Remove compatibility workarounds
[x264.git] / x264.h
blob585a75718b06308547662feff9696998c4e9f0a2
1 /*****************************************************************************
2 * x264.h: x264 public header
3 *****************************************************************************
4 * Copyright (C) 2003-2019 x264 project
6 * Authors: Laurent Aimar <fenrir@via.ecp.fr>
7 * Loren Merritt <lorenm@u.washington.edu>
8 * Fiona Glaser <fiona@x264.com>
10 * This program is free software; you can redistribute it and/or modify
11 * it under the terms of the GNU General Public License as published by
12 * the Free Software Foundation; either version 2 of the License, or
13 * (at your option) any later version.
15 * This program is distributed in the hope that it will be useful,
16 * but WITHOUT ANY WARRANTY; without even the implied warranty of
17 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
18 * GNU General Public License for more details.
20 * You should have received a copy of the GNU General Public License
21 * along with this program; if not, write to the Free Software
22 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02111, USA.
24 * This program is also available under a commercial proprietary license.
25 * For more information, contact us at licensing@x264.com.
26 *****************************************************************************/
28 #ifndef X264_X264_H
29 #define X264_X264_H
31 #ifdef __cplusplus
32 extern "C" {
33 #endif
35 #if !defined(_STDINT_H) && !defined(_STDINT_H_) && !defined(_STDINT_H_INCLUDED) && !defined(_STDINT) &&\
36 !defined(_SYS_STDINT_H_) && !defined(_INTTYPES_H) && !defined(_INTTYPES_H_) && !defined(_INTTYPES)
37 # ifdef _MSC_VER
38 # pragma message("You must include stdint.h or inttypes.h before x264.h")
39 # else
40 # warning You must include stdint.h or inttypes.h before x264.h
41 # endif
42 #endif
44 #include <stdarg.h>
46 #include "x264_config.h"
48 #define X264_BUILD 157
50 /* Application developers planning to link against a shared library version of
51 * libx264 from a Microsoft Visual Studio or similar development environment
52 * will need to define X264_API_IMPORTS before including this header.
53 * This clause does not apply to MinGW, similar development environments, or non
54 * Windows platforms. */
55 #ifdef X264_API_IMPORTS
56 #define X264_API __declspec(dllimport)
57 #else
58 #define X264_API
59 #endif
61 /* x264_t:
62 * opaque handler for encoder */
63 typedef struct x264_t x264_t;
65 /****************************************************************************
66 * NAL structure and functions
67 ****************************************************************************/
69 enum nal_unit_type_e
71 NAL_UNKNOWN = 0,
72 NAL_SLICE = 1,
73 NAL_SLICE_DPA = 2,
74 NAL_SLICE_DPB = 3,
75 NAL_SLICE_DPC = 4,
76 NAL_SLICE_IDR = 5, /* ref_idc != 0 */
77 NAL_SEI = 6, /* ref_idc == 0 */
78 NAL_SPS = 7,
79 NAL_PPS = 8,
80 NAL_AUD = 9,
81 NAL_FILLER = 12,
82 /* ref_idc == 0 for 6,9,10,11,12 */
84 enum nal_priority_e
86 NAL_PRIORITY_DISPOSABLE = 0,
87 NAL_PRIORITY_LOW = 1,
88 NAL_PRIORITY_HIGH = 2,
89 NAL_PRIORITY_HIGHEST = 3,
92 /* The data within the payload is already NAL-encapsulated; the ref_idc and type
93 * are merely in the struct for easy access by the calling application.
94 * All data returned in an x264_nal_t, including the data in p_payload, is no longer
95 * valid after the next call to x264_encoder_encode. Thus it must be used or copied
96 * before calling x264_encoder_encode or x264_encoder_headers again. */
97 typedef struct x264_nal_t
99 int i_ref_idc; /* nal_priority_e */
100 int i_type; /* nal_unit_type_e */
101 int b_long_startcode;
102 int i_first_mb; /* If this NAL is a slice, the index of the first MB in the slice. */
103 int i_last_mb; /* If this NAL is a slice, the index of the last MB in the slice. */
105 /* Size of payload (including any padding) in bytes. */
106 int i_payload;
107 /* If param->b_annexb is set, Annex-B bytestream with startcode.
108 * Otherwise, startcode is replaced with a 4-byte size.
109 * This size is the size used in mp4/similar muxing; it is equal to i_payload-4 */
110 uint8_t *p_payload;
112 /* Size of padding in bytes. */
113 int i_padding;
114 } x264_nal_t;
116 /****************************************************************************
117 * Encoder parameters
118 ****************************************************************************/
119 /* CPU flags */
121 /* x86 */
122 #define X264_CPU_MMX (1<<0)
123 #define X264_CPU_MMX2 (1<<1) /* MMX2 aka MMXEXT aka ISSE */
124 #define X264_CPU_MMXEXT X264_CPU_MMX2
125 #define X264_CPU_SSE (1<<2)
126 #define X264_CPU_SSE2 (1<<3)
127 #define X264_CPU_LZCNT (1<<4)
128 #define X264_CPU_SSE3 (1<<5)
129 #define X264_CPU_SSSE3 (1<<6)
130 #define X264_CPU_SSE4 (1<<7) /* SSE4.1 */
131 #define X264_CPU_SSE42 (1<<8) /* SSE4.2 */
132 #define X264_CPU_AVX (1<<9) /* Requires OS support even if YMM registers aren't used */
133 #define X264_CPU_XOP (1<<10) /* AMD XOP */
134 #define X264_CPU_FMA4 (1<<11) /* AMD FMA4 */
135 #define X264_CPU_FMA3 (1<<12)
136 #define X264_CPU_BMI1 (1<<13)
137 #define X264_CPU_BMI2 (1<<14)
138 #define X264_CPU_AVX2 (1<<15)
139 #define X264_CPU_AVX512 (1<<16) /* AVX-512 {F, CD, BW, DQ, VL}, requires OS support */
140 /* x86 modifiers */
141 #define X264_CPU_CACHELINE_32 (1<<17) /* avoid memory loads that span the border between two cachelines */
142 #define X264_CPU_CACHELINE_64 (1<<18) /* 32/64 is the size of a cacheline in bytes */
143 #define X264_CPU_SSE2_IS_SLOW (1<<19) /* avoid most SSE2 functions on Athlon64 */
144 #define X264_CPU_SSE2_IS_FAST (1<<20) /* a few functions are only faster on Core2 and Phenom */
145 #define X264_CPU_SLOW_SHUFFLE (1<<21) /* The Conroe has a slow shuffle unit (relative to overall SSE performance) */
146 #define X264_CPU_STACK_MOD4 (1<<22) /* if stack is only mod4 and not mod16 */
147 #define X264_CPU_SLOW_ATOM (1<<23) /* The Atom is terrible: slow SSE unaligned loads, slow
148 * SIMD multiplies, slow SIMD variable shifts, slow pshufb,
149 * cacheline split penalties -- gather everything here that
150 * isn't shared by other CPUs to avoid making half a dozen
151 * new SLOW flags. */
152 #define X264_CPU_SLOW_PSHUFB (1<<24) /* such as on the Intel Atom */
153 #define X264_CPU_SLOW_PALIGNR (1<<25) /* such as on the AMD Bobcat */
155 /* PowerPC */
156 #define X264_CPU_ALTIVEC 0x0000001
158 /* ARM and AArch64 */
159 #define X264_CPU_ARMV6 0x0000001
160 #define X264_CPU_NEON 0x0000002 /* ARM NEON */
161 #define X264_CPU_FAST_NEON_MRC 0x0000004 /* Transfer from NEON to ARM register is fast (Cortex-A9) */
162 #define X264_CPU_ARMV8 0x0000008
164 /* MIPS */
165 #define X264_CPU_MSA 0x0000001 /* MIPS MSA */
167 /* Analyse flags */
168 #define X264_ANALYSE_I4x4 0x0001 /* Analyse i4x4 */
169 #define X264_ANALYSE_I8x8 0x0002 /* Analyse i8x8 (requires 8x8 transform) */
170 #define X264_ANALYSE_PSUB16x16 0x0010 /* Analyse p16x8, p8x16 and p8x8 */
171 #define X264_ANALYSE_PSUB8x8 0x0020 /* Analyse p8x4, p4x8, p4x4 */
172 #define X264_ANALYSE_BSUB16x16 0x0100 /* Analyse b16x8, b8x16 and b8x8 */
174 #define X264_DIRECT_PRED_NONE 0
175 #define X264_DIRECT_PRED_SPATIAL 1
176 #define X264_DIRECT_PRED_TEMPORAL 2
177 #define X264_DIRECT_PRED_AUTO 3
178 #define X264_ME_DIA 0
179 #define X264_ME_HEX 1
180 #define X264_ME_UMH 2
181 #define X264_ME_ESA 3
182 #define X264_ME_TESA 4
183 #define X264_CQM_FLAT 0
184 #define X264_CQM_JVT 1
185 #define X264_CQM_CUSTOM 2
186 #define X264_RC_CQP 0
187 #define X264_RC_CRF 1
188 #define X264_RC_ABR 2
189 #define X264_QP_AUTO 0
190 #define X264_AQ_NONE 0
191 #define X264_AQ_VARIANCE 1
192 #define X264_AQ_AUTOVARIANCE 2
193 #define X264_AQ_AUTOVARIANCE_BIASED 3
194 #define X264_B_ADAPT_NONE 0
195 #define X264_B_ADAPT_FAST 1
196 #define X264_B_ADAPT_TRELLIS 2
197 #define X264_WEIGHTP_NONE 0
198 #define X264_WEIGHTP_SIMPLE 1
199 #define X264_WEIGHTP_SMART 2
200 #define X264_B_PYRAMID_NONE 0
201 #define X264_B_PYRAMID_STRICT 1
202 #define X264_B_PYRAMID_NORMAL 2
203 #define X264_KEYINT_MIN_AUTO 0
204 #define X264_KEYINT_MAX_INFINITE (1<<30)
206 /* AVC-Intra flavors */
207 #define X264_AVCINTRA_FLAVOR_PANASONIC 0
208 #define X264_AVCINTRA_FLAVOR_SONY 1
210 static const char * const x264_direct_pred_names[] = { "none", "spatial", "temporal", "auto", 0 };
211 static const char * const x264_motion_est_names[] = { "dia", "hex", "umh", "esa", "tesa", 0 };
212 static const char * const x264_b_pyramid_names[] = { "none", "strict", "normal", 0 };
213 static const char * const x264_overscan_names[] = { "undef", "show", "crop", 0 };
214 static const char * const x264_vidformat_names[] = { "component", "pal", "ntsc", "secam", "mac", "undef", 0 };
215 static const char * const x264_fullrange_names[] = { "off", "on", 0 };
216 static const char * const x264_colorprim_names[] = { "", "bt709", "undef", "", "bt470m", "bt470bg", "smpte170m", "smpte240m", "film", "bt2020", "smpte428",
217 "smpte431", "smpte432", 0 };
218 static const char * const x264_transfer_names[] = { "", "bt709", "undef", "", "bt470m", "bt470bg", "smpte170m", "smpte240m", "linear", "log100", "log316",
219 "iec61966-2-4", "bt1361e", "iec61966-2-1", "bt2020-10", "bt2020-12", "smpte2084", "smpte428", "arib-std-b67", 0 };
220 static const char * const x264_colmatrix_names[] = { "GBR", "bt709", "undef", "", "fcc", "bt470bg", "smpte170m", "smpte240m", "YCgCo", "bt2020nc", "bt2020c",
221 "smpte2085", "chroma-derived-nc", "chroma-derived-c", "ICtCp", 0 };
222 static const char * const x264_nal_hrd_names[] = { "none", "vbr", "cbr", 0 };
223 static const char * const x264_avcintra_flavor_names[] = { "panasonic", "sony", 0 };
225 /* Colorspace type */
226 #define X264_CSP_MASK 0x00ff /* */
227 #define X264_CSP_NONE 0x0000 /* Invalid mode */
228 #define X264_CSP_I400 0x0001 /* monochrome 4:0:0 */
229 #define X264_CSP_I420 0x0002 /* yuv 4:2:0 planar */
230 #define X264_CSP_YV12 0x0003 /* yvu 4:2:0 planar */
231 #define X264_CSP_NV12 0x0004 /* yuv 4:2:0, with one y plane and one packed u+v */
232 #define X264_CSP_NV21 0x0005 /* yuv 4:2:0, with one y plane and one packed v+u */
233 #define X264_CSP_I422 0x0006 /* yuv 4:2:2 planar */
234 #define X264_CSP_YV16 0x0007 /* yvu 4:2:2 planar */
235 #define X264_CSP_NV16 0x0008 /* yuv 4:2:2, with one y plane and one packed u+v */
236 #define X264_CSP_YUYV 0x0009 /* yuyv 4:2:2 packed */
237 #define X264_CSP_UYVY 0x000a /* uyvy 4:2:2 packed */
238 #define X264_CSP_V210 0x000b /* 10-bit yuv 4:2:2 packed in 32 */
239 #define X264_CSP_I444 0x000c /* yuv 4:4:4 planar */
240 #define X264_CSP_YV24 0x000d /* yvu 4:4:4 planar */
241 #define X264_CSP_BGR 0x000e /* packed bgr 24bits */
242 #define X264_CSP_BGRA 0x000f /* packed bgr 32bits */
243 #define X264_CSP_RGB 0x0010 /* packed rgb 24bits */
244 #define X264_CSP_MAX 0x0011 /* end of list */
245 #define X264_CSP_VFLIP 0x1000 /* the csp is vertically flipped */
246 #define X264_CSP_HIGH_DEPTH 0x2000 /* the csp has a depth of 16 bits per pixel component */
248 /* Slice type */
249 #define X264_TYPE_AUTO 0x0000 /* Let x264 choose the right type */
250 #define X264_TYPE_IDR 0x0001
251 #define X264_TYPE_I 0x0002
252 #define X264_TYPE_P 0x0003
253 #define X264_TYPE_BREF 0x0004 /* Non-disposable B-frame */
254 #define X264_TYPE_B 0x0005
255 #define X264_TYPE_KEYFRAME 0x0006 /* IDR or I depending on b_open_gop option */
256 #define IS_X264_TYPE_I(x) ((x)==X264_TYPE_I || (x)==X264_TYPE_IDR || (x)==X264_TYPE_KEYFRAME)
257 #define IS_X264_TYPE_B(x) ((x)==X264_TYPE_B || (x)==X264_TYPE_BREF)
259 /* Log level */
260 #define X264_LOG_NONE (-1)
261 #define X264_LOG_ERROR 0
262 #define X264_LOG_WARNING 1
263 #define X264_LOG_INFO 2
264 #define X264_LOG_DEBUG 3
266 /* Threading */
267 #define X264_THREADS_AUTO 0 /* Automatically select optimal number of threads */
268 #define X264_SYNC_LOOKAHEAD_AUTO (-1) /* Automatically select optimal lookahead thread buffer size */
270 /* HRD */
271 #define X264_NAL_HRD_NONE 0
272 #define X264_NAL_HRD_VBR 1
273 #define X264_NAL_HRD_CBR 2
275 /* Zones: override ratecontrol or other options for specific sections of the video.
276 * See x264_encoder_reconfig() for which options can be changed.
277 * If zones overlap, whichever comes later in the list takes precedence. */
278 typedef struct x264_zone_t
280 int i_start, i_end; /* range of frame numbers */
281 int b_force_qp; /* whether to use qp vs bitrate factor */
282 int i_qp;
283 float f_bitrate_factor;
284 struct x264_param_t *param;
285 } x264_zone_t;
287 typedef struct x264_param_t
289 /* CPU flags */
290 unsigned int cpu;
291 int i_threads; /* encode multiple frames in parallel */
292 int i_lookahead_threads; /* multiple threads for lookahead analysis */
293 int b_sliced_threads; /* Whether to use slice-based threading. */
294 int b_deterministic; /* whether to allow non-deterministic optimizations when threaded */
295 int b_cpu_independent; /* force canonical behavior rather than cpu-dependent optimal algorithms */
296 int i_sync_lookahead; /* threaded lookahead buffer */
298 /* Video Properties */
299 int i_width;
300 int i_height;
301 int i_csp; /* CSP of encoded bitstream */
302 int i_bitdepth;
303 int i_level_idc;
304 int i_frame_total; /* number of frames to encode if known, else 0 */
306 /* NAL HRD
307 * Uses Buffering and Picture Timing SEIs to signal HRD
308 * The HRD in H.264 was not designed with VFR in mind.
309 * It is therefore not recommendeded to use NAL HRD with VFR.
310 * Furthermore, reconfiguring the VBV (via x264_encoder_reconfig)
311 * will currently generate invalid HRD. */
312 int i_nal_hrd;
314 struct
316 /* they will be reduced to be 0 < x <= 65535 and prime */
317 int i_sar_height;
318 int i_sar_width;
320 int i_overscan; /* 0=undef, 1=no overscan, 2=overscan */
322 /* see h264 annex E for the values of the following */
323 int i_vidformat;
324 int b_fullrange;
325 int i_colorprim;
326 int i_transfer;
327 int i_colmatrix;
328 int i_chroma_loc; /* both top & bottom */
329 } vui;
331 /* Bitstream parameters */
332 int i_frame_reference; /* Maximum number of reference frames */
333 int i_dpb_size; /* Force a DPB size larger than that implied by B-frames and reference frames.
334 * Useful in combination with interactive error resilience. */
335 int i_keyint_max; /* Force an IDR keyframe at this interval */
336 int i_keyint_min; /* Scenecuts closer together than this are coded as I, not IDR. */
337 int i_scenecut_threshold; /* how aggressively to insert extra I frames */
338 int b_intra_refresh; /* Whether or not to use periodic intra refresh instead of IDR frames. */
340 int i_bframe; /* how many b-frame between 2 references pictures */
341 int i_bframe_adaptive;
342 int i_bframe_bias;
343 int i_bframe_pyramid; /* Keep some B-frames as references: 0=off, 1=strict hierarchical, 2=normal */
344 int b_open_gop;
345 int b_bluray_compat;
346 int i_avcintra_class;
347 int i_avcintra_flavor;
349 int b_deblocking_filter;
350 int i_deblocking_filter_alphac0; /* [-6, 6] -6 light filter, 6 strong */
351 int i_deblocking_filter_beta; /* [-6, 6] idem */
353 int b_cabac;
354 int i_cabac_init_idc;
356 int b_interlaced;
357 int b_constrained_intra;
359 int i_cqm_preset;
360 char *psz_cqm_file; /* filename (in UTF-8) of CQM file, JM format */
361 uint8_t cqm_4iy[16]; /* used only if i_cqm_preset == X264_CQM_CUSTOM */
362 uint8_t cqm_4py[16];
363 uint8_t cqm_4ic[16];
364 uint8_t cqm_4pc[16];
365 uint8_t cqm_8iy[64];
366 uint8_t cqm_8py[64];
367 uint8_t cqm_8ic[64];
368 uint8_t cqm_8pc[64];
370 /* Log */
371 void (*pf_log)( void *, int i_level, const char *psz, va_list );
372 void *p_log_private;
373 int i_log_level;
374 int b_full_recon; /* fully reconstruct frames, even when not necessary for encoding. Implied by psz_dump_yuv */
375 char *psz_dump_yuv; /* filename (in UTF-8) for reconstructed frames */
377 /* Encoder analyser parameters */
378 struct
380 unsigned int intra; /* intra partitions */
381 unsigned int inter; /* inter partitions */
383 int b_transform_8x8;
384 int i_weighted_pred; /* weighting for P-frames */
385 int b_weighted_bipred; /* implicit weighting for B-frames */
386 int i_direct_mv_pred; /* spatial vs temporal mv prediction */
387 int i_chroma_qp_offset;
389 int i_me_method; /* motion estimation algorithm to use (X264_ME_*) */
390 int i_me_range; /* integer pixel motion estimation search range (from predicted mv) */
391 int i_mv_range; /* maximum length of a mv (in pixels). -1 = auto, based on level */
392 int i_mv_range_thread; /* minimum space between threads. -1 = auto, based on number of threads. */
393 int i_subpel_refine; /* subpixel motion estimation quality */
394 int b_chroma_me; /* chroma ME for subpel and mode decision in P-frames */
395 int b_mixed_references; /* allow each mb partition to have its own reference number */
396 int i_trellis; /* trellis RD quantization */
397 int b_fast_pskip; /* early SKIP detection on P-frames */
398 int b_dct_decimate; /* transform coefficient thresholding on P-frames */
399 int i_noise_reduction; /* adaptive pseudo-deadzone */
400 float f_psy_rd; /* Psy RD strength */
401 float f_psy_trellis; /* Psy trellis strength */
402 int b_psy; /* Toggle all psy optimizations */
404 int b_mb_info; /* Use input mb_info data in x264_picture_t */
405 int b_mb_info_update; /* Update the values in mb_info according to the results of encoding. */
407 /* the deadzone size that will be used in luma quantization */
408 int i_luma_deadzone[2]; /* {inter, intra} */
410 int b_psnr; /* compute and print PSNR stats */
411 int b_ssim; /* compute and print SSIM stats */
412 } analyse;
414 /* Rate control parameters */
415 struct
417 int i_rc_method; /* X264_RC_* */
419 int i_qp_constant; /* 0=lossless */
420 int i_qp_min; /* min allowed QP value */
421 int i_qp_max; /* max allowed QP value */
422 int i_qp_step; /* max QP step between frames */
424 int i_bitrate;
425 float f_rf_constant; /* 1pass VBR, nominal QP */
426 float f_rf_constant_max; /* In CRF mode, maximum CRF as caused by VBV */
427 float f_rate_tolerance;
428 int i_vbv_max_bitrate;
429 int i_vbv_buffer_size;
430 float f_vbv_buffer_init; /* <=1: fraction of buffer_size. >1: kbit */
431 float f_ip_factor;
432 float f_pb_factor;
434 /* VBV filler: force CBR VBV and use filler bytes to ensure hard-CBR.
435 * Implied by NAL-HRD CBR. */
436 int b_filler;
438 int i_aq_mode; /* psy adaptive QP. (X264_AQ_*) */
439 float f_aq_strength;
440 int b_mb_tree; /* Macroblock-tree ratecontrol. */
441 int i_lookahead;
443 /* 2pass */
444 int b_stat_write; /* Enable stat writing in psz_stat_out */
445 char *psz_stat_out; /* output filename (in UTF-8) of the 2pass stats file */
446 int b_stat_read; /* Read stat from psz_stat_in and use it */
447 char *psz_stat_in; /* input filename (in UTF-8) of the 2pass stats file */
449 /* 2pass params (same as ffmpeg ones) */
450 float f_qcompress; /* 0.0 => cbr, 1.0 => constant qp */
451 float f_qblur; /* temporally blur quants */
452 float f_complexity_blur; /* temporally blur complexity */
453 x264_zone_t *zones; /* ratecontrol overrides */
454 int i_zones; /* number of zone_t's */
455 char *psz_zones; /* alternate method of specifying zones */
456 } rc;
458 /* Cropping Rectangle parameters: added to those implicitly defined by
459 non-mod16 video resolutions. */
460 struct
462 unsigned int i_left;
463 unsigned int i_top;
464 unsigned int i_right;
465 unsigned int i_bottom;
466 } crop_rect;
468 /* frame packing arrangement flag */
469 int i_frame_packing;
471 /* alternative transfer SEI */
472 int i_alternative_transfer;
474 /* Muxing parameters */
475 int b_aud; /* generate access unit delimiters */
476 int b_repeat_headers; /* put SPS/PPS before each keyframe */
477 int b_annexb; /* if set, place start codes (4 bytes) before NAL units,
478 * otherwise place size (4 bytes) before NAL units. */
479 int i_sps_id; /* SPS and PPS id number */
480 int b_vfr_input; /* VFR input. If 1, use timebase and timestamps for ratecontrol purposes.
481 * If 0, use fps only. */
482 int b_pulldown; /* use explicity set timebase for CFR */
483 uint32_t i_fps_num;
484 uint32_t i_fps_den;
485 uint32_t i_timebase_num; /* Timebase numerator */
486 uint32_t i_timebase_den; /* Timebase denominator */
488 int b_tff;
490 /* Pulldown:
491 * The correct pic_struct must be passed with each input frame.
492 * The input timebase should be the timebase corresponding to the output framerate. This should be constant.
493 * e.g. for 3:2 pulldown timebase should be 1001/30000
494 * The PTS passed with each frame must be the PTS of the frame after pulldown is applied.
495 * Frame doubling and tripling require b_vfr_input set to zero (see H.264 Table D-1)
497 * Pulldown changes are not clearly defined in H.264. Therefore, it is the calling app's responsibility to manage this.
500 int b_pic_struct;
502 /* Fake Interlaced.
504 * Used only when b_interlaced=0. Setting this flag makes it possible to flag the stream as PAFF interlaced yet
505 * encode all frames progessively. It is useful for encoding 25p and 30p Blu-Ray streams.
508 int b_fake_interlaced;
510 /* Don't optimize header parameters based on video content, e.g. ensure that splitting an input video, compressing
511 * each part, and stitching them back together will result in identical SPS/PPS. This is necessary for stitching
512 * with container formats that don't allow multiple SPS/PPS. */
513 int b_stitchable;
515 int b_opencl; /* use OpenCL when available */
516 int i_opencl_device; /* specify count of GPU devices to skip, for CLI users */
517 void *opencl_device_id; /* pass explicit cl_device_id as void*, for API users */
518 char *psz_clbin_file; /* filename (in UTF-8) of the compiled OpenCL kernel cache file */
520 /* Slicing parameters */
521 int i_slice_max_size; /* Max size per slice in bytes; includes estimated NAL overhead. */
522 int i_slice_max_mbs; /* Max number of MBs per slice; overrides i_slice_count. */
523 int i_slice_min_mbs; /* Min number of MBs per slice */
524 int i_slice_count; /* Number of slices per frame: forces rectangular slices. */
525 int i_slice_count_max; /* Absolute cap on slices per frame; stops applying slice-max-size
526 * and slice-max-mbs if this is reached. */
528 /* Optional callback for freeing this x264_param_t when it is done being used.
529 * Only used when the x264_param_t sits in memory for an indefinite period of time,
530 * i.e. when an x264_param_t is passed to x264_t in an x264_picture_t or in zones.
531 * Not used when x264_encoder_reconfig is called directly. */
532 void (*param_free)( void* );
534 /* Optional low-level callback for low-latency encoding. Called for each output NAL unit
535 * immediately after the NAL unit is finished encoding. This allows the calling application
536 * to begin processing video data (e.g. by sending packets over a network) before the frame
537 * is done encoding.
539 * This callback MUST do the following in order to work correctly:
540 * 1) Have available an output buffer of at least size nal->i_payload*3/2 + 5 + 64.
541 * 2) Call x264_nal_encode( h, dst, nal ), where dst is the output buffer.
542 * After these steps, the content of nal is valid and can be used in the same way as if
543 * the NAL unit were output by x264_encoder_encode.
545 * This does not need to be synchronous with the encoding process: the data pointed to
546 * by nal (both before and after x264_nal_encode) will remain valid until the next
547 * x264_encoder_encode call. The callback must be re-entrant.
549 * This callback does not work with frame-based threads; threads must be disabled
550 * or sliced-threads enabled. This callback also does not work as one would expect
551 * with HRD -- since the buffering period SEI cannot be calculated until the frame
552 * is finished encoding, it will not be sent via this callback.
554 * Note also that the NALs are not necessarily returned in order when sliced threads is
555 * enabled. Accordingly, the variable i_first_mb and i_last_mb are available in
556 * x264_nal_t to help the calling application reorder the slices if necessary.
558 * When this callback is enabled, x264_encoder_encode does not return valid NALs;
559 * the calling application is expected to acquire all output NALs through the callback.
561 * It is generally sensible to combine this callback with a use of slice-max-mbs or
562 * slice-max-size.
564 * The opaque pointer is the opaque pointer from the input frame associated with this
565 * NAL unit. This helps distinguish between nalu_process calls from different sources,
566 * e.g. if doing multiple encodes in one process.
568 void (*nalu_process)( x264_t *h, x264_nal_t *nal, void *opaque );
569 } x264_param_t;
571 void x264_nal_encode( x264_t *h, uint8_t *dst, x264_nal_t *nal );
573 /****************************************************************************
574 * H.264 level restriction information
575 ****************************************************************************/
577 typedef struct x264_level_t
579 uint8_t level_idc;
580 uint32_t mbps; /* max macroblock processing rate (macroblocks/sec) */
581 uint32_t frame_size; /* max frame size (macroblocks) */
582 uint32_t dpb; /* max decoded picture buffer (mbs) */
583 uint32_t bitrate; /* max bitrate (kbit/sec) */
584 uint32_t cpb; /* max vbv buffer (kbit) */
585 uint16_t mv_range; /* max vertical mv component range (pixels) */
586 uint8_t mvs_per_2mb; /* max mvs per 2 consecutive mbs. */
587 uint8_t slice_rate; /* ?? */
588 uint8_t mincr; /* min compression ratio */
589 uint8_t bipred8x8; /* limit bipred to >=8x8 */
590 uint8_t direct8x8; /* limit b_direct to >=8x8 */
591 uint8_t frame_only; /* forbid interlacing */
592 } x264_level_t;
594 /* all of the levels defined in the standard, terminated by .level_idc=0 */
595 X264_API extern const x264_level_t x264_levels[];
597 /****************************************************************************
598 * Basic parameter handling functions
599 ****************************************************************************/
601 /* x264_param_default:
602 * fill x264_param_t with default values and do CPU detection */
603 void x264_param_default( x264_param_t * );
605 /* x264_param_parse:
606 * set one parameter by name.
607 * returns 0 on success, or returns one of the following errors.
608 * note: BAD_VALUE occurs only if it can't even parse the value,
609 * numerical range is not checked until x264_encoder_open() or
610 * x264_encoder_reconfig().
611 * value=NULL means "true" for boolean options, but is a BAD_VALUE for non-booleans. */
612 #define X264_PARAM_BAD_NAME (-1)
613 #define X264_PARAM_BAD_VALUE (-2)
614 int x264_param_parse( x264_param_t *, const char *name, const char *value );
616 /****************************************************************************
617 * Advanced parameter handling functions
618 ****************************************************************************/
620 /* These functions expose the full power of x264's preset-tune-profile system for
621 * easy adjustment of large numbers of internal parameters.
623 * In order to replicate x264CLI's option handling, these functions MUST be called
624 * in the following order:
625 * 1) x264_param_default_preset
626 * 2) Custom user options (via param_parse or directly assigned variables)
627 * 3) x264_param_apply_fastfirstpass
628 * 4) x264_param_apply_profile
630 * Additionally, x264CLI does not apply step 3 if the preset chosen is "placebo"
631 * or --slow-firstpass is set. */
633 /* x264_param_default_preset:
634 * The same as x264_param_default, but also use the passed preset and tune
635 * to modify the default settings.
636 * (either can be NULL, which implies no preset or no tune, respectively)
638 * Currently available presets are, ordered from fastest to slowest: */
639 static const char * const x264_preset_names[] = { "ultrafast", "superfast", "veryfast", "faster", "fast", "medium", "slow", "slower", "veryslow", "placebo", 0 };
641 /* The presets can also be indexed numerically, as in:
642 * x264_param_default_preset( &param, "3", ... )
643 * with ultrafast mapping to "0" and placebo mapping to "9". This mapping may
644 * of course change if new presets are added in between, but will always be
645 * ordered from fastest to slowest.
647 * Warning: the speed of these presets scales dramatically. Ultrafast is a full
648 * 100 times faster than placebo!
650 * Currently available tunings are: */
651 static const char * const x264_tune_names[] = { "film", "animation", "grain", "stillimage", "psnr", "ssim", "fastdecode", "zerolatency", 0 };
653 /* Multiple tunings can be used if separated by a delimiter in ",./-+",
654 * however multiple psy tunings cannot be used.
655 * film, animation, grain, stillimage, psnr, and ssim are psy tunings.
657 * returns 0 on success, negative on failure (e.g. invalid preset/tune name). */
658 int x264_param_default_preset( x264_param_t *, const char *preset, const char *tune );
660 /* x264_param_apply_fastfirstpass:
661 * If first-pass mode is set (rc.b_stat_read == 0, rc.b_stat_write == 1),
662 * modify the encoder settings to disable options generally not useful on
663 * the first pass. */
664 void x264_param_apply_fastfirstpass( x264_param_t * );
666 /* x264_param_apply_profile:
667 * Applies the restrictions of the given profile.
668 * Currently available profiles are, from most to least restrictive: */
669 static const char * const x264_profile_names[] = { "baseline", "main", "high", "high10", "high422", "high444", 0 };
671 /* (can be NULL, in which case the function will do nothing)
673 * Does NOT guarantee that the given profile will be used: if the restrictions
674 * of "High" are applied to settings that are already Baseline-compatible, the
675 * stream will remain baseline. In short, it does not increase settings, only
676 * decrease them.
678 * returns 0 on success, negative on failure (e.g. invalid profile name). */
679 int x264_param_apply_profile( x264_param_t *, const char *profile );
681 /****************************************************************************
682 * Picture structures and functions
683 ****************************************************************************/
685 /* x264_chroma_format:
686 * Specifies the chroma formats that x264 supports encoding. When this
687 * value is non-zero, then it represents a X264_CSP_* that is the only
688 * chroma format that x264 supports encoding. If the value is 0 then
689 * there are no restrictions. */
690 X264_API extern const int x264_chroma_format;
692 enum pic_struct_e
694 PIC_STRUCT_AUTO = 0, // automatically decide (default)
695 PIC_STRUCT_PROGRESSIVE = 1, // progressive frame
696 // "TOP" and "BOTTOM" are not supported in x264 (PAFF only)
697 PIC_STRUCT_TOP_BOTTOM = 4, // top field followed by bottom
698 PIC_STRUCT_BOTTOM_TOP = 5, // bottom field followed by top
699 PIC_STRUCT_TOP_BOTTOM_TOP = 6, // top field, bottom field, top field repeated
700 PIC_STRUCT_BOTTOM_TOP_BOTTOM = 7, // bottom field, top field, bottom field repeated
701 PIC_STRUCT_DOUBLE = 8, // double frame
702 PIC_STRUCT_TRIPLE = 9, // triple frame
705 typedef struct x264_hrd_t
707 double cpb_initial_arrival_time;
708 double cpb_final_arrival_time;
709 double cpb_removal_time;
711 double dpb_output_time;
712 } x264_hrd_t;
714 /* Arbitrary user SEI:
715 * Payload size is in bytes and the payload pointer must be valid.
716 * Payload types and syntax can be found in Annex D of the H.264 Specification.
717 * SEI payload alignment bits as described in Annex D must be included at the
718 * end of the payload if needed.
719 * The payload should not be NAL-encapsulated.
720 * Payloads are written first in order of input, apart from in the case when HRD
721 * is enabled where payloads are written after the Buffering Period SEI. */
723 typedef struct x264_sei_payload_t
725 int payload_size;
726 int payload_type;
727 uint8_t *payload;
728 } x264_sei_payload_t;
730 typedef struct x264_sei_t
732 int num_payloads;
733 x264_sei_payload_t *payloads;
734 /* In: optional callback to free each payload AND x264_sei_payload_t when used. */
735 void (*sei_free)( void* );
736 } x264_sei_t;
738 typedef struct x264_image_t
740 int i_csp; /* Colorspace */
741 int i_plane; /* Number of image planes */
742 int i_stride[4]; /* Strides for each plane */
743 uint8_t *plane[4]; /* Pointers to each plane */
744 } x264_image_t;
746 typedef struct x264_image_properties_t
748 /* All arrays of data here are ordered as follows:
749 * each array contains one offset per macroblock, in raster scan order. In interlaced
750 * mode, top-field MBs and bottom-field MBs are interleaved at the row level.
751 * Macroblocks are 16x16 blocks of pixels (with respect to the luma plane). For the
752 * purposes of calculating the number of macroblocks, width and height are rounded up to
753 * the nearest 16. If in interlaced mode, height is rounded up to the nearest 32 instead. */
755 /* In: an array of quantizer offsets to be applied to this image during encoding.
756 * These are added on top of the decisions made by x264.
757 * Offsets can be fractional; they are added before QPs are rounded to integer.
758 * Adaptive quantization must be enabled to use this feature. Behavior if quant
759 * offsets differ between encoding passes is undefined. */
760 float *quant_offsets;
761 /* In: optional callback to free quant_offsets when used.
762 * Useful if one wants to use a different quant_offset array for each frame. */
763 void (*quant_offsets_free)( void* );
765 /* In: optional array of flags for each macroblock.
766 * Allows specifying additional information for the encoder such as which macroblocks
767 * remain unchanged. Usable flags are listed below.
768 * x264_param_t.analyse.b_mb_info must be set to use this, since x264 needs to track
769 * extra data internally to make full use of this information.
771 * Out: if b_mb_info_update is set, x264 will update this array as a result of encoding.
773 * For "MBINFO_CONSTANT", it will remove this flag on any macroblock whose decoded
774 * pixels have changed. This can be useful for e.g. noting which areas of the
775 * frame need to actually be blitted. Note: this intentionally ignores the effects
776 * of deblocking for the current frame, which should be fine unless one needs exact
777 * pixel-perfect accuracy.
779 * Results for MBINFO_CONSTANT are currently only set for P-frames, and are not
780 * guaranteed to enumerate all blocks which haven't changed. (There may be false
781 * negatives, but no false positives.)
783 uint8_t *mb_info;
784 /* In: optional callback to free mb_info when used. */
785 void (*mb_info_free)( void* );
787 /* The macroblock is constant and remains unchanged from the previous frame. */
788 #define X264_MBINFO_CONSTANT (1<<0)
789 /* More flags may be added in the future. */
791 /* Out: SSIM of the the frame luma (if x264_param_t.b_ssim is set) */
792 double f_ssim;
793 /* Out: Average PSNR of the frame (if x264_param_t.b_psnr is set) */
794 double f_psnr_avg;
795 /* Out: PSNR of Y, U, and V (if x264_param_t.b_psnr is set) */
796 double f_psnr[3];
798 /* Out: Average effective CRF of the encoded frame */
799 double f_crf_avg;
800 } x264_image_properties_t;
802 typedef struct x264_picture_t
804 /* In: force picture type (if not auto)
805 * If x264 encoding parameters are violated in the forcing of picture types,
806 * x264 will correct the input picture type and log a warning.
807 * Out: type of the picture encoded */
808 int i_type;
809 /* In: force quantizer for != X264_QP_AUTO */
810 int i_qpplus1;
811 /* In: pic_struct, for pulldown/doubling/etc...used only if b_pic_struct=1.
812 * use pic_struct_e for pic_struct inputs
813 * Out: pic_struct element associated with frame */
814 int i_pic_struct;
815 /* Out: whether this frame is a keyframe. Important when using modes that result in
816 * SEI recovery points being used instead of IDR frames. */
817 int b_keyframe;
818 /* In: user pts, Out: pts of encoded picture (user)*/
819 int64_t i_pts;
820 /* Out: frame dts. When the pts of the first frame is close to zero,
821 * initial frames may have a negative dts which must be dealt with by any muxer */
822 int64_t i_dts;
823 /* In: custom encoding parameters to be set from this frame forwards
824 (in coded order, not display order). If NULL, continue using
825 parameters from the previous frame. Some parameters, such as
826 aspect ratio, can only be changed per-GOP due to the limitations
827 of H.264 itself; in this case, the caller must force an IDR frame
828 if it needs the changed parameter to apply immediately. */
829 x264_param_t *param;
830 /* In: raw image data */
831 /* Out: reconstructed image data. x264 may skip part of the reconstruction process,
832 e.g. deblocking, in frames where it isn't necessary. To force complete
833 reconstruction, at a small speed cost, set b_full_recon. */
834 x264_image_t img;
835 /* In: optional information to modify encoder decisions for this frame
836 * Out: information about the encoded frame */
837 x264_image_properties_t prop;
838 /* Out: HRD timing information. Output only when i_nal_hrd is set. */
839 x264_hrd_t hrd_timing;
840 /* In: arbitrary user SEI (e.g subtitles, AFDs) */
841 x264_sei_t extra_sei;
842 /* private user data. copied from input to output frames. */
843 void *opaque;
844 } x264_picture_t;
846 /* x264_picture_init:
847 * initialize an x264_picture_t. Needs to be done if the calling application
848 * allocates its own x264_picture_t as opposed to using x264_picture_alloc. */
849 void x264_picture_init( x264_picture_t *pic );
851 /* x264_picture_alloc:
852 * alloc data for a picture. You must call x264_picture_clean on it.
853 * returns 0 on success, or -1 on malloc failure or invalid colorspace. */
854 int x264_picture_alloc( x264_picture_t *pic, int i_csp, int i_width, int i_height );
856 /* x264_picture_clean:
857 * free associated resource for a x264_picture_t allocated with
858 * x264_picture_alloc ONLY */
859 void x264_picture_clean( x264_picture_t *pic );
861 /****************************************************************************
862 * Encoder functions
863 ****************************************************************************/
865 /* Force a link error in the case of linking against an incompatible API version.
866 * Glue #defines exist to force correct macro expansion; the final output of the macro
867 * is x264_encoder_open_##X264_BUILD (for purposes of dlopen). */
868 #define x264_encoder_glue1(x,y) x##y
869 #define x264_encoder_glue2(x,y) x264_encoder_glue1(x,y)
870 #define x264_encoder_open x264_encoder_glue2(x264_encoder_open_,X264_BUILD)
872 /* x264_encoder_open:
873 * create a new encoder handler, all parameters from x264_param_t are copied */
874 x264_t *x264_encoder_open( x264_param_t * );
876 /* x264_encoder_reconfig:
877 * various parameters from x264_param_t are copied.
878 * this takes effect immediately, on whichever frame is encoded next;
879 * due to delay, this may not be the next frame passed to encoder_encode.
880 * if the change should apply to some particular frame, use x264_picture_t->param instead.
881 * returns 0 on success, negative on parameter validation error.
882 * not all parameters can be changed; see the actual function for a detailed breakdown.
884 * since not all parameters can be changed, moving from preset to preset may not always
885 * fully copy all relevant parameters, but should still work usably in practice. however,
886 * more so than for other presets, many of the speed shortcuts used in ultrafast cannot be
887 * switched out of; using reconfig to switch between ultrafast and other presets is not
888 * recommended without a more fine-grained breakdown of parameters to take this into account. */
889 int x264_encoder_reconfig( x264_t *, x264_param_t * );
890 /* x264_encoder_parameters:
891 * copies the current internal set of parameters to the pointer provided
892 * by the caller. useful when the calling application needs to know
893 * how x264_encoder_open has changed the parameters, or the current state
894 * of the encoder after multiple x264_encoder_reconfig calls.
895 * note that the data accessible through pointers in the returned param struct
896 * (e.g. filenames) should not be modified by the calling application. */
897 void x264_encoder_parameters( x264_t *, x264_param_t * );
898 /* x264_encoder_headers:
899 * return the SPS and PPS that will be used for the whole stream.
900 * *pi_nal is the number of NAL units outputted in pp_nal.
901 * returns the number of bytes in the returned NALs.
902 * returns negative on error.
903 * the payloads of all output NALs are guaranteed to be sequential in memory. */
904 int x264_encoder_headers( x264_t *, x264_nal_t **pp_nal, int *pi_nal );
905 /* x264_encoder_encode:
906 * encode one picture.
907 * *pi_nal is the number of NAL units outputted in pp_nal.
908 * returns the number of bytes in the returned NALs.
909 * returns negative on error and zero if no NAL units returned.
910 * the payloads of all output NALs are guaranteed to be sequential in memory. */
911 int x264_encoder_encode( x264_t *, x264_nal_t **pp_nal, int *pi_nal, x264_picture_t *pic_in, x264_picture_t *pic_out );
912 /* x264_encoder_close:
913 * close an encoder handler */
914 void x264_encoder_close( x264_t * );
915 /* x264_encoder_delayed_frames:
916 * return the number of currently delayed (buffered) frames
917 * this should be used at the end of the stream, to know when you have all the encoded frames. */
918 int x264_encoder_delayed_frames( x264_t * );
919 /* x264_encoder_maximum_delayed_frames( x264_t * ):
920 * return the maximum number of delayed (buffered) frames that can occur with the current
921 * parameters. */
922 int x264_encoder_maximum_delayed_frames( x264_t * );
923 /* x264_encoder_intra_refresh:
924 * If an intra refresh is not in progress, begin one with the next P-frame.
925 * If an intra refresh is in progress, begin one as soon as the current one finishes.
926 * Requires that b_intra_refresh be set.
928 * Useful for interactive streaming where the client can tell the server that packet loss has
929 * occurred. In this case, keyint can be set to an extremely high value so that intra refreshes
930 * only occur when calling x264_encoder_intra_refresh.
932 * In multi-pass encoding, if x264_encoder_intra_refresh is called differently in each pass,
933 * behavior is undefined.
935 * Should not be called during an x264_encoder_encode. */
936 void x264_encoder_intra_refresh( x264_t * );
937 /* x264_encoder_invalidate_reference:
938 * An interactive error resilience tool, designed for use in a low-latency one-encoder-few-clients
939 * system. When the client has packet loss or otherwise incorrectly decodes a frame, the encoder
940 * can be told with this command to "forget" the frame and all frames that depend on it, referencing
941 * only frames that occurred before the loss. This will force a keyframe if no frames are left to
942 * reference after the aforementioned "forgetting".
944 * It is strongly recommended to use a large i_dpb_size in this case, which allows the encoder to
945 * keep around extra, older frames to fall back on in case more recent frames are all invalidated.
946 * Unlike increasing i_frame_reference, this does not increase the number of frames used for motion
947 * estimation and thus has no speed impact. It is also recommended to set a very large keyframe
948 * interval, so that keyframes are not used except as necessary for error recovery.
950 * x264_encoder_invalidate_reference is not currently compatible with the use of B-frames or intra
951 * refresh.
953 * In multi-pass encoding, if x264_encoder_invalidate_reference is called differently in each pass,
954 * behavior is undefined.
956 * Should not be called during an x264_encoder_encode, but multiple calls can be made simultaneously.
958 * Returns 0 on success, negative on failure. */
959 int x264_encoder_invalidate_reference( x264_t *, int64_t pts );
961 #ifdef __cplusplus
963 #endif
965 #endif