FFmpeg  4.0
libkvazaar.c
Go to the documentation of this file.
1 /*
2  * libkvazaar encoder
3  *
4  * Copyright (c) 2015 Tampere University of Technology
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 #include <kvazaar.h>
24 #include <stdint.h>
25 #include <string.h>
26 
27 #include "libavutil/attributes.h"
28 #include "libavutil/avassert.h"
29 #include "libavutil/dict.h"
30 #include "libavutil/error.h"
31 #include "libavutil/imgutils.h"
32 #include "libavutil/internal.h"
33 #include "libavutil/log.h"
34 #include "libavutil/mem.h"
35 #include "libavutil/pixdesc.h"
36 #include "libavutil/opt.h"
37 
38 #include "avcodec.h"
39 #include "internal.h"
40 
41 typedef struct LibkvazaarContext {
42  const AVClass *class;
43 
44  const kvz_api *api;
45  kvz_encoder *encoder;
46  kvz_config *config;
47 
48  char *kvz_params;
50 
52 {
53  LibkvazaarContext *const ctx = avctx->priv_data;
54  const kvz_api *const api = ctx->api = kvz_api_get(8);
55  kvz_config *cfg = NULL;
56  kvz_encoder *enc = NULL;
57 
58  /* Kvazaar requires width and height to be multiples of eight. */
59  if (avctx->width % 8 || avctx->height % 8) {
60  av_log(avctx, AV_LOG_ERROR,
61  "Video dimensions are not a multiple of 8 (%dx%d).\n",
62  avctx->width, avctx->height);
63  return AVERROR(ENOSYS);
64  }
65 
66  ctx->config = cfg = api->config_alloc();
67  if (!cfg) {
68  av_log(avctx, AV_LOG_ERROR,
69  "Could not allocate kvazaar config structure.\n");
70  return AVERROR(ENOMEM);
71  }
72 
73  if (!api->config_init(cfg)) {
74  av_log(avctx, AV_LOG_ERROR,
75  "Could not initialize kvazaar config structure.\n");
76  return AVERROR_BUG;
77  }
78 
79  cfg->width = avctx->width;
80  cfg->height = avctx->height;
81 
82  if (avctx->ticks_per_frame > INT_MAX / avctx->time_base.num) {
83  av_log(avctx, AV_LOG_ERROR,
84  "Could not set framerate for kvazaar: integer overflow\n");
85  return AVERROR(EINVAL);
86  }
87  cfg->framerate_num = avctx->time_base.den;
88  cfg->framerate_denom = avctx->time_base.num * avctx->ticks_per_frame;
89  cfg->target_bitrate = avctx->bit_rate;
90  cfg->vui.sar_width = avctx->sample_aspect_ratio.num;
91  cfg->vui.sar_height = avctx->sample_aspect_ratio.den;
92 
93  if (ctx->kvz_params) {
94  AVDictionary *dict = NULL;
95  if (!av_dict_parse_string(&dict, ctx->kvz_params, "=", ",", 0)) {
96  AVDictionaryEntry *entry = NULL;
97  while ((entry = av_dict_get(dict, "", entry, AV_DICT_IGNORE_SUFFIX))) {
98  if (!api->config_parse(cfg, entry->key, entry->value)) {
99  av_log(avctx, AV_LOG_WARNING, "Invalid option: %s=%s.\n",
100  entry->key, entry->value);
101  }
102  }
103  av_dict_free(&dict);
104  }
105  }
106 
107  ctx->encoder = enc = api->encoder_open(cfg);
108  if (!enc) {
109  av_log(avctx, AV_LOG_ERROR, "Could not open kvazaar encoder.\n");
110  return AVERROR_BUG;
111  }
112 
113  if (avctx->flags & AV_CODEC_FLAG_GLOBAL_HEADER) {
114  kvz_data_chunk *data_out = NULL;
115  kvz_data_chunk *chunk = NULL;
116  uint32_t len_out;
117  uint8_t *p;
118 
119  if (!api->encoder_headers(enc, &data_out, &len_out))
120  return AVERROR(ENOMEM);
121 
122  avctx->extradata = p = av_mallocz(len_out + AV_INPUT_BUFFER_PADDING_SIZE);
123  if (!p) {
124  ctx->api->chunk_free(data_out);
125  return AVERROR(ENOMEM);
126  }
127 
128  avctx->extradata_size = len_out;
129 
130  for (chunk = data_out; chunk != NULL; chunk = chunk->next) {
131  memcpy(p, chunk->data, chunk->len);
132  p += chunk->len;
133  }
134 
135  ctx->api->chunk_free(data_out);
136  }
137 
138  return 0;
139 }
140 
142 {
143  LibkvazaarContext *ctx = avctx->priv_data;
144 
145  if (ctx->api) {
146  ctx->api->encoder_close(ctx->encoder);
147  ctx->api->config_destroy(ctx->config);
148  }
149 
150  if (avctx->extradata)
151  av_freep(&avctx->extradata);
152 
153  return 0;
154 }
155 
157  AVPacket *avpkt,
158  const AVFrame *frame,
159  int *got_packet_ptr)
160 {
161  LibkvazaarContext *ctx = avctx->priv_data;
162  kvz_picture *input_pic = NULL;
163  kvz_picture *recon_pic = NULL;
164  kvz_frame_info frame_info;
165  kvz_data_chunk *data_out = NULL;
166  uint32_t len_out = 0;
167  int retval = 0;
168 
169  *got_packet_ptr = 0;
170 
171  if (frame) {
172  if (frame->width != ctx->config->width ||
173  frame->height != ctx->config->height) {
174  av_log(avctx, AV_LOG_ERROR,
175  "Changing video dimensions during encoding is not supported. "
176  "(changed from %dx%d to %dx%d)\n",
177  ctx->config->width, ctx->config->height,
178  frame->width, frame->height);
179  retval = AVERROR_INVALIDDATA;
180  goto done;
181  }
182 
183  if (frame->format != avctx->pix_fmt) {
184  av_log(avctx, AV_LOG_ERROR,
185  "Changing pixel format during encoding is not supported. "
186  "(changed from %s to %s)\n",
188  av_get_pix_fmt_name(frame->format));
189  retval = AVERROR_INVALIDDATA;
190  goto done;
191  }
192 
193  // Allocate input picture for kvazaar.
194  input_pic = ctx->api->picture_alloc(frame->width, frame->height);
195  if (!input_pic) {
196  av_log(avctx, AV_LOG_ERROR, "Failed to allocate picture.\n");
197  retval = AVERROR(ENOMEM);
198  goto done;
199  }
200 
201  // Copy pixels from frame to input_pic.
202  {
203  int dst_linesizes[4] = {
204  frame->width,
205  frame->width / 2,
206  frame->width / 2,
207  0
208  };
209  av_image_copy(input_pic->data, dst_linesizes,
210  frame->data, frame->linesize,
211  frame->format, frame->width, frame->height);
212  }
213 
214  input_pic->pts = frame->pts;
215  }
216 
217  retval = ctx->api->encoder_encode(ctx->encoder,
218  input_pic,
219  &data_out, &len_out,
220  &recon_pic, NULL,
221  &frame_info);
222  if (!retval) {
223  av_log(avctx, AV_LOG_ERROR, "Failed to encode frame.\n");
224  retval = AVERROR_INVALIDDATA;
225  goto done;
226  }
227  else
228  retval = 0; /* kvazaar returns 1 on success */
229 
230  if (data_out) {
231  kvz_data_chunk *chunk = NULL;
232  uint64_t written = 0;
233 
234  retval = ff_alloc_packet2(avctx, avpkt, len_out, len_out);
235  if (retval < 0) {
236  av_log(avctx, AV_LOG_ERROR, "Failed to allocate output packet.\n");
237  goto done;
238  }
239 
240  for (chunk = data_out; chunk != NULL; chunk = chunk->next) {
241  av_assert0(written + chunk->len <= len_out);
242  memcpy(avpkt->data + written, chunk->data, chunk->len);
243  written += chunk->len;
244  }
245 
246  avpkt->pts = recon_pic->pts;
247  avpkt->dts = recon_pic->dts;
248  avpkt->flags = 0;
249  // IRAP VCL NAL unit types span the range
250  // [BLA_W_LP (16), RSV_IRAP_VCL23 (23)].
251  if (frame_info.nal_unit_type >= KVZ_NAL_BLA_W_LP &&
252  frame_info.nal_unit_type <= KVZ_NAL_RSV_IRAP_VCL23) {
253  avpkt->flags |= AV_PKT_FLAG_KEY;
254  }
255 
256  *got_packet_ptr = 1;
257  }
258 
259 done:
260  ctx->api->picture_free(input_pic);
261  ctx->api->picture_free(recon_pic);
262  ctx->api->chunk_free(data_out);
263  return retval;
264 }
265 
266 static const enum AVPixelFormat pix_fmts[] = {
269 };
270 
271 #define OFFSET(x) offsetof(LibkvazaarContext, x)
272 #define VE AV_OPT_FLAG_VIDEO_PARAM | AV_OPT_FLAG_ENCODING_PARAM
273 static const AVOption options[] = {
274  { "kvazaar-params", "Set kvazaar parameters as a comma-separated list of key=value pairs.",
275  OFFSET(kvz_params), AV_OPT_TYPE_STRING, { .str = NULL }, 0, 0, VE },
276  { NULL },
277 };
278 
279 static const AVClass class = {
280  .class_name = "libkvazaar",
281  .item_name = av_default_item_name,
282  .option = options,
284 };
285 
286 static const AVCodecDefault defaults[] = {
287  { "b", "0" },
288  { NULL },
289 };
290 
292  .name = "libkvazaar",
293  .long_name = NULL_IF_CONFIG_SMALL("libkvazaar H.265 / HEVC"),
294  .type = AVMEDIA_TYPE_VIDEO,
295  .id = AV_CODEC_ID_HEVC,
296  .capabilities = AV_CODEC_CAP_DELAY,
297  .pix_fmts = pix_fmts,
298 
299  .priv_class = &class,
300  .priv_data_size = sizeof(LibkvazaarContext),
301  .defaults = defaults,
302 
304  .encode2 = libkvazaar_encode,
305  .close = libkvazaar_close,
306 
308 
309  .wrapper_name = "libkvazaar",
310 };
#define FF_CODEC_CAP_INIT_CLEANUP
The codec allows calling the close function for deallocation even if the init function returned a fai...
Definition: internal.h:48
#define NULL
Definition: coverity.c:32
#define AVERROR_INVALIDDATA
Invalid data found when processing input.
Definition: error.h:59
This structure describes decoded (raw) audio or video data.
Definition: frame.h:218
AVOption.
Definition: opt.h:246
misc image utilities
#define AV_LOG_WARNING
Something somehow does not look correct.
Definition: log.h:182
int64_t bit_rate
the average bitrate
Definition: avcodec.h:1568
#define LIBAVUTIL_VERSION_INT
Definition: version.h:85
Memory handling functions.
static av_cold int init(AVCodecContext *avctx)
Definition: avrndec.c:35
int num
Numerator.
Definition: rational.h:59
const char * av_default_item_name(void *ptr)
Return the context name.
Definition: log.c:191
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:1896
enum AVPixelFormat pix_fmt
Pixel format, see AV_PIX_FMT_xxx.
Definition: avcodec.h:1727
static av_cold int libkvazaar_close(AVCodecContext *avctx)
Definition: libkvazaar.c:141
AVCodec ff_libkvazaar_encoder
Definition: libkvazaar.c:291
#define VE
Definition: libkvazaar.c:272
AVCodec.
Definition: avcodec.h:3408
Macro definitions for various function/variable attributes.
AVRational time_base
This is the fundamental unit of time (in seconds) in terms of which frame timestamps are represented...
Definition: avcodec.h:1640
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
#define av_assert0(cond)
assert() equivalent, that is always enabled.
Definition: avassert.h:37
Public dictionary API.
int ff_alloc_packet2(AVCodecContext *avctx, AVPacket *avpkt, int64_t size, int64_t min_size)
Check AVPacket size and/or allocate data.
Definition: encode.c:32
#define OFFSET(x)
Definition: libkvazaar.c:271
#define FF_CODEC_CAP_INIT_THREADSAFE
The codec does not modify any global variables in the init function, allowing to call the init functi...
Definition: internal.h:40
uint8_t
#define av_cold
Definition: attributes.h:82
AVOptions.
int64_t pts
Presentation timestamp in time_base units (time when frame should be shown to user).
Definition: frame.h:311
uint8_t * extradata
some codecs need / can use extradata like Huffman tables.
Definition: avcodec.h:1618
static AVFrame * frame
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:1430
kvz_encoder * encoder
Definition: libkvazaar.c:45
kvz_config * config
Definition: libkvazaar.c:46
#define av_log(a,...)
const kvz_api * api
Definition: libkvazaar.c:44
#define AV_PKT_FLAG_KEY
The packet contains a keyframe.
Definition: avcodec.h:1462
int width
Definition: frame.h:276
#define AV_LOG_ERROR
Something went wrong and cannot losslessly be recovered.
Definition: log.h:176
error code definitions
#define AVERROR(e)
Definition: error.h:43
#define NULL_IF_CONFIG_SMALL(x)
Return NULL if CONFIG_SMALL is true, otherwise the argument without modification. ...
Definition: internal.h:186
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:1598
simple assert() macros that are a bit more flexible than ISO C assert().
void * av_mallocz(size_t size)
Allocate a memory block with alignment suitable for all memory accesses (including vectors if availab...
Definition: mem.c:236
const char * name
Name of the codec implementation.
Definition: avcodec.h:3415
void av_image_copy(uint8_t *dst_data[4], int dst_linesizes[4], const uint8_t *src_data[4], const int src_linesizes[4], enum AVPixelFormat pix_fmt, int width, int height)
Copy image in src_data to dst_data.
Definition: imgutils.c:387
int flags
A combination of AV_PKT_FLAG values.
Definition: avcodec.h:1436
static const AVOption options[]
Definition: libkvazaar.c:273
common internal API header
int width
picture width / height.
Definition: avcodec.h:1690
AVFormatContext * ctx
Definition: movenc.c:48
int ticks_per_frame
For some codecs, the time base is closer to the field rate than the frame rate.
Definition: avcodec.h:1649
static const AVCodecDefault defaults[]
Definition: libkvazaar.c:286
int format
format of the frame, -1 if unknown or unset Values correspond to enum AVPixelFormat for video frames...
Definition: frame.h:291
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
Libavcodec external API header.
int linesize[AV_NUM_DATA_POINTERS]
For video, size in bytes of each picture line.
Definition: frame.h:249
main external API structure.
Definition: avcodec.h:1518
int extradata_size
Definition: avcodec.h:1619
#define AVERROR_BUG
Internal bug, also see AVERROR_BUG2.
Definition: error.h:50
Describe the class of an AVClass context structure.
Definition: log.h:67
static int libkvazaar_encode(AVCodecContext *avctx, AVPacket *avpkt, const AVFrame *frame, int *got_packet_ptr)
Definition: libkvazaar.c:156
static av_cold int libkvazaar_init(AVCodecContext *avctx)
Definition: libkvazaar.c:51
static enum AVPixelFormat pix_fmts[]
Definition: libkvazaar.c:266
uint8_t * data[AV_NUM_DATA_POINTERS]
pointer to the picture/channel planes.
Definition: frame.h:232
#define AV_CODEC_FLAG_GLOBAL_HEADER
Place global headers in extradata instead of every keyframe.
Definition: avcodec.h:882
planar YUV 4:2:0, 12bpp, (1 Cr & Cb sample per 2x2 Y samples)
Definition: pixfmt.h:62
common internal api header.
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:773
void * priv_data
Definition: avcodec.h:1545
char * value
Definition: dict.h:87
int64_t dts
Decompression timestamp in AVStream->time_base units; the time at which the packet is decompressed...
Definition: avcodec.h:1429
int height
Definition: frame.h:276
#define av_freep(p)
#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
const char * av_get_pix_fmt_name(enum AVPixelFormat pix_fmt)
Return the short name for a pixel format, NULL in case pix_fmt is unknown.
Definition: pixdesc.c:2279
AVPixelFormat
Pixel format.
Definition: pixfmt.h:60
This structure stores compressed data.
Definition: avcodec.h:1407
int64_t pts
Presentation timestamp in AVStream->time_base units; the time at which the decompressed packet will b...
Definition: avcodec.h:1423