ffmpeg/libavcodec/libx265.c

   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
  29 #include "libavutil/internal.h"
  30 #include "libavutil/common.h"
  31 #include "libavutil/opt.h"
  32 #include "libavutil/pixdesc.h"
  33 #include "avcodec.h"
  34 #include "internal.h"
  35
  36 typedef struct libx265Context {
  37     const AVClass *class;
  38
  39     x265_encoder *encoder;
  40     x265_param   *params;
  41
  42     char *preset;
  43     char *tune;
  44     char *x265_opts;
  45 } libx265Context;
  46
  47 static int is_keyframe(NalUnitType naltype)
  48 {
  49     switch (naltype) {
  50     case NAL_UNIT_CODED_SLICE_BLA_W_LP:
  51     case NAL_UNIT_CODED_SLICE_BLA_W_RADL:
  52     case NAL_UNIT_CODED_SLICE_BLA_N_LP:
  53     case NAL_UNIT_CODED_SLICE_IDR_W_RADL:
  54     case NAL_UNIT_CODED_SLICE_IDR_N_LP:
  55     case NAL_UNIT_CODED_SLICE_CRA:
  56         return 1;
  57     default:
  58         return 0;
  59     }
  60 }
  61
  62 static av_cold int libx265_encode_close(AVCodecContext *avctx)
  63 {
  64     libx265Context *ctx = avctx->priv_data;
  65
  66     av_frame_free(&avctx->coded_frame);
  67
  68     x265_param_free(ctx->params);
  69
  70     if (ctx->encoder)
  71         x265_encoder_close(ctx->encoder);
  72
  73     return 0;
  74 }
  75
  76 static av_cold int libx265_encode_init(AVCodecContext *avctx)
  77 {
  78     libx265Context *ctx = avctx->priv_data;
  79     x265_nal *nal;
  80     char sar[12];
  81     int sar_num, sar_den;
  82     int nnal;
  83
  84     if (avctx->strict_std_compliance > FF_COMPLIANCE_EXPERIMENTAL &&
  85         !av_pix_fmt_desc_get(avctx->pix_fmt)->log2_chroma_w) {
  86         av_log(avctx, AV_LOG_ERROR,
  87                "4:2:2 and 4:4:4 support is not fully defined for HEVC yet. "
  88                "Set -strict experimental to encode anyway.\n");
  89         return AVERROR(ENOSYS);
  90     }
  91
  92     avctx->coded_frame = av_frame_alloc();
  93     if (!avctx->coded_frame) {
  94         av_log(avctx, AV_LOG_ERROR, "Could not allocate frame.\n");
  95         return AVERROR(ENOMEM);
  96     }
  97
  98     ctx->params = x265_param_alloc();
  99     if (!ctx->params) {
 100         av_log(avctx, AV_LOG_ERROR, "Could not allocate x265 param structure.\n");
 101         return AVERROR(ENOMEM);
 102     }
 103
 104     if (x265_param_default_preset(ctx->params, ctx->preset, ctx->tune) < 0) {
 105         av_log(avctx, AV_LOG_ERROR, "Invalid preset or tune.\n");
 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 & CODEC_FLAG_PSNR);
 115
 116     if (avctx->sample_aspect_ratio.num > 0 && avctx->sample_aspect_ratio.den > 0) {
 117         av_reduce(&sar_num, &sar_den,
 118                   avctx->sample_aspect_ratio.num,
 119                   avctx->sample_aspect_ratio.den, 65535);
 120         snprintf(sar, sizeof(sar), "%d:%d", sar_num, sar_den);
 121         if (x265_param_parse(ctx->params, "sar", sar) == X265_PARAM_BAD_VALUE) {
 122             av_log(avctx, AV_LOG_ERROR, "Invalid SAR: %d:%d.\n", sar_num, sar_den);
 123             return AVERROR_INVALIDDATA;
 124         }
 125     }
 126
 127     switch (avctx->pix_fmt) {
 128     case AV_PIX_FMT_YUV420P:
 129     case AV_PIX_FMT_YUV420P10:
 130         ctx->params->internalCsp = X265_CSP_I420;
 131         break;
 132     case AV_PIX_FMT_YUV422P:
 133     case AV_PIX_FMT_YUV422P10:
 134         ctx->params->internalCsp = X265_CSP_I422;
 135         break;
 136     case AV_PIX_FMT_YUV444P:
 137     case AV_PIX_FMT_YUV444P10:
 138         ctx->params->internalCsp = X265_CSP_I444;
 139         break;
 140     }
 141
 142     if (avctx->bit_rate > 0) {
 143         ctx->params->rc.bitrate         = avctx->bit_rate / 1000;
 144         ctx->params->rc.rateControlMode = X265_RC_ABR;
 145     }
 146
 147     if (!(avctx->flags & CODEC_FLAG_GLOBAL_HEADER))
 148         ctx->params->bRepeatHeaders = 1;
 149
 150     if (ctx->x265_opts) {
 151         AVDictionary *dict    = NULL;
 152         AVDictionaryEntry *en = NULL;
 153
 154         if (!av_dict_parse_string(&dict, ctx->x265_opts, "=", ":", 0)) {
 155             while ((en = av_dict_get(dict, "", en, AV_DICT_IGNORE_SUFFIX))) {
 156                 int parse_ret = x265_param_parse(ctx->params, en->key, en->value);
 157
 158                 switch (parse_ret) {
 159                 case X265_PARAM_BAD_NAME:
 160                     av_log(avctx, AV_LOG_WARNING,
 161                           "Unknown option: %s.\n", en->key);
 162                     break;
 163                 case X265_PARAM_BAD_VALUE:
 164                     av_log(avctx, AV_LOG_WARNING,
 165                           "Invalid value for %s: %s.\n", en->key, en->value);
 166                     break;
 167                 default:
 168                     break;
 169                 }
 170             }
 171             av_dict_free(&dict);
 172         }
 173     }
 174
 175     ctx->encoder = x265_encoder_open(ctx->params);
 176     if (!ctx->encoder) {
 177         av_log(avctx, AV_LOG_ERROR, "Cannot open libx265 encoder.\n");
 178         libx265_encode_close(avctx);
 179         return AVERROR_INVALIDDATA;
 180     }
 181
 182     if (avctx->flags & CODEC_FLAG_GLOBAL_HEADER) {
 183         avctx->extradata_size = x265_encoder_headers(ctx->encoder, &nal, &nnal);
 184         if (avctx->extradata_size <= 0) {
 185             av_log(avctx, AV_LOG_ERROR, "Cannot encode headers.\n");
 186             libx265_encode_close(avctx);
 187             return AVERROR_INVALIDDATA;
 188         }
 189
 190         avctx->extradata = av_malloc(avctx->extradata_size + FF_INPUT_BUFFER_PADDING_SIZE);
 191         if (!avctx->extradata) {
 192             av_log(avctx, AV_LOG_ERROR,
 193                    "Cannot allocate HEVC header of size %d.\n", avctx->extradata_size);
 194             libx265_encode_close(avctx);
 195             return AVERROR(ENOMEM);
 196         }
 197
 198         memcpy(avctx->extradata, nal[0].payload, avctx->extradata_size);
 199     }
 200
 201     return 0;
 202 }
 203
 204 static int libx265_encode_frame(AVCodecContext *avctx, AVPacket *pkt,
 205                                 const AVFrame *pic, int *got_packet)
 206 {
 207     libx265Context *ctx = avctx->priv_data;
 208     x265_picture x265pic;
 209     x265_picture x265pic_out = { { 0 } };
 210     x265_nal *nal;
 211     uint8_t *dst;
 212     int payload = 0;
 213     int nnal;
 214     int ret;
 215     int i;
 216
 217     x265_picture_init(ctx->params, &x265pic);
 218
 219     if (pic) {
 220         for (i = 0; i < 3; i++) {
 221            x265pic.planes[i] = pic->data[i];
 222            x265pic.stride[i] = pic->linesize[i];
 223         }
 224
 225         x265pic.pts      = pic->pts;
 226         x265pic.bitDepth = av_pix_fmt_desc_get(avctx->pix_fmt)->comp[0].depth_minus1 + 1;
 227     }
 228
 229     ret = x265_encoder_encode(ctx->encoder, &nal, &nnal,
 230                               pic ? &x265pic : NULL, &x265pic_out);
 231     if (ret < 0)
 232         return AVERROR_EXTERNAL;
 233
 234     if (!nnal)
 235         return 0;
 236
 237     for (i = 0; i < nnal; i++)
 238         payload += nal[i].sizeBytes;
 239
 240     ret = ff_alloc_packet(pkt, payload);
 241     if (ret < 0) {
 242         av_log(avctx, AV_LOG_ERROR, "Error getting output packet.\n");
 243         return ret;
 244     }
 245     dst = pkt->data;
 246
 247     for (i = 0; i < nnal; i++) {
 248         memcpy(dst, nal[i].payload, nal[i].sizeBytes);
 249         dst += nal[i].sizeBytes;
 250
 251         if (is_keyframe(nal[i].type))
 252             pkt->flags |= AV_PKT_FLAG_KEY;
 253     }
 254
 255     pkt->pts = x265pic_out.pts;
 256     pkt->dts = x265pic_out.dts;
 257
 258     *got_packet = 1;
 259     return 0;
 260 }
 261
 262 static const enum AVPixelFormat x265_csp_eight[] = {
 263     AV_PIX_FMT_YUV420P,
 264     AV_PIX_FMT_YUV422P,
 265     AV_PIX_FMT_YUV444P,
 266     AV_PIX_FMT_NONE
 267 };
 268
 269 static const enum AVPixelFormat x265_csp_twelve[] = {
 270     AV_PIX_FMT_YUV420P,
 271     AV_PIX_FMT_YUV422P,
 272     AV_PIX_FMT_YUV444P,
 273     AV_PIX_FMT_YUV420P10,
 274     AV_PIX_FMT_YUV422P10,
 275     AV_PIX_FMT_YUV444P10,
 276     AV_PIX_FMT_NONE
 277 };
 278
 279 static av_cold void libx265_encode_init_csp(AVCodec *codec)
 280 {
 281     if (x265_max_bit_depth == 8)
 282         codec->pix_fmts = x265_csp_eight;
 283     else if (x265_max_bit_depth == 12)
 284         codec->pix_fmts = x265_csp_twelve;
 285 }
 286
 287 #define OFFSET(x) offsetof(libx265Context, x)
 288 #define VE AV_OPT_FLAG_VIDEO_PARAM | AV_OPT_FLAG_ENCODING_PARAM
 289 static const AVOption options[] = {
 290     { "preset",      "set the x265 preset",                                                         OFFSET(preset),    AV_OPT_TYPE_STRING, { 0 }, 0, 0, VE },
 291     { "tune",        "set the x265 tune parameter",                                                 OFFSET(tune),      AV_OPT_TYPE_STRING, { 0 }, 0, 0, VE },
 292     { "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 },
 293     { NULL }
 294 };
 295
 296 static const AVClass class = {
 297     .class_name = "libx265",
 298     .item_name  = av_default_item_name,
 299     .option     = options,
 300     .version    = LIBAVUTIL_VERSION_INT,
 301 };
 302
 303 static const AVCodecDefault x265_defaults[] = {
 304     { "b", "0" },
 305     { NULL },
 306 };
 307
 308 AVCodec ff_libx265_encoder = {
 309     .name             = "libx265",
 310     .long_name        = NULL_IF_CONFIG_SMALL("libx265 H.265 / HEVC"),
 311     .type             = AVMEDIA_TYPE_VIDEO,
 312     .id               = AV_CODEC_ID_HEVC,
 313     .init             = libx265_encode_init,
 314     .init_static_data = libx265_encode_init_csp,
 315     .encode2          = libx265_encode_frame,
 316     .close            = libx265_encode_close,
 317     .priv_data_size   = sizeof(libx265Context),
 318     .priv_class       = &class,
 319     .defaults         = x265_defaults,
 320     .capabilities     = CODEC_CAP_DELAY | CODEC_CAP_AUTO_THREADS,
 321 };