1 /*****************************************************************************
2 * common.h: h264 encoder
3 *****************************************************************************
4 * Copyright (C) 2003 Laurent Aimar
5 * $Id: common.h,v 1.1 2004/06/03 19:27:06 fenrir Exp $
7 * Authors: Laurent Aimar <fenrir@via.ecp.fr>
9 * This program is free software; you can redistribute it and/or modify
10 * it under the terms of the GNU General Public License as published by
11 * the Free Software Foundation; either version 2 of the License, or
12 * (at your option) any later version.
14 * This program is distributed in the hope that it will be useful,
15 * but WITHOUT ANY WARRANTY; without even the implied warranty of
16 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
17 * GNU General Public License for more details.
19 * You should have received a copy of the GNU General Public License
20 * along with this program; if not, write to the Free Software
21 * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111, USA.
22 *****************************************************************************/
36 #define snprintf _snprintf
37 #define X264_VERSION "" // no configure script for msvc
43 #define pthread_t HANDLE
44 #define pthread_create(t,u,f,d) *(t)=CreateThread(NULL,0,f,d,0,NULL)
45 #define pthread_join(t,s) { WaitForSingleObject(t,INFINITE); \
47 #define HAVE_PTHREAD 1
49 #elif defined(SYS_BEOS)
50 #include <kernel/OS.h>
51 #define pthread_t thread_id
52 #define pthread_create(t,u,f,d) { *(t)=spawn_thread(f,"",10,d); \
53 resume_thread(*(t)); }
54 #define pthread_join(t,s) wait_for_thread(t,(long*)s)
55 #define HAVE_PTHREAD 1
57 #elif defined(HAVE_PTHREAD)
61 /****************************************************************************
63 ****************************************************************************/
64 #define X264_MIN(a,b) ( (a)<(b) ? (a) : (b) )
65 #define X264_MAX(a,b) ( (a)>(b) ? (a) : (b) )
66 #define X264_MIN3(a,b,c) X264_MIN((a),X264_MIN((b),(c)))
67 #define X264_MAX3(a,b,c) X264_MAX((a),X264_MAX((b),(c)))
68 #define X264_MIN4(a,b,c,d) X264_MIN((a),X264_MIN3((b),(c),(d)))
69 #define X264_MAX4(a,b,c,d) X264_MAX((a),X264_MAX3((b),(c),(d)))
70 #define XCHG(type,a,b) { type t = a; a = b; b = t; }
71 #define FIX8(f) ((int)(f*(1<<8)+.5))
73 #if defined(__GNUC__) && (__GNUC__ > 3 || __GNUC__ == 3 && __GNUC_MINOR__ > 0)
74 #define UNUSED __attribute__((unused))
79 #define X264_BFRAME_MAX 16
80 #define X264_SLICE_MAX 4
81 #define X264_NAL_MAX (4 + X264_SLICE_MAX)
83 /****************************************************************************
85 ****************************************************************************/
98 /****************************************************************************
100 ****************************************************************************/
101 /* x264_malloc : will do or emulate a memalign
102 * XXX you HAVE TO use x264_free for buffer allocated
105 void *x264_malloc( int );
106 void *x264_realloc( void *p, int i_size );
107 void x264_free( void * );
109 /* x264_slurp_file: malloc space for the whole file and read it */
110 char *x264_slurp_file( const char *filename );
112 /* mdate: return the current date in microsecond */
113 int64_t x264_mdate( void );
115 /* x264_param2string: return a (malloced) string containing most of
116 * the encoding options */
117 char *x264_param2string( x264_param_t *p, int b_res );
120 void x264_log( x264_t *h, int i_level, const char *psz_fmt, ... );
122 static inline int x264_clip3( int v, int i_min, int i_max )
124 return ( (v < i_min) ? i_min : (v > i_max) ? i_max : v );
127 static inline float x264_clip3f( float v, float f_min, float f_max )
129 return ( (v < f_min) ? f_min : (v > f_max) ? f_max : v );
132 static inline int x264_median( int a, int b, int c )
138 max = b; /* no need to do 'b > max' (more consuming than always doing affectation) */
145 return a + b + c - min - max;
149 /****************************************************************************
151 ****************************************************************************/
161 static const char slice_type_to_char[] = { 'P', 'B', 'I', 'S', 'S' };
179 int i_idr_pic_id; /* -1 if nal_type != 5 */
182 int i_delta_poc_bottom;
185 int i_redundant_pic_cnt;
187 int b_direct_spatial_mv_pred;
189 int b_num_ref_idx_override;
190 int i_num_ref_idx_l0_active;
191 int i_num_ref_idx_l1_active;
193 int b_ref_pic_list_reordering_l0;
194 int b_ref_pic_list_reordering_l1;
198 } ref_pic_list_order[2][16];
200 int i_cabac_init_idc;
207 /* deblocking filter */
208 int i_disable_deblocking_filter_idc;
209 int i_alpha_c0_offset;
212 } x264_slice_header_t;
216 #define X264_SCAN8_SIZE (6*8)
217 #define X264_SCAN8_0 (4+1*8)
219 static const int x264_scan8[16+2*4] =
222 4+1*8, 5+1*8, 4+2*8, 5+2*8,
223 6+1*8, 7+1*8, 6+2*8, 7+2*8,
224 4+3*8, 5+3*8, 4+4*8, 5+4*8,
225 6+3*8, 7+3*8, 6+4*8, 7+4*8,
245 typedef struct x264_ratecontrol_t x264_ratecontrol_t;
246 typedef struct x264_vlc_table_t x264_vlc_table_t;
250 /* encoder parameters */
253 x264_t *thread[X264_SLICE_MAX];
255 /* bitstream output */
259 x264_nal_t nal[X264_NAL_MAX];
260 int i_bitstream; /* size of p_bitstream */
261 uint8_t *p_bitstream; /* will hold data for all nal */
265 /* frame number/poc */
268 int i_frame_offset; /* decoding only */
269 int i_frame_num; /* decoding only */
270 int i_poc_msb; /* decoding only */
271 int i_poc_lsb; /* decoding only */
272 int i_poc; /* decoding only */
274 int i_thread_num; /* threads only */
275 int i_nal_type; /* threads only */
276 int i_nal_ref_idc; /* threads only */
278 /* We use only one SPS and one PPS */
279 x264_sps_t sps_array[1];
281 x264_pps_t pps_array[1];
285 int dequant4_mf[4][6][4][4];
286 int dequant8_mf[2][6][8][8];
287 int quant4_mf[4][6][4][4];
288 int quant8_mf[2][6][8][8];
289 int unquant4_mf[4][52][16];
290 int unquant8_mf[2][52][64];
292 uint32_t nr_residual_sum[2][64];
293 uint32_t nr_offset[2][64];
294 uint32_t nr_count[2];
297 x264_slice_header_t sh;
304 /* Frames to be encoded (whose types have been decided) */
305 x264_frame_t *current[X264_BFRAME_MAX+3];
306 /* Temporary buffer (frames types not yet decided) */
307 x264_frame_t *next[X264_BFRAME_MAX+3];
309 x264_frame_t *unused[X264_BFRAME_MAX+3];
310 /* For adaptive B decision */
311 x264_frame_t *last_nonb;
313 /* frames used for reference +1 for decoding + sentinels */
314 x264_frame_t *reference[16+2+1+2];
316 int i_last_idr; /* Frame number of the last IDR */
318 int i_input; /* Number of input frames already accepted */
320 int i_max_dpb; /* Number of frames allocated in the decoded picture buffer */
323 int i_delay; /* Number of frames buffered for B reordering */
324 int b_have_lowres; /* Whether 1/2 resolution luma planes are being used */
327 /* current frame being encoded */
330 /* frame being reconstructed */
333 /* references lists */
335 x264_frame_t *fref0[16+3]; /* ref list 0 */
337 x264_frame_t *fref1[16+3]; /* ref list 1 */
338 int b_ref_reorder[2];
342 /* Current MB DCT coeffs */
345 DECLARE_ALIGNED( int, luma16x16_dc[16], 16 );
346 DECLARE_ALIGNED( int, chroma_dc[2][4], 16 );
347 // FIXME merge with union
348 DECLARE_ALIGNED( int, luma8x8[4][64], 16 );
351 DECLARE_ALIGNED( int, residual_ac[15], 16 );
352 DECLARE_ALIGNED( int, luma4x4[16], 16 );
356 /* MB table and cache for current frame/mb */
359 int i_mb_count; /* number of mbs in a frame */
373 /* Search parameters */
378 int b_noise_reduction;
380 /* Allowed qpel MV range to stay within the picture + emulated edge pixels */
383 /* Subpel MV range for motion search.
384 * same mv_min/max but includes levels' i_mv_range. */
387 /* Fullpel MV range for motion search */
391 /* neighboring MBs */
392 unsigned int i_neighbour;
393 unsigned int i_neighbour8[4]; /* neighbours of each 8x8 or 4x4 block that are available */
394 unsigned int i_neighbour4[16]; /* at the time the block is coded */
397 int i_mb_type_topleft;
398 int i_mb_type_topright;
401 int8_t *type; /* mb type */
402 int8_t *qp; /* mb qp */
403 int16_t *cbp; /* mb cbp: 0x0?: luma, 0x?0: chroma, 0x100: luma dc, 0x0200 and 0x0400: chroma dc (all set for PCM)*/
404 int8_t (*intra4x4_pred_mode)[7]; /* intra4x4 pred mode. for non I4x4 set to I_PRED_4x4_DC(2) */
405 uint8_t (*non_zero_count)[16+4+4]; /* nzc. for I_PCM set to 16 */
406 int8_t *chroma_pred_mode; /* chroma_pred_mode. cabac only. for non intra I_PRED_CHROMA_DC(0) */
407 int16_t (*mv[2])[2]; /* mb mv. set to 0 for intra mb */
408 int16_t (*mvd[2])[2]; /* mb mv difference with predict. set to 0 if intra. cabac only */
409 int8_t *ref[2]; /* mb ref. set to -1 if non used (intra or Lx only) */
410 int16_t (*mvr[2][16])[2]; /* 16x16 mv for each possible ref */
411 int8_t *skipbp; /* block pattern for SKIP or DIRECT (sub)mbs. B-frames + cabac only */
412 int8_t *mb_transform_size; /* transform_size_8x8_flag of each mb */
417 int i_sub_partition[4];
423 int i_intra16x16_pred_mode;
424 int i_chroma_pred_mode;
428 /* space for p_fenc and p_fdec */
429 #define FENC_STRIDE 16
430 #define FDEC_STRIDE 32
431 DECLARE_ALIGNED( uint8_t, fenc_buf[24*FENC_STRIDE], 16 );
432 DECLARE_ALIGNED( uint8_t, fdec_buf[27*FDEC_STRIDE], 16 );
434 /* pointer over mb of the frame to be compressed */
437 /* pointer over mb of the frame to be reconstrucated */
440 /* pointer over mb of the references */
441 uint8_t *p_fref[2][16][4+2]; /* last: lN, lH, lV, lHV, cU, cV */
442 uint16_t *p_integral[2][16];
451 /* real intra4x4_pred_mode if I_4X4 or I_8X8, I_PRED_4x4_DC if mb available, -1 if not */
452 int intra4x4_pred_mode[X264_SCAN8_SIZE];
454 /* i_non_zero_count if availble else 0x80 */
455 int non_zero_count[X264_SCAN8_SIZE];
457 /* -1 if unused, -2 if unavaible */
458 int8_t ref[2][X264_SCAN8_SIZE];
460 /* 0 if non avaible */
461 int16_t mv[2][X264_SCAN8_SIZE][2];
462 int16_t mvd[2][X264_SCAN8_SIZE][2];
464 /* 1 if SKIP or DIRECT. set only for B-frames + CABAC */
465 int8_t skip[X264_SCAN8_SIZE];
467 int16_t direct_mv[2][X264_SCAN8_SIZE][2];
468 int8_t direct_ref[2][X264_SCAN8_SIZE];
470 /* number of neighbors (top and left) that used 8x8 dct */
471 int i_neighbour_transform_size;
472 int b_transform_8x8_allowed;
476 int i_qp; /* current qp */
477 int i_last_qp; /* last qp */
478 int i_last_dqp; /* last delta qp */
479 int b_variable_qp; /* whether qp is allowed to vary per macroblock */
481 int b_direct_auto_read; /* take stats for --direct auto from the 2pass log */
482 int b_direct_auto_write; /* analyse direct modes, to use and/or save */
484 /* B_direct and weighted prediction */
485 int dist_scale_factor[16][16];
486 int bipred_weight[16][16];
487 /* maps fref1[0]'s ref indices into the current list0 */
488 int map_col_to_list0_buf[2]; // for negative indices
489 int map_col_to_list0[16];
492 /* rate control encoding only */
493 x264_ratecontrol_t *rc;
498 /* Current frame stats */
501 /* Headers bits (MV+Ref+MB Block Type */
503 /* Texture bits (Intra/Predicted) */
513 int i_mb_count_8x8dct[2];
514 int i_mb_count_size[7];
515 int i_mb_count_ref[16];
516 /* Estimated (SATD) cost as Intra/Predicted frame */
517 /* XXX: both omit the cost of MBs coded as P_SKIP */
520 /* Adaptive direct mv pred */
521 int i_direct_score[2];
524 /* Cummulated stats */
527 int i_slice_count[5];
528 int64_t i_slice_size[5];
531 int64_t i_sqe_global[5];
532 float f_psnr_average[5];
533 float f_psnr_mean_y[5];
534 float f_psnr_mean_u[5];
535 float f_psnr_mean_v[5];
537 int64_t i_mb_count[5][19];
538 int64_t i_mb_count_8x8dct[2];
539 int64_t i_mb_count_size[2][7];
540 int64_t i_mb_count_ref[2][16];
542 int i_direct_score[2];
543 int i_direct_frames[2];
547 /* CPU functions dependants */
548 x264_predict_t predict_16x16[4+3];
549 x264_predict_t predict_8x8c[4+3];
550 x264_predict8x8_t predict_8x8[9+3];
551 x264_predict_t predict_4x4[9+3];
553 x264_pixel_function_t pixf;
554 x264_mc_functions_t mc;
555 x264_dct_function_t dctf;
556 x264_csp_function_t csp;
557 x264_quant_function_t quantf;
558 x264_deblock_function_t loopf;
560 /* vlc table for decoding purpose only */
561 x264_vlc_table_t *x264_coeff_token_lookup[5];
562 x264_vlc_table_t *x264_level_prefix_lookup;
563 x264_vlc_table_t *x264_total_zeros_lookup[15];
564 x264_vlc_table_t *x264_total_zeros_dc_lookup[3];
565 x264_vlc_table_t *x264_run_before_lookup[7];
568 struct visualize_t *visualize;
572 // included at the end because it needs x264_t
573 #include "macroblock.h"