tools/sofa2wavs: fix build on Windows
[ffmpeg.git] / libavformat / dss.c
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/channel_layout.h"
23 #include "libavutil/intreadwrite.h"
24 #include "libavutil/mem.h"
25
26 #include "avformat.h"
27 #include "demux.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
56 int packet_size;
57 int dss_header_size;
58 } DSSDemuxContext;
59
60 static int dss_probe(const AVProbeData *p)
61 {
62 if ( AV_RL32(p->buf) != MKTAG(0x2, 'd', 's', 's')
63 && AV_RL32(p->buf) != MKTAG(0x3, 'd', 's', 's'))
64 return 0;
65
66 return AVPROBE_SCORE_MAX;
67 }
68
69 static int dss_read_metadata_date(AVFormatContext *s, unsigned int offset,
70 const char *key)
71 {
72 AVIOContext *pb = s->pb;
73 char datetime[64], string[DSS_TIME_SIZE + 1] = { 0 };
74 int y, month, d, h, minute, sec;
75 int ret;
76
77 avio_seek(pb, offset, SEEK_SET);
78
79 ret = avio_read(s->pb, string, DSS_TIME_SIZE);
80 if (ret < DSS_TIME_SIZE)
81 return ret < 0 ? ret : AVERROR_EOF;
82
83 if (sscanf(string, "%2d%2d%2d%2d%2d%2d", &y, &month, &d, &h, &minute, &sec) != 6)
84 return AVERROR_INVALIDDATA;
85 /* We deal with a two-digit year here, so set the default date to 2000
86 * and hope it will never be used in the next century. */
87 snprintf(datetime, sizeof(datetime), "%.4d-%.2d-%.2dT%.2d:%.2d:%.2d",
88 y + 2000, month, d, h, minute, sec);
89 return av_dict_set(&s->metadata, key, datetime, 0);
90 }
91
92 static int dss_read_metadata_string(AVFormatContext *s, unsigned int offset,
93 unsigned int size, const char *key)
94 {
95 AVIOContext *pb = s->pb;
96 char *value;
97 int ret;
98
99 avio_seek(pb, offset, SEEK_SET);
100
101 value = av_mallocz(size + 1);
102 if (!value)
103 return AVERROR(ENOMEM);
104
105 ret = avio_read(s->pb, value, size);
106 if (ret < size) {
107 av_free(value);
108 return ret < 0 ? ret : AVERROR_EOF;
109 }
110
111 return av_dict_set(&s->metadata, key, value, AV_DICT_DONT_STRDUP_VAL);
112 }
113
114 static int dss_read_header(AVFormatContext *s)
115 {
116 DSSDemuxContext *ctx = s->priv_data;
117 AVIOContext *pb = s->pb;
118 AVStream *st;
119 int64_t ret64;
120 int ret, version;
121
122 st = avformat_new_stream(s, NULL);
123 if (!st)
124 return AVERROR(ENOMEM);
125
126 version = avio_r8(pb);
127 ctx->dss_header_size = version * DSS_BLOCK_SIZE;
128
129 ret = dss_read_metadata_string(s, DSS_HEAD_OFFSET_AUTHOR,
130 DSS_AUTHOR_SIZE, "author");
131 if (ret)
132 return ret;
133
134 ret = dss_read_metadata_date(s, DSS_HEAD_OFFSET_END_TIME, "date");
135 if (ret)
136 return ret;
137
138 ret = dss_read_metadata_string(s, DSS_HEAD_OFFSET_COMMENT,
139 DSS_COMMENT_SIZE, "comment");
140 if (ret)
141 return ret;
142
143 avio_seek(pb, DSS_HEAD_OFFSET_ACODEC, SEEK_SET);
144 ctx->audio_codec = avio_r8(pb);
145
146 if (ctx->audio_codec == DSS_ACODEC_DSS_SP) {
147 st->codecpar->codec_id = AV_CODEC_ID_DSS_SP;
148 st->codecpar->sample_rate = 11025;
149 s->bit_rate = 8 * (DSS_FRAME_SIZE - 1) * st->codecpar->sample_rate
150 * 512 / (506 * 264);
151 } else if (ctx->audio_codec == DSS_ACODEC_G723_1) {
152 st->codecpar->codec_id = AV_CODEC_ID_G723_1;
153 st->codecpar->sample_rate = 8000;
154 } else {
155 avpriv_request_sample(s, "Support for codec %x in DSS",
156 ctx->audio_codec);
157 return AVERROR_PATCHWELCOME;
158 }
159
160 st->codecpar->codec_type = AVMEDIA_TYPE_AUDIO;
161 st->codecpar->ch_layout = (AVChannelLayout)AV_CHANNEL_LAYOUT_MONO;
162
163 avpriv_set_pts_info(st, 64, 1, st->codecpar->sample_rate);
164 st->start_time = 0;
165
166 /* Jump over header */
167
168 if ((ret64 = avio_seek(pb, ctx->dss_header_size, SEEK_SET)) < 0)
169 return (int)ret64;
170
171 ctx->counter = 0;
172 ctx->swap = 0;
173
174 return 0;
175 }
176
177 static void dss_skip_audio_header(AVFormatContext *s, AVPacket *pkt)
178 {
179 DSSDemuxContext *ctx = s->priv_data;
180 AVIOContext *pb = s->pb;
181
182 avio_skip(pb, DSS_AUDIO_BLOCK_HEADER_SIZE);
183 ctx->counter += DSS_BLOCK_SIZE - DSS_AUDIO_BLOCK_HEADER_SIZE;
184 }
185
186 static void dss_sp_byte_swap(DSSDemuxContext *ctx, uint8_t *data)
187 {
188 int i;
189
190 if (ctx->swap) {
191 for (i = 0; i < DSS_FRAME_SIZE - 2; i += 2)
192 data[i] = data[i + 4];
193
194 /* Zero the padding. */
195 data[DSS_FRAME_SIZE] = 0;
196 data[1] = ctx->dss_sp_swap_byte;
197 } else {
198 ctx->dss_sp_swap_byte = data[DSS_FRAME_SIZE - 2];
199 }
200
201 /* make sure byte 40 is always 0 */
202 data[DSS_FRAME_SIZE - 2] = 0;
203 ctx->swap ^= 1;
204 }
205
206 static int dss_sp_read_packet(AVFormatContext *s, AVPacket *pkt)
207 {
208 DSSDemuxContext *ctx = s->priv_data;
209 int read_size, ret, offset = 0, buff_offset = 0;
210 int64_t pos = avio_tell(s->pb);
211
212 if (ctx->counter == 0)
213 dss_skip_audio_header(s, pkt);
214
215 if (ctx->swap) {
216 read_size = DSS_FRAME_SIZE - 2;
217 buff_offset = 3;
218 } else
219 read_size = DSS_FRAME_SIZE;
220
221 ret = av_new_packet(pkt, DSS_FRAME_SIZE);
222 if (ret < 0)
223 return ret;
224
225 pkt->duration = 264;
226 pkt->pos = pos;
227 pkt->stream_index = 0;
228
229 if (ctx->counter < read_size) {
230 ret = avio_read(s->pb, pkt->data + buff_offset,
231 ctx->counter);
232 if (ret < ctx->counter)
233 goto error_eof;
234
235 offset = ctx->counter;
236 dss_skip_audio_header(s, pkt);
237 }
238 ctx->counter -= read_size;
239
240 /* This will write one byte into pkt's padding if buff_offset == 3 */
241 ret = avio_read(s->pb, pkt->data + offset + buff_offset,
242 read_size - offset);
243 if (ret < read_size - offset)
244 goto error_eof;
245
246 dss_sp_byte_swap(ctx, pkt->data);
247
248 if (ctx->dss_sp_swap_byte < 0) {
249 return AVERROR(EAGAIN);
250 }
251
252 return 0;
253
254 error_eof:
255 return ret < 0 ? ret : AVERROR_EOF;
256 }
257
258 static int dss_723_1_read_packet(AVFormatContext *s, AVPacket *pkt)
259 {
260 DSSDemuxContext *ctx = s->priv_data;
261 AVStream *st = s->streams[0];
262 int size, byte, ret, offset;
263 int64_t pos = avio_tell(s->pb);
264
265 if (ctx->counter == 0)
266 dss_skip_audio_header(s, pkt);
267
268 /* We make one byte-step here. Don't forget to add offset. */
269 byte = avio_r8(s->pb);
270 if (byte == 0xff)
271 return AVERROR_INVALIDDATA;
272
273 size = frame_size[byte & 3];
274
275 ctx->packet_size = size;
276 ctx->counter--;
277
278 ret = av_new_packet(pkt, size);
279 if (ret < 0)
280 return ret;
281 pkt->pos = pos;
282
283 pkt->data[0] = byte;
284 offset = 1;
285 pkt->duration = 240;
286 s->bit_rate = 8LL * size-- * st->codecpar->sample_rate * 512 / (506 * pkt->duration);
287
288 pkt->stream_index = 0;
289
290 if (ctx->counter < size) {
291 ret = avio_read(s->pb, pkt->data + offset,
292 ctx->counter);
293 if (ret < ctx->counter)
294 return ret < 0 ? ret : AVERROR_EOF;
295
296 offset += ctx->counter;
297 size -= ctx->counter;
298 ctx->counter = 0;
299 dss_skip_audio_header(s, pkt);
300 }
301 ctx->counter -= size;
302
303 ret = avio_read(s->pb, pkt->data + offset, size);
304 if (ret < size)
305 return ret < 0 ? ret : AVERROR_EOF;
306
307 return 0;
308 }
309
310 static int dss_read_packet(AVFormatContext *s, AVPacket *pkt)
311 {
312 DSSDemuxContext *ctx = s->priv_data;
313
314 if (ctx->audio_codec == DSS_ACODEC_DSS_SP)
315 return dss_sp_read_packet(s, pkt);
316 else
317 return dss_723_1_read_packet(s, pkt);
318 }
319
320 static int dss_read_seek(AVFormatContext *s, int stream_index,
321 int64_t timestamp, int flags)
322 {
323 DSSDemuxContext *ctx = s->priv_data;
324 int64_t ret, seekto;
325 uint8_t header[DSS_AUDIO_BLOCK_HEADER_SIZE];
326 int offset;
327
328 if (ctx->audio_codec == DSS_ACODEC_DSS_SP)
329 seekto = timestamp / 264 * 41 / 506 * 512;
330 else
331 seekto = timestamp / 240 * ctx->packet_size / 506 * 512;
332
333 if (seekto < 0)
334 seekto = 0;
335
336 seekto += ctx->dss_header_size;
337
338 ret = avio_seek(s->pb, seekto, SEEK_SET);
339 if (ret < 0)
340 return ret;
341
342 avio_read(s->pb, header, DSS_AUDIO_BLOCK_HEADER_SIZE);
343 ctx->swap = !!(header[0] & 0x80);
344 offset = 2*header[1] + 2*ctx->swap;
345 if (offset < DSS_AUDIO_BLOCK_HEADER_SIZE)
346 return AVERROR_INVALIDDATA;
347 if (offset == DSS_AUDIO_BLOCK_HEADER_SIZE) {
348 ctx->counter = 0;
349 offset = avio_skip(s->pb, -DSS_AUDIO_BLOCK_HEADER_SIZE);
350 } else {
351 ctx->counter = DSS_BLOCK_SIZE - offset;
352 offset = avio_skip(s->pb, offset - DSS_AUDIO_BLOCK_HEADER_SIZE);
353 }
354 ctx->dss_sp_swap_byte = -1;
355 return 0;
356 }
357
358
359 const FFInputFormat ff_dss_demuxer = {
360 .p.name = "dss",
361 .p.long_name = NULL_IF_CONFIG_SMALL("Digital Speech Standard (DSS)"),
362 .p.extensions = "dss",
363 .priv_data_size = sizeof(DSSDemuxContext),
364 .read_probe = dss_probe,
365 .read_header = dss_read_header,
366 .read_packet = dss_read_packet,
367 .read_seek = dss_read_seek,
368 };