, 0);
-        put_rac(c, state, 0); //no transparency plane
+            put_symbol(c, state, f->bits_per_raw_sample, 0);
+        put_rac(c, state, f->chroma_planes);
+        put_symbol(c, state, f->chroma_h_shift, 0);
+        put_symbol(c, state, f->chroma_v_shift, 0);
+        put_rac(c, state, f->transparency);
 
         write_quant_tables(c, f->quant_table);
     }else{
@@ -659,6 +673,8 @@ static av_cold int common_init(AVCodecContext *avctx){
     s->avctx= avctx;
     s->flags= avctx->flags;
 
+    avcodec_get_frame_defaults(&s->picture);
+
     ff_dsputil_init(&s->dsp, avctx);
 
     s->width = avctx->width;
@@ -678,6 +694,8 @@ static int init_slice_state(FFV1Context *f){
 
     for(i=0; i<f->slice_count; i++){
         FFV1Context *fs= f->slice_context[i];
+        fs->plane_count= f->plane_count;
+        fs->transparency= f->transparency;
         for(j=0; j<f->plane_count; j++){
             PlaneContext * const p= &fs->plane[j];
 
@@ -726,7 +744,7 @@ static av_cold int init_slice_contexts(FFV1Context *f){
         fs->slice_x     = sxs;
         fs->slice_y     = sys;
 
-        fs->sample_buffer = av_malloc(9 * (fs->width+6) * sizeof(*fs->sample_buffer));
+        fs->sample_buffer = av_malloc(3*4 * (fs->width+6) * sizeof(*fs->sample_buffer));
         if (!fs->sample_buffer)
             return AVERROR(ENOMEM);
     }
@@ -767,11 +785,11 @@ static int write_extra_header(FFV1Context *f){
         }
     }
     put_symbol(c, state, f->colorspace, 0); //YUV cs type
-    put_symbol(c, state, f->avctx->bits_per_raw_sample, 0);
-    put_rac(c, state, 1); //chroma planes
-        put_symbol(c, state, f->chroma_h_shift, 0);
-        put_symbol(c, state, f->chroma_v_shift, 0);
-    put_rac(c, state, 0); //no transparency plane
+    put_symbol(c, state, f->bits_per_raw_sample, 0);
+    put_rac(c, state, f->chroma_planes);
+    put_symbol(c, state, f->chroma_h_shift, 0);
+    put_symbol(c, state, f->chroma_v_shift, 0);
+    put_rac(c, state, f->transparency);
     put_symbol(c, state, f->num_h_slices-1, 0);
     put_symbol(c, state, f->num_v_slices-1, 0);
 
@@ -858,10 +876,75 @@ static av_cold int encode_init(AVCodecContext *avctx)
         for(i=1; i<256; i++)
             s->state_transition[i]=ver2_state[i];
 
-    s->plane_count=2;
+    s->plane_count=3;
+    switch(avctx->pix_fmt){
+    case PIX_FMT_YUV444P9:
+    case PIX_FMT_YUV422P9:
+    case PIX_FMT_YUV420P9:
+        if (!avctx->bits_per_raw_sample)
+            s->bits_per_raw_sample = 9;
+    case PIX_FMT_YUV444P10:
+    case PIX_FMT_YUV420P10:
+    case PIX_FMT_YUV422P10:
+        s->packed_at_lsb = 1;
+        if (!avctx->bits_per_raw_sample && !s->bits_per_raw_sample)
+            s->bits_per_raw_sample = 10;
+    case PIX_FMT_GRAY16:
+    case PIX_FMT_YUV444P16:
+    case PIX_FMT_YUV422P16:
+    case PIX_FMT_YUV420P16:
+        if (!avctx->bits_per_raw_sample && !s->bits_per_raw_sample) {
+            s->bits_per_raw_sample = 16;
+        } else if (!s->bits_per_raw_sample){
+            s->bits_per_raw_sample = avctx->bits_per_raw_sample;
+        }
+        if(s->bits_per_raw_sample <=8){
+            av_log(avctx, AV_LOG_ERROR, "bits_per_raw_sample invalid\n");
+            return -1;
+        }
+        if(!s->ac){
+            av_log(avctx, AV_LOG_ERROR, "bits_per_raw_sample of more than 8 needs -coder 1 currently\n");
+            return -1;
+        }
+        s->version= FFMAX(s->version, 1);
+    case PIX_FMT_GRAY8:
+    case PIX_FMT_YUV444P:
+    case PIX_FMT_YUV440P:
+    case PIX_FMT_YUV422P:
+    case PIX_FMT_YUV420P:
+    case PIX_FMT_YUV411P:
+    case PIX_FMT_YUV410P:
+        s->chroma_planes= av_pix_fmt_descriptors[avctx->pix_fmt].nb_components < 3 ? 0 : 1;
+        s->colorspace= 0;
+        break;
+    case PIX_FMT_YUVA444P:
+    case PIX_FMT_YUVA420P:
+        s->chroma_planes= 1;
+        s->colorspace= 0;
+        s->transparency= 1;
+        break;
+    case PIX_FMT_RGB32:
+        s->colorspace= 1;
+        s->transparency= 1;
+        break;
+    case PIX_FMT_0RGB32:
+        s->colorspace= 1;
+        break;
+    default:
+        av_log(avctx, AV_LOG_ERROR, "format not supported\n");
+        return -1;
+    }
+    if (s->transparency) {
+        av_log(avctx, AV_LOG_WARNING, "Storing alpha plane, this will require a recent FFV1 decoder to playback!\n");
+    }
+    if (avctx->context_model > 1U) {
+        av_log(avctx, AV_LOG_ERROR, "Invalid context model %d, valid values are 0 and 1\n", avctx->context_model);
+        return AVERROR(EINVAL);
+    }
+
     for(i=0; i<256; i++){
         s->quant_table_count=2;
-        if(avctx->bits_per_raw_sample <=8){
+        if(s->bits_per_raw_sample <=8){
             s->quant_tables[0][0][i]=           quant11[i];
             s->quant_tables[0][1][i]=        11*quant11[i];
             s->quant_tables[0][2][i]=     11*11*quant11[i];
@@ -897,33 +980,8 @@ static av_cold int encode_init(AVCodecContext *avctx)
         return AVERROR(ENOMEM);
 
     avctx->coded_frame= &s->picture;
-    switch(avctx->pix_fmt){
-    case PIX_FMT_YUV444P16:
-    case PIX_FMT_YUV422P16:
-    case PIX_FMT_YUV420P16:
-        if(avctx->bits_per_raw_sample <=8){
-            av_log(avctx, AV_LOG_ERROR, "bits_per_raw_sample invalid\n");
-            return -1;
-        }
-        if(!s->ac){
-            av_log(avctx, AV_LOG_ERROR, "bits_per_raw_sample of more than 8 needs -coder 1 currently\n");
-            return -1;
-        }
-        s->version= FFMAX(s->version, 1);
-    case PIX_FMT_YUV444P:
-    case PIX_FMT_YUV422P:
-    case PIX_FMT_YUV420P:
-    case PIX_FMT_YUV411P:
-    case PIX_FMT_YUV410P:
-        s->colorspace= 0;
-        break;
-    case PIX_FMT_RGB32:
-        s->colorspace= 1;
-        break;
-    default:
-        av_log(avctx, AV_LOG_ERROR, "format not supported\n");
-        return -1;
-    }
+    if(!s->transparency)
+        s->plane_count= 2;
     avcodec_get_chroma_sub_sample(avctx->pix_fmt, &s->chroma_h_shift, &s->chroma_v_shift);
 
     s->picture_number=0;
@@ -1041,12 +1099,12 @@ static void clear_state(FFV1Context *f){
                 }else
                 memset(p->state, 128, CONTEXT_SIZE*p->context_count);
             }else{
-            for(j=0; j<p->context_count; j++){
+                for(j=0; j<p->context_count; j++){
                     p->vlc_state[j].drift= 0;
                     p->vlc_state[j].error_sum= 4; //FFMAX((RANGE + 32)/64, 2);
                     p->vlc_state[j].bias= 0;
                     p->vlc_state[j].count= 1;
-            }
+                }
             }
         }
     }
@@ -1061,6 +1119,7 @@ static int encode_slice(AVCodecContext *c, void *arg){
     int x= fs->slice_x;
     int y= fs->slice_y;
     AVFrame * const p= &f->picture;
+    const int ps= (f->bits_per_raw_sample>8)+1;
 
     if(f->colorspace==0){
         const int chroma_width = -((-width )>>f->chroma_h_shift);
@@ -1068,12 +1127,16 @@ static int encode_slice(AVCodecContext *c, void *arg){
         const int cx= x>>f->chroma_h_shift;
         const int cy= y>>f->chroma_v_shift;
 
-        encode_plane(fs, p->data[0] + x + y*p->linesize[0], width, height, p->linesize[0], 0);
+        encode_plane(fs, p->data[0] + ps*x + y*p->linesize[0], width, height, p->linesize[0], 0);
 
-        encode_plane(fs, p->data[1] + cx+cy*p->linesize[1], chroma_width, chroma_height, p->linesize[1], 1);
-        encode_plane(fs, p->data[2] + cx+cy*p->linesize[2], chroma_width, chroma_height, p->linesize[2], 1);
+        if (f->chroma_planes){
+            encode_plane(fs, p->data[1] + ps*cx+cy*p->linesize[1], chroma_width, chroma_height, p->linesize[1], 1);
+            encode_plane(fs, p->data[2] + ps*cx+cy*p->linesize[2], chroma_width, chroma_height, p->linesize[2], 1);
+        }
+        if (fs->transparency)
+            encode_plane(fs, p->data[3] + ps*x + y*p->linesize[3], width, height, p->linesize[3], 2);
     }else{
-        encode_rgb_frame(fs, (uint32_t*)(p->data[0]) + x + y*(p->linesize[0]/4), width, height, p->linesize[0]/4);
+        encode_rgb_frame(fs, (uint32_t*)(p->data[0]) + ps*x + y*(p->linesize[0]/4), width, height, p->linesize[0]/4);
     }
     emms_c();
 
@@ -1091,12 +1154,9 @@ static int encode_frame(AVCodecContext *avctx, AVPacket *pkt,
     uint8_t *buf_p;
     int i, ret;
 
-    if (!pkt->data &&
-        (ret = av_new_packet(pkt, avctx->width*avctx->height*((8*2+1+1)*4)/8
-                                  + FF_MIN_BUFFER_SIZE)) < 0) {
-        av_log(avctx, AV_LOG_ERROR, "Error getting output packet.\n");
+    if ((ret = ff_alloc_packet2(avctx, pkt, avctx->width*avctx->height*((8*2+1+1)*4)/8
+                                  + FF_MIN_BUFFER_SIZE)) < 0)
         return ret;
-    }
 
     ff_init_range_encoder(c, pkt->data, pkt->size);
     ff_build_rac_states(c, 0.05*(1LL<<32), 256-8);
@@ -1339,8 +1399,14 @@ static void decode_plane(FFV1Context *s, uint8_t *src, int w, int h, int stride,
             }
         }else{
             decode_line(s, w, sample, plane_index, s->avctx->bits_per_raw_sample);
-            for(x=0; x<w; x++){
-                ((uint16_t*)(src + stride*y))[x]= sample[1][x] << (16 - s->avctx->bits_per_raw_sample);
+            if(s->packed_at_lsb){
+                for(x=0; x<w; x++){
+                    ((uint16_t*)(src + stride*y))[x]= sample[1][x];
+                }
+            }else{
+                for(x=0; x<w; x++){
+                    ((uint16_t*)(src + stride*y))[x]= sample[1][x] << (16 - s->avctx->bits_per_raw_sample);
+                }
             }
         }
 //STOP_TIMER("decode-line")}
@@ -1349,18 +1415,18 @@ static void decode_plane(FFV1Context *s, uint8_t *src, int w, int h, int stride,
 
 static void decode_rgb_frame(FFV1Context *s, uint32_t *src, int w, int h, int stride){
     int x, y, p;
-    int16_t *sample[3][2];
-    for(x=0; x<3; x++){
+    int16_t *sample[4][2];
+    for(x=0; x<4; x++){
         sample[x][0] = s->sample_buffer +  x*2   *(w+6) + 3;
         sample[x][1] = s->sample_buffer + (x*2+1)*(w+6) + 3;
     }
 
     s->run_index=0;
 
-    memset(s->sample_buffer, 0, 6*(w+6)*sizeof(*s->sample_buffer));
+    memset(s->sample_buffer, 0, 8*(w+6)*sizeof(*s->sample_buffer));
 
     for(y=0; y<h; y++){
-        for(p=0; p<3; p++){
+        for(p=0; p<3 + s->transparency; p++){
             int16_t *temp = sample[p][0]; //FIXME try a normal buffer
 
             sample[p][0]= sample[p][1];
@@ -1368,12 +1434,13 @@ static void decode_rgb_frame(FFV1Context *s, uint32_t *src, int w, int h, int st
 
             sample[p][1][-1]= sample[p][0][0  ];
             sample[p][0][ w]= sample[p][0][w-1];
-            decode_line(s, w, sample[p], FFMIN(p, 1), 9);
+            decode_line(s, w, sample[p], (p+1)/2, 9);
         }
         for(x=0; x<w; x++){
             int g= sample[0][1][x];
             int b= sample[1][1][x];
             int r= sample[2][1][x];
+            int a= sample[3][1][x];
 
 //            assert(g>=0 && b>=0 && r>=0);
 //            assert(g<256 && b<512 && r<512);
@@ -1384,7 +1451,7 @@ static void decode_rgb_frame(FFV1Context *s, uint32_t *src, int w, int h, int st
             b += g;
             r += g;
 
-            src[x + stride*y]= b + (g<<8) + (r<<16) + (0xFF<<24);
+            src[x + stride*y]= b + (g<<8) + (r<<16) + (a<<24);
         }
     }
 }
@@ -1396,6 +1463,7 @@ static int decode_slice(AVCodecContext *c, void *arg){
     int height= fs->slice_height;
     int x= fs->slice_x;
     int y= fs->slice_y;
+    const int ps= (c->bits_per_raw_sample>8)+1;
     AVFrame * const p= &f->picture;
 
     av_assert1(width && height);
@@ -1404,12 +1472,16 @@ static int decode_slice(AVCodecContext *c, void *arg){
         const int chroma_height= -((-height)>>f->chroma_v_shift);
         const int cx= x>>f->chroma_h_shift;
         const int cy= y>>f->chroma_v_shift;
-        decode_plane(fs, p->data[0] + x + y*p->linesize[0], width, height, p->linesize[0], 0);
+        decode_plane(fs, p->data[0] + ps*x + y*p->linesize[0], width, height, p->linesize[0], 0);
 
-        decode_plane(fs, p->data[1] + cx+cy*p->linesize[1], chroma_width, chroma_height, p->linesize[1], 1);
-        decode_plane(fs, p->data[2] + cx+cy*p->linesize[1], chroma_width, chroma_height, p->linesize[2], 1);
+        if (f->chroma_planes){
+            decode_plane(fs, p->data[1] + ps*cx+cy*p->linesize[1], chroma_width, chroma_height, p->linesize[1], 1);
+            decode_plane(fs, p->data[2] + ps*cx+cy*p->linesize[2], chroma_width, chroma_height, p->linesize[2], 1);
+        }
+        if (fs->transparency)
+            decode_plane(fs, p->data[3] + ps*x + y*p->linesize[3], width, height, p->linesize[3], 2);
     }else{
-        decode_rgb_frame(fs, (uint32_t*)p->data[0] + x + y*(p->linesize[0]/4), width, height, p->linesize[0]/4);
+        decode_rgb_frame(fs, (uint32_t*)p->data[0] + ps*x + y*(p->linesize[0]/4), width, height, p->linesize[0]/4);
     }
 
     emms_c();
@@ -1482,8 +1554,8 @@ static int read_extra_header(FFV1Context *f){
     get_rac(c, state); //no chroma = false
     f->chroma_h_shift= get_symbol(c, state, 0);
     f->chroma_v_shift= get_symbol(c, state, 0);
-    get_rac(c, state); //transparency plane
-    f->plane_count= 2;
+    f->transparency= get_rac(c, state);
+    f->plane_count= 2 + f->transparency;
     f->num_h_slices= 1 + get_symbol(c, state, 0);
     f->num_v_slices= 1 + get_symbol(c, state, 0);
     if(f->num_h_slices > (unsigned)f->width || f->num_v_slices > (unsigned)f->height){
@@ -1536,17 +1608,23 @@ static int read_header(FFV1Context *f){
         f->colorspace= get_symbol(c, state, 0); //YUV cs type
         if(f->version>0)
             f->avctx->bits_per_raw_sample= get_symbol(c, state, 0);
-        get_rac(c, state); //no chroma = false
+        f->chroma_planes= get_rac(c, state);
         f->chroma_h_shift= get_symbol(c, state, 0);
         f->chroma_v_shift= get_symbol(c, state, 0);
-        get_rac(c, state); //transparency plane
-        f->plane_count= 2;
+        f->transparency= get_rac(c, state);
+        f->plane_count= 2 + f->transparency;
     }
 
     if(f->colorspace==0){
-        if(f->avctx->bits_per_raw_sample<=8){
+        if(!f->transparency && !f->chroma_planes){
+            if (f->avctx->bits_per_raw_sample<=8)
+                f->avctx->pix_fmt= PIX_FMT_GRAY8;
+            else
+                f->avctx->pix_fmt= PIX_FMT_GRAY16;
+        }else if(f->avctx->bits_per_raw_sample<=8 && !f->transparency){
             switch(16*f->chroma_h_shift + f->chroma_v_shift){
             case 0x00: f->avctx->pix_fmt= PIX_FMT_YUV444P; break;
+            case 0x01: f->avctx->pix_fmt= PIX_FMT_YUV440P; break;
             case 0x10: f->avctx->pix_fmt= PIX_FMT_YUV422P; break;
             case 0x11: f->avctx->pix_fmt= PIX_FMT_YUV420P; break;
             case 0x20: f->avctx->pix_fmt= PIX_FMT_YUV411P; break;
@@ -1555,7 +1633,35 @@ static int read_header(FFV1Context *f){
                 av_log(f->avctx, AV_LOG_ERROR, "format not supported\n");
                 return -1;
             }
-        }else{
+        }else if(f->avctx->bits_per_raw_sample<=8 && f->transparency){
+            switch(16*f->chroma_h_shift + f->chroma_v_shift){
+            case 0x00: f->avctx->pix_fmt= PIX_FMT_YUVA444P; break;
+            case 0x11: f->avctx->pix_fmt= PIX_FMT_YUVA420P; break;
+            default:
+                av_log(f->avctx, AV_LOG_ERROR, "format not supported\n");
+                return -1;
+            }
+        }else if(f->avctx->bits_per_raw_sample==9) {
+            f->packed_at_lsb=1;
+            switch(16*f->chroma_h_shift + f->chroma_v_shift){
+            case 0x00: f->avctx->pix_fmt= PIX_FMT_YUV444P9; break;
+            case 0x10: f->avctx->pix_fmt= PIX_FMT_YUV422P9; break;
+            case 0x11: f->avctx->pix_fmt= PIX_FMT_YUV420P9; break;
+            default:
+                av_log(f->avctx, AV_LOG_ERROR, "format not supported\n");
+                return -1;
+            }
+        }else if(f->avctx->bits_per_raw_sample==10) {
+            f->packed_at_lsb=1;
+            switch(16*f->chroma_h_shift + f->chroma_v_shift){
+            case 0x00: f->avctx->pix_fmt= PIX_FMT_YUV444P10; break;
+            case 0x10: f->avctx->pix_fmt= PIX_FMT_YUV422P10; break;
+            case 0x11: f->avctx->pix_fmt= PIX_FMT_YUV420P10; break;
+            default:
+                av_log(f->avctx, AV_LOG_ERROR, "format not supported\n");
+                return -1;
+            }
+        }else {
             switch(16*f->chroma_h_shift + f->chroma_v_shift){
             case 0x00: f->avctx->pix_fmt= PIX_FMT_YUV444P16; break;
             case 0x10: f->avctx->pix_fmt= PIX_FMT_YUV422P16; break;
@@ -1570,7 +1676,8 @@ static int read_header(FFV1Context *f){
             av_log(f->avctx, AV_LOG_ERROR, "chroma subsampling not supported in this colorspace\n");
             return -1;
         }
-        f->avctx->pix_fmt= PIX_FMT_RGB32;
+        if(f->transparency) f->avctx->pix_fmt= PIX_FMT_RGB32;
+        else                f->avctx->pix_fmt= PIX_FMT_0RGB32;
     }else{
         av_log(f->avctx, AV_LOG_ERROR, "colorspace not supported\n");
         return -1;
@@ -1592,6 +1699,7 @@ static int read_header(FFV1Context *f){
     for(j=0; j<f->slice_count; j++){
         FFV1Context *fs= f->slice_context[j];
         fs->ac= f->ac;
+        fs->packed_at_lsb= f->packed_at_lsb;
 
         if(f->version >= 2){
             fs->slice_x     = get_symbol(c, state, 0)   *f->width ;
@@ -1758,11 +1866,7 @@ AVCodec ff_ffv1_encoder = {
     .encode2        = encode_frame,
     .close          = common_end,
     .capabilities   = CODEC_CAP_SLICE_THREADS,
-    .pix_fmts       = (const enum PixelFormat[]){
-        PIX_FMT_YUV420P, PIX_FMT_YUV444P, PIX_FMT_YUV422P, PIX_FMT_YUV411P,
-        PIX_FMT_YUV410P, PIX_FMT_RGB32, PIX_FMT_YUV420P16, PIX_FMT_YUV422P16,
-        PIX_FMT_YUV444P16, PIX_FMT_NONE
-    },
+    .pix_fmts= (const enum PixelFormat[]){PIX_FMT_YUV420P, PIX_FMT_YUVA420P, PIX_FMT_YUV444P, PIX_FMT_YUVA444P, PIX_FMT_YUV440P, PIX_FMT_YUV422P, PIX_FMT_YUV411P, PIX_FMT_YUV410P, PIX_FMT_0RGB32, PIX_FMT_RGB32, PIX_FMT_YUV420P16, PIX_FMT_YUV422P16, PIX_FMT_YUV444P16, PIX_FMT_YUV444P9, PIX_FMT_YUV422P9, PIX_FMT_YUV420P9, PIX_FMT_YUV420P10, PIX_FMT_YUV422P10, PIX_FMT_YUV444P10, PIX_FMT_GRAY16, PIX_FMT_GRAY8, PIX_FMT_NONE},
     .long_name      = NULL_IF_CONFIG_SMALL("FFmpeg video codec #1"),
 };
 #endif