blob: 083eb4ad43b2f9cb1b25f35f53ea2dd47dc1aae4
1 | /* |
2 | * Digital Speech Standard (DSS) demuxer |
3 | * Copyright (c) 2014 Oleksij Rempel <linux@rempel-privat.de> |
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 "libavutil/attributes.h" |
23 | #include "libavutil/bswap.h" |
24 | #include "libavutil/channel_layout.h" |
25 | #include "libavutil/intreadwrite.h" |
26 | |
27 | #include "avformat.h" |
28 | #include "internal.h" |
29 | |
30 | #define DSS_HEAD_OFFSET_AUTHOR 0xc |
31 | #define DSS_AUTHOR_SIZE 16 |
32 | |
33 | #define DSS_HEAD_OFFSET_START_TIME 0x26 |
34 | #define DSS_HEAD_OFFSET_END_TIME 0x32 |
35 | #define DSS_TIME_SIZE 12 |
36 | |
37 | #define DSS_HEAD_OFFSET_ACODEC 0x2a4 |
38 | #define DSS_ACODEC_DSS_SP 0x0 /* SP mode */ |
39 | #define DSS_ACODEC_G723_1 0x2 /* LP mode */ |
40 | |
41 | #define DSS_HEAD_OFFSET_COMMENT 0x31e |
42 | #define DSS_COMMENT_SIZE 64 |
43 | |
44 | #define DSS_BLOCK_SIZE 512 |
45 | #define DSS_AUDIO_BLOCK_HEADER_SIZE 6 |
46 | #define DSS_FRAME_SIZE 42 |
47 | |
48 | static const uint8_t frame_size[4] = { 24, 20, 4, 1 }; |
49 | |
50 | typedef struct DSSDemuxContext { |
51 | unsigned int audio_codec; |
52 | int counter; |
53 | int swap; |
54 | int dss_sp_swap_byte; |
55 | int8_t *dss_sp_buf; |
56 | |
57 | int packet_size; |
58 | int dss_header_size; |
59 | } DSSDemuxContext; |
60 | |
61 | static int dss_probe(AVProbeData *p) |
62 | { |
63 | if ( AV_RL32(p->buf) != MKTAG(0x2, 'd', 's', 's') |
64 | && AV_RL32(p->buf) != MKTAG(0x3, 'd', 's', 's')) |
65 | return 0; |
66 | |
67 | return AVPROBE_SCORE_MAX; |
68 | } |
69 | |
70 | static int dss_read_metadata_date(AVFormatContext *s, unsigned int offset, |
71 | const char *key) |
72 | { |
73 | AVIOContext *pb = s->pb; |
74 | char datetime[64], string[DSS_TIME_SIZE + 1] = { 0 }; |
75 | int y, month, d, h, minute, sec; |
76 | int ret; |
77 | |
78 | avio_seek(pb, offset, SEEK_SET); |
79 | |
80 | ret = avio_read(s->pb, string, DSS_TIME_SIZE); |
81 | if (ret < DSS_TIME_SIZE) |
82 | return ret < 0 ? ret : AVERROR_EOF; |
83 | |
84 | if (sscanf(string, "%2d%2d%2d%2d%2d%2d", &y, &month, &d, &h, &minute, &sec) != 6) |
85 | return AVERROR_INVALIDDATA; |
86 | /* We deal with a two-digit year here, so set the default date to 2000 |
87 | * and hope it will never be used in the next century. */ |
88 | snprintf(datetime, sizeof(datetime), "%.4d-%.2d-%.2dT%.2d:%.2d:%.2d", |
89 | y + 2000, month, d, h, minute, sec); |
90 | return av_dict_set(&s->metadata, key, datetime, 0); |
91 | } |
92 | |
93 | static int dss_read_metadata_string(AVFormatContext *s, unsigned int offset, |
94 | unsigned int size, const char *key) |
95 | { |
96 | AVIOContext *pb = s->pb; |
97 | char *value; |
98 | int ret; |
99 | |
100 | avio_seek(pb, offset, SEEK_SET); |
101 | |
102 | value = av_mallocz(size + 1); |
103 | if (!value) |
104 | return AVERROR(ENOMEM); |
105 | |
106 | ret = avio_read(s->pb, value, size); |
107 | if (ret < size) { |
108 | ret = ret < 0 ? ret : AVERROR_EOF; |
109 | goto exit; |
110 | } |
111 | |
112 | ret = av_dict_set(&s->metadata, key, value, 0); |
113 | |
114 | exit: |
115 | av_free(value); |
116 | return ret; |
117 | } |
118 | |
119 | static int dss_read_header(AVFormatContext *s) |
120 | { |
121 | DSSDemuxContext *ctx = s->priv_data; |
122 | AVIOContext *pb = s->pb; |
123 | AVStream *st; |
124 | int ret, version; |
125 | |
126 | st = avformat_new_stream(s, NULL); |
127 | if (!st) |
128 | return AVERROR(ENOMEM); |
129 | |
130 | version = avio_r8(pb); |
131 | ctx->dss_header_size = version * DSS_BLOCK_SIZE; |
132 | |
133 | ret = dss_read_metadata_string(s, DSS_HEAD_OFFSET_AUTHOR, |
134 | DSS_AUTHOR_SIZE, "author"); |
135 | if (ret) |
136 | return ret; |
137 | |
138 | ret = dss_read_metadata_date(s, DSS_HEAD_OFFSET_END_TIME, "date"); |
139 | if (ret) |
140 | return ret; |
141 | |
142 | ret = dss_read_metadata_string(s, DSS_HEAD_OFFSET_COMMENT, |
143 | DSS_COMMENT_SIZE, "comment"); |
144 | if (ret) |
145 | return ret; |
146 | |
147 | avio_seek(pb, DSS_HEAD_OFFSET_ACODEC, SEEK_SET); |
148 | ctx->audio_codec = avio_r8(pb); |
149 | |
150 | if (ctx->audio_codec == DSS_ACODEC_DSS_SP) { |
151 | st->codecpar->codec_id = AV_CODEC_ID_DSS_SP; |
152 | st->codecpar->sample_rate = 11025; |
153 | } else if (ctx->audio_codec == DSS_ACODEC_G723_1) { |
154 | st->codecpar->codec_id = AV_CODEC_ID_G723_1; |
155 | st->codecpar->sample_rate = 8000; |
156 | } else { |
157 | avpriv_request_sample(s, "Support for codec %x in DSS", |
158 | ctx->audio_codec); |
159 | return AVERROR_PATCHWELCOME; |
160 | } |
161 | |
162 | st->codecpar->codec_type = AVMEDIA_TYPE_AUDIO; |
163 | st->codecpar->channel_layout = AV_CH_LAYOUT_MONO; |
164 | st->codecpar->channels = 1; |
165 | |
166 | avpriv_set_pts_info(st, 64, 1, st->codecpar->sample_rate); |
167 | st->start_time = 0; |
168 | |
169 | /* Jump over header */ |
170 | |
171 | if (avio_seek(pb, ctx->dss_header_size, SEEK_SET) != ctx->dss_header_size) |
172 | return AVERROR(EIO); |
173 | |
174 | ctx->counter = 0; |
175 | ctx->swap = 0; |
176 | |
177 | ctx->dss_sp_buf = av_malloc(DSS_FRAME_SIZE + 1); |
178 | if (!ctx->dss_sp_buf) |
179 | return AVERROR(ENOMEM); |
180 | |
181 | return 0; |
182 | } |
183 | |
184 | static void dss_skip_audio_header(AVFormatContext *s, AVPacket *pkt) |
185 | { |
186 | DSSDemuxContext *ctx = s->priv_data; |
187 | AVIOContext *pb = s->pb; |
188 | |
189 | avio_skip(pb, DSS_AUDIO_BLOCK_HEADER_SIZE); |
190 | ctx->counter += DSS_BLOCK_SIZE - DSS_AUDIO_BLOCK_HEADER_SIZE; |
191 | } |
192 | |
193 | static void dss_sp_byte_swap(DSSDemuxContext *ctx, |
194 | uint8_t *dst, const uint8_t *src) |
195 | { |
196 | int i; |
197 | |
198 | if (ctx->swap) { |
199 | for (i = 3; i < DSS_FRAME_SIZE; i += 2) |
200 | dst[i] = src[i]; |
201 | |
202 | for (i = 0; i < DSS_FRAME_SIZE - 2; i += 2) |
203 | dst[i] = src[i + 4]; |
204 | |
205 | dst[1] = ctx->dss_sp_swap_byte; |
206 | } else { |
207 | memcpy(dst, src, DSS_FRAME_SIZE); |
208 | ctx->dss_sp_swap_byte = src[DSS_FRAME_SIZE - 2]; |
209 | } |
210 | |
211 | /* make sure byte 40 is always 0 */ |
212 | dst[DSS_FRAME_SIZE - 2] = 0; |
213 | ctx->swap ^= 1; |
214 | } |
215 | |
216 | static int dss_sp_read_packet(AVFormatContext *s, AVPacket *pkt) |
217 | { |
218 | DSSDemuxContext *ctx = s->priv_data; |
219 | AVStream *st = s->streams[0]; |
220 | int read_size, ret, offset = 0, buff_offset = 0; |
221 | int64_t pos = avio_tell(s->pb); |
222 | |
223 | if (ctx->counter == 0) |
224 | dss_skip_audio_header(s, pkt); |
225 | |
226 | if (ctx->swap) { |
227 | read_size = DSS_FRAME_SIZE - 2; |
228 | buff_offset = 3; |
229 | } else |
230 | read_size = DSS_FRAME_SIZE; |
231 | |
232 | ctx->counter -= read_size; |
233 | ctx->packet_size = DSS_FRAME_SIZE - 1; |
234 | |
235 | ret = av_new_packet(pkt, DSS_FRAME_SIZE); |
236 | if (ret < 0) |
237 | return ret; |
238 | |
239 | pkt->duration = 264; |
240 | pkt->pos = pos; |
241 | pkt->stream_index = 0; |
242 | s->bit_rate = 8LL * ctx->packet_size * st->codecpar->sample_rate * 512 / (506 * pkt->duration); |
243 | |
244 | if (ctx->counter < 0) { |
245 | int size2 = ctx->counter + read_size; |
246 | |
247 | ret = avio_read(s->pb, ctx->dss_sp_buf + offset + buff_offset, |
248 | size2 - offset); |
249 | if (ret < size2 - offset) |
250 | goto error_eof; |
251 | |
252 | dss_skip_audio_header(s, pkt); |
253 | offset = size2; |
254 | } |
255 | |
256 | ret = avio_read(s->pb, ctx->dss_sp_buf + offset + buff_offset, |
257 | read_size - offset); |
258 | if (ret < read_size - offset) |
259 | goto error_eof; |
260 | |
261 | dss_sp_byte_swap(ctx, pkt->data, ctx->dss_sp_buf); |
262 | |
263 | if (ctx->dss_sp_swap_byte < 0) { |
264 | ret = AVERROR(EAGAIN); |
265 | goto error_eof; |
266 | } |
267 | |
268 | return pkt->size; |
269 | |
270 | error_eof: |
271 | av_packet_unref(pkt); |
272 | return ret < 0 ? ret : AVERROR_EOF; |
273 | } |
274 | |
275 | static int dss_723_1_read_packet(AVFormatContext *s, AVPacket *pkt) |
276 | { |
277 | DSSDemuxContext *ctx = s->priv_data; |
278 | AVStream *st = s->streams[0]; |
279 | int size, byte, ret, offset; |
280 | int64_t pos = avio_tell(s->pb); |
281 | |
282 | if (ctx->counter == 0) |
283 | dss_skip_audio_header(s, pkt); |
284 | |
285 | /* We make one byte-step here. Don't forget to add offset. */ |
286 | byte = avio_r8(s->pb); |
287 | if (byte == 0xff) |
288 | return AVERROR_INVALIDDATA; |
289 | |
290 | size = frame_size[byte & 3]; |
291 | |
292 | ctx->packet_size = size; |
293 | ctx->counter -= size; |
294 | |
295 | ret = av_new_packet(pkt, size); |
296 | if (ret < 0) |
297 | return ret; |
298 | pkt->pos = pos; |
299 | |
300 | pkt->data[0] = byte; |
301 | offset = 1; |
302 | pkt->duration = 240; |
303 | s->bit_rate = 8LL * size * st->codecpar->sample_rate * 512 / (506 * pkt->duration); |
304 | |
305 | pkt->stream_index = 0; |
306 | |
307 | if (ctx->counter < 0) { |
308 | int size2 = ctx->counter + size; |
309 | |
310 | ret = avio_read(s->pb, pkt->data + offset, |
311 | size2 - offset); |
312 | if (ret < size2 - offset) { |
313 | av_packet_unref(pkt); |
314 | return ret < 0 ? ret : AVERROR_EOF; |
315 | } |
316 | |
317 | dss_skip_audio_header(s, pkt); |
318 | offset = size2; |
319 | } |
320 | |
321 | ret = avio_read(s->pb, pkt->data + offset, size - offset); |
322 | if (ret < size - offset) { |
323 | av_packet_unref(pkt); |
324 | return ret < 0 ? ret : AVERROR_EOF; |
325 | } |
326 | |
327 | return pkt->size; |
328 | } |
329 | |
330 | static int dss_read_packet(AVFormatContext *s, AVPacket *pkt) |
331 | { |
332 | DSSDemuxContext *ctx = s->priv_data; |
333 | |
334 | if (ctx->audio_codec == DSS_ACODEC_DSS_SP) |
335 | return dss_sp_read_packet(s, pkt); |
336 | else |
337 | return dss_723_1_read_packet(s, pkt); |
338 | } |
339 | |
340 | static int dss_read_close(AVFormatContext *s) |
341 | { |
342 | DSSDemuxContext *ctx = s->priv_data; |
343 | |
344 | av_freep(&ctx->dss_sp_buf); |
345 | |
346 | return 0; |
347 | } |
348 | |
349 | static int dss_read_seek(AVFormatContext *s, int stream_index, |
350 | int64_t timestamp, int flags) |
351 | { |
352 | DSSDemuxContext *ctx = s->priv_data; |
353 | int64_t ret, seekto; |
354 | uint8_t header[DSS_AUDIO_BLOCK_HEADER_SIZE]; |
355 | int offset; |
356 | |
357 | if (ctx->audio_codec == DSS_ACODEC_DSS_SP) |
358 | seekto = timestamp / 264 * 41 / 506 * 512; |
359 | else |
360 | seekto = timestamp / 240 * ctx->packet_size / 506 * 512; |
361 | |
362 | if (seekto < 0) |
363 | seekto = 0; |
364 | |
365 | seekto += ctx->dss_header_size; |
366 | |
367 | ret = avio_seek(s->pb, seekto, SEEK_SET); |
368 | if (ret < 0) |
369 | return ret; |
370 | |
371 | avio_read(s->pb, header, DSS_AUDIO_BLOCK_HEADER_SIZE); |
372 | ctx->swap = !!(header[0] & 0x80); |
373 | offset = 2*header[1] + 2*ctx->swap; |
374 | if (offset < DSS_AUDIO_BLOCK_HEADER_SIZE) |
375 | return AVERROR_INVALIDDATA; |
376 | if (offset == DSS_AUDIO_BLOCK_HEADER_SIZE) { |
377 | ctx->counter = 0; |
378 | offset = avio_skip(s->pb, -DSS_AUDIO_BLOCK_HEADER_SIZE); |
379 | } else { |
380 | ctx->counter = DSS_BLOCK_SIZE - offset; |
381 | offset = avio_skip(s->pb, offset - DSS_AUDIO_BLOCK_HEADER_SIZE); |
382 | } |
383 | ctx->dss_sp_swap_byte = -1; |
384 | return 0; |
385 | } |
386 | |
387 | |
388 | AVInputFormat ff_dss_demuxer = { |
389 | .name = "dss", |
390 | .long_name = NULL_IF_CONFIG_SMALL("Digital Speech Standard (DSS)"), |
391 | .priv_data_size = sizeof(DSSDemuxContext), |
392 | .read_probe = dss_probe, |
393 | .read_header = dss_read_header, |
394 | .read_packet = dss_read_packet, |
395 | .read_close = dss_read_close, |
396 | .read_seek = dss_read_seek, |
397 | .extensions = "dss" |
398 | }; |
399 |