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