1 /*****************************************************************************
2 * x264.h: h264 encoder library
3 *****************************************************************************
4 * Copyright (C) 2003-2008 x264 Project
6 * Authors: Laurent Aimar <fenrir@via.ecp.fr>
7 * Loren Merritt <lorenm@u.washington.edu>
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., 51 Franklin Street, Fifth Floor, Boston, MA 02111, USA.
22 *****************************************************************************/
27 #if !defined(_STDINT_H) && !defined(_STDINT_H_) && \
28 !defined(_INTTYPES_H) && !defined(_INTTYPES_H_)
30 # pragma message("You must include stdint.h or inttypes.h before x264.h")
32 # warning You must include stdint.h or inttypes.h before x264.h
41 * opaque handler for encoder */
42 typedef struct x264_t x264_t;
44 /****************************************************************************
45 * Initialisation structure and function.
46 ****************************************************************************/
49 #define X264_CPU_CACHELINE_32 0x000001 /* avoid memory loads that span the border between two cachelines */
50 #define X264_CPU_CACHELINE_64 0x000002 /* 32/64 is the size of a cacheline in bytes */
51 #define X264_CPU_ALTIVEC 0x000004
52 #define X264_CPU_MMX 0x000008
53 #define X264_CPU_MMXEXT 0x000010 /* MMX2 aka MMXEXT aka ISSE */
54 #define X264_CPU_SSE 0x000020
55 #define X264_CPU_SSE2 0x000040
56 #define X264_CPU_SSE2_IS_SLOW 0x000080 /* avoid most SSE2 functions on Athlon64 */
57 #define X264_CPU_SSE2_IS_FAST 0x000100 /* a few functions are only faster on Core2 and Phenom */
58 #define X264_CPU_SSE3 0x000200
59 #define X264_CPU_SSSE3 0x000400
60 #define X264_CPU_SHUFFLE_IS_FAST 0x000800 /* Penryn, Nehalem, and Phenom have fast shuffle units */
61 #define X264_CPU_STACK_MOD4 0x001000 /* if stack is only mod4 and not mod16 */
62 #define X264_CPU_SSE4 0x002000 /* SSE4.1 */
63 #define X264_CPU_SSE42 0x004000 /* SSE4.2 */
64 #define X264_CPU_SSE_MISALIGN 0x008000 /* Phenom support for misaligned SSE instruction arguments */
65 #define X264_CPU_LZCNT 0x010000 /* Phenom support for "leading zero count" instruction. */
66 #define X264_CPU_ARMV6 0x020000
67 #define X264_CPU_NEON 0x040000 /* ARM NEON */
68 #define X264_CPU_FAST_NEON_MRC 0x080000 /* Transfer from NEON to ARM register is fast (Cortex-A9) */
72 #define X264_ANALYSE_I4x4 0x0001 /* Analyse i4x4 */
73 #define X264_ANALYSE_I8x8 0x0002 /* Analyse i8x8 (requires 8x8 transform) */
74 #define X264_ANALYSE_PSUB16x16 0x0010 /* Analyse p16x8, p8x16 and p8x8 */
75 #define X264_ANALYSE_PSUB8x8 0x0020 /* Analyse p8x4, p4x8, p4x4 */
76 #define X264_ANALYSE_BSUB16x16 0x0100 /* Analyse b16x8, b8x16 and b8x8 */
77 #define X264_DIRECT_PRED_NONE 0
78 #define X264_DIRECT_PRED_SPATIAL 1
79 #define X264_DIRECT_PRED_TEMPORAL 2
80 #define X264_DIRECT_PRED_AUTO 3
85 #define X264_ME_TESA 4
86 #define X264_CQM_FLAT 0
87 #define X264_CQM_JVT 1
88 #define X264_CQM_CUSTOM 2
92 #define X264_AQ_NONE 0
93 #define X264_AQ_VARIANCE 1
94 #define X264_AQ_AUTOVARIANCE 2
95 #define X264_B_ADAPT_NONE 0
96 #define X264_B_ADAPT_FAST 1
97 #define X264_B_ADAPT_TRELLIS 2
98 #define X264_WEIGHTP_NONE 0
99 #define X264_WEIGHTP_BLIND 1
100 #define X264_WEIGHTP_SMART 2
101 #define X264_B_PYRAMID_NONE 0
102 #define X264_B_PYRAMID_STRICT 1
103 #define X264_B_PYRAMID_NORMAL 2
105 static const char * const x264_direct_pred_names[] = { "none", "spatial", "temporal", "auto", 0 };
106 static const char * const x264_motion_est_names[] = { "dia", "hex", "umh", "esa", "tesa", 0 };
107 static const char * const x264_b_pyramid_names[] = { "none", "strict", "normal", 0 };
108 static const char * const x264_overscan_names[] = { "undef", "show", "crop", 0 };
109 static const char * const x264_vidformat_names[] = { "component", "pal", "ntsc", "secam", "mac", "undef", 0 };
110 static const char * const x264_fullrange_names[] = { "off", "on", 0 };
111 static const char * const x264_colorprim_names[] = { "", "bt709", "undef", "", "bt470m", "bt470bg", "smpte170m", "smpte240m", "film", 0 };
112 static const char * const x264_transfer_names[] = { "", "bt709", "undef", "", "bt470m", "bt470bg", "smpte170m", "smpte240m", "linear", "log100", "log316", 0 };
113 static const char * const x264_colmatrix_names[] = { "GBR", "bt709", "undef", "", "fcc", "bt470bg", "smpte170m", "smpte240m", "YCgCo", 0 };
116 * legacy only; nothing other than I420 is really supported. */
117 #define X264_CSP_MASK 0x00ff /* */
118 #define X264_CSP_NONE 0x0000 /* Invalid mode */
119 #define X264_CSP_I420 0x0001 /* yuv 4:2:0 planar */
120 #define X264_CSP_I422 0x0002 /* yuv 4:2:2 planar */
121 #define X264_CSP_I444 0x0003 /* yuv 4:4:4 planar */
122 #define X264_CSP_YV12 0x0004 /* yuv 4:2:0 planar */
123 #define X264_CSP_YUYV 0x0005 /* yuv 4:2:2 packed */
124 #define X264_CSP_RGB 0x0006 /* rgb 24bits */
125 #define X264_CSP_BGR 0x0007 /* bgr 24bits */
126 #define X264_CSP_BGRA 0x0008 /* bgr 32bits */
127 #define X264_CSP_MAX 0x0009 /* end of list */
128 #define X264_CSP_VFLIP 0x1000 /* */
131 #define X264_TYPE_AUTO 0x0000 /* Let x264 choose the right type */
132 #define X264_TYPE_IDR 0x0001
133 #define X264_TYPE_I 0x0002
134 #define X264_TYPE_P 0x0003
135 #define X264_TYPE_BREF 0x0004 /* Non-disposable B-frame */
136 #define X264_TYPE_B 0x0005
137 #define IS_X264_TYPE_I(x) ((x)==X264_TYPE_I || (x)==X264_TYPE_IDR)
138 #define IS_X264_TYPE_B(x) ((x)==X264_TYPE_B || (x)==X264_TYPE_BREF)
141 #define X264_LOG_NONE (-1)
142 #define X264_LOG_ERROR 0
143 #define X264_LOG_WARNING 1
144 #define X264_LOG_INFO 2
145 #define X264_LOG_DEBUG 3
148 #define X264_THREADS_AUTO 0 /* Automatically select optimal number of threads */
149 #define X264_SYNC_LOOKAHEAD_AUTO (-1) /* Automatically select optimal lookahead thread buffer size */
151 /* Zones: override ratecontrol or other options for specific sections of the video.
152 * See x264_encoder_reconfig() for which options can be changed.
153 * If zones overlap, whichever comes later in the list takes precedence. */
156 int i_start, i_end; /* range of frame numbers */
157 int b_force_qp; /* whether to use qp vs bitrate factor */
159 float f_bitrate_factor;
160 struct x264_param_t *param;
163 typedef struct x264_param_t
167 int i_threads; /* encode multiple frames in parallel */
168 int b_sliced_threads; /* Whether to use slice-based threading. */
169 int b_deterministic; /* whether to allow non-deterministic optimizations when threaded */
170 int i_sync_lookahead; /* threaded lookahead buffer */
172 /* Video Properties */
175 int i_csp; /* CSP of encoded bitstream, only i420 supported */
177 int i_frame_total; /* number of frames to encode if known, else 0 */
181 /* they will be reduced to be 0 < x <= 65535 and prime */
185 int i_overscan; /* 0=undef, 1=no overscan, 2=overscan */
187 /* see h264 annex E for the values of the following */
193 int i_chroma_loc; /* both top & bottom */
199 /* Bitstream parameters */
200 int i_frame_reference; /* Maximum number of reference frames */
201 int i_keyint_max; /* Force an IDR keyframe at this interval */
202 int i_keyint_min; /* Scenecuts closer together than this are coded as I, not IDR. */
203 int i_scenecut_threshold; /* how aggressively to insert extra I frames */
204 int b_intra_refresh; /* Whether or not to use periodic intra refresh instead of IDR frames. */
206 int i_bframe; /* how many b-frame between 2 references pictures */
207 int i_bframe_adaptive;
209 int i_bframe_pyramid; /* Keep some B-frames as references: 0=off, 1=strict hierarchical, 2=normal */
211 int b_deblocking_filter;
212 int i_deblocking_filter_alphac0; /* [-6, 6] -6 light filter, 6 strong */
213 int i_deblocking_filter_beta; /* [-6, 6] idem */
216 int i_cabac_init_idc;
219 int b_constrained_intra;
222 char *psz_cqm_file; /* JM format */
223 uint8_t cqm_4iy[16]; /* used only if i_cqm_preset == X264_CQM_CUSTOM */
231 void (*pf_log)( void *, int i_level, const char *psz, va_list );
235 char *psz_dump_yuv; /* filename for reconstructed frames */
237 /* Encoder analyser parameters */
240 unsigned int intra; /* intra partitions */
241 unsigned int inter; /* inter partitions */
244 int i_weighted_pred; /* weighting for P-frames */
245 int b_weighted_bipred; /* implicit weighting for B-frames */
246 int i_direct_mv_pred; /* spatial vs temporal mv prediction */
247 int i_chroma_qp_offset;
249 int i_me_method; /* motion estimation algorithm to use (X264_ME_*) */
250 int i_me_range; /* integer pixel motion estimation search range (from predicted mv) */
251 int i_mv_range; /* maximum length of a mv (in pixels). -1 = auto, based on level */
252 int i_mv_range_thread; /* minimum space between threads. -1 = auto, based on number of threads. */
253 int i_subpel_refine; /* subpixel motion estimation quality */
254 int b_chroma_me; /* chroma ME for subpel and mode decision in P-frames */
255 int b_mixed_references; /* allow each mb partition in P-frames to have it's own reference number */
256 int i_trellis; /* trellis RD quantization */
257 int b_fast_pskip; /* early SKIP detection on P-frames */
258 int b_dct_decimate; /* transform coefficient thresholding on P-frames */
259 int i_noise_reduction; /* adaptive pseudo-deadzone */
260 float f_psy_rd; /* Psy RD strength */
261 float f_psy_trellis; /* Psy trellis strength */
262 int b_psy; /* Toggle all psy optimizations */
264 /* the deadzone size that will be used in luma quantization */
265 int i_luma_deadzone[2]; /* {inter, intra} */
267 int b_psnr; /* compute and print PSNR stats */
268 int b_ssim; /* compute and print SSIM stats */
271 /* Rate control parameters */
274 int i_rc_method; /* X264_RC_* */
276 int i_qp_constant; /* 0-51 */
277 int i_qp_min; /* min allowed QP value */
278 int i_qp_max; /* max allowed QP value */
279 int i_qp_step; /* max QP step between frames */
282 float f_rf_constant; /* 1pass VBR, nominal QP */
283 float f_rate_tolerance;
284 int i_vbv_max_bitrate;
285 int i_vbv_buffer_size;
286 float f_vbv_buffer_init; /* <=1: fraction of buffer_size. >1: kbit */
290 int i_aq_mode; /* psy adaptive QP. (X264_AQ_*) */
292 int b_mb_tree; /* Macroblock-tree ratecontrol. */
296 int b_stat_write; /* Enable stat writing in psz_stat_out */
298 int b_stat_read; /* Read stat from psz_stat_in and use it */
301 /* 2pass params (same as ffmpeg ones) */
302 float f_qcompress; /* 0.0 => cbr, 1.0 => constant qp */
303 float f_qblur; /* temporally blur quants */
304 float f_complexity_blur; /* temporally blur complexity */
305 x264_zone_t *zones; /* ratecontrol overrides */
306 int i_zones; /* number of zone_t's */
307 char *psz_zones; /* alternate method of specifying zones */
310 /* Muxing parameters */
311 int b_aud; /* generate access unit delimiters */
312 int b_repeat_headers; /* put SPS/PPS before each keyframe */
313 int b_annexb; /* if set, place start codes (4 bytes) before NAL units,
314 * otherwise place size (4 bytes) before NAL units. */
315 int i_sps_id; /* SPS and PPS id number */
316 int b_vfr_input; /* VFR input */
317 int i_timebase_num; /* Timebase numerator */
318 int i_timebase_den; /* Timebase denominator */
320 /* Slicing parameters */
321 int i_slice_max_size; /* Max size per slice in bytes; includes estimated NAL overhead. */
322 int i_slice_max_mbs; /* Max number of MBs per slice; overrides i_slice_count. */
323 int i_slice_count; /* Number of slices per frame: forces rectangular slices. */
325 /* Optional callback for freeing this x264_param_t when it is done being used.
326 * Only used when the x264_param_t sits in memory for an indefinite period of time,
327 * i.e. when an x264_param_t is passed to x264_t in an x264_picture_t or in zones.
328 * Not used when x264_encoder_reconfig is called directly. */
329 void (*param_free)( void* );
334 int mbps; /* max macroblock processing rate (macroblocks/sec) */
335 int frame_size; /* max frame size (macroblocks) */
336 int dpb; /* max decoded picture buffer (bytes) */
337 int bitrate; /* max bitrate (kbit/sec) */
338 int cpb; /* max vbv buffer (kbit) */
339 int mv_range; /* max vertical mv component range (pixels) */
340 int mvs_per_2mb; /* max mvs per 2 consecutive mbs. */
341 int slice_rate; /* ?? */
342 int bipred8x8; /* limit bipred to >=8x8 */
343 int direct8x8; /* limit b_direct to >=8x8 */
344 int frame_only; /* forbid interlacing */
347 /* all of the levels defined in the standard, terminated by .level_idc=0 */
348 extern const x264_level_t x264_levels[];
350 /* x264_param_default:
351 * fill x264_param_t with default values and do CPU detection */
352 void x264_param_default( x264_param_t * );
355 * set one parameter by name.
356 * returns 0 on success, or returns one of the following errors.
357 * note: BAD_VALUE occurs only if it can't even parse the value,
358 * numerical range is not checked until x264_encoder_open() or
359 * x264_encoder_reconfig().
360 * value=NULL means "true" for boolean options, but is a BAD_VALUE for non-booleans. */
361 #define X264_PARAM_BAD_NAME (-1)
362 #define X264_PARAM_BAD_VALUE (-2)
363 int x264_param_parse( x264_param_t *, const char *name, const char *value );
365 /****************************************************************************
366 * Picture structures and functions.
367 ****************************************************************************/
379 /* In: force picture type (if not auto)
380 * If x264 encoding parameters are violated in the forcing of picture types,
381 * x264 will correct the input picture type and log a warning.
382 * The quality of frametype decisions may suffer if a great deal of fine-grained
383 * mixing of auto and forced frametypes is done.
384 * Out: type of the picture encoded */
386 /* In: force quantizer for > 0 */
388 /* Out: whether this frame is a keyframe. Important when using modes that result in
389 * SEI recovery points being used instead of IDR frames. */
391 /* In: user pts, Out: pts of encoded picture (user)*/
393 /* Out: frame dts. Since the pts of the first frame is always zero,
394 * initial frames may have a negative dts which must be dealt with by any muxer */
396 /* In: custom encoding parameters to be set from this frame forwards
397 (in coded order, not display order). If NULL, continue using
398 parameters from the previous frame. Some parameters, such as
399 aspect ratio, can only be changed per-GOP due to the limitations
400 of H.264 itself; in this case, the caller must force an IDR frame
401 if it needs the changed parameter to apply immediately. */
405 /* private user data. libx264 doesn't touch this,
406 not even copy it from input to output frames. */
410 /* x264_picture_alloc:
411 * alloc data for a picture. You must call x264_picture_clean on it.
412 * returns 0 on success, or -1 on malloc failure. */
413 int x264_picture_alloc( x264_picture_t *pic, int i_csp, int i_width, int i_height );
415 /* x264_picture_clean:
416 * free associated resource for a x264_picture_t allocated with
417 * x264_picture_alloc ONLY */
418 void x264_picture_clean( x264_picture_t *pic );
420 /****************************************************************************
421 * NAL structure and functions:
422 ****************************************************************************/
431 NAL_SLICE_IDR = 5, /* ref_idc != 0 */
432 NAL_SEI = 6, /* ref_idc == 0 */
436 /* ref_idc == 0 for 6,9,10,11,12 */
440 NAL_PRIORITY_DISPOSABLE = 0,
441 NAL_PRIORITY_LOW = 1,
442 NAL_PRIORITY_HIGH = 2,
443 NAL_PRIORITY_HIGHEST = 3,
446 /* The data within the payload is already NAL-encapsulated; the ref_idc and type
447 * are merely in the struct for easy access by the calling application.
448 * All data returned in an x264_nal_t, including the data in p_payload, is no longer
449 * valid after the next call to x264_encoder_encode. Thus it must be used or copied
450 * before calling x264_encoder_encode or x264_encoder_headers again. */
453 int i_ref_idc; /* nal_priority_e */
454 int i_type; /* nal_unit_type_e */
456 /* Size of payload in bytes. */
458 /* If param->b_annexb is set, Annex-B bytestream with 4-byte startcode.
459 * Otherwise, startcode is replaced with a 4-byte size.
460 * This size is the size used in mp4/similar muxing; it is equal to i_payload-4 */
464 /****************************************************************************
466 ****************************************************************************/
468 /* Force a link error in the case of linking against an incompatible API version.
469 * Glue #defines exist to force correct macro expansion; the final output of the macro
470 * is x264_encoder_open_##X264_BUILD (for purposes of dlopen). */
471 #define x264_encoder_glue1(x,y) x##y
472 #define x264_encoder_glue2(x,y) x264_encoder_glue1(x,y)
473 #define x264_encoder_open x264_encoder_glue2(x264_encoder_open_,X264_BUILD)
475 /* x264_encoder_open:
476 * create a new encoder handler, all parameters from x264_param_t are copied */
477 x264_t *x264_encoder_open( x264_param_t * );
479 /* x264_encoder_reconfig:
480 * analysis-related parameters from x264_param_t are copied.
481 * this takes effect immediately, on whichever frame is encoded next;
482 * due to delay, this may not be the next frame passed to encoder_encode.
483 * if the change should apply to some particular frame, use x264_picture_t->param instead.
484 * returns 0 on success, negative on parameter validation error. */
485 int x264_encoder_reconfig( x264_t *, x264_param_t * );
486 /* x264_encoder_parameters:
487 * copies the current internal set of parameters to the pointer provided
488 * by the caller. useful when the calling application needs to know
489 * how x264_encoder_open has changed the parameters, or the current state
490 * of the encoder after multiple x264_encoder_reconfig calls.
491 * note that the data accessible through pointers in the returned param struct
492 * (e.g. filenames) should not be modified by the calling application. */
493 void x264_encoder_parameters( x264_t *, x264_param_t * );
494 /* x264_encoder_headers:
495 * return the SPS and PPS that will be used for the whole stream.
496 * if i_nal > 0, returns the total size of all NAL payloads.
497 * returns negative on error.
498 * the payloads of all output NALs are guaranteed to be sequential in memory. */
499 int x264_encoder_headers( x264_t *, x264_nal_t **, int * );
500 /* x264_encoder_encode:
501 * encode one picture.
502 * if i_nal > 0, returns the total size of all NAL payloads.
503 * returns negative on error, zero if no NAL units returned.
504 * the payloads of all output NALs are guaranteed to be sequential in memory. */
505 int x264_encoder_encode ( x264_t *, x264_nal_t **, int *, x264_picture_t *, x264_picture_t * );
506 /* x264_encoder_close:
507 * close an encoder handler */
508 void x264_encoder_close ( x264_t * );
509 /* x264_encoder_delayed_frames:
510 * return the number of currently delayed (buffered) frames
511 * this should be used at the end of the stream, to know when you have all the encoded frames. */
512 int x264_encoder_delayed_frames( x264_t * );