FFmpeg
 All Data Structures Namespaces Files Functions Variables Typedefs Enumerations Enumerator Macros Groups Pages
libx265.c
Go to the documentation of this file.
1 /*
2  * libx265 encoder
3  *
4  * Copyright (c) 2013-2014 Derek Buitenhuis
5  *
6  * This file is part of FFmpeg.
7  *
8  * FFmpeg is free software; you can redistribute it and/or
9  * modify it under the terms of the GNU Lesser General Public
10  * License as published by the Free Software Foundation; either
11  * version 2.1 of the License, or (at your option) any later version.
12  *
13  * FFmpeg 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 GNU
16  * Lesser General Public License for more details.
17  *
18  * You should have received a copy of the GNU Lesser General Public
19  * License along with FFmpeg; if not, write to the Free Software
20  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
21  */
22 
23 #if defined(_MSC_VER)
24 #define X265_API_IMPORTS 1
25 #endif
26 
27 #include <x265.h>
28 #include <float.h>
29 
30 #include "libavutil/internal.h"
31 #include "libavutil/common.h"
32 #include "libavutil/opt.h"
33 #include "libavutil/pixdesc.h"
34 #include "avcodec.h"
35 #include "internal.h"
36 
37 typedef struct libx265Context {
38  const AVClass *class;
39 
40  x265_encoder *encoder;
41  x265_param *params;
42  const x265_api *api;
43 
44  float crf;
45  char *preset;
46  char *tune;
47  char *x265_opts;
49 
50 static int is_keyframe(NalUnitType naltype)
51 {
52  switch (naltype) {
53  case NAL_UNIT_CODED_SLICE_BLA_W_LP:
54  case NAL_UNIT_CODED_SLICE_BLA_W_RADL:
55  case NAL_UNIT_CODED_SLICE_BLA_N_LP:
56  case NAL_UNIT_CODED_SLICE_IDR_W_RADL:
57  case NAL_UNIT_CODED_SLICE_IDR_N_LP:
58  case NAL_UNIT_CODED_SLICE_CRA:
59  return 1;
60  default:
61  return 0;
62  }
63 }
64 
66 {
67  libx265Context *ctx = avctx->priv_data;
68 
69  ctx->api->param_free(ctx->params);
70 
71  if (ctx->encoder)
72  ctx->api->encoder_close(ctx->encoder);
73 
74  return 0;
75 }
76 
78 {
79  libx265Context *ctx = avctx->priv_data;
80 
81  ctx->api = x265_api_get(av_pix_fmt_desc_get(avctx->pix_fmt)->comp[0].depth);
82  if (!ctx->api)
83  ctx->api = x265_api_get(0);
84 
85  ctx->params = ctx->api->param_alloc();
86  if (!ctx->params) {
87  av_log(avctx, AV_LOG_ERROR, "Could not allocate x265 param structure.\n");
88  return AVERROR(ENOMEM);
89  }
90 
91  if (ctx->api->param_default_preset(ctx->params, ctx->preset, ctx->tune) < 0) {
92  int i;
93 
94  av_log(avctx, AV_LOG_ERROR, "Error setting preset/tune %s/%s.\n", ctx->preset, ctx->tune);
95  av_log(avctx, AV_LOG_INFO, "Possible presets:");
96  for (i = 0; x265_preset_names[i]; i++)
97  av_log(avctx, AV_LOG_INFO, " %s", x265_preset_names[i]);
98 
99  av_log(avctx, AV_LOG_INFO, "\n");
100  av_log(avctx, AV_LOG_INFO, "Possible tunes:");
101  for (i = 0; x265_tune_names[i]; i++)
102  av_log(avctx, AV_LOG_INFO, " %s", x265_tune_names[i]);
103 
104  av_log(avctx, AV_LOG_INFO, "\n");
105 
106  return AVERROR(EINVAL);
107  }
108 
109  ctx->params->frameNumThreads = avctx->thread_count;
110  ctx->params->fpsNum = avctx->time_base.den;
111  ctx->params->fpsDenom = avctx->time_base.num * avctx->ticks_per_frame;
112  ctx->params->sourceWidth = avctx->width;
113  ctx->params->sourceHeight = avctx->height;
114  ctx->params->bEnablePsnr = !!(avctx->flags & AV_CODEC_FLAG_PSNR);
115 
116  if ((avctx->color_primaries <= AVCOL_PRI_BT2020 &&
118  (avctx->color_trc <= AVCOL_TRC_BT2020_12 &&
119  avctx->color_trc != AVCOL_TRC_UNSPECIFIED) ||
120  (avctx->colorspace <= AVCOL_SPC_BT2020_CL &&
121  avctx->colorspace != AVCOL_SPC_UNSPECIFIED)) {
122 
123  ctx->params->vui.bEnableVideoSignalTypePresentFlag = 1;
124  ctx->params->vui.bEnableColorDescriptionPresentFlag = 1;
125 
126  // x265 validates the parameters internally
127  ctx->params->vui.colorPrimaries = avctx->color_primaries;
128  ctx->params->vui.transferCharacteristics = avctx->color_trc;
129  ctx->params->vui.matrixCoeffs = avctx->colorspace;
130  }
131 
132  if (avctx->sample_aspect_ratio.num > 0 && avctx->sample_aspect_ratio.den > 0) {
133  char sar[12];
134  int sar_num, sar_den;
135 
136  av_reduce(&sar_num, &sar_den,
137  avctx->sample_aspect_ratio.num,
138  avctx->sample_aspect_ratio.den, 65535);
139  snprintf(sar, sizeof(sar), "%d:%d", sar_num, sar_den);
140  if (ctx->api->param_parse(ctx->params, "sar", sar) == X265_PARAM_BAD_VALUE) {
141  av_log(avctx, AV_LOG_ERROR, "Invalid SAR: %d:%d.\n", sar_num, sar_den);
142  return AVERROR_INVALIDDATA;
143  }
144  }
145 
146  switch (avctx->pix_fmt) {
147  case AV_PIX_FMT_YUV420P:
150  ctx->params->internalCsp = X265_CSP_I420;
151  break;
152  case AV_PIX_FMT_YUV422P:
155  ctx->params->internalCsp = X265_CSP_I422;
156  break;
157  case AV_PIX_FMT_GBRP:
158  case AV_PIX_FMT_GBRP10:
159  case AV_PIX_FMT_GBRP12:
160  ctx->params->vui.matrixCoeffs = AVCOL_SPC_RGB;
161  ctx->params->vui.bEnableVideoSignalTypePresentFlag = 1;
162  ctx->params->vui.bEnableColorDescriptionPresentFlag = 1;
163  case AV_PIX_FMT_YUV444P:
166  ctx->params->internalCsp = X265_CSP_I444;
167  break;
168  case AV_PIX_FMT_GRAY8:
169  if (ctx->api->api_build_number < 85) {
170  av_log(avctx, AV_LOG_ERROR,
171  "libx265 version is %d, must be at least 85 for gray encoding.\n",
172  ctx->api->api_build_number);
173  return AVERROR_INVALIDDATA;
174  }
175  ctx->params->internalCsp = X265_CSP_I400;
176  break;
177  }
178 
179  if (ctx->crf >= 0) {
180  char crf[6];
181 
182  snprintf(crf, sizeof(crf), "%2.2f", ctx->crf);
183  if (ctx->api->param_parse(ctx->params, "crf", crf) == X265_PARAM_BAD_VALUE) {
184  av_log(avctx, AV_LOG_ERROR, "Invalid crf: %2.2f.\n", ctx->crf);
185  return AVERROR(EINVAL);
186  }
187  } else if (avctx->bit_rate > 0) {
188  ctx->params->rc.bitrate = avctx->bit_rate / 1000;
189  ctx->params->rc.rateControlMode = X265_RC_ABR;
190  }
191 
192  if (!(avctx->flags & AV_CODEC_FLAG_GLOBAL_HEADER))
193  ctx->params->bRepeatHeaders = 1;
194 
195  if (ctx->x265_opts) {
196  AVDictionary *dict = NULL;
197  AVDictionaryEntry *en = NULL;
198 
199  if (!av_dict_parse_string(&dict, ctx->x265_opts, "=", ":", 0)) {
200  while ((en = av_dict_get(dict, "", en, AV_DICT_IGNORE_SUFFIX))) {
201  int parse_ret = ctx->api->param_parse(ctx->params, en->key, en->value);
202 
203  switch (parse_ret) {
204  case X265_PARAM_BAD_NAME:
205  av_log(avctx, AV_LOG_WARNING,
206  "Unknown option: %s.\n", en->key);
207  break;
208  case X265_PARAM_BAD_VALUE:
209  av_log(avctx, AV_LOG_WARNING,
210  "Invalid value for %s: %s.\n", en->key, en->value);
211  break;
212  default:
213  break;
214  }
215  }
216  av_dict_free(&dict);
217  }
218  }
219 
220  ctx->encoder = ctx->api->encoder_open(ctx->params);
221  if (!ctx->encoder) {
222  av_log(avctx, AV_LOG_ERROR, "Cannot open libx265 encoder.\n");
223  libx265_encode_close(avctx);
224  return AVERROR_INVALIDDATA;
225  }
226 
227  if (avctx->flags & AV_CODEC_FLAG_GLOBAL_HEADER) {
228  x265_nal *nal;
229  int nnal;
230 
231  avctx->extradata_size = ctx->api->encoder_headers(ctx->encoder, &nal, &nnal);
232  if (avctx->extradata_size <= 0) {
233  av_log(avctx, AV_LOG_ERROR, "Cannot encode headers.\n");
234  libx265_encode_close(avctx);
235  return AVERROR_INVALIDDATA;
236  }
237 
239  if (!avctx->extradata) {
240  av_log(avctx, AV_LOG_ERROR,
241  "Cannot allocate HEVC header of size %d.\n", avctx->extradata_size);
242  libx265_encode_close(avctx);
243  return AVERROR(ENOMEM);
244  }
245 
246  memcpy(avctx->extradata, nal[0].payload, avctx->extradata_size);
247  }
248 
249  return 0;
250 }
251 
253  const AVFrame *pic, int *got_packet)
254 {
255  libx265Context *ctx = avctx->priv_data;
256  x265_picture x265pic;
257  x265_picture x265pic_out = { 0 };
258  x265_nal *nal;
259  uint8_t *dst;
260  int payload = 0;
261  int nnal;
262  int ret;
263  int i;
264 
265  ctx->api->picture_init(ctx->params, &x265pic);
266 
267  if (pic) {
268  for (i = 0; i < 3; i++) {
269  x265pic.planes[i] = pic->data[i];
270  x265pic.stride[i] = pic->linesize[i];
271  }
272 
273  x265pic.pts = pic->pts;
274  x265pic.bitDepth = av_pix_fmt_desc_get(avctx->pix_fmt)->comp[0].depth;
275 
276  x265pic.sliceType = pic->pict_type == AV_PICTURE_TYPE_I ? X265_TYPE_I :
277  pic->pict_type == AV_PICTURE_TYPE_P ? X265_TYPE_P :
278  pic->pict_type == AV_PICTURE_TYPE_B ? X265_TYPE_B :
279  X265_TYPE_AUTO;
280  }
281 
282  ret = ctx->api->encoder_encode(ctx->encoder, &nal, &nnal,
283  pic ? &x265pic : NULL, &x265pic_out);
284  if (ret < 0)
285  return AVERROR_EXTERNAL;
286 
287  if (!nnal)
288  return 0;
289 
290  for (i = 0; i < nnal; i++)
291  payload += nal[i].sizeBytes;
292 
293  ret = ff_alloc_packet(pkt, payload);
294  if (ret < 0) {
295  av_log(avctx, AV_LOG_ERROR, "Error getting output packet.\n");
296  return ret;
297  }
298  dst = pkt->data;
299 
300  for (i = 0; i < nnal; i++) {
301  memcpy(dst, nal[i].payload, nal[i].sizeBytes);
302  dst += nal[i].sizeBytes;
303 
304  if (is_keyframe(nal[i].type))
305  pkt->flags |= AV_PKT_FLAG_KEY;
306  }
307 
308  pkt->pts = x265pic_out.pts;
309  pkt->dts = x265pic_out.dts;
310 
311 #if FF_API_CODED_FRAME
313  switch (x265pic_out.sliceType) {
314  case X265_TYPE_IDR:
315  case X265_TYPE_I:
317  break;
318  case X265_TYPE_P:
320  break;
321  case X265_TYPE_B:
323  break;
324  }
326 #endif
327 
328  *got_packet = 1;
329  return 0;
330 }
331 
332 static const enum AVPixelFormat x265_csp_eight[] = {
339 };
340 
341 static const enum AVPixelFormat x265_csp_ten[] = {
352 };
353 
354 static const enum AVPixelFormat x265_csp_twelve[] = {
369 };
370 
372 {
373  if (x265_api_get(12))
374  codec->pix_fmts = x265_csp_twelve;
375  else if (x265_api_get(10))
376  codec->pix_fmts = x265_csp_ten;
377  else if (x265_api_get(8))
378  codec->pix_fmts = x265_csp_eight;
379 }
380 
381 #define OFFSET(x) offsetof(libx265Context, x)
382 #define VE AV_OPT_FLAG_VIDEO_PARAM | AV_OPT_FLAG_ENCODING_PARAM
383 static const AVOption options[] = {
384  { "crf", "set the x265 crf", OFFSET(crf), AV_OPT_TYPE_FLOAT, { .dbl = -1 }, -1, FLT_MAX, VE },
385  { "preset", "set the x265 preset", OFFSET(preset), AV_OPT_TYPE_STRING, { 0 }, 0, 0, VE },
386  { "tune", "set the x265 tune parameter", OFFSET(tune), AV_OPT_TYPE_STRING, { 0 }, 0, 0, VE },
387  { "x265-params", "set the x265 configuration using a :-separated list of key=value parameters", OFFSET(x265_opts), AV_OPT_TYPE_STRING, { 0 }, 0, 0, VE },
388  { NULL }
389 };
390 
391 static const AVClass class = {
392  .class_name = "libx265",
393  .item_name = av_default_item_name,
394  .option = options,
396 };
397 
398 static const AVCodecDefault x265_defaults[] = {
399  { "b", "0" },
400  { NULL },
401 };
402 
404  .name = "libx265",
405  .long_name = NULL_IF_CONFIG_SMALL("libx265 H.265 / HEVC"),
406  .type = AVMEDIA_TYPE_VIDEO,
407  .id = AV_CODEC_ID_HEVC,
408  .init = libx265_encode_init,
409  .init_static_data = libx265_encode_init_csp,
410  .encode2 = libx265_encode_frame,
411  .close = libx265_encode_close,
412  .priv_data_size = sizeof(libx265Context),
413  .priv_class = &class,
414  .defaults = x265_defaults,
416 };
ITU-R BT2020 for 12-bit system.
Definition: pixfmt.h:426
#define NULL
Definition: coverity.c:32
static const AVOption options[]
Definition: libx265.c:383
#define AVERROR_INVALIDDATA
Invalid data found when processing input.
Definition: error.h:59
const AVPixFmtDescriptor * av_pix_fmt_desc_get(enum AVPixelFormat pix_fmt)
Definition: pixdesc.c:2266
This structure describes decoded (raw) audio or video data.
Definition: frame.h:184
AVOption.
Definition: opt.h:245
static int is_keyframe(NalUnitType naltype)
Definition: libx265.c:50
planar YUV 4:4:4, 24bpp, (1 Cr & Cb sample per 1x1 Y samples)
Definition: pixfmt.h:67
#define AV_LOG_WARNING
Something somehow does not look correct.
Definition: log.h:182
int64_t bit_rate
the average bitrate
Definition: avcodec.h:1741
#define LIBAVUTIL_VERSION_INT
Definition: version.h:85
float crf
Definition: libx265.c:44
planar GBR 4:4:4 24bpp
Definition: pixfmt.h:180
int num
Numerator.
Definition: rational.h:59
#define AV_PIX_FMT_GBRP10
Definition: pixfmt.h:357
AVRational sample_aspect_ratio
sample aspect ratio (0 if unknown) That is the width of a pixel divided by the height of the pixel...
Definition: avcodec.h:2087
enum AVPixelFormat pix_fmt
Pixel format, see AV_PIX_FMT_xxx.
Definition: avcodec.h:1904
x265_param * params
Definition: libx265.c:41
#define AV_PIX_FMT_YUV420P12
Definition: pixfmt.h:345
static AVPacket pkt
#define AV_CODEC_CAP_AUTO_THREADS
Codec supports avctx->thread_count == 0 (auto).
Definition: avcodec.h:1034
AVCodec.
Definition: avcodec.h:3600
order of coefficients is actually GBR, also IEC 61966-2-1 (sRGB)
Definition: pixfmt.h:437
AVRational time_base
This is the fundamental unit of time (in seconds) in terms of which frame timestamps are represented...
Definition: avcodec.h:1813
const char * class_name
The name of the class; usually it is the same name as the context structure type to which the AVClass...
Definition: log.h:72
#define AV_CODEC_CAP_DELAY
Encoder or decoder requires flushing with NULL input at the end in order to give the complete and cor...
Definition: avcodec.h:984
AVComponentDescriptor comp[4]
Parameters that describe how pixels are packed.
Definition: pixdesc.h:117
uint8_t
#define av_cold
Definition: attributes.h:82
#define av_malloc(s)
AVOptions.
#define VE
Definition: libx265.c:382
int64_t pts
Presentation timestamp in time_base units (time when frame should be shown to user).
Definition: frame.h:268
uint8_t * extradata
some codecs need / can use extradata like Huffman tables.
Definition: avcodec.h:1791
AVDictionaryEntry * av_dict_get(const AVDictionary *m, const char *key, const AVDictionaryEntry *prev, int flags)
Get a dictionary entry with matching key.
Definition: dict.c:40
uint8_t * data
Definition: avcodec.h:1601
char * x265_opts
Definition: libx265.c:47
int av_reduce(int *dst_num, int *dst_den, int64_t num, int64_t den, int64_t max)
Reduce a fraction.
Definition: rational.c:35
#define AV_PIX_FMT_YUV422P12
Definition: pixfmt.h:346
#define av_log(a,...)
#define AV_PKT_FLAG_KEY
The packet contains a keyframe.
Definition: avcodec.h:1633
#define AV_LOG_ERROR
Something went wrong and cannot losslessly be recovered.
Definition: log.h:176
av_default_item_name
#define AVERROR(e)
Definition: error.h:43
const x265_api * api
Definition: libx265.c:42
static int libx265_encode_frame(AVCodecContext *avctx, AVPacket *pkt, const AVFrame *pic, int *got_packet)
Definition: libx265.c:252
#define NULL_IF_CONFIG_SMALL(x)
Return NULL if CONFIG_SMALL is true, otherwise the argument without modification. ...
Definition: internal.h:176
void av_dict_free(AVDictionary **pm)
Free all the memory allocated for an AVDictionary struct and all keys and values. ...
Definition: dict.c:203
int flags
AV_CODEC_FLAG_*.
Definition: avcodec.h:1771
const char * name
Name of the codec implementation.
Definition: avcodec.h:3607
#define AV_PIX_FMT_YUV444P10
Definition: pixfmt.h:344
int flags
A combination of AV_PKT_FLAG values.
Definition: avcodec.h:1607
planar YUV 4:2:2, 16bpp, (1 Cr & Cb sample per 2x1 Y samples)
Definition: pixfmt.h:66
common internal API header
static av_cold void libx265_encode_init_csp(AVCodec *codec)
Definition: libx265.c:371
enum AVPixelFormat * pix_fmts
array of supported pixel formats, or NULL if unknown, array is terminated by -1
Definition: avcodec.h:3621
enum AVPictureType pict_type
Picture type of the frame.
Definition: frame.h:258
AVCodec ff_libx265_encoder
Definition: libx265.c:403
static av_cold int libx265_encode_init(AVCodecContext *avctx)
Definition: libx265.c:77
int width
picture width / height.
Definition: avcodec.h:1863
AVFormatContext * ctx
Definition: movenc.c:48
#define AV_CODEC_FLAG_PSNR
error[?] variables will be set during encoding.
Definition: avcodec.h:865
enum AVColorPrimaries color_primaries
Chromaticity coordinates of the source primaries.
Definition: avcodec.h:2392
int ticks_per_frame
For some codecs, the time base is closer to the field rate than the frame rate.
Definition: avcodec.h:1822
#define OFFSET(x)
Definition: libx265.c:381
int thread_count
thread count is used to decide how many independent tasks should be passed to execute() ...
Definition: avcodec.h:3107
char * preset
Definition: libx265.c:45
int av_dict_parse_string(AVDictionary **pm, const char *str, const char *key_val_sep, const char *pairs_sep, int flags)
Parse the key/value pairs list and add the parsed entries to a dictionary.
Definition: dict.c:180
#define AV_LOG_INFO
Standard information.
Definition: log.h:187
Libavcodec external API header.
int linesize[AV_NUM_DATA_POINTERS]
For video, size in bytes of each picture line.
Definition: frame.h:215
main external API structure.
Definition: avcodec.h:1676
GLint GLenum type
Definition: opengl_enc.c:105
int extradata_size
Definition: avcodec.h:1792
#define AV_PIX_FMT_YUV420P10
Definition: pixfmt.h:341
Describe the class of an AVClass context structure.
Definition: log.h:67
enum AVColorSpace colorspace
YUV colorspace type.
Definition: avcodec.h:2406
enum AVColorTransferCharacteristic color_trc
Color Transfer Characteristic.
Definition: avcodec.h:2399
static enum AVPixelFormat x265_csp_ten[]
Definition: libx265.c:341
#define snprintf
Definition: snprintf.h:34
char * tune
Definition: libx265.c:46
#define AV_PIX_FMT_GBRP12
Definition: pixfmt.h:358
#define AV_PIX_FMT_YUV422P10
Definition: pixfmt.h:342
#define AV_PIX_FMT_YUV444P12
Definition: pixfmt.h:348
uint8_t * data[AV_NUM_DATA_POINTERS]
pointer to the picture/channel planes.
Definition: frame.h:198
#define AV_CODEC_FLAG_GLOBAL_HEADER
Place global headers in extradata instead of every keyframe.
Definition: avcodec.h:882
ITU-R BT2020 constant luminance system.
Definition: pixfmt.h:447
preset
Definition: vf_curves.c:46
planar YUV 4:2:0, 12bpp, (1 Cr & Cb sample per 2x2 Y samples)
Definition: pixfmt.h:62
Y , 8bpp.
Definition: pixfmt.h:70
#define FF_DISABLE_DEPRECATION_WARNINGS
Definition: internal.h:80
common internal api header.
common internal and external API header
static enum AVPixelFormat x265_csp_twelve[]
Definition: libx265.c:354
Bi-dir predicted.
Definition: avutil.h:270
attribute_deprecated AVFrame * coded_frame
the picture in the bitstream
Definition: avcodec.h:3098
char * key
Definition: dict.h:86
int den
Denominator.
Definition: rational.h:60
#define AV_INPUT_BUFFER_PADDING_SIZE
Required number of additionally allocated bytes at the end of the input bitstream for decoding...
Definition: avcodec.h:734
void * priv_data
Definition: avcodec.h:1718
static const AVCodecDefault x265_defaults[]
Definition: libx265.c:398
char * value
Definition: dict.h:87
#define FF_ENABLE_DEPRECATION_WARNINGS
Definition: internal.h:81
attribute_deprecated int ff_alloc_packet(AVPacket *avpkt, int size)
Definition: utils.c:1763
static av_cold int libx265_encode_close(AVCodecContext *avctx)
Definition: libx265.c:65
static enum AVPixelFormat x265_csp_eight[]
Definition: libx265.c:332
int64_t dts
Decompression timestamp in AVStream->time_base units; the time at which the packet is decompressed...
Definition: avcodec.h:1600
ITU-R BT2020.
Definition: pixfmt.h:400
#define AV_DICT_IGNORE_SUFFIX
Return first entry in a dictionary whose first part corresponds to the search key, ignoring the suffix of the found key string.
Definition: dict.h:70
static const AVCodecDefault defaults[]
Definition: dcaenc.c:975
int depth
Number of bits in the component.
Definition: pixdesc.h:58
#define AVERROR_EXTERNAL
Generic error in an external library.
Definition: error.h:57
AVPixelFormat
Pixel format.
Definition: pixfmt.h:60
This structure stores compressed data.
Definition: avcodec.h:1578
int64_t pts
Presentation timestamp in AVStream->time_base units; the time at which the decompressed packet will b...
Definition: avcodec.h:1594
Predicted.
Definition: avutil.h:269
x265_encoder * encoder
Definition: libx265.c:40