009cc97471de7d817ecc6a8e12438b405830338d
[libav.git] / libavcodec / x264.c
1 /*
2 * H.264 encoding using the x264 library
3 * Copyright (C) 2005 Mans Rullgard <mru@inprovide.com>
4 *
5 * This file is part of FFmpeg.
6 *
7 * FFmpeg is free software; you can redistribute it and/or
8 * modify it under the terms of the GNU Lesser General Public
9 * License as published by the Free Software Foundation; either
10 * version 2.1 of the License, or (at your option) any later version.
11 *
12 * FFmpeg is distributed in the hope that it will be useful,
13 * but WITHOUT ANY WARRANTY; without even the implied warranty of
14 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
15 * Lesser General Public License for more details.
16 *
17 * You should have received a copy of the GNU Lesser General Public
18 * License along with FFmpeg; if not, write to the Free Software
19 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
20 */
21
22 #include "avcodec.h"
23 #include <x264.h>
24 #include <math.h>
25 #include <stdio.h>
26 #include <stdlib.h>
27 #include <string.h>
28
29 typedef struct X264Context {
30 x264_param_t params;
31 x264_t *enc;
32 x264_picture_t pic;
33 AVFrame out_pic;
34 } X264Context;
35
36 static void
37 X264_log(void *p, int level, const char *fmt, va_list args)
38 {
39 static const int level_map[] = {
40 [X264_LOG_ERROR] = AV_LOG_ERROR,
41 [X264_LOG_WARNING] = AV_LOG_ERROR,
42 [X264_LOG_INFO] = AV_LOG_INFO,
43 [X264_LOG_DEBUG] = AV_LOG_DEBUG
44 };
45
46 if(level < 0 || level > X264_LOG_DEBUG)
47 return;
48
49 av_vlog(p, level_map[level], fmt, args);
50 }
51
52
53 static int
54 encode_nals(uint8_t *buf, int size, x264_nal_t *nals, int nnal)
55 {
56 uint8_t *p = buf;
57 int i;
58
59 for(i = 0; i < nnal; i++){
60 int s = x264_nal_encode(p, &size, 1, nals + i);
61 if(s < 0)
62 return -1;
63 p += s;
64 }
65
66 return p - buf;
67 }
68
69 extern int
70 X264_frame(AVCodecContext *ctx, uint8_t *buf, int bufsize, void *data)
71 {
72 X264Context *x4 = ctx->priv_data;
73 AVFrame *frame = data;
74 x264_nal_t *nal;
75 int nnal, i;
76 x264_picture_t pic_out;
77
78 x4->pic.img.i_csp = X264_CSP_I420;
79 x4->pic.img.i_plane = 3;
80
81 if (frame) {
82 for(i = 0; i < 3; i++){
83 x4->pic.img.plane[i] = frame->data[i];
84 x4->pic.img.i_stride[i] = frame->linesize[i];
85 }
86
87 x4->pic.i_pts = frame->pts;
88 x4->pic.i_type = X264_TYPE_AUTO;
89 }
90
91 if(x264_encoder_encode(x4->enc, &nal, &nnal, frame? &x4->pic: NULL,
92 &pic_out))
93 return -1;
94
95 bufsize = encode_nals(buf, bufsize, nal, nnal);
96 if(bufsize < 0)
97 return -1;
98
99 /* FIXME: dts */
100 x4->out_pic.pts = pic_out.i_pts;
101
102 switch(pic_out.i_type){
103 case X264_TYPE_IDR:
104 case X264_TYPE_I:
105 x4->out_pic.pict_type = FF_I_TYPE;
106 break;
107 case X264_TYPE_P:
108 x4->out_pic.pict_type = FF_P_TYPE;
109 break;
110 case X264_TYPE_B:
111 case X264_TYPE_BREF:
112 x4->out_pic.pict_type = FF_B_TYPE;
113 break;
114 }
115
116 x4->out_pic.key_frame = pic_out.i_type == X264_TYPE_IDR;
117 x4->out_pic.quality = (pic_out.i_qpplus1 - 1) * FF_QP2LAMBDA;
118
119 return bufsize;
120 }
121
122 static int
123 X264_close(AVCodecContext *avctx)
124 {
125 X264Context *x4 = avctx->priv_data;
126
127 if(x4->enc)
128 x264_encoder_close(x4->enc);
129
130 return 0;
131 }
132
133 extern int
134 X264_init(AVCodecContext *avctx)
135 {
136 X264Context *x4 = avctx->priv_data;
137
138 x264_param_default(&x4->params);
139
140 x4->params.pf_log = X264_log;
141 x4->params.p_log_private = avctx;
142
143 x4->params.i_keyint_max = avctx->gop_size;
144 x4->params.rc.i_bitrate = avctx->bit_rate / 1000;
145 x4->params.rc.i_vbv_buffer_size = avctx->rc_buffer_size / 1000;
146 x4->params.rc.i_vbv_max_bitrate = avctx->rc_max_rate / 1000;
147 x4->params.rc.b_stat_write = (avctx->flags & CODEC_FLAG_PASS1);
148 if(avctx->flags & CODEC_FLAG_PASS2) x4->params.rc.b_stat_read = 1;
149 else{
150 if(avctx->crf){
151 x4->params.rc.i_rc_method = X264_RC_CRF;
152 x4->params.rc.f_rf_constant = avctx->crf;
153 }else if(avctx->cqp > -1){
154 x4->params.rc.i_rc_method = X264_RC_CQP;
155 x4->params.rc.i_qp_constant = avctx->cqp;
156 }
157 }
158
159 // if neither crf nor cqp modes are selected we have to enable the RC
160 // we do it this way because we cannot check if the bitrate has been set
161 if(!(avctx->crf || (avctx->cqp > -1))) x4->params.rc.i_rc_method = X264_RC_ABR;
162
163 x4->params.i_bframe = avctx->max_b_frames;
164 x4->params.b_cabac = avctx->coder_type == FF_CODER_TYPE_AC;
165 x4->params.b_bframe_adaptive = avctx->b_frame_strategy;
166 x4->params.i_bframe_bias = avctx->bframebias;
167 x4->params.b_bframe_pyramid = (avctx->flags2 & CODEC_FLAG2_BPYRAMID);
168 avctx->has_b_frames= (avctx->flags2 & CODEC_FLAG2_BPYRAMID) ? 2 : !!avctx->max_b_frames;
169
170 x4->params.i_keyint_min = avctx->keyint_min;
171 if(x4->params.i_keyint_min > x4->params.i_keyint_max)
172 x4->params.i_keyint_min = x4->params.i_keyint_max;
173
174 x4->params.i_scenecut_threshold = avctx->scenechange_threshold;
175
176 x4->params.b_deblocking_filter = (avctx->flags & CODEC_FLAG_LOOP_FILTER);
177 x4->params.i_deblocking_filter_alphac0 = avctx->deblockalpha;
178 x4->params.i_deblocking_filter_beta = avctx->deblockbeta;
179
180 x4->params.rc.i_qp_min = avctx->qmin;
181 x4->params.rc.i_qp_max = avctx->qmax;
182 x4->params.rc.i_qp_step = avctx->max_qdiff;
183
184 x4->params.rc.f_qcompress = avctx->qcompress; /* 0.0 => cbr, 1.0 => constant qp */
185 x4->params.rc.f_qblur = avctx->qblur; /* temporally blur quants */
186 x4->params.rc.f_complexity_blur = avctx->complexityblur;
187
188 x4->params.i_frame_reference = avctx->refs;
189
190 x4->params.i_width = avctx->width;
191 x4->params.i_height = avctx->height;
192 x4->params.vui.i_sar_width = avctx->sample_aspect_ratio.num;
193 x4->params.vui.i_sar_height = avctx->sample_aspect_ratio.den;
194 x4->params.i_fps_num = avctx->time_base.den;
195 x4->params.i_fps_den = avctx->time_base.num;
196
197 x4->params.analyse.inter = 0;
198 if(avctx->partitions){
199 if(avctx->partitions & X264_PART_I4X4)
200 x4->params.analyse.inter |= X264_ANALYSE_I4x4;
201 if(avctx->partitions & X264_PART_I8X8)
202 x4->params.analyse.inter |= X264_ANALYSE_I8x8;
203 if(avctx->partitions & X264_PART_P8X8)
204 x4->params.analyse.inter |= X264_ANALYSE_PSUB16x16;
205 if(avctx->partitions & X264_PART_P4X4)
206 x4->params.analyse.inter |= X264_ANALYSE_PSUB8x8;
207 if(avctx->partitions & X264_PART_B8X8)
208 x4->params.analyse.inter |= X264_ANALYSE_BSUB16x16;
209 }
210
211 x4->params.analyse.i_direct_mv_pred = avctx->directpred;
212
213 x4->params.analyse.b_weighted_bipred = (avctx->flags2 & CODEC_FLAG2_WPRED);
214
215 if(avctx->me_method == ME_EPZS)
216 x4->params.analyse.i_me_method = X264_ME_DIA;
217 else if(avctx->me_method == ME_HEX)
218 x4->params.analyse.i_me_method = X264_ME_HEX;
219 else if(avctx->me_method == ME_UMH)
220 x4->params.analyse.i_me_method = X264_ME_UMH;
221 else if(avctx->me_method == ME_FULL)
222 x4->params.analyse.i_me_method = X264_ME_ESA;
223 else x4->params.analyse.i_me_method = X264_ME_HEX;
224
225 x4->params.analyse.i_me_range = avctx->me_range;
226 x4->params.analyse.i_subpel_refine = avctx->me_subpel_quality;
227
228 x4->params.analyse.b_bframe_rdo = (avctx->flags2 & CODEC_FLAG2_BRDO);
229 x4->params.analyse.b_mixed_references =
230 (avctx->flags2 & CODEC_FLAG2_MIXED_REFS);
231 x4->params.analyse.b_chroma_me = (avctx->me_cmp & FF_CMP_CHROMA);
232 x4->params.analyse.b_transform_8x8 = (avctx->flags2 & CODEC_FLAG2_8X8DCT);
233 x4->params.analyse.b_fast_pskip = (avctx->flags2 & CODEC_FLAG2_FASTPSKIP);
234
235 x4->params.analyse.i_trellis = avctx->trellis;
236 x4->params.analyse.i_noise_reduction = avctx->noise_reduction;
237
238 if(avctx->level > 0) x4->params.i_level_idc = avctx->level;
239
240 x4->params.rc.f_rate_tolerance =
241 (float)avctx->bit_rate_tolerance/avctx->bit_rate;
242
243 if((avctx->rc_buffer_size != 0) &&
244 (avctx->rc_initial_buffer_occupancy <= avctx->rc_buffer_size)){
245 x4->params.rc.f_vbv_buffer_init =
246 (float)avctx->rc_initial_buffer_occupancy/avctx->rc_buffer_size;
247 }
248 else x4->params.rc.f_vbv_buffer_init = 0.9;
249
250 x4->params.rc.f_ip_factor = 1/fabs(avctx->i_quant_factor);
251 x4->params.rc.f_pb_factor = avctx->b_quant_factor;
252 x4->params.analyse.i_chroma_qp_offset = avctx->chromaoffset;
253 x4->params.rc.psz_rc_eq = avctx->rc_eq;
254
255 x4->params.analyse.b_psnr = (avctx->flags & CODEC_FLAG_PSNR);
256 x4->params.i_log_level = X264_LOG_DEBUG;
257
258 x4->params.b_aud = (avctx->flags2 & CODEC_FLAG2_AUD);
259
260 x4->params.i_threads = avctx->thread_count;
261
262 if(avctx->flags & CODEC_FLAG_GLOBAL_HEADER){
263 x4->params.b_repeat_headers = 0;
264 }
265
266 x4->enc = x264_encoder_open(&x4->params);
267 if(!x4->enc)
268 return -1;
269
270 avctx->coded_frame = &x4->out_pic;
271
272 if(avctx->flags & CODEC_FLAG_GLOBAL_HEADER){
273 x264_nal_t *nal;
274 int nnal, i, s = 0;
275
276 x264_encoder_headers(x4->enc, &nal, &nnal);
277
278 /* 5 bytes NAL header + worst case escaping */
279 for(i = 0; i < nnal; i++)
280 s += 5 + nal[i].i_payload * 4 / 3;
281
282 avctx->extradata = av_malloc(s);
283 avctx->extradata_size = encode_nals(avctx->extradata, s, nal, nnal);
284 }
285
286 return 0;
287 }
288
289 AVCodec x264_encoder = {
290 .name = "h264",
291 .type = CODEC_TYPE_VIDEO,
292 .id = CODEC_ID_H264,
293 .priv_data_size = sizeof(X264Context),
294 .init = X264_init,
295 .encode = X264_frame,
296 .close = X264_close,
297 .capabilities = CODEC_CAP_DELAY,
298 .pix_fmts = (enum PixelFormat[]) { PIX_FMT_YUV420P, -1 }
299 };