blob: ead6d2b4243a83d80c3931c03605f27a5bb3fab0
1 | /* |
2 | * 4X Technologies .4xm File Demuxer (no muxer) |
3 | * Copyright (c) 2003 The FFmpeg project |
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 |
24 | * 4X Technologies file demuxer |
25 | * by Mike Melanson (melanson@pcisys.net) |
26 | * for more information on the .4xm file format, visit: |
27 | * http://www.pcisys.net/~melanson/codecs/ |
28 | */ |
29 | |
30 | #include "libavutil/intreadwrite.h" |
31 | #include "libavutil/intfloat.h" |
32 | #include "libavcodec/internal.h" |
33 | #include "avformat.h" |
34 | #include "internal.h" |
35 | |
36 | #define RIFF_TAG MKTAG('R', 'I', 'F', 'F') |
37 | #define FOURXMV_TAG MKTAG('4', 'X', 'M', 'V') |
38 | #define LIST_TAG MKTAG('L', 'I', 'S', 'T') |
39 | #define HEAD_TAG MKTAG('H', 'E', 'A', 'D') |
40 | #define TRK__TAG MKTAG('T', 'R', 'K', '_') |
41 | #define MOVI_TAG MKTAG('M', 'O', 'V', 'I') |
42 | #define VTRK_TAG MKTAG('V', 'T', 'R', 'K') |
43 | #define STRK_TAG MKTAG('S', 'T', 'R', 'K') |
44 | #define std__TAG MKTAG('s', 't', 'd', '_') |
45 | #define name_TAG MKTAG('n', 'a', 'm', 'e') |
46 | #define vtrk_TAG MKTAG('v', 't', 'r', 'k') |
47 | #define strk_TAG MKTAG('s', 't', 'r', 'k') |
48 | #define ifrm_TAG MKTAG('i', 'f', 'r', 'm') |
49 | #define pfrm_TAG MKTAG('p', 'f', 'r', 'm') |
50 | #define cfrm_TAG MKTAG('c', 'f', 'r', 'm') |
51 | #define ifr2_TAG MKTAG('i', 'f', 'r', '2') |
52 | #define pfr2_TAG MKTAG('p', 'f', 'r', '2') |
53 | #define cfr2_TAG MKTAG('c', 'f', 'r', '2') |
54 | #define snd__TAG MKTAG('s', 'n', 'd', '_') |
55 | |
56 | #define vtrk_SIZE 0x44 |
57 | #define strk_SIZE 0x28 |
58 | |
59 | #define GET_LIST_HEADER() \ |
60 | fourcc_tag = avio_rl32(pb); \ |
61 | size = avio_rl32(pb); \ |
62 | if (fourcc_tag != LIST_TAG) \ |
63 | return AVERROR_INVALIDDATA; \ |
64 | fourcc_tag = avio_rl32(pb); |
65 | |
66 | typedef struct AudioTrack { |
67 | int sample_rate; |
68 | int bits; |
69 | int channels; |
70 | int stream_index; |
71 | int adpcm; |
72 | int64_t audio_pts; |
73 | } AudioTrack; |
74 | |
75 | typedef struct FourxmDemuxContext { |
76 | int video_stream_index; |
77 | int track_count; |
78 | AudioTrack *tracks; |
79 | |
80 | int64_t video_pts; |
81 | AVRational fps; |
82 | } FourxmDemuxContext; |
83 | |
84 | static int fourxm_probe(AVProbeData *p) |
85 | { |
86 | if ((AV_RL32(&p->buf[0]) != RIFF_TAG) || |
87 | (AV_RL32(&p->buf[8]) != FOURXMV_TAG)) |
88 | return 0; |
89 | |
90 | return AVPROBE_SCORE_MAX; |
91 | } |
92 | |
93 | static int parse_vtrk(AVFormatContext *s, |
94 | FourxmDemuxContext *fourxm, uint8_t *buf, int size, |
95 | int left) |
96 | { |
97 | AVStream *st; |
98 | /* check that there is enough data */ |
99 | if (size != vtrk_SIZE || left < size + 8) { |
100 | return AVERROR_INVALIDDATA; |
101 | } |
102 | |
103 | /* allocate a new AVStream */ |
104 | st = avformat_new_stream(s, NULL); |
105 | if (!st) |
106 | return AVERROR(ENOMEM); |
107 | |
108 | avpriv_set_pts_info(st, 60, fourxm->fps.den, fourxm->fps.num); |
109 | |
110 | fourxm->video_stream_index = st->index; |
111 | |
112 | st->codecpar->codec_type = AVMEDIA_TYPE_VIDEO; |
113 | st->codecpar->codec_id = AV_CODEC_ID_4XM; |
114 | |
115 | st->codecpar->extradata = av_mallocz(4 + AV_INPUT_BUFFER_PADDING_SIZE); |
116 | if (!st->codecpar->extradata) |
117 | return AVERROR(ENOMEM); |
118 | st->codecpar->extradata_size = 4; |
119 | AV_WL32(st->codecpar->extradata, AV_RL32(buf + 16)); |
120 | st->codecpar->width = AV_RL32(buf + 36); |
121 | st->codecpar->height = AV_RL32(buf + 40); |
122 | |
123 | return 0; |
124 | } |
125 | |
126 | |
127 | static int parse_strk(AVFormatContext *s, |
128 | FourxmDemuxContext *fourxm, uint8_t *buf, int size, |
129 | int left) |
130 | { |
131 | AVStream *st; |
132 | int track; |
133 | /* check that there is enough data */ |
134 | if (size != strk_SIZE || left < size + 8) |
135 | return AVERROR_INVALIDDATA; |
136 | |
137 | track = AV_RL32(buf + 8); |
138 | if ((unsigned)track >= UINT_MAX / sizeof(AudioTrack) - 1) { |
139 | av_log(s, AV_LOG_ERROR, "current_track too large\n"); |
140 | return AVERROR_INVALIDDATA; |
141 | } |
142 | |
143 | if (track + 1 > fourxm->track_count) { |
144 | if (av_reallocp_array(&fourxm->tracks, track + 1, sizeof(AudioTrack))) |
145 | return AVERROR(ENOMEM); |
146 | memset(&fourxm->tracks[fourxm->track_count], 0, |
147 | sizeof(AudioTrack) * (track + 1 - fourxm->track_count)); |
148 | fourxm->track_count = track + 1; |
149 | } |
150 | fourxm->tracks[track].adpcm = AV_RL32(buf + 12); |
151 | fourxm->tracks[track].channels = AV_RL32(buf + 36); |
152 | fourxm->tracks[track].sample_rate = AV_RL32(buf + 40); |
153 | fourxm->tracks[track].bits = AV_RL32(buf + 44); |
154 | fourxm->tracks[track].audio_pts = 0; |
155 | |
156 | if (fourxm->tracks[track].channels <= 0 || |
157 | fourxm->tracks[track].channels > FF_SANE_NB_CHANNELS || |
158 | fourxm->tracks[track].sample_rate <= 0 || |
159 | fourxm->tracks[track].bits <= 0 || |
160 | fourxm->tracks[track].bits > INT_MAX / FF_SANE_NB_CHANNELS) { |
161 | av_log(s, AV_LOG_ERROR, "audio header invalid\n"); |
162 | return AVERROR_INVALIDDATA; |
163 | } |
164 | if (!fourxm->tracks[track].adpcm && fourxm->tracks[track].bits<8) { |
165 | av_log(s, AV_LOG_ERROR, "bits unspecified for non ADPCM\n"); |
166 | return AVERROR_INVALIDDATA; |
167 | } |
168 | |
169 | if (fourxm->tracks[track].sample_rate > INT64_MAX / fourxm->tracks[track].bits / fourxm->tracks[track].channels) { |
170 | av_log(s, AV_LOG_ERROR, "Overflow during bit rate calculation %d * %d * %d\n", |
171 | fourxm->tracks[track].sample_rate, fourxm->tracks[track].bits, fourxm->tracks[track].channels); |
172 | return AVERROR_INVALIDDATA; |
173 | } |
174 | |
175 | /* allocate a new AVStream */ |
176 | st = avformat_new_stream(s, NULL); |
177 | if (!st) |
178 | return AVERROR(ENOMEM); |
179 | |
180 | st->id = track; |
181 | avpriv_set_pts_info(st, 60, 1, fourxm->tracks[track].sample_rate); |
182 | |
183 | fourxm->tracks[track].stream_index = st->index; |
184 | |
185 | st->codecpar->codec_type = AVMEDIA_TYPE_AUDIO; |
186 | st->codecpar->codec_tag = 0; |
187 | st->codecpar->channels = fourxm->tracks[track].channels; |
188 | st->codecpar->sample_rate = fourxm->tracks[track].sample_rate; |
189 | st->codecpar->bits_per_coded_sample = fourxm->tracks[track].bits; |
190 | st->codecpar->bit_rate = (int64_t)st->codecpar->channels * |
191 | st->codecpar->sample_rate * |
192 | st->codecpar->bits_per_coded_sample; |
193 | st->codecpar->block_align = st->codecpar->channels * |
194 | st->codecpar->bits_per_coded_sample; |
195 | |
196 | if (fourxm->tracks[track].adpcm){ |
197 | st->codecpar->codec_id = AV_CODEC_ID_ADPCM_4XM; |
198 | } else if (st->codecpar->bits_per_coded_sample == 8) { |
199 | st->codecpar->codec_id = AV_CODEC_ID_PCM_U8; |
200 | } else |
201 | st->codecpar->codec_id = AV_CODEC_ID_PCM_S16LE; |
202 | |
203 | return 0; |
204 | } |
205 | |
206 | static int fourxm_read_header(AVFormatContext *s) |
207 | { |
208 | AVIOContext *pb = s->pb; |
209 | unsigned int fourcc_tag; |
210 | unsigned int size; |
211 | int header_size; |
212 | FourxmDemuxContext *fourxm = s->priv_data; |
213 | unsigned char *header; |
214 | int i, ret; |
215 | |
216 | fourxm->track_count = 0; |
217 | fourxm->tracks = NULL; |
218 | fourxm->fps = (AVRational){1,1}; |
219 | |
220 | /* skip the first 3 32-bit numbers */ |
221 | avio_skip(pb, 12); |
222 | |
223 | /* check for LIST-HEAD */ |
224 | GET_LIST_HEADER(); |
225 | header_size = size - 4; |
226 | if (fourcc_tag != HEAD_TAG || header_size < 0) |
227 | return AVERROR_INVALIDDATA; |
228 | |
229 | /* allocate space for the header and load the whole thing */ |
230 | header = av_malloc(header_size); |
231 | if (!header) |
232 | return AVERROR(ENOMEM); |
233 | if (avio_read(pb, header, header_size) != header_size) { |
234 | av_free(header); |
235 | return AVERROR(EIO); |
236 | } |
237 | |
238 | /* take the lazy approach and search for any and all vtrk and strk chunks */ |
239 | for (i = 0; i < header_size - 8; i++) { |
240 | fourcc_tag = AV_RL32(&header[i]); |
241 | size = AV_RL32(&header[i + 4]); |
242 | if (size > header_size - i - 8 && (fourcc_tag == vtrk_TAG || fourcc_tag == strk_TAG)) { |
243 | av_log(s, AV_LOG_ERROR, "chunk larger than array %d>%d\n", size, header_size - i - 8); |
244 | return AVERROR_INVALIDDATA; |
245 | } |
246 | |
247 | if (fourcc_tag == std__TAG) { |
248 | if (header_size - i < 16) { |
249 | av_log(s, AV_LOG_ERROR, "std TAG truncated\n"); |
250 | ret = AVERROR_INVALIDDATA; |
251 | goto fail; |
252 | } |
253 | fourxm->fps = av_d2q(av_int2float(AV_RL32(&header[i + 12])), 10000); |
254 | } else if (fourcc_tag == vtrk_TAG) { |
255 | if ((ret = parse_vtrk(s, fourxm, header + i, size, |
256 | header_size - i)) < 0) |
257 | goto fail; |
258 | |
259 | i += 8 + size; |
260 | } else if (fourcc_tag == strk_TAG) { |
261 | if ((ret = parse_strk(s, fourxm, header + i, size, |
262 | header_size - i)) < 0) |
263 | goto fail; |
264 | |
265 | i += 8 + size; |
266 | } |
267 | } |
268 | |
269 | /* skip over the LIST-MOVI chunk (which is where the stream should be */ |
270 | GET_LIST_HEADER(); |
271 | if (fourcc_tag != MOVI_TAG) { |
272 | ret = AVERROR_INVALIDDATA; |
273 | goto fail; |
274 | } |
275 | |
276 | av_free(header); |
277 | /* initialize context members */ |
278 | fourxm->video_pts = -1; /* first frame will push to 0 */ |
279 | |
280 | return 0; |
281 | fail: |
282 | av_freep(&fourxm->tracks); |
283 | av_free(header); |
284 | return ret; |
285 | } |
286 | |
287 | static int fourxm_read_packet(AVFormatContext *s, |
288 | AVPacket *pkt) |
289 | { |
290 | FourxmDemuxContext *fourxm = s->priv_data; |
291 | AVIOContext *pb = s->pb; |
292 | unsigned int fourcc_tag; |
293 | unsigned int size; |
294 | int ret = 0; |
295 | unsigned int track_number; |
296 | int packet_read = 0; |
297 | unsigned char header[8]; |
298 | int audio_frame_count; |
299 | |
300 | while (!packet_read) { |
301 | if ((ret = avio_read(s->pb, header, 8)) < 0) |
302 | return ret; |
303 | fourcc_tag = AV_RL32(&header[0]); |
304 | size = AV_RL32(&header[4]); |
305 | if (avio_feof(pb)) |
306 | return AVERROR(EIO); |
307 | switch (fourcc_tag) { |
308 | case LIST_TAG: |
309 | /* this is a good time to bump the video pts */ |
310 | fourxm->video_pts++; |
311 | |
312 | /* skip the LIST-* tag and move on to the next fourcc */ |
313 | avio_rl32(pb); |
314 | break; |
315 | |
316 | case ifrm_TAG: |
317 | case pfrm_TAG: |
318 | case cfrm_TAG: |
319 | case ifr2_TAG: |
320 | case pfr2_TAG: |
321 | case cfr2_TAG: |
322 | /* allocate 8 more bytes than 'size' to account for fourcc |
323 | * and size */ |
324 | if (size + 8 < size || av_new_packet(pkt, size + 8)) |
325 | return AVERROR(EIO); |
326 | pkt->stream_index = fourxm->video_stream_index; |
327 | pkt->pts = fourxm->video_pts; |
328 | pkt->pos = avio_tell(s->pb); |
329 | memcpy(pkt->data, header, 8); |
330 | ret = avio_read(s->pb, &pkt->data[8], size); |
331 | |
332 | if (ret < 0) { |
333 | av_packet_unref(pkt); |
334 | } else { |
335 | packet_read = 1; |
336 | av_shrink_packet(pkt, ret + 8); |
337 | } |
338 | break; |
339 | |
340 | case snd__TAG: |
341 | track_number = avio_rl32(pb); |
342 | avio_skip(pb, 4); |
343 | size -= 8; |
344 | |
345 | if (track_number < fourxm->track_count && |
346 | fourxm->tracks[track_number].channels > 0) { |
347 | ret = av_get_packet(s->pb, pkt, size); |
348 | if (ret < 0) |
349 | return AVERROR(EIO); |
350 | pkt->stream_index = |
351 | fourxm->tracks[track_number].stream_index; |
352 | pkt->pts = fourxm->tracks[track_number].audio_pts; |
353 | packet_read = 1; |
354 | |
355 | /* pts accounting */ |
356 | audio_frame_count = size; |
357 | if (fourxm->tracks[track_number].adpcm) |
358 | audio_frame_count -= 2 * (fourxm->tracks[track_number].channels); |
359 | audio_frame_count /= fourxm->tracks[track_number].channels; |
360 | if (fourxm->tracks[track_number].adpcm) { |
361 | audio_frame_count *= 2; |
362 | } else |
363 | audio_frame_count /= |
364 | (fourxm->tracks[track_number].bits / 8); |
365 | fourxm->tracks[track_number].audio_pts += audio_frame_count; |
366 | } else { |
367 | avio_skip(pb, size); |
368 | } |
369 | break; |
370 | |
371 | default: |
372 | avio_skip(pb, size); |
373 | break; |
374 | } |
375 | } |
376 | return ret; |
377 | } |
378 | |
379 | static int fourxm_read_close(AVFormatContext *s) |
380 | { |
381 | FourxmDemuxContext *fourxm = s->priv_data; |
382 | |
383 | av_freep(&fourxm->tracks); |
384 | |
385 | return 0; |
386 | } |
387 | |
388 | AVInputFormat ff_fourxm_demuxer = { |
389 | .name = "4xm", |
390 | .long_name = NULL_IF_CONFIG_SMALL("4X Technologies"), |
391 | .priv_data_size = sizeof(FourxmDemuxContext), |
392 | .read_probe = fourxm_probe, |
393 | .read_header = fourxm_read_header, |
394 | .read_packet = fourxm_read_packet, |
395 | .read_close = fourxm_read_close, |
396 | }; |
397 |