2 * 3GPP TS 26.245 Timed Text decoder
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/avstring.h"
25 #include "libavutil/common.h"
26 #include "libavutil/bprint.h"
27 #include "libavutil/intreadwrite.h"
28 #include "libavutil/mem.h"
30 #define STYLE_FLAG_BOLD 1
31 #define STYLE_FLAG_ITALIC 2
32 #define STYLE_FLAG_UNDERLINE 4
34 static int text_to_ass(AVBPrint *buf, const char *text, const char *text_end,
35 char **style_start, char **style_end,
36 uint8_t **style_flags, int style_entries)
39 while (text < text_end) {
40 for (i = 0; i < style_entries; i++) {
41 if (*style_flags[i] && text == style_start[i]) {
42 if (*style_flags[i] & STYLE_FLAG_BOLD)
43 av_bprintf(buf, "{\\b1}");
44 if (*style_flags[i] & STYLE_FLAG_ITALIC)
45 av_bprintf(buf, "{\\i1}");
46 if (*style_flags[i] & STYLE_FLAG_UNDERLINE)
47 av_bprintf(buf, "{\\u1}");
55 av_bprintf(buf, "\\N");
58 av_bprint_chars(buf, *text, 1);
62 for (i = 0; i < style_entries; i++) {
63 if (*style_flags[i] && text == style_end[i]) {
64 if (*style_flags[i] & STYLE_FLAG_BOLD)
65 av_bprintf(buf, "{\\b0}");
66 if (*style_flags[i] & STYLE_FLAG_ITALIC)
67 av_bprintf(buf, "{\\i0}");
68 if (*style_flags[i] & STYLE_FLAG_UNDERLINE)
69 av_bprintf(buf, "{\\u0}");
78 static int mov_text_init(AVCodecContext *avctx) {
80 * TODO: Handle the default text style.
81 * NB: Most players ignore styles completely, with the result that
82 * it's very common to find files where the default style is broken
83 * and respecting it results in a worse experience than ignoring it.
85 return ff_ass_subtitle_header_default(avctx);
88 static int mov_text_decode_frame(AVCodecContext *avctx,
89 void *data, int *got_sub_ptr, AVPacket *avpkt)
91 AVSubtitle *sub = data;
92 int ret, ts_start, ts_end;
94 char *ptr = avpkt->data;
97 int text_length, tsmb_type, style_entries;
98 uint64_t tsmb_size, tracksize;
99 char **style_start = { 0, };
100 char **style_end = { 0, };
101 uint8_t **style_flags = { 0, };
103 int index, i, size_var;
107 if (!ptr || avpkt->size < 2)
108 return AVERROR_INVALIDDATA;
111 * A packet of size two with value zero is an empty subtitle
112 * used to mark the end of the previous non-empty subtitle.
113 * We can just drop them here as we have duration information
114 * already. If the value is non-zero, then it's technically a
117 if (avpkt->size == 2)
118 return AV_RB16(ptr) == 0 ? 0 : AVERROR_INVALIDDATA;
121 * The first two bytes of the packet are the length of the text string
122 * In complex cases, there are style descriptors appended to the string
123 * so we can't just assume the packet size is the string size.
125 text_length = AV_RB16(ptr);
126 end = ptr + FFMIN(2 + text_length, avpkt->size);
129 ts_start = av_rescale_q(avpkt->pts,
131 (AVRational){1,100});
132 ts_end = av_rescale_q(avpkt->pts + avpkt->duration,
134 (AVRational){1,100});
137 tracksize = 2 + text_length;
138 // Note that the spec recommends lines be no longer than 2048 characters.
139 av_bprint_init(&buf, 0, AV_BPRINT_SIZE_UNLIMITED);
140 if (text_length + 2 != avpkt->size) {
141 while (tracksize + 8 <= avpkt->size) {
142 // A box is a minimum of 8 bytes.
143 tsmb = ptr + tracksize - 2;
144 tsmb_size = AV_RB32(tsmb);
146 tsmb_type = AV_RB32(tsmb);
149 if (tsmb_size == 1) {
150 if (tracksize + 16 > avpkt->size)
152 tsmb_size = AV_RB64(tsmb);
157 //size_var is equal to 10 or 18 depending on the size of box
159 if (tracksize + tsmb_size > avpkt->size)
162 if (tsmb_type == MKBETAG('s','t','y','l')) {
163 if (tracksize + size_var > avpkt->size)
165 style_entries = AV_RB16(tsmb);
168 // A single style record is of length 12 bytes.
169 if (tracksize + size_var + style_entries * 12 > avpkt->size)
172 for(i = 0; i < style_entries; i++) {
173 style_pos = ptr + AV_RB16(tsmb);
175 av_dynarray_add(&style_start, &index, style_pos);
177 style_pos = ptr + AV_RB16(tsmb);
179 av_dynarray_add(&style_end, &index, style_pos);
181 // fontID = AV_RB16(tsmb);
185 return AVERROR(ENOMEM);
186 *flag = AV_RB8(tsmb);
188 av_dynarray_add(&style_flags, &index, flag);
189 //fontsize=AV_RB8(tsmb);
194 text_to_ass(&buf, ptr, end, style_start, style_end, style_flags, style_entries);
196 for(i = 0; i < style_entries; i++) {
197 av_freep(&style_flags[i]);
199 av_freep(&style_start);
200 av_freep(&style_end);
201 av_freep(&style_flags);
203 tracksize = tracksize + tsmb_size;
206 text_to_ass(&buf, ptr, end, NULL, NULL, 0, 0);
208 ret = ff_ass_add_rect_bprint(sub, &buf, ts_start, ts_end - ts_start);
209 av_bprint_finalize(&buf, NULL);
212 *got_sub_ptr = sub->num_rects > 0;
216 AVCodec ff_movtext_decoder = {
218 .long_name = NULL_IF_CONFIG_SMALL("3GPP Timed Text subtitle"),
219 .type = AVMEDIA_TYPE_SUBTITLE,
220 .id = AV_CODEC_ID_MOV_TEXT,
221 .init = mov_text_init,
222 .decode = mov_text_decode_frame,