Mercurial > libavcodec.hg
annotate pgssubdec.c @ 10553:86c7e3c6de00 libavcodec
Update libtheora wrapper to use the 1.0 API
author | conrad |
---|---|
date | Sun, 22 Nov 2009 21:08:37 +0000 |
parents | e85fbfe709d8 |
children | 8a4984c5cacc |
rev | line source |
---|---|
10083 | 1 /* |
2 * PGS subtitle decoder | |
3 * Copyright (c) 2009 Stephen Backway | |
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 /** | |
23 * @file libavcodec/pgssubdec.c | |
24 * PGS subtitle decoder | |
25 */ | |
26 | |
27 #include "avcodec.h" | |
28 #include "dsputil.h" | |
29 #include "colorspace.h" | |
30 #include "bytestream.h" | |
31 | |
32 //#define DEBUG_PACKET_CONTENTS | |
33 | |
34 #define RGBA(r,g,b,a) (((a) << 24) | ((r) << 16) | ((g) << 8) | (b)) | |
35 | |
36 enum SegmentType { | |
37 PALETTE_SEGMENT = 0x14, | |
38 PICTURE_SEGMENT = 0x15, | |
39 PRESENTATION_SEGMENT = 0x16, | |
40 WINDOW_SEGMENT = 0x17, | |
41 DISPLAY_SEGMENT = 0x80, | |
42 }; | |
43 | |
44 typedef struct PGSSubPresentation { | |
45 int x; | |
46 int y; | |
47 int video_w; | |
48 int video_h; | |
49 int id_number; | |
50 } PGSSubPresentation; | |
51 | |
52 typedef struct PGSSubPicture { | |
53 int w; | |
54 int h; | |
55 uint8_t *rle; | |
56 unsigned int rle_buffer_size, rle_data_len; | |
57 } PGSSubPicture; | |
58 | |
59 typedef struct PGSSubContext { | |
60 PGSSubPresentation presentation; | |
61 uint32_t clut[256]; | |
62 PGSSubPicture picture; | |
63 } PGSSubContext; | |
64 | |
65 static av_cold int init_decoder(AVCodecContext *avctx) | |
66 { | |
67 avctx->pix_fmt = PIX_FMT_RGB32; | |
68 | |
69 return 0; | |
70 } | |
71 | |
72 static av_cold int close_decoder(AVCodecContext *avctx) | |
73 { | |
74 PGSSubContext *ctx = avctx->priv_data; | |
75 | |
76 av_freep(&ctx->picture.rle); | |
77 ctx->picture.rle_buffer_size = 0; | |
78 | |
79 return 0; | |
80 } | |
81 | |
82 /** | |
83 * Decodes the RLE data. | |
84 * | |
85 * The subtitle is stored as an Run Length Encoded image. | |
86 * | |
87 * @param avctx contains the current codec context | |
88 * @param sub pointer to the processed subtitle data | |
89 * @param buf pointer to the RLE data to process | |
90 * @param buf_size size of the RLE data to process | |
91 */ | |
92 static int decode_rle(AVCodecContext *avctx, AVSubtitle *sub, | |
93 const uint8_t *buf, unsigned int buf_size) | |
94 { | |
95 const uint8_t *rle_bitmap_end; | |
96 int pixel_count, line_count; | |
97 | |
98 rle_bitmap_end = buf + buf_size; | |
99 | |
100 sub->rects[0]->pict.data[0] = av_malloc(sub->rects[0]->w * sub->rects[0]->h); | |
101 | |
102 if (!sub->rects[0]->pict.data[0]) | |
103 return -1; | |
104 | |
105 pixel_count = 0; | |
106 line_count = 0; | |
107 | |
108 while (buf < rle_bitmap_end && line_count < sub->rects[0]->h) { | |
109 uint8_t flags, color; | |
110 int run; | |
111 | |
112 color = bytestream_get_byte(&buf); | |
113 run = 1; | |
114 | |
115 if (color == 0x00) { | |
116 flags = bytestream_get_byte(&buf); | |
117 run = flags & 0x3f; | |
118 if (flags & 0x40) | |
119 run = (run << 8) + bytestream_get_byte(&buf); | |
120 color = flags & 0x80 ? bytestream_get_byte(&buf) : 0; | |
121 } | |
122 | |
123 if (run > 0 && pixel_count + run <= sub->rects[0]->w * sub->rects[0]->h) { | |
124 memset(sub->rects[0]->pict.data[0] + pixel_count, color, run); | |
125 pixel_count += run; | |
126 } else if (!run) { | |
127 /* | |
128 * New Line. Check if correct pixels decoded, if not display warning | |
129 * and adjust bitmap pointer to correct new line position. | |
130 */ | |
131 if (pixel_count % sub->rects[0]->w > 0) | |
132 av_log(avctx, AV_LOG_ERROR, "Decoded %d pixels, when line should be %d pixels\n", | |
133 pixel_count % sub->rects[0]->w, sub->rects[0]->w); | |
134 line_count++; | |
135 } | |
136 } | |
137 | |
138 dprintf(avctx, "Pixel Count = %d, Area = %d\n", pixel_count, sub->rects[0]->w * sub->rects[0]->h); | |
139 | |
140 return 0; | |
141 } | |
142 | |
143 /** | |
144 * Parses the picture segment packet. | |
145 * | |
146 * The picture segment contains details on the sequence id, | |
147 * width, height and Run Length Encoded (RLE) bitmap data. | |
148 * | |
149 * @param avctx contains the current codec context | |
150 * @param buf pointer to the packet to process | |
151 * @param buf_size size of packet to process | |
152 * @todo TODO: Enable support for RLE data over multiple packets | |
153 */ | |
154 static int parse_picture_segment(AVCodecContext *avctx, | |
155 const uint8_t *buf, int buf_size) | |
156 { | |
157 PGSSubContext *ctx = avctx->priv_data; | |
158 | |
159 uint8_t sequence_desc; | |
160 unsigned int rle_bitmap_len, width, height; | |
161 | |
162 /* skip 3 unknown bytes: Object ID (2 bytes), Version Number */ | |
163 buf += 3; | |
164 | |
165 /* Read the Sequence Description to determine if start of RLE data or appended to previous RLE */ | |
166 sequence_desc = bytestream_get_byte(&buf); | |
167 | |
168 if (!(sequence_desc & 0x80)) { | |
169 av_log(avctx, AV_LOG_ERROR, "Decoder does not support object data over multiple packets.\n"); | |
170 return -1; | |
171 } | |
172 | |
173 /* Decode rle bitmap length */ | |
174 rle_bitmap_len = bytestream_get_be24(&buf); | |
175 | |
176 /* Check to ensure we have enough data for rle_bitmap_length if just a single packet */ | |
177 if (rle_bitmap_len > buf_size - 7) { | |
178 av_log(avctx, AV_LOG_ERROR, "Not enough RLE data for specified length of %d.\n", rle_bitmap_len); | |
179 return -1; | |
180 } | |
181 | |
182 ctx->picture.rle_data_len = rle_bitmap_len; | |
183 | |
184 /* Get bitmap dimensions from data */ | |
185 width = bytestream_get_be16(&buf); | |
186 height = bytestream_get_be16(&buf); | |
187 | |
188 /* Make sure the bitmap is not too large */ | |
189 if (ctx->presentation.video_w < width || ctx->presentation.video_h < height) { | |
190 av_log(avctx, AV_LOG_ERROR, "Bitmap dimensions larger then video.\n"); | |
191 return -1; | |
192 } | |
193 | |
194 ctx->picture.w = width; | |
195 ctx->picture.h = height; | |
196 | |
197 av_fast_malloc(&ctx->picture.rle, &ctx->picture.rle_buffer_size, rle_bitmap_len); | |
198 | |
199 if (!ctx->picture.rle) | |
200 return -1; | |
201 | |
202 memcpy(ctx->picture.rle, buf, rle_bitmap_len); | |
203 | |
204 return 0; | |
205 } | |
206 | |
207 /** | |
208 * Parses the palette segment packet. | |
209 * | |
210 * The palette segment contains details of the palette, | |
211 * a maximum of 256 colors can be defined. | |
212 * | |
213 * @param avctx contains the current codec context | |
214 * @param buf pointer to the packet to process | |
215 * @param buf_size size of packet to process | |
216 */ | |
217 static void parse_palette_segment(AVCodecContext *avctx, | |
218 const uint8_t *buf, int buf_size) | |
219 { | |
220 PGSSubContext *ctx = avctx->priv_data; | |
221 | |
222 const uint8_t *buf_end = buf + buf_size; | |
223 const uint8_t *cm = ff_cropTbl + MAX_NEG_CROP; | |
224 int color_id; | |
225 int y, cb, cr, alpha; | |
226 int r, g, b, r_add, g_add, b_add; | |
227 | |
228 /* Skip two null bytes */ | |
229 buf += 2; | |
230 | |
231 while (buf < buf_end) { | |
232 color_id = bytestream_get_byte(&buf); | |
233 y = bytestream_get_byte(&buf); | |
234 cb = bytestream_get_byte(&buf); | |
235 cr = bytestream_get_byte(&buf); | |
236 alpha = bytestream_get_byte(&buf); | |
237 | |
238 YUV_TO_RGB1(cb, cr); | |
239 YUV_TO_RGB2(r, g, b, y); | |
240 | |
241 dprintf(avctx, "Color %d := (%d,%d,%d,%d)\n", color_id, r, g, b, alpha); | |
242 | |
243 /* Store color in palette */ | |
244 ctx->clut[color_id] = RGBA(r,g,b,alpha); | |
245 } | |
246 } | |
247 | |
248 /** | |
249 * Parses the presentation segment packet. | |
250 * | |
251 * The presentation segment contains details on the video | |
252 * width, video height, x & y subtitle position. | |
253 * | |
254 * @param avctx contains the current codec context | |
255 * @param buf pointer to the packet to process | |
256 * @param buf_size size of packet to process | |
257 * @todo TODO: Implement cropping | |
258 * @todo TODO: Implement forcing of subtitles | |
259 * @todo TODO: Blanking of subtitle | |
260 */ | |
261 static void parse_presentation_segment(AVCodecContext *avctx, | |
262 const uint8_t *buf, int buf_size) | |
263 { | |
264 PGSSubContext *ctx = avctx->priv_data; | |
265 | |
266 int x, y; | |
267 uint8_t block; | |
268 | |
269 ctx->presentation.video_w = bytestream_get_be16(&buf); | |
270 ctx->presentation.video_h = bytestream_get_be16(&buf); | |
271 | |
272 dprintf(avctx, "Video Dimensions %dx%d\n", | |
273 ctx->presentation.video_w, ctx->presentation.video_h); | |
274 | |
275 /* Skip 1 bytes of unknown, frame rate? */ | |
276 buf++; | |
277 | |
278 ctx->presentation.id_number = bytestream_get_be16(&buf); | |
279 | |
280 /* Next byte is the state. */ | |
281 block = bytestream_get_byte(&buf);; | |
282 if (block == 0x80) { | |
283 /* | |
284 * Skip 7 bytes of unknown: | |
285 * palette_update_flag (0x80), | |
286 * palette_id_to_use, | |
287 * Object Number (if > 0 determines if more data to process), | |
288 * object_id_ref (2 bytes), | |
289 * window_id_ref, | |
290 * composition_flag (0x80 - object cropped, 0x40 - object forced) | |
291 */ | |
292 buf += 7; | |
293 | |
294 x = bytestream_get_be16(&buf); | |
295 y = bytestream_get_be16(&buf); | |
296 | |
297 /* TODO If cropping, cropping_x, cropping_y, cropping_width, cropping_height (all 2 bytes).*/ | |
298 | |
299 dprintf(avctx, "Subtitle Placement x=%d, y=%d\n", x, y); | |
300 | |
301 if (x > ctx->presentation.video_w || y > ctx->presentation.video_h) { | |
302 av_log(avctx, AV_LOG_ERROR, "Subtitle out of video bounds. x = %d, y = %d, video width = %d, video height = %d.\n", | |
303 x, y, ctx->presentation.video_w, ctx->presentation.video_h); | |
304 x = 0; y = 0; | |
305 } | |
306 | |
307 /* Fill in dimensions */ | |
308 ctx->presentation.x = x; | |
309 ctx->presentation.y = y; | |
310 } else if (block == 0x00) { | |
311 /* TODO: Blank context as subtitle should not be displayed. | |
312 * If the subtitle is blanked now the subtitle is not | |
313 * on screen long enough to read, due to a delay in | |
314 * initial display timing. | |
315 */ | |
316 } | |
317 } | |
318 | |
319 /** | |
320 * Parses the display segment packet. | |
321 * | |
322 * The display segment controls the updating of the display. | |
323 * | |
324 * @param avctx contains the current codec context | |
325 * @param data pointer to the data pertaining the subtitle to display | |
326 * @param buf pointer to the packet to process | |
327 * @param buf_size size of packet to process | |
328 * @todo TODO: Fix start time, relies on correct PTS, currently too late | |
329 * | |
330 * @todo TODO: Fix end time, normally cleared by a second display | |
331 * @todo segment, which is currently ignored as it clears | |
332 * @todo the subtitle too early. | |
333 */ | |
334 static int display_end_segment(AVCodecContext *avctx, void *data, | |
335 const uint8_t *buf, int buf_size) | |
336 { | |
337 AVSubtitle *sub = data; | |
338 PGSSubContext *ctx = avctx->priv_data; | |
339 | |
340 /* | |
341 * The end display time is a timeout value and is only reached | |
342 * if the next subtitle is later then timeout or subtitle has | |
343 * not been cleared by a subsequent empty display command. | |
344 */ | |
345 | |
10087 | 346 memset(sub, 0, sizeof(*sub)); |
10083 | 347 sub->start_display_time = 0; |
348 sub->end_display_time = 20000; | |
349 sub->format = 0; | |
350 | |
10089 | 351 sub->rects = av_mallocz(sizeof(*sub->rects)); |
352 sub->rects[0] = av_mallocz(sizeof(*sub->rects[0])); | |
353 sub->num_rects = 1; | |
10083 | 354 |
355 sub->rects[0]->x = ctx->presentation.x; | |
356 sub->rects[0]->y = ctx->presentation.y; | |
357 sub->rects[0]->w = ctx->picture.w; | |
358 sub->rects[0]->h = ctx->picture.h; | |
359 sub->rects[0]->type = SUBTITLE_BITMAP; | |
360 | |
361 /* Process bitmap */ | |
362 sub->rects[0]->pict.linesize[0] = ctx->picture.w; | |
363 | |
364 if (ctx->picture.rle) | |
365 if(decode_rle(avctx, sub, ctx->picture.rle, ctx->picture.rle_data_len) < 0) | |
366 return 0; | |
367 | |
368 /* Allocate memory for colors */ | |
369 sub->rects[0]->nb_colors = 256; | |
10092
e85fbfe709d8
Always allocate a buffer of AVPALETTE_SIZE for palette in the
cehoyos
parents:
10089
diff
changeset
|
370 sub->rects[0]->pict.data[1] = av_mallocz(AVPALETTE_SIZE); |
10083 | 371 |
372 memcpy(sub->rects[0]->pict.data[1], ctx->clut, sub->rects[0]->nb_colors * sizeof(uint32_t)); | |
373 | |
374 return 1; | |
375 } | |
376 | |
377 static int decode(AVCodecContext *avctx, void *data, int *data_size, | |
378 AVPacket *avpkt) | |
379 { | |
380 const uint8_t *buf = avpkt->data; | |
381 int buf_size = avpkt->size; | |
382 | |
383 const uint8_t *buf_end; | |
384 uint8_t segment_type; | |
385 int segment_length; | |
386 | |
387 #ifdef DEBUG_PACKET_CONTENTS | |
388 int i; | |
389 | |
390 av_log(avctx, AV_LOG_INFO, "PGS sub packet:\n"); | |
391 | |
392 for (i = 0; i < buf_size; i++) { | |
393 av_log(avctx, AV_LOG_INFO, "%02x ", buf[i]); | |
394 if (i % 16 == 15) | |
395 av_log(avctx, AV_LOG_INFO, "\n"); | |
396 } | |
397 | |
398 if (i & 15) | |
399 av_log(avctx, AV_LOG_INFO, "\n"); | |
400 #endif | |
401 | |
402 *data_size = 0; | |
403 | |
404 /* Ensure that we have received at a least a segment code and segment length */ | |
405 if (buf_size < 3) | |
406 return -1; | |
407 | |
408 buf_end = buf + buf_size; | |
409 | |
410 /* Step through buffer to identify segments */ | |
411 while (buf < buf_end) { | |
412 segment_type = bytestream_get_byte(&buf); | |
413 segment_length = bytestream_get_be16(&buf); | |
414 | |
415 dprintf(avctx, "Segment Length %d, Segment Type %x\n", segment_length, segment_type); | |
416 | |
417 if (segment_type != DISPLAY_SEGMENT && segment_length > buf_end - buf) | |
418 break; | |
419 | |
420 switch (segment_type) { | |
421 case PALETTE_SEGMENT: | |
422 parse_palette_segment(avctx, buf, segment_length); | |
423 break; | |
424 case PICTURE_SEGMENT: | |
425 parse_picture_segment(avctx, buf, segment_length); | |
426 break; | |
427 case PRESENTATION_SEGMENT: | |
428 parse_presentation_segment(avctx, buf, segment_length); | |
429 break; | |
430 case WINDOW_SEGMENT: | |
431 /* | |
432 * Window Segment Structure (No new information provided): | |
433 * 2 bytes: Unkown, | |
434 * 2 bytes: X position of subtitle, | |
435 * 2 bytes: Y position of subtitle, | |
436 * 2 bytes: Width of subtitle, | |
437 * 2 bytes: Height of subtitle. | |
438 */ | |
439 break; | |
440 case DISPLAY_SEGMENT: | |
441 *data_size = display_end_segment(avctx, data, buf, segment_length); | |
442 break; | |
443 default: | |
444 av_log(avctx, AV_LOG_ERROR, "Unknown subtitle segment type 0x%x, length %d\n", | |
445 segment_type, segment_length); | |
446 break; | |
447 } | |
448 | |
449 buf += segment_length; | |
450 } | |
451 | |
452 return buf_size; | |
453 } | |
454 | |
455 AVCodec pgssub_decoder = { | |
456 "pgssub", | |
457 CODEC_TYPE_SUBTITLE, | |
458 CODEC_ID_HDMV_PGS_SUBTITLE, | |
459 sizeof(PGSSubContext), | |
460 init_decoder, | |
461 NULL, | |
462 close_decoder, | |
463 decode, | |
464 .long_name = NULL_IF_CONFIG_SMALL("HDMV Presentation Graphic Stream subtitles"), | |
465 }; |