1 /*****************************************************************************
2 * depth.c: bit-depth conversion video filter
3 *****************************************************************************
4 * Copyright (C) 2010-2012 x264 project
6 * Authors: Oskar Arvidsson <oskar@irock.se>
8 * This program is free software; you can redistribute it and/or modify
9 * it under the terms of the GNU General Public License as published by
10 * the Free Software Foundation; either version 2 of the License, or
11 * (at your option) any later version.
13 * This program is distributed in the hope that it will be useful,
14 * but WITHOUT ANY WARRANTY; without even the implied warranty of
15 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
16 * GNU General Public License for more details.
18 * You should have received a copy of the GNU General Public License
19 * along with this program; if not, write to the Free Software
20 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02111, USA.
22 * This program is also available under a commercial proprietary license.
23 * For more information, contact us at licensing@x264.com.
24 *****************************************************************************/
28 #define FAIL_IF_ERROR( cond, ... ) FAIL_IF_ERR( cond, NAME, __VA_ARGS__ )
30 cli_vid_filter_t depth_filter;
35 cli_vid_filter_t prev_filter;
43 static int depth_filter_csp_is_supported( int csp )
45 int csp_mask = csp & X264_CSP_MASK;
46 return csp_mask == X264_CSP_I420 ||
47 csp_mask == X264_CSP_I422 ||
48 csp_mask == X264_CSP_I444 ||
49 csp_mask == X264_CSP_YV12 ||
50 csp_mask == X264_CSP_YV16 ||
51 csp_mask == X264_CSP_YV24 ||
52 csp_mask == X264_CSP_NV12 ||
53 csp_mask == X264_CSP_NV16;
56 static int csp_num_interleaved( int csp, int plane )
58 int csp_mask = csp & X264_CSP_MASK;
59 return ( (csp_mask == X264_CSP_NV12 || csp_mask == X264_CSP_NV16) && plane == 1 ) ? 2 : 1;
62 /* The dithering algorithm is based on Sierra-2-4A error diffusion. It has been
63 * written in such a way so that if the source has been upconverted using the
64 * same algorithm as used in scale_image, dithering down to the source bit
65 * depth again is lossless. */
66 #define DITHER_PLANE( pitch ) \
67 static void dither_plane_##pitch( pixel *dst, int dst_stride, uint16_t *src, int src_stride, \
68 int width, int height, int16_t *errors ) \
70 const int lshift = 16-BIT_DEPTH; \
71 const int rshift = 2*BIT_DEPTH-16; \
72 const int pixel_max = (1 << BIT_DEPTH)-1; \
73 const int half = 1 << (16-BIT_DEPTH); \
74 memset( errors, 0, (width+1) * sizeof(int16_t) ); \
75 for( int y = 0; y < height; y++, src += src_stride, dst += dst_stride ) \
78 for( int x = 0; x < width; x++ ) \
80 err = err*2 + errors[x] + errors[x+1]; \
81 dst[x*pitch] = x264_clip3( (((src[x*pitch]+half)<<2)+err)*pixel_max >> 18, 0, pixel_max ); \
82 errors[x] = err = src[x*pitch] - (dst[x*pitch] << lshift) - (dst[x*pitch] >> rshift); \
90 static void dither_image( cli_image_t *out, cli_image_t *img, int16_t *error_buf )
92 int csp_mask = img->csp & X264_CSP_MASK;
93 for( int i = 0; i < img->planes; i++ )
95 int num_interleaved = csp_num_interleaved( img->csp, i );
96 int height = x264_cli_csps[csp_mask].height[i] * img->height;
97 int width = x264_cli_csps[csp_mask].width[i] * img->width / num_interleaved;
99 #define CALL_DITHER_PLANE( pitch, off ) \
100 dither_plane_##pitch( ((pixel*)out->plane[i])+off, out->stride[i]/sizeof(pixel), \
101 ((uint16_t*)img->plane[i])+off, img->stride[i]/2, width, height, error_buf )
103 if( num_interleaved == 1 )
105 CALL_DITHER_PLANE( 1, 0 );
109 CALL_DITHER_PLANE( 2, 0 );
110 CALL_DITHER_PLANE( 2, 1 );
115 static void scale_image( cli_image_t *output, cli_image_t *img )
117 /* this function mimics how swscale does upconversion. 8-bit is converted
118 * to 16-bit through left shifting the orginal value with 8 and then adding
119 * the original value to that. This effectively keeps the full color range
120 * while also being fast. for n-bit we basically do the same thing, but we
121 * discard the lower 16-n bits. */
122 int csp_mask = img->csp & X264_CSP_MASK;
123 const int shift = 16-BIT_DEPTH;
124 for( int i = 0; i < img->planes; i++ )
126 uint8_t *src = img->plane[i];
127 uint16_t *dst = (uint16_t*)output->plane[i];
128 int height = x264_cli_csps[csp_mask].height[i] * img->height;
129 int width = x264_cli_csps[csp_mask].width[i] * img->width;
131 for( int j = 0; j < height; j++ )
133 for( int k = 0; k < width; k++ )
134 dst[k] = ((src[k] << 8) + src[k]) >> shift;
136 src += img->stride[i];
137 dst += output->stride[i]/2;
142 static int get_frame( hnd_t handle, cli_pic_t *output, int frame )
144 depth_hnd_t *h = handle;
146 if( h->prev_filter.get_frame( h->prev_hnd, output, frame ) )
149 if( h->bit_depth < 16 && output->img.csp & X264_CSP_HIGH_DEPTH )
151 dither_image( &h->buffer.img, &output->img, h->error_buf );
152 output->img = h->buffer.img;
154 else if( h->bit_depth > 8 && !(output->img.csp & X264_CSP_HIGH_DEPTH) )
156 scale_image( &h->buffer.img, &output->img );
157 output->img = h->buffer.img;
162 static int release_frame( hnd_t handle, cli_pic_t *pic, int frame )
164 depth_hnd_t *h = handle;
165 return h->prev_filter.release_frame( h->prev_hnd, pic, frame );
168 static void free_filter( hnd_t handle )
170 depth_hnd_t *h = handle;
171 h->prev_filter.free( h->prev_hnd );
172 x264_cli_pic_clean( &h->buffer );
176 static int init( hnd_t *handle, cli_vid_filter_t *filter, video_info_t *info,
177 x264_param_t *param, char *opt_string )
180 int change_fmt = (info->csp ^ param->i_csp) & X264_CSP_HIGH_DEPTH;
181 int csp = ~(~info->csp ^ change_fmt);
182 int bit_depth = 8*x264_cli_csp_depth_factor( csp );
186 static const char *optlist[] = { "bit_depth", NULL };
187 char **opts = x264_split_options( opt_string, optlist );
191 char *str_bit_depth = x264_get_option( "bit_depth", opts );
192 bit_depth = x264_otoi( str_bit_depth, -1 );
194 ret = bit_depth < 8 || bit_depth > 16;
195 csp = bit_depth > 8 ? csp | X264_CSP_HIGH_DEPTH : csp & ~X264_CSP_HIGH_DEPTH;
196 change_fmt = (info->csp ^ csp) & X264_CSP_HIGH_DEPTH;
197 x264_free_string_array( opts );
203 FAIL_IF_ERROR( bit_depth != BIT_DEPTH, "this build supports only bit depth %d\n", BIT_DEPTH )
204 FAIL_IF_ERROR( ret, "unsupported bit depth conversion.\n" )
206 /* only add the filter to the chain if it's needed */
207 if( change_fmt || bit_depth != 8 * x264_cli_csp_depth_factor( csp ) )
209 FAIL_IF_ERROR( !depth_filter_csp_is_supported(csp), "unsupported colorspace.\n" )
210 depth_hnd_t *h = x264_malloc( sizeof(depth_hnd_t) + (info->width+1)*sizeof(int16_t) );
215 h->error_buf = (int16_t*)(h + 1);
217 h->bit_depth = bit_depth;
218 h->prev_hnd = *handle;
219 h->prev_filter = *filter;
221 if( x264_cli_pic_alloc( &h->buffer, h->dst_csp, info->width, info->height ) )
228 *filter = depth_filter;
229 info->csp = h->dst_csp;
235 cli_vid_filter_t depth_filter = { NAME, NULL, init, get_frame, release_frame, free_filter, NULL };