1 /*
2 * Copyright (c) 2010, Google, Inc.
3 *
4 * This file is part of FFmpeg.
5 *
6 * FFmpeg is free software; you can redistribute it and/or
7 * modify it under the terms of the GNU Lesser General Public
8 * License as published by the Free Software Foundation; either
9 * version 2.1 of the License, or (at your option) any later version.
10 *
11 * FFmpeg is distributed in the hope that it will be useful,
12 * but WITHOUT ANY WARRANTY; without even the implied warranty of
13 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
14 * Lesser General Public License for more details.
15 *
16 * You should have received a copy of the GNU Lesser General Public
17 * License along with FFmpeg; if not, write to the Free Software
18 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
19 */
20
21 /**
22 * @file
23 * VP8 encoder support via libvpx
24 */
25
26 #define VPX_DISABLE_CTRL_TYPECHECKS 1
27 #define VPX_CODEC_DISABLE_COMPAT 1
28 #include <vpx/vpx_encoder.h>
29 #include <vpx/vp8cx.h>
30
42
43 /**
44 * Portion of struct vpx_codec_cx_pkt from vpx_encoder.h.
45 * One encoded frame returned from the library.
46 */
48 void *
buf;
/**< compressed data buffer */
49 size_t sz;
/**< length of compressed data */
52 int64_t
pts;
/**< time stamp to show frame
53 (in timebase units) */
54 unsigned long duration;
/**< duration to show frame
55 (in timebase units) */
56 uint32_t
flags;
/**< flags for this frame */
58 int have_sse;
/**< true if we have pending sse[] */
61 };
62
63 typedef struct VP8EncoderContext {
65 struct vpx_codec_ctx encoder;
66 struct vpx_image rawimg;
67 struct vpx_codec_ctx encoder_alpha;
68 struct vpx_image rawimg_alpha;
70 struct vpx_fixed_buf twopass_stats;
73 int have_sse;
/**< true if we have pending sse[] */
76
78 /**
79 * VP8 specific flags, see VP8F_* below.
80 */
82 #define VP8F_ERROR_RESILIENT 0x00000001 ///< Enable measures appropriate for streaming over lossy links
83 #define VP8F_AUTO_ALT_REF 0x00000002
///< Enable automatic alternate reference frame generation
84
86
90
92
100
101 // VP9-only
110
111 /** String mappings for enum vp8e_enc_control_id */
113 [VP8E_SET_CPUUSED] = "VP8E_SET_CPUUSED",
114 [VP8E_SET_ENABLEAUTOALTREF] = "VP8E_SET_ENABLEAUTOALTREF",
115 [VP8E_SET_NOISE_SENSITIVITY] = "VP8E_SET_NOISE_SENSITIVITY",
116 [VP8E_SET_STATIC_THRESHOLD] = "VP8E_SET_STATIC_THRESHOLD",
117 [VP8E_SET_TOKEN_PARTITIONS] = "VP8E_SET_TOKEN_PARTITIONS",
118 [VP8E_SET_ARNR_MAXFRAMES] = "VP8E_SET_ARNR_MAXFRAMES",
119 [VP8E_SET_ARNR_STRENGTH] = "VP8E_SET_ARNR_STRENGTH",
120 [VP8E_SET_ARNR_TYPE] = "VP8E_SET_ARNR_TYPE",
121 [VP8E_SET_TUNING] = "VP8E_SET_TUNING",
122 [VP8E_SET_CQ_LEVEL] = "VP8E_SET_CQ_LEVEL",
123 [VP8E_SET_MAX_INTRA_BITRATE_PCT] = "VP8E_SET_MAX_INTRA_BITRATE_PCT",
124 #if CONFIG_LIBVPX_VP9_ENCODER
125 [VP9E_SET_LOSSLESS] = "VP9E_SET_LOSSLESS",
126 [VP9E_SET_TILE_COLUMNS] = "VP9E_SET_TILE_COLUMNS",
127 [VP9E_SET_TILE_ROWS] = "VP9E_SET_TILE_ROWS",
128 [VP9E_SET_FRAME_PARALLEL_DECODING] = "VP9E_SET_FRAME_PARALLEL_DECODING",
129 [VP9E_SET_AQ_MODE] = "VP9E_SET_AQ_MODE",
130 #if VPX_ENCODER_ABI_VERSION > 8
131 [VP9E_SET_COLOR_SPACE] = "VP9E_SET_COLOR_SPACE",
132 #endif
133 #if VPX_ENCODER_ABI_VERSION >= 11
134 [VP9E_SET_COLOR_RANGE] = "VP9E_SET_COLOR_RANGE",
135 #endif
136 #endif
137 };
138
140 {
142 const char *error = vpx_codec_error(&ctx->
encoder);
143 const char *detail = vpx_codec_error_detail(&ctx->
encoder);
144
146 if (detail)
148 }
149
151 const struct vpx_codec_enc_cfg *cfg)
152 {
155
156 av_log(avctx, level,
"vpx_codec_enc_cfg\n");
157 av_log(avctx, level,
"generic settings\n"
158 " %*s%u\n %*s%u\n %*s%u\n %*s%u\n %*s%u\n"
159 #if CONFIG_LIBVPX_VP9_ENCODER && defined(VPX_IMG_FMT_HIGHBITDEPTH)
160 " %*s%u\n %*s%u\n"
161 #endif
162 " %*s{%u/%u}\n %*s%u\n %*s%d\n %*s%u\n",
163 width, "g_usage:", cfg->g_usage,
164 width, "g_threads:", cfg->g_threads,
165 width, "g_profile:", cfg->g_profile,
166 width, "g_w:", cfg->g_w,
167 width, "g_h:", cfg->g_h,
168 #if CONFIG_LIBVPX_VP9_ENCODER && defined(VPX_IMG_FMT_HIGHBITDEPTH)
169 width, "g_bit_depth:", cfg->g_bit_depth,
170 width, "g_input_bit_depth:", cfg->g_input_bit_depth,
171 #endif
172 width, "g_timebase:", cfg->g_timebase.num, cfg->g_timebase.den,
173 width, "g_error_resilient:", cfg->g_error_resilient,
174 width, "g_pass:", cfg->g_pass,
175 width, "g_lag_in_frames:", cfg->g_lag_in_frames);
176 av_log(avctx, level,
"rate control settings\n"
177 " %*s%u\n %*s%u\n %*s%u\n %*s%u\n"
179 width, "rc_dropframe_thresh:", cfg->rc_dropframe_thresh,
180 width, "rc_resize_allowed:", cfg->rc_resize_allowed,
181 width, "rc_resize_up_thresh:", cfg->rc_resize_up_thresh,
182 width, "rc_resize_down_thresh:", cfg->rc_resize_down_thresh,
183 width, "rc_end_usage:", cfg->rc_end_usage,
184 width, "rc_twopass_stats_in:", cfg->rc_twopass_stats_in.buf, cfg->rc_twopass_stats_in.sz,
185 width, "rc_target_bitrate:", cfg->rc_target_bitrate);
186 av_log(avctx, level,
"quantizer settings\n"
187 " %*s%u\n %*s%u\n",
188 width, "rc_min_quantizer:", cfg->rc_min_quantizer,
189 width, "rc_max_quantizer:", cfg->rc_max_quantizer);
190 av_log(avctx, level,
"bitrate tolerance\n"
191 " %*s%u\n %*s%u\n",
192 width, "rc_undershoot_pct:", cfg->rc_undershoot_pct,
193 width, "rc_overshoot_pct:", cfg->rc_overshoot_pct);
194 av_log(avctx, level,
"decoder buffer model\n"
195 " %*s%u\n %*s%u\n %*s%u\n",
196 width, "rc_buf_sz:", cfg->rc_buf_sz,
197 width, "rc_buf_initial_sz:", cfg->rc_buf_initial_sz,
198 width, "rc_buf_optimal_sz:", cfg->rc_buf_optimal_sz);
199 av_log(avctx, level,
"2 pass rate control settings\n"
200 " %*s%u\n %*s%u\n %*s%u\n",
201 width, "rc_2pass_vbr_bias_pct:", cfg->rc_2pass_vbr_bias_pct,
202 width, "rc_2pass_vbr_minsection_pct:", cfg->rc_2pass_vbr_minsection_pct,
203 width, "rc_2pass_vbr_maxsection_pct:", cfg->rc_2pass_vbr_maxsection_pct);
204 av_log(avctx, level,
"keyframing settings\n"
205 " %*s%d\n %*s%u\n %*s%u\n",
206 width, "kf_mode:", cfg->kf_mode,
207 width, "kf_min_dist:", cfg->kf_min_dist,
208 width, "kf_max_dist:", cfg->kf_max_dist);
209 av_log(avctx, level,
"\n");
210 }
211
213 {
215
216 while (*p)
218 *p = cx_frame;
220 }
221
223 {
228 }
229
231 {
233
234 while (p) {
237 p = list;
238 }
239 }
240
242 enum vp8e_enc_control_id
id,
int val)
243 {
247 int res;
248
251
252 res = vpx_codec_control(&ctx->
encoder,
id, val);
253 if (res != VPX_CODEC_OK) {
254 snprintf(buf,
sizeof(buf),
"Failed to set %s codec control",
257 }
258
259 return res == VPX_CODEC_OK ? 0 :
AVERROR(EINVAL);
260 }
261
263 {
265
266 vpx_codec_destroy(&ctx->
encoder);
272 return 0;
273 }
274
275 #if CONFIG_LIBVPX_VP9_ENCODER
277 struct vpx_codec_enc_cfg *enccfg, vpx_codec_flags_t *
flags,
278 vpx_img_fmt_t *img_fmt)
279 {
280 #ifdef VPX_IMG_FMT_HIGHBITDEPTH
281 enccfg->g_bit_depth = enccfg->g_input_bit_depth = 8;
282 #endif
285 enccfg->g_profile = 0;
286 *img_fmt = VPX_IMG_FMT_I420;
287 return 0;
289 enccfg->g_profile = 1;
290 *img_fmt = VPX_IMG_FMT_I422;
291 return 0;
292 #if VPX_IMAGE_ABI_VERSION >= 3
294 enccfg->g_profile = 1;
295 *img_fmt = VPX_IMG_FMT_I440;
296 return 0;
297 #endif
299 enccfg->g_profile = 1;
300 *img_fmt = VPX_IMG_FMT_I444;
301 return 0;
302 #ifdef VPX_IMG_FMT_HIGHBITDEPTH
305 if (codec_caps & VPX_CODEC_CAP_HIGHBITDEPTH) {
306 enccfg->g_bit_depth = enccfg->g_input_bit_depth =
308 enccfg->g_profile = 2;
309 *img_fmt = VPX_IMG_FMT_I42016;
310 *flags |= VPX_CODEC_USE_HIGHBITDEPTH;
311 return 0;
312 }
313 break;
316 if (codec_caps & VPX_CODEC_CAP_HIGHBITDEPTH) {
317 enccfg->g_bit_depth = enccfg->g_input_bit_depth =
319 enccfg->g_profile = 3;
320 *img_fmt = VPX_IMG_FMT_I42216;
321 *flags |= VPX_CODEC_USE_HIGHBITDEPTH;
322 return 0;
323 }
324 break;
325 #if VPX_IMAGE_ABI_VERSION >= 3
328 if (codec_caps & VPX_CODEC_CAP_HIGHBITDEPTH) {
329 enccfg->g_bit_depth = enccfg->g_input_bit_depth =
331 enccfg->g_profile = 3;
332 *img_fmt = VPX_IMG_FMT_I44016;
333 *flags |= VPX_CODEC_USE_HIGHBITDEPTH;
334 return 0;
335 }
336 break;
337 #endif
340 if (codec_caps & VPX_CODEC_CAP_HIGHBITDEPTH) {
341 enccfg->g_bit_depth = enccfg->g_input_bit_depth =
343 enccfg->g_profile = 3;
344 *img_fmt = VPX_IMG_FMT_I44416;
345 *flags |= VPX_CODEC_USE_HIGHBITDEPTH;
346 return 0;
347 }
348 break;
349 #endif
350 default:
351 break;
352 }
355 }
356
357 #if VPX_ENCODER_ABI_VERSION > 8
359 {
360 enum vpx_color_space vpx_cs;
361
371 default:
374 return;
375 }
377 }
378 #endif
379
380 #if VPX_ENCODER_ABI_VERSION >= 11
382 {
383 enum vpx_color_range vpx_cr;
388 default:
391 return;
392 }
393
395 }
396 #endif
397 #endif
398
400 const struct vpx_codec_iface *iface)
401 {
403 struct vpx_codec_enc_cfg enccfg = { 0 };
404 struct vpx_codec_enc_cfg enccfg_alpha;
407 int res;
408 vpx_img_fmt_t img_fmt = VPX_IMG_FMT_I420;
409 #if CONFIG_LIBVPX_VP9_ENCODER
410 vpx_codec_caps_t codec_caps = vpx_codec_get_caps(iface);
411 #endif
412
415
418
419 if ((res = vpx_codec_enc_config_default(iface, &enccfg, 0)) != VPX_CODEC_OK) {
421 vpx_codec_err_to_string(res));
423 }
424
425 #if CONFIG_LIBVPX_VP9_ENCODER
427 if (
set_pix_fmt(avctx, codec_caps, &enccfg, &flags, &img_fmt))
429 }
430 #endif
431
436 }
437
439
440 enccfg.g_w = avctx->
width;
441 enccfg.g_h = avctx->
height;
446
448 enccfg.g_pass = VPX_RC_FIRST_PASS;
450 enccfg.g_pass = VPX_RC_LAST_PASS;
451 else
452 enccfg.g_pass = VPX_RC_ONE_PASS;
453
456 enccfg.rc_end_usage = VPX_CBR;
457 }
else if (ctx->
crf >= 0) {
458 enccfg.rc_end_usage = VPX_CQ;
459 #if CONFIG_LIBVPX_VP9_ENCODER
461 enccfg.rc_end_usage = VPX_Q;
462 #endif
463 }
464
468 #if CONFIG_LIBVPX_VP9_ENCODER
469 } else if (enccfg.rc_end_usage == VPX_Q) {
470 #endif
471 } else {
472 if (enccfg.rc_end_usage == VPX_CQ) {
473 enccfg.rc_target_bitrate = 1000000;
474 } else {
475 avctx->
bit_rate = enccfg.rc_target_bitrate * 1000;
477 "Neither bitrate nor constrained quality specified, using default bitrate of %dkbit/sec\n",
478 enccfg.rc_target_bitrate);
479 }
480 }
481
483 enccfg.rc_min_quantizer =
484 enccfg.rc_max_quantizer = 0;
485 } else {
486 if (avctx->
qmin >= 0)
487 enccfg.rc_min_quantizer = avctx->
qmin;
488 if (avctx->
qmax >= 0)
489 enccfg.rc_max_quantizer = avctx->
qmax;
490 }
491
492 if (enccfg.rc_end_usage == VPX_CQ
493 #if CONFIG_LIBVPX_VP9_ENCODER
494 || enccfg.rc_end_usage == VPX_Q
495 #endif
496 ) {
497 if (ctx->
crf < enccfg.rc_min_quantizer || ctx->
crf > enccfg.rc_max_quantizer) {
499 "CQ level %d must be between minimum and maximum quantizer value (%d-%d)\n",
500 ctx->
crf, enccfg.rc_min_quantizer, enccfg.rc_max_quantizer);
502 }
503 }
504
505 #if FF_API_PRIVATE_OPT
510 #endif
512
513 //0-100 (0 => CBR, 100 => VBR)
516 enccfg.rc_2pass_vbr_minsection_pct =
519 enccfg.rc_2pass_vbr_maxsection_pct =
521
523 enccfg.rc_buf_sz =
526 enccfg.rc_buf_initial_sz =
528 enccfg.rc_buf_optimal_sz = enccfg.rc_buf_sz * 5 / 6;
529 #if FF_API_MPV_OPT
533 "deprecated, use the undershoot-pct private option instead.\n");
535 }
537 #endif
542
543 //_enc_init() will balk if kf_min_dist differs from max w/VPX_KF_AUTO
547 enccfg.kf_max_dist = avctx->
gop_size;
548
549 if (enccfg.g_pass == VPX_RC_FIRST_PASS)
550 enccfg.g_lag_in_frames = 0;
551 else if (enccfg.g_pass == VPX_RC_LAST_PASS) {
552 int decode_size, ret;
553
557 }
558
561 if (ret < 0) {
566 return ret;
567 }
570 if (decode_size < 0) {
573 }
574
577 }
578
579 /* 0-3: For non-zero values the encoder increasingly optimizes for reduced
580 complexity playback on low powered devices at the expense of encode
581 quality. */
583 enccfg.g_profile = avctx->
profile;
584
586
588 /* Construct Encoder Context */
589 res = vpx_codec_enc_init(&ctx->
encoder, iface, &enccfg, flags);
590 if (res != VPX_CODEC_OK) {
593 }
594
596 enccfg_alpha = enccfg;
597 res = vpx_codec_enc_init(&ctx->
encoder_alpha, iface, &enccfg_alpha, flags);
598 if (res != VPX_CODEC_OK) {
601 }
602 }
603
604 //codec control failures are currently treated only as warnings
619
621 #if FF_API_PRIVATE_OPT
626 #endif
629 }
630 #if FF_API_MPV_OPT
634 "use the static-thresh private option instead.\n");
636 }
638 #endif
644
645 #if CONFIG_LIBVPX_VP9_ENCODER
657 #if VPX_ENCODER_ABI_VERSION > 8
658 set_colorspace(avctx);
659 #endif
660 #if VPX_ENCODER_ABI_VERSION >= 11
661 set_color_range(avctx);
662 #endif
663 }
664 #endif
665
667
668 //provide dummy value to initialize wrapper, values will be updated each _encode()
670 (unsigned char*)1);
671 #if CONFIG_LIBVPX_VP9_ENCODER && defined(VPX_IMG_FMT_HIGHBITDEPTH)
673 ctx->
rawimg.bit_depth = enccfg.g_bit_depth;
674 #endif
675
678 (unsigned char*)1);
679
681 if (!cpb_props)
683
684 if (enccfg.rc_end_usage == VPX_CBR ||
685 enccfg.g_pass != VPX_RC_ONE_PASS) {
689 }
691
692 return 0;
693 }
694
696 const struct vpx_codec_cx_pkt *
src,
697 const struct vpx_codec_cx_pkt *src_alpha,
699 {
700 dst->
pts = src->data.frame.pts;
701 dst->
duration = src->data.frame.duration;
702 dst->
flags = src->data.frame.flags;
703 dst->
sz = src->data.frame.sz;
704 dst->
buf = src->data.frame.buf;
706 /* For alt-ref frame, don't store PSNR or increment frame_number */
707 if (!(dst->
flags & VPX_FRAME_IS_INVISIBLE)) {
711 /* associate last-seen SSE to the frame. */
712 /* Transfers ownership from ctx to dst. */
713 /* WARNING! This makes the assumption that PSNR_PKT comes
714 just before the frame it refers to! */
715 memcpy(dst->
sse, ctx->
sse,
sizeof(dst->
sse));
717 }
718 } else {
720 }
721 if (src_alpha) {
722 dst->
buf_alpha = src_alpha->data.frame.buf;
723 dst->
sz_alpha = src_alpha->data.frame.sz;
724 } else {
727 }
728 }
729
730 /**
731 * Store coded frame information in format suitable for return from encode2().
732 *
733 * Write information from @a cx_frame to @a pkt
734 * @return packet data size on success
735 * @return a negative AVERROR on error
736 */
739 {
742 if (ret >= 0) {
743 int pict_type;
746 #if FF_API_CODED_FRAME
751 #endif
752
753 if (!!(cx_frame->
flags & VPX_FRAME_IS_KEY)) {
755 #if FF_API_CODED_FRAME
759 #endif
761 } else {
763 #if FF_API_CODED_FRAME
767 #endif
768 }
769
771 cx_frame->
have_sse ? 3 : 0, pict_type);
772
774 int i;
775 /* Beware of the Y/U/V/all order! */
776 #if FF_API_CODED_FRAME
783 #endif
784 for (i = 0; i < 3; ++i) {
785 avctx->
error[i] += cx_frame->
sse[i + 1];
786 }
788 }
793 if(!side_data) {
797 }
800 }
801 } else {
802 return ret;
803 }
805 }
806
807 /**
808 * Queue multiple output frames from the encoder, returning the front-most.
809 * In cases where vpx_codec_get_cx_data() returns more than 1 frame append
810 * the frame queue. Return the head frame if available.
811 * @return Stored frame size
812 * @return AVERROR(EINVAL) on output size error
813 * @return AVERROR(ENOMEM) on coded frame queue data allocation error
814 */
816 {
818 const struct vpx_codec_cx_pkt *
pkt;
819 const struct vpx_codec_cx_pkt *pkt_alpha =
NULL;
820 const void *iter =
NULL;
821 const void *iter_alpha =
NULL;
823
826 /* return the leading frame if we've already begun queueing */
828 if (size < 0)
832 }
833
834 /* consume all available output from the encoder before returning. buffers
835 are only good through the next vpx_codec call */
836 while ((pkt = vpx_codec_get_cx_data(&ctx->
encoder, &iter)) &&
839 switch (pkt->kind) {
840 case VPX_CODEC_CX_FRAME_PKT:
841 if (!size) {
843
844 /* avoid storing the frame when the list is empty and we haven't yet
845 provided a frame for output */
847 cx_pktcpy(&cx_frame, pkt, pkt_alpha, ctx);
849 if (size < 0)
851 } else {
854
855 if (!cx_frame) {
857 "Frame queue element alloc failed\n");
859 }
860 cx_pktcpy(cx_frame, pkt, pkt_alpha, ctx);
862
863 if (!cx_frame->
buf) {
869 }
870 memcpy(cx_frame->
buf, pkt->data.frame.buf, pkt->data.frame.sz);
879 }
880 memcpy(cx_frame->
buf_alpha, pkt_alpha->data.frame.buf, pkt_alpha->data.frame.sz);
881 }
883 }
884 break;
885 case VPX_CODEC_STATS_PKT: {
887 int err;
889 stats->sz +
890 pkt->data.twopass_stats.sz)) < 0) {
891 stats->sz = 0;
893 return err;
894 }
895 memcpy((
uint8_t*)stats->buf + stats->sz,
896 pkt->data.twopass_stats.buf, pkt->data.twopass_stats.sz);
897 stats->sz += pkt->data.twopass_stats.sz;
898 break;
899 }
900 case VPX_CODEC_PSNR_PKT:
902 ctx->
sse[0] = pkt->data.psnr.sse[0];
903 ctx->
sse[1] = pkt->data.psnr.sse[1];
904 ctx->
sse[2] = pkt->data.psnr.sse[2];
905 ctx->
sse[3] = pkt->data.psnr.sse[3];
907 break;
908 case VPX_CODEC_CUSTOM_PKT:
909 //ignore unsupported/unrecognized packet types
910 break;
911 }
912 }
913
915 }
916
919 {
921 struct vpx_image *rawimg =
NULL;
922 struct vpx_image *rawimg_alpha =
NULL;
923 int64_t timestamp = 0;
924 int res, coded_size;
925 vpx_enc_frame_flags_t flags = 0;
926
927 if (frame) {
929 rawimg->planes[VPX_PLANE_Y] = frame->
data[0];
930 rawimg->planes[VPX_PLANE_U] = frame->
data[1];
931 rawimg->planes[VPX_PLANE_V] = frame->
data[2];
932 rawimg->stride[VPX_PLANE_Y] = frame->
linesize[0];
933 rawimg->stride[VPX_PLANE_U] = frame->
linesize[1];
934 rawimg->stride[VPX_PLANE_V] = frame->
linesize[2];
938 rawimg_alpha->planes[VPX_PLANE_Y] = frame->
data[3];
941 if (!u_plane || !v_plane) {
945 }
947 rawimg_alpha->planes[VPX_PLANE_U] = u_plane;
949 rawimg_alpha->planes[VPX_PLANE_V] = v_plane;
950 rawimg_alpha->stride[VPX_PLANE_Y] = frame->
linesize[0];
951 rawimg_alpha->stride[VPX_PLANE_U] = frame->
linesize[1];
952 rawimg_alpha->stride[VPX_PLANE_V] = frame->
linesize[2];
953 }
954 timestamp = frame->
pts;
956 flags |= VPX_EFLAG_FORCE_KF;
957 }
958
959 res = vpx_codec_encode(&ctx->
encoder, rawimg, timestamp,
961 if (res != VPX_CODEC_OK) {
964 }
965
967 res = vpx_codec_encode(&ctx->
encoder_alpha, rawimg_alpha, timestamp,
969 if (res != VPX_CODEC_OK) {
972 }
973 }
974
976
979
983 b64_size);
985 }
988 }
989
990 if (rawimg_alpha) {
991 av_freep(&rawimg_alpha->planes[VPX_PLANE_U]);
992 av_freep(&rawimg_alpha->planes[VPX_PLANE_V]);
993 }
994
995 *got_packet = !!coded_size;
996 return 0;
997 }
998
999 #define OFFSET(x) offsetof(VP8Context, x)
1000 #define VE AV_OPT_FLAG_VIDEO_PARAM | AV_OPT_FLAG_ENCODING_PARAM
1001
1002 #ifndef VPX_ERROR_RESILIENT_DEFAULT
1003 #define VPX_ERROR_RESILIENT_DEFAULT 1
1004 #define VPX_ERROR_RESILIENT_PARTITIONS 2
1005 #endif
1006
1007 #define COMMON_OPTIONS \
1008 { "auto-alt-ref", "Enable use of alternate reference " \
1009 "frames (2-pass only)", OFFSET(auto_alt_ref), AV_OPT_TYPE_BOOL, {.i64 = -1}, -1, 1, VE}, \
1010 { "lag-in-frames", "Number of frames to look ahead for " \
1011 "alternate reference frame selection", OFFSET(lag_in_frames), AV_OPT_TYPE_INT, {.i64 = -1}, -1, INT_MAX, VE}, \
1012 { "arnr-maxframes", "altref noise reduction max frame count", OFFSET(arnr_max_frames), AV_OPT_TYPE_INT, {.i64 = -1}, -1, INT_MAX, VE}, \
1013 { "arnr-strength", "altref noise reduction filter strength", OFFSET(arnr_strength), AV_OPT_TYPE_INT, {.i64 = -1}, -1, INT_MAX, VE}, \
1014 { "arnr-type", "altref noise reduction filter type", OFFSET(arnr_type), AV_OPT_TYPE_INT, {.i64 = -1}, -1, INT_MAX, VE, "arnr_type"}, \
1015 { "backward", NULL, 0, AV_OPT_TYPE_CONST, {.i64 = 1}, 0, 0, VE, "arnr_type" }, \
1016 { "forward", NULL, 0, AV_OPT_TYPE_CONST, {.i64 = 2}, 0, 0, VE, "arnr_type" }, \
1017 { "centered", NULL, 0, AV_OPT_TYPE_CONST, {.i64 = 3}, 0, 0, VE, "arnr_type" }, \
1018 { "tune", "Tune the encoding to a specific scenario", OFFSET(tune), AV_OPT_TYPE_INT, {.i64 = -1}, -1, INT_MAX, VE, "tune"}, \
1019 { "psnr", NULL, 0, AV_OPT_TYPE_CONST, {.i64 = VP8_TUNE_PSNR}, 0, 0, VE, "tune"}, \
1020 { "ssim", NULL, 0, AV_OPT_TYPE_CONST, {.i64 = VP8_TUNE_SSIM}, 0, 0, VE, "tune"}, \
1021 { "deadline", "Time to spend encoding, in microseconds.", OFFSET(deadline), AV_OPT_TYPE_INT, {.i64 = VPX_DL_GOOD_QUALITY}, INT_MIN, INT_MAX, VE, "quality"}, \
1022 { "best", NULL, 0, AV_OPT_TYPE_CONST, {.i64 = VPX_DL_BEST_QUALITY}, 0, 0, VE, "quality"}, \
1023 { "good", NULL, 0, AV_OPT_TYPE_CONST, {.i64 = VPX_DL_GOOD_QUALITY}, 0, 0, VE, "quality"}, \
1024 { "realtime", NULL, 0, AV_OPT_TYPE_CONST, {.i64 = VPX_DL_REALTIME}, 0, 0, VE, "quality"}, \
1025 { "error-resilient", "Error resilience configuration", OFFSET(error_resilient), AV_OPT_TYPE_FLAGS, {.i64 = 0}, INT_MIN, INT_MAX, VE, "er"}, \
1026 { "max-intra-rate", "Maximum I-frame bitrate (pct) 0=unlimited", OFFSET(max_intra_rate), AV_OPT_TYPE_INT, {.i64 = -1}, -1, INT_MAX, VE}, \
1027 { "default", "Improve resiliency against losses of whole frames", 0, AV_OPT_TYPE_CONST, {.i64 = VPX_ERROR_RESILIENT_DEFAULT}, 0, 0, VE, "er"}, \
1028 { "partitions", "The frame partitions are independently decodable " \
1029 "by the bool decoder, meaning that partitions can be decoded even " \
1030 "though earlier partitions have been lost. Note that intra predicition" \
1031 " is still done over the partition boundary.", 0, AV_OPT_TYPE_CONST, {.i64 = VPX_ERROR_RESILIENT_PARTITIONS}, 0, 0, VE, "er"}, \
1032 { "crf", "Select the quality for constant quality mode", offsetof(VP8Context, crf), AV_OPT_TYPE_INT, {.i64 = -1}, -1, 63, VE }, \
1033 { "static-thresh", "A change threshold on blocks below which they will be skipped by the encoder", OFFSET(static_thresh), AV_OPT_TYPE_INT, { .i64 = 0 }, 0, INT_MAX, VE }, \
1034 { "drop-threshold", "Frame drop threshold", offsetof(VP8Context, drop_threshold), AV_OPT_TYPE_INT, {.i64 = 0 }, INT_MIN, INT_MAX, VE }, \
1035 { "noise-sensitivity", "Noise sensitivity", OFFSET(noise_sensitivity), AV_OPT_TYPE_INT, {.i64 = 0 }, 0, 4, VE}, \
1036 { "undershoot-pct", "Datarate undershoot (min) target (%)", OFFSET(rc_undershoot_pct), AV_OPT_TYPE_INT, { .i64 = -1 }, -1, 100, VE }, \
1037 { "overshoot-pct", "Datarate overshoot (max) target (%)", OFFSET(rc_overshoot_pct), AV_OPT_TYPE_INT, { .i64 = -1 }, -1, 1000, VE }, \
1038
1039 #define LEGACY_OPTIONS \
1040 {"speed", "", offsetof(VP8Context, cpu_used), AV_OPT_TYPE_INT, {.i64 = 1}, -16, 16, VE}, \
1041 {"quality", "", offsetof(VP8Context, deadline), AV_OPT_TYPE_INT, {.i64 = VPX_DL_GOOD_QUALITY}, INT_MIN, INT_MAX, VE, "quality"}, \
1042 {"vp8flags", "", offsetof(VP8Context, flags), AV_OPT_TYPE_FLAGS, {.i64 = 0}, 0, UINT_MAX, VE, "flags"}, \
1043 {"error_resilient", "enable error resilience", 0, AV_OPT_TYPE_CONST, {.i64 = VP8F_ERROR_RESILIENT}, INT_MIN, INT_MAX, VE, "flags"}, \
1044 {"altref", "enable use of alternate reference frames (VP8/2-pass only)", 0, AV_OPT_TYPE_CONST, {.i64 = VP8F_AUTO_ALT_REF}, INT_MIN, INT_MAX, VE, "flags"}, \
1045 {"arnr_max_frames", "altref noise reduction max frame count", offsetof(VP8Context, arnr_max_frames), AV_OPT_TYPE_INT, {.i64 = 0}, 0, 15, VE}, \
1046 {"arnr_strength", "altref noise reduction filter strength", offsetof(VP8Context, arnr_strength), AV_OPT_TYPE_INT, {.i64 = 3}, 0, 6, VE}, \
1047 {"arnr_type", "altref noise reduction filter type", offsetof(VP8Context, arnr_type), AV_OPT_TYPE_INT, {.i64 = 3}, 1, 3, VE}, \
1048 {"rc_lookahead", "Number of frames to look ahead for alternate reference frame selection", offsetof(VP8Context, lag_in_frames), AV_OPT_TYPE_INT, {.i64 = 25}, 0, 25, VE}, \
1049
1050 #if CONFIG_LIBVPX_VP8_ENCODER
1051 static const AVOption vp8_options[] = {
1056 };
1057 #endif
1058
1059 #if CONFIG_LIBVPX_VP9_ENCODER
1060 static const AVOption vp9_options[] = {
1064 {
"tile-columns",
"Number of tile columns to use, log2",
OFFSET(tile_columns),
AV_OPT_TYPE_INT, {.i64 = -1}, -1, 6,
VE},
1065 {
"tile-rows",
"Number of tile rows to use, log2",
OFFSET(tile_rows),
AV_OPT_TYPE_INT, {.i64 = -1}, -1, 2,
VE},
1066 {
"frame-parallel",
"Enable frame parallel decodability features",
OFFSET(frame_parallel),
AV_OPT_TYPE_BOOL,{.i64 = -1}, -1, 1,
VE},
1067 {
"aq-mode",
"adaptive quantization mode",
OFFSET(aq_mode),
AV_OPT_TYPE_INT, {.i64 = -1}, -1, 3,
VE,
"aq_mode"},
1069 {
"variance",
"Variance based Aq", 0,
AV_OPT_TYPE_CONST, {.i64 = 1}, 0, 0,
VE,
"aq_mode" },
1070 {
"complexity",
"Complexity based Aq", 0,
AV_OPT_TYPE_CONST, {.i64 = 2}, 0, 0,
VE,
"aq_mode" },
1074 };
1075 #endif
1076
1077 #undef COMMON_OPTIONS
1078 #undef LEGACY_OPTIONS
1079
1081 { "qmin", "-1" },
1082 { "qmax", "-1" },
1083 { "g", "-1" },
1084 { "keyint_min", "-1" },
1086 };
1087
1088 #if CONFIG_LIBVPX_VP8_ENCODER
1090 {
1091 return vpx_init(avctx, vpx_codec_vp8_cx());
1092 }
1093
1094 static const AVClass class_vp8 = {
1097 .option = vp8_options,
1099 };
1100
1101 AVCodec ff_libvpx_vp8_encoder = {
1112 .priv_class = &class_vp8,
1114 };
1115 #endif /* CONFIG_LIBVPX_VP8_ENCODER */
1116
1117 #if CONFIG_LIBVPX_VP9_ENCODER
1119 {
1120 return vpx_init(avctx, vpx_codec_vp9_cx());
1121 }
1122
1123 static const AVClass class_vp9 = {
1126 .option = vp9_options,
1128 };
1129
1130 AVCodec ff_libvpx_vp9_encoder = {
1131 .
name =
"libvpx-vp9",
1141 .priv_class = &class_vp9,
1144 };
1145 #endif /* CONFIG_LIBVPX_VP9_ENCODER */
also ITU-R BT1361 / IEC 61966-2-4 xvYCC709 / SMPTE RP177 Annex B
struct vpx_image rawimg_alpha
const char const char void * val
planar YUV 4:4:0,20bpp, (1 Cr & Cb sample per 1x2 Y samples), little-endian
#define AVERROR_INVALIDDATA
Invalid data found when processing input.
This structure describes decoded (raw) audio or video data.
int ff_side_data_set_encoder_stats(AVPacket *pkt, int quality, int64_t *error, int error_count, int pict_type)
planar YUV 4:2:0, 15bpp, (1 Cr & Cb sample per 2x2 Y samples), little-endian
struct vpx_codec_ctx encoder
uint64_t error[AV_NUM_DATA_POINTERS]
error
int64_t av_rescale_rnd(int64_t a, int64_t b, int64_t c, enum AVRounding rnd)
Rescale a 64-bit integer with specified rounding.
planar YUV 4:4:4, 24bpp, (1 Cr & Cb sample per 1x1 Y samples)
#define AV_LOG_WARNING
Something somehow does not look correct.
int64_t bit_rate
the average bitrate
#define LIBAVUTIL_VERSION_INT
static av_cold int init(AVCodecContext *avctx)
int max_bitrate
Maximum bitrate of the stream, in bits per second.
also ITU-R BT601-6 625 / ITU-R BT1358 625 / ITU-R BT1700 625 PAL & SECAM / IEC 61966-2-4 xvYCC601 ...
int rc_initial_buffer_occupancy
Number of bits which should be loaded into the rc buffer before decoding starts.
enum AVColorRange color_range
MPEG vs JPEG YUV range.
also ITU-R BT601-6 525 / ITU-R BT1358 525 / ITU-R BT1700 NTSC / functionally identical to above ...
enum AVPixelFormat pix_fmt
Pixel format, see AV_PIX_FMT_xxx.
size_t sz
length of compressed data
static int sse(MpegEncContext *s, uint8_t *src1, uint8_t *src2, int w, int h, int stride)
char * stats_in
pass2 encoding statistics input buffer Concatenated stuff from stats_out of pass1 should be placed he...
#define AV_CODEC_CAP_AUTO_THREADS
Codec supports avctx->thread_count == 0 (auto).
struct vpx_codec_ctx encoder_alpha
order of coefficients is actually GBR, also IEC 61966-2-1 (sRGB)
int min_bitrate
Minimum bitrate of the stream, in bits per second.
AVRational time_base
This is the fundamental unit of time (in seconds) in terms of which frame timestamps are represented...
#define VP8F_AUTO_ALT_REF
Enable automatic alternate reference frame generation.
const char * class_name
The name of the class; usually it is the same name as the context structure type to which the AVClass...
#define AV_CODEC_CAP_DELAY
Encoder or decoder requires flushing with NULL input at the end in order to give the complete and cor...
#define av_assert0(cond)
assert() equivalent, that is always enabled.
planar YUV 4:2:0, 20bpp, (1 Cr & Cb sample per 2x2 Y & A samples)
attribute_deprecated float rc_buffer_aggressivity
static av_cold int codecctl_int(AVCodecContext *avctx, enum vp8e_enc_control_id id, int val)
int64_t pts
time stamp to show frame (in timebase units)
static void coded_frame_add(void *list, struct FrameListData *cx_frame)
planar YUV 4:4:4,36bpp, (1 Cr & Cb sample per 1x1 Y samples), little-endian
int64_t pts
Presentation timestamp in time_base units (time when frame should be shown to user).
#define AV_LOG_VERBOSE
Detailed information.
int buffer_size
The size of the buffer to which the ratecontrol is applied, in bits.
char * stats_out
pass1 encoding statistics output buffer
attribute_deprecated uint64_t error[AV_NUM_DATA_POINTERS]
attribute_deprecated int frame_skip_threshold
#define AV_PKT_FLAG_KEY
The packet contains a keyframe.
planar YUV 4:4:4, 30bpp, (1 Cr & Cb sample per 1x1 Y samples), little-endian
#define AV_LOG_ERROR
Something went wrong and cannot losslessly be recovered.
static int storeframe(AVCodecContext *avctx, struct FrameListData *cx_frame, AVPacket *pkt)
Store coded frame information in format suitable for return from encode2().
int qmax
maximum quantizer
#define NULL_IF_CONFIG_SMALL(x)
Return NULL if CONFIG_SMALL is true, otherwise the argument without modification. ...
#define AV_LOG_DEBUG
Stuff which is only useful for libav* developers.
planar YUV 4:2:2, 20bpp, (1 Cr & Cb sample per 2x1 Y samples), little-endian
int flags
AV_CODEC_FLAG_*.
Round to nearest and halfway cases away from zero.
simple assert() macros that are a bit more flexible than ISO C assert().
const char * name
Name of the codec implementation.
char * av_base64_encode(char *out, int out_size, const uint8_t *in, int in_size)
Encode data to base64 and null-terminate.
int flags
A combination of AV_PKT_FLAG values.
planar YUV 4:2:2, 16bpp, (1 Cr & Cb sample per 2x1 Y samples)
int rc_buffer_size
decoder bitstream buffer size
int64_t rc_min_rate
minimum bitrate
common internal API header
static void cx_pktcpy(struct FrameListData *dst, const struct vpx_codec_cx_pkt *src, const struct vpx_codec_cx_pkt *src_alpha, VP8Context *ctx)
static av_cold void dump_enc_cfg(AVCodecContext *avctx, const struct vpx_codec_enc_cfg *cfg)
enum AVPictureType pict_type
Picture type of the frame.
#define AV_BASE64_SIZE(x)
Calculate the output size needed to base64-encode x bytes to a null-terminated string.
int width
picture width / height.
#define FF_PROFILE_UNKNOWN
ITU-R BT2020 non-constant luminance system.
attribute_deprecated int noise_reduction
#define AV_CODEC_FLAG_PSNR
error[?] variables will be set during encoding.
#define AV_CODEC_FLAG_PASS1
Use internal 2pass ratecontrol in first pass mode.
static av_cold int vp8_free(AVCodecContext *avctx)
int ticks_per_frame
For some codecs, the time base is closer to the field rate than the frame rate.
planar YUV 4:2:0,18bpp, (1 Cr & Cb sample per 2x2 Y samples), little-endian
static av_cold void log_encoder_error(AVCodecContext *avctx, const char *desc)
attribute_deprecated int mb_threshold
int thread_count
thread count is used to decide how many independent tasks should be passed to execute() ...
struct FrameListData * next
the normal 2^n-1 "JPEG" YUV ranges
#define VP8F_ERROR_RESILIENT
Enable measures appropriate for streaming over lossy links.
static av_cold int vpx_init(AVCodecContext *avctx, const struct vpx_codec_iface *iface)
int flags
VP8 specific flags, see VP8F_* below.
This structure describes the bitrate properties of an encoded bitstream.
static const AVCodecDefault defaults[]
#define AV_LOG_INFO
Standard information.
Libavcodec external API header.
av_cold void ff_vp9_init_static(AVCodec *codec)
int linesize[AV_NUM_DATA_POINTERS]
For video, size in bytes of each picture line.
static av_cold int vp9_init(AVFormatContext *ctx, int st_index, PayloadContext *data)
main external API structure.
static int queue_frames(AVCodecContext *avctx, AVPacket *pkt_out)
Queue multiple output frames from the encoder, returning the front-most.
void av_packet_unref(AVPacket *pkt)
Wipe the packet.
int qmin
minimum quantizer
planar YUV 4:4:0,24bpp, (1 Cr & Cb sample per 1x2 Y samples), little-endian
Data found in BlockAdditional element of matroska container.
Describe the class of an AVClass context structure.
enum AVColorSpace colorspace
YUV colorspace type.
uint32_t flags
flags for this frame
int ff_alloc_packet2(AVCodecContext *avctx, AVPacket *avpkt, int64_t size, int64_t min_size)
Check AVPacket size and/or allocate data.
static av_cold void free_coded_frame(struct FrameListData *cx_frame)
float qcompress
amount of qscale change between easy & hard scenes (0.0-1.0)
static int set_pix_fmt(AVCodecContext *avctx, struct vpx_image *img)
static int vp8_encode(AVCodecContext *avctx, AVPacket *pkt, const AVFrame *frame, int *got_packet)
static enum AVPixelFormat pix_fmts[]
void * buf
compressed data buffer
int have_sse
true if we have pending sse[]
uint8_t * data[AV_NUM_DATA_POINTERS]
pointer to the picture/channel planes.
the normal 219*2^(n-8) "MPEG" YUV ranges
int gop_size
the number of pictures in a group of pictures, or 0 for intra_only
static const char *const ctlidstr[]
String mappings for enum vp8e_enc_control_id.
struct vpx_fixed_buf twopass_stats
static av_cold int vp8_init(AVFormatContext *s, int st_index, PayloadContext *vp8)
static const AVProfile profiles[]
planar YUV 4:2:0, 12bpp, (1 Cr & Cb sample per 2x2 Y samples)
planar YUV 4:2:2,24bpp, (1 Cr & Cb sample per 2x1 Y samples), little-endian
#define FF_DISABLE_DEPRECATION_WARNINGS
common internal api header.
common internal and external API header
int av_reallocp(void *ptr, size_t size)
Allocate or reallocate a block of memory.
attribute_deprecated AVFrame * coded_frame
the picture in the bitstream
AVCPBProperties * ff_add_cpb_side_data(AVCodecContext *avctx)
Add a CPB properties side data to an encoding context.
static av_cold void free_frame_list(struct FrameListData *list)
#define AV_CODEC_FLAG_PASS2
Use internal 2pass ratecontrol in second pass mode.
int slices
Number of slices.
#define FF_ENABLE_DEPRECATION_WARNINGS
Portion of struct vpx_codec_cx_pkt from vpx_encoder.h.
int avg_bitrate
Average bitrate of the stream, in bits per second.
int key_frame
1 -> keyframe, 0-> not
unsigned long duration
duration to show frame (in timebase units)
int av_base64_decode(uint8_t *out, const char *in_str, int out_size)
Decode a base64-encoded string.
int have_sse
true if we have pending sse[]
int64_t dts
Decompression timestamp in AVStream->time_base units; the time at which the packet is decompressed...
planar YUV 4:4:0 (1 Cr & Cb sample per 1x2 Y samples)
uint8_t * av_packet_new_side_data(AVPacket *pkt, enum AVPacketSideDataType type, int size)
Allocate new information of a packet.
const AVProfile ff_vp9_profiles[]
AVPixelFormat
Pixel format.
This structure stores compressed data.
struct FrameListData * coded_frame_list
int64_t pts
Presentation timestamp in AVStream->time_base units; the time at which the decompressed packet will b...
int64_t rc_max_rate
maximum bitrate
int keyint_min
minimum GOP size