2 * 3GPP TS 26.245 Timed Text encoder
3 * Copyright (c) 2012 Philip Langdale <philipl@overt.org>
5 * This file is part of FFmpeg.
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.
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.
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
24 #include "libavutil/avassert.h"
25 #include "libavutil/avstring.h"
26 #include "libavutil/intreadwrite.h"
27 #include "libavutil/mem.h"
28 #include "libavutil/common.h"
29 #include "ass_split.h"
32 #define STYLE_FLAG_BOLD (1<<0)
33 #define STYLE_FLAG_ITALIC (1<<1)
34 #define STYLE_FLAG_UNDERLINE (1<<2)
35 #define STYLE_RECORD_SIZE 12
38 #define av_bprint_append_any(buf, data, size) av_bprint_append_data(buf, ((const char*)data), size)
47 ASSSplitContext *ass_ctx;
49 StyleBox **style_attributes;
50 StyleBox *style_attributes_temp;
52 uint8_t style_box_flag;
55 uint16_t style_entries;
56 uint16_t style_fontID;
57 uint8_t style_fontsize;
63 static av_cold int mov_text_encode_init(AVCodecContext *avctx)
66 * For now, we'll use a fixed default style. When we add styling
67 * support, this will be generated from the ASS style.
69 static const uint8_t text_sample_entry[] = {
70 0x00, 0x00, 0x00, 0x00, // uint32_t displayFlags
71 0x01, // int8_t horizontal-justification
72 0xFF, // int8_t vertical-justification
73 0x00, 0x00, 0x00, 0x00, // uint8_t background-color-rgba[4]
75 0x00, 0x00, // int16_t top
76 0x00, 0x00, // int16_t left
77 0x00, 0x00, // int16_t bottom
78 0x00, 0x00, // int16_t right
81 0x00, 0x00, // uint16_t startChar
82 0x00, 0x00, // uint16_t endChar
83 0x00, 0x01, // uint16_t font-ID
84 0x00, // uint8_t face-style-flags
85 0x12, // uint8_t font-size
86 0xFF, 0xFF, 0xFF, 0xFF, // uint8_t text-color-rgba[4]
89 0x00, 0x00, 0x00, 0x12, // uint32_t size
90 'f', 't', 'a', 'b', // uint8_t name[4]
91 0x00, 0x01, // uint16_t entry-count
93 0x00, 0x01, // uint16_t font-ID
94 0x05, // uint8_t font-name-length
95 'S', 'e', 'r', 'i', 'f',// uint8_t font[font-name-length]
100 MovTextContext *s = avctx->priv_data;
102 avctx->extradata_size = sizeof text_sample_entry;
103 avctx->extradata = av_mallocz(avctx->extradata_size + FF_INPUT_BUFFER_PADDING_SIZE);
104 if (!avctx->extradata)
105 return AVERROR(ENOMEM);
107 av_bprint_init(&s->buffer, 0, AV_BPRINT_SIZE_UNLIMITED);
109 memcpy(avctx->extradata, text_sample_entry, avctx->extradata_size);
111 s->ass_ctx = ff_ass_split(avctx->subtitle_header);
112 return s->ass_ctx ? 0 : AVERROR_INVALIDDATA;
115 static void mov_text_style_cb(void *priv, const char style, int close)
117 MovTextContext *s = priv;
119 if (s->style_box_flag == 0) { //first style entry
121 s->style_attributes_temp = av_malloc(sizeof(*s->style_attributes_temp));
123 if (!s->style_attributes_temp) {
124 av_bprint_clear(&s->buffer);
125 s->style_box_flag = 0;
129 s->style_attributes_temp->style_flag = 0;
130 s->style_attributes_temp->style_start = AV_RB16(&s->text_pos);
132 if (s->style_attributes_temp->style_flag) { //break the style record here and start a new one
133 s->style_attributes_temp->style_end = AV_RB16(&s->text_pos);
134 av_dynarray_add(&s->style_attributes, &s->count, s->style_attributes_temp);
136 s->style_attributes_temp = av_malloc(sizeof(*s->style_attributes_temp));
138 if (!s->style_attributes_temp) {
139 av_bprint_clear(&s->buffer);
140 s->style_box_flag = 0;
144 s->style_attributes_temp->style_flag = s->style_attributes[s->count - 1]->style_flag;
145 s->style_attributes_temp->style_start = AV_RB16(&s->text_pos);
147 s->style_attributes_temp->style_flag = 0;
148 s->style_attributes_temp->style_start = AV_RB16(&s->text_pos);
153 s->style_attributes_temp->style_flag |= STYLE_FLAG_BOLD;
156 s->style_attributes_temp->style_flag |= STYLE_FLAG_ITALIC;
159 s->style_attributes_temp->style_flag |= STYLE_FLAG_UNDERLINE;
163 s->style_attributes_temp->style_end = AV_RB16(&s->text_pos);
164 av_dynarray_add(&s->style_attributes, &s->count, s->style_attributes_temp);
166 s->style_attributes_temp = av_malloc(sizeof(*s->style_attributes_temp));
168 if (!s->style_attributes_temp) {
169 av_bprint_clear(&s->buffer);
170 s->style_box_flag = 0;
174 s->style_attributes_temp->style_flag = s->style_attributes[s->count - 1]->style_flag;
177 s->style_attributes_temp->style_flag &= ~STYLE_FLAG_BOLD;
180 s->style_attributes_temp->style_flag &= ~STYLE_FLAG_ITALIC;
183 s->style_attributes_temp->style_flag &= ~STYLE_FLAG_UNDERLINE;
186 if (s->style_attributes_temp->style_flag) { //start of new style record
187 s->style_attributes_temp->style_start = AV_RB16(&s->text_pos);
190 s->style_box_flag = 1;
193 static void mov_text_text_cb(void *priv, const char *text, int len)
195 MovTextContext *s = priv;
196 av_bprint_append_data(&s->buffer, text, len);
200 static void mov_text_new_line_cb(void *priv, int forced)
202 MovTextContext *s = priv;
203 av_bprint_append_data(&s->buffer, "\n", 1);
207 static const ASSCodesCallbacks mov_text_callbacks = {
208 .text = mov_text_text_cb,
209 .new_line = mov_text_new_line_cb,
210 .style = mov_text_style_cb,
213 static int mov_text_encode_frame(AVCodecContext *avctx, unsigned char *buf,
214 int bufsize, const AVSubtitle *sub)
216 MovTextContext *s = avctx->priv_data;
218 int i, j, num, length;
222 s->style_box_flag = 0;
223 s->style_entries = 0;
225 for (i = 0; i < sub->num_rects; i++) {
227 if (sub->rects[i]->type != SUBTITLE_ASS) {
228 av_log(avctx, AV_LOG_ERROR, "Only SUBTITLE_ASS type supported.\n");
229 return AVERROR(ENOSYS);
232 dialog = ff_ass_split_dialog(s->ass_ctx, sub->rects[i]->ass, 0, &num);
233 for (; dialog && num--; dialog++) {
234 ff_ass_split_override_codes(&mov_text_callbacks, s, dialog->text);
236 if (s->style_box_flag) {
237 s->tsmb_size = s->count * STYLE_RECORD_SIZE + SIZE_ADD; //size of one style record is 12 bytes
238 s->tsmb_size = AV_RB32(&s->tsmb_size);
239 s->tsmb_type = MKTAG('s','t','y','l');
240 s->style_entries = AV_RB16(&s->count);
241 s->style_fontID = 0x00 | 0x01<<8;
242 s->style_fontsize = 0x12;
243 s->style_color = MKTAG(0xFF, 0xFF, 0xFF, 0xFF);
244 /*The above three attributes are hard coded for now
245 but will come from ASS style in the future*/
246 av_bprint_append_any(&s->buffer, &s->tsmb_size, 4);
247 av_bprint_append_any(&s->buffer, &s->tsmb_type, 4);
248 av_bprint_append_any(&s->buffer, &s->style_entries, 2);
249 for (j = 0; j < s->count; j++) {
250 av_bprint_append_any(&s->buffer, &s->style_attributes[j]->style_start, 2);
251 av_bprint_append_any(&s->buffer, &s->style_attributes[j]->style_end, 2);
252 av_bprint_append_any(&s->buffer, &s->style_fontID, 2);
253 av_bprint_append_any(&s->buffer, &s->style_attributes[j]->style_flag, 1);
254 av_bprint_append_any(&s->buffer, &s->style_fontsize, 1);
255 av_bprint_append_any(&s->buffer, &s->style_color, 4);
257 for (j = 0; j < s->count; j++) {
258 av_freep(&s->style_attributes[j]);
260 av_freep(&s->style_attributes);
261 av_freep(&s->style_attributes_temp);
265 AV_WB16(buf, s->text_pos);
268 if (!av_bprint_is_complete(&s->buffer)) {
269 length = AVERROR(ENOMEM);
273 if (!s->buffer.len) {
278 if (s->buffer.len > bufsize - 3) {
279 av_log(avctx, AV_LOG_ERROR, "Buffer too small for ASS event.\n");
280 length = AVERROR(EINVAL);
284 memcpy(buf, s->buffer.str, s->buffer.len);
285 length = s->buffer.len + 2;
288 av_bprint_clear(&s->buffer);
292 static int mov_text_encode_close(AVCodecContext *avctx)
294 MovTextContext *s = avctx->priv_data;
295 ff_ass_split_free(s->ass_ctx);
296 av_bprint_finalize(&s->buffer, NULL);
300 AVCodec ff_movtext_encoder = {
302 .long_name = NULL_IF_CONFIG_SMALL("3GPP Timed Text subtitle"),
303 .type = AVMEDIA_TYPE_SUBTITLE,
304 .id = AV_CODEC_ID_MOV_TEXT,
305 .priv_data_size = sizeof(MovTextContext),
306 .init = mov_text_encode_init,
307 .encode_sub = mov_text_encode_frame,
308 .close = mov_text_encode_close,