libavformat/westwood.c

   1 /*
   2  * Westwood Studios Multimedia Formats Demuxer (VQA, AUD)
   3  * Copyright (c) 2003 The ffmpeg Project
   4  *
   5  * This library is free software; you can redistribute it and/or
   6  * modify it under the terms of the GNU Lesser General Public
   7  * License as published by the Free Software Foundation; either
   8  * version 2 of the License, or (at your option) any later version.
   9  *
  10  * This library is distributed in the hope that it will be useful,
  11  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  12  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  13  * Lesser General Public License for more details.
  14  *
  15  * You should have received a copy of the GNU Lesser General Public
  16  * License along with this library; if not, write to the Free Software
  17  * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
  18  */
  19
  20 /**
  21  * @file westwood.c
  22  * Westwood Studios VQA & AUD file demuxers
  23  * by Mike Melanson (melanson@pcisys.net)
  24  * for more information on the Westwood file formats, visit:
  25  *   http://www.pcisys.net/~melanson/codecs/
  26  *   http://www.geocities.com/SiliconValley/8682/aud3.txt
  27  *
  28  * Implementation note: There is no definite file signature for AUD files.
  29  * The demuxer uses a probabilistic strategy for content detection. This
  30  * entails performing sanity checks on certain header values in order to
  31  * qualify a file. Refer to wsaud_probe() for the precise parameters.
  32  */
  33
  34 #include "avformat.h"
  35
  36 #define LE_16(x)  ((((uint8_t*)(x))[1] << 8) | ((uint8_t*)(x))[0])
  37 #define LE_32(x)  ((((uint8_t*)(x))[3] << 24) | \
  38                    (((uint8_t*)(x))[2] << 16) | \
  39                    (((uint8_t*)(x))[1] << 8) | \
  40                     ((uint8_t*)(x))[0])
  41 #define BE_32(x)  ((((uint8_t*)(x))[0] << 24) | \
  42                    (((uint8_t*)(x))[1] << 16) | \
  43                    (((uint8_t*)(x))[2] << 8) | \
  44                     ((uint8_t*)(x))[3])
  45
  46 #define AUD_HEADER_SIZE 12
  47 #define AUD_CHUNK_PREAMBLE_SIZE 8
  48 #define AUD_CHUNK_SIGNATURE 0x0000DEAF
  49
  50 #define FOURCC_TAG( ch0, ch1, ch2, ch3 ) \
  51         ( (long)(unsigned char)(ch3) | \
  52         ( (long)(unsigned char)(ch2) << 8 ) | \
  53         ( (long)(unsigned char)(ch1) << 16 ) | \
  54         ( (long)(unsigned char)(ch0) << 24 ) )
  55
  56 #define FORM_TAG FOURCC_TAG('F', 'O', 'R', 'M')
  57 #define WVQA_TAG FOURCC_TAG('W', 'V', 'Q', 'A')
  58 #define VQHD_TAG FOURCC_TAG('V', 'Q', 'H', 'D')
  59 #define FINF_TAG FOURCC_TAG('F', 'I', 'N', 'F')
  60 #define SND0_TAG FOURCC_TAG('S', 'N', 'D', '0')
  61 #define SND2_TAG FOURCC_TAG('S', 'N', 'D', '2')
  62 #define VQFR_TAG FOURCC_TAG('V', 'Q', 'F', 'R')
  63
  64 /* don't know what these tags are for, but acknowledge their existence */
  65 #define CINF_TAG FOURCC_TAG('C', 'I', 'N', 'F')
  66 #define CINH_TAG FOURCC_TAG('C', 'I', 'N', 'H')
  67 #define CIND_TAG FOURCC_TAG('C', 'I', 'N', 'D')
  68 #define PINF_TAG FOURCC_TAG('P', 'I', 'N', 'F')
  69 #define PINH_TAG FOURCC_TAG('P', 'I', 'N', 'H')
  70 #define PIND_TAG FOURCC_TAG('P', 'I', 'N', 'D')
  71
  72 #define VQA_HEADER_SIZE 0x2A
  73 #define VQA_FRAMERATE 15
  74 #define VQA_VIDEO_PTS_INC (90000 / VQA_FRAMERATE)
  75 #define VQA_PREAMBLE_SIZE 8
  76
  77 typedef struct WsAudDemuxContext {
  78     int audio_samplerate;
  79     int audio_channels;
  80     int audio_bits;
  81     int audio_type;
  82     int audio_stream_index;
  83     int64_t audio_frame_counter;
  84 } WsAudDemuxContext;
  85
  86 typedef struct WsVqaDemuxContext {
  87     int audio_samplerate;
  88     int audio_channels;
  89     int audio_bits;
  90
  91     int audio_stream_index;
  92     int video_stream_index;
  93
  94     int64_t audio_frame_counter;
  95     int64_t video_pts;
  96 } WsVqaDemuxContext;
  97
  98 static int wsaud_probe(AVProbeData *p)
  99 {
 100     int field;
 101
 102     /* Probabilistic content detection strategy: There is no file signature
 103      * so perform sanity checks on various header parameters:
 104      *   8000 <= sample rate (16 bits) <= 48000  ==> 40001 acceptable numbers
 105      *   compression type (8 bits) = 1 or 99     ==> 2 acceptable numbers
 106      * There is a total of 24 bits. The number space contains 2^24 =
 107      * 16777216 numbers. There are 40001 * 2 = 80002 acceptable combinations
 108      * of numbers. There is a 80002/16777216 = 0.48% chance of a false
 109      * positive.
 110      */
 111
 112     if (p->buf_size < AUD_HEADER_SIZE)
 113         return 0;
 114
 115     /* check sample rate */
 116     field = LE_16(&p->buf[0]);
 117     if ((field < 8000) || (field > 48000))
 118         return 0;
 119
 120     /* note: only check for WS IMA (type 99) right now since there is no
 121      * support for type 1 */
 122     if (p->buf[11] != 99)
 123         return 0;
 124
 125     /* return 1/2 certainty since this file check is a little sketchy */
 126     return AVPROBE_SCORE_MAX / 2;
 127 }
 128
 129 static int wsaud_read_header(AVFormatContext *s,
 130                              AVFormatParameters *ap)
 131 {
 132     WsAudDemuxContext *wsaud = (WsAudDemuxContext *)s->priv_data;
 133     ByteIOContext *pb = &s->pb;
 134     AVStream *st;
 135     unsigned char header[AUD_HEADER_SIZE];
 136
 137     if (get_buffer(pb, header, AUD_HEADER_SIZE) != AUD_HEADER_SIZE)
 138         return -EIO;
 139     wsaud->audio_samplerate = LE_16(&header[0]);
 140     if (header[11] == 99)
 141         wsaud->audio_type = CODEC_ID_ADPCM_IMA_WS;
 142     else
 143         return AVERROR_INVALIDDATA;
 144
 145     /* flag 0 indicates stereo */
 146     wsaud->audio_channels = (header[10] & 0x1) + 1;
 147     /* flag 1 indicates 16 bit audio */
 148     wsaud->audio_bits = (((header[10] & 0x2) >> 1) + 1) * 8;
 149
 150     /* set the pts reference the same as the sample rate */
 151     s->pts_num = 1;
 152     s->pts_den = wsaud->audio_samplerate;
 153
 154     /* initialize the audio decoder stream */
 155     st = av_new_stream(s, 0);
 156     if (!st)
 157         return AVERROR_NOMEM;
 158     st->codec.codec_type = CODEC_TYPE_AUDIO;
 159     st->codec.codec_id = wsaud->audio_type;
 160     st->codec.codec_tag = 0;  /* no tag */
 161     st->codec.channels = wsaud->audio_channels;
 162     st->codec.sample_rate = wsaud->audio_samplerate;
 163     st->codec.bits_per_sample = wsaud->audio_bits;
 164     st->codec.bit_rate = st->codec.channels * st->codec.sample_rate *
 165         st->codec.bits_per_sample / 4;
 166     st->codec.block_align = st->codec.channels * st->codec.bits_per_sample;
 167
 168     wsaud->audio_stream_index = st->index;
 169     wsaud->audio_frame_counter = 0;
 170
 171     return 0;
 172 }
 173
 174 static int wsaud_read_packet(AVFormatContext *s,
 175                              AVPacket *pkt)
 176 {
 177     WsAudDemuxContext *wsaud = (WsAudDemuxContext *)s->priv_data;
 178     ByteIOContext *pb = &s->pb;
 179     unsigned char preamble[AUD_CHUNK_PREAMBLE_SIZE];
 180     unsigned int chunk_size;
 181     int ret = 0;
 182
 183     if (get_buffer(pb, preamble, AUD_CHUNK_PREAMBLE_SIZE) !=
 184         AUD_CHUNK_PREAMBLE_SIZE)
 185         return -EIO;
 186
 187     /* validate the chunk */
 188     if (LE_32(&preamble[4]) != AUD_CHUNK_SIGNATURE)
 189         return AVERROR_INVALIDDATA;
 190
 191     chunk_size = LE_16(&preamble[0]);
 192     if (av_new_packet(pkt, chunk_size))
 193         return -EIO;
 194     pkt->stream_index = wsaud->audio_stream_index;
 195     pkt->pts = wsaud->audio_frame_counter;
 196     pkt->pts /= wsaud->audio_samplerate;
 197     if ((ret = get_buffer(pb, pkt->data, chunk_size)) != chunk_size) {
 198         av_free_packet(pkt);
 199         ret = -EIO;
 200     }
 201
 202     /* 2 samples/byte, 1 or 2 samples per frame depending on stereo */
 203     wsaud->audio_frame_counter += (chunk_size * 2) / wsaud->audio_channels;
 204
 205     return ret;
 206 }
 207
 208 static int wsaud_read_close(AVFormatContext *s)
 209 {
 210 //    WsAudDemuxContext *wsaud = (WsAudDemuxContext *)s->priv_data;
 211
 212     return 0;
 213 }
 214
 215
 216 static int wsvqa_probe(AVProbeData *p)
 217 {
 218     /* need 12 bytes to qualify */
 219     if (p->buf_size < 12)
 220         return 0;
 221
 222     /* check for the VQA signatures */
 223     if ((BE_32(&p->buf[0]) != FORM_TAG) ||
 224         (BE_32(&p->buf[8]) != WVQA_TAG))
 225         return 0;
 226
 227     return AVPROBE_SCORE_MAX;
 228 }
 229
 230 static int wsvqa_read_header(AVFormatContext *s,
 231                              AVFormatParameters *ap)
 232 {
 233     WsVqaDemuxContext *wsvqa = (WsVqaDemuxContext *)s->priv_data;
 234     ByteIOContext *pb = &s->pb;
 235     AVStream *st;
 236     unsigned char *header;
 237     unsigned char scratch[VQA_PREAMBLE_SIZE];
 238     unsigned int chunk_tag;
 239     unsigned int chunk_size;
 240
 241     /* set the pts reference (1 pts = 1/90000) */
 242     s->pts_num = 1;
 243     s->pts_den = 90000;
 244
 245     /* initialize the video decoder stream */
 246     st = av_new_stream(s, 0);
 247     if (!st)
 248         return AVERROR_NOMEM;
 249     wsvqa->video_stream_index = st->index;
 250     st->codec.codec_type = CODEC_TYPE_VIDEO;
 251     st->codec.codec_id = CODEC_ID_WS_VQA;
 252     st->codec.codec_tag = 0;  /* no fourcc */
 253
 254     /* skip to the start of the VQA header */
 255     url_fseek(pb, 20, SEEK_SET);
 256
 257     /* the VQA header needs to go to the decoder */
 258     st->codec.extradata_size = VQA_HEADER_SIZE;
 259     st->codec.extradata = av_malloc(VQA_HEADER_SIZE);
 260     header = (unsigned char *)st->codec.extradata;
 261     if (get_buffer(pb, st->codec.extradata, VQA_HEADER_SIZE) !=
 262         VQA_HEADER_SIZE) {
 263         av_free(st->codec.extradata);
 264         return -EIO;
 265     }
 266     st->codec.width = LE_16(&header[6]);
 267     st->codec.height = LE_16(&header[8]);
 268
 269     /* initialize the audio decoder stream is sample rate is non-zero */
 270     if (LE_16(&header[24])) {
 271         st = av_new_stream(s, 0);
 272         if (!st)
 273             return AVERROR_NOMEM;
 274         st->codec.codec_type = CODEC_TYPE_AUDIO;
 275         st->codec.codec_id = CODEC_ID_ADPCM_IMA_WS;
 276         st->codec.codec_tag = 0;  /* no tag */
 277         st->codec.sample_rate = LE_16(&header[24]);
 278         st->codec.channels = header[26];
 279         st->codec.bits_per_sample = 16;
 280         st->codec.bit_rate = st->codec.channels * st->codec.sample_rate *
 281             st->codec.bits_per_sample / 4;
 282         st->codec.block_align = st->codec.channels * st->codec.bits_per_sample;
 283
 284         wsvqa->audio_stream_index = st->index;
 285         wsvqa->audio_samplerate = st->codec.sample_rate;
 286         wsvqa->audio_channels = st->codec.channels;
 287         wsvqa->audio_frame_counter = 0;
 288     }
 289
 290     /* there are 0 or more chunks before the FINF chunk; iterate until
 291      * FINF has been skipped and the file will be ready to be demuxed */
 292     do {
 293         if (get_buffer(pb, scratch, VQA_PREAMBLE_SIZE) != VQA_PREAMBLE_SIZE) {
 294             av_free(st->codec.extradata);
 295             return -EIO;
 296         }
 297         chunk_tag = BE_32(&scratch[0]);
 298         chunk_size = BE_32(&scratch[4]);
 299
 300         /* catch any unknown header tags, for curiousity */
 301         switch (chunk_tag) {
 302         case CINF_TAG:
 303         case CINH_TAG:
 304         case CIND_TAG:
 305         case PINF_TAG:
 306         case PINH_TAG:
 307         case PIND_TAG:
 308         case FINF_TAG:
 309             break;
 310
 311         default:
 312             av_log (s, AV_LOG_ERROR, " note: unknown chunk seen (%c%c%c%c)\n",
 313                 scratch[0], scratch[1],
 314                 scratch[2], scratch[3]);
 315             break;
 316         }
 317
 318         url_fseek(pb, chunk_size, SEEK_CUR);
 319     } while (chunk_tag != FINF_TAG);
 320
 321     wsvqa->video_pts = wsvqa->audio_frame_counter = 0;
 322
 323     return 0;
 324 }
 325
 326 static int wsvqa_read_packet(AVFormatContext *s,
 327                              AVPacket *pkt)
 328 {
 329     WsVqaDemuxContext *wsvqa = (WsVqaDemuxContext *)s->priv_data;
 330     ByteIOContext *pb = &s->pb;
 331     int ret = 0;
 332     unsigned char preamble[VQA_PREAMBLE_SIZE];
 333     unsigned int chunk_type;
 334     unsigned int chunk_size;
 335     int skip_byte;
 336
 337     if (get_buffer(pb, preamble, VQA_PREAMBLE_SIZE) != VQA_PREAMBLE_SIZE)
 338         return -EIO;
 339
 340     chunk_type = BE_32(&preamble[0]);
 341     chunk_size = BE_32(&preamble[4]);
 342     skip_byte = chunk_size & 0x01;
 343
 344     if ((chunk_type == SND2_TAG) || (chunk_type == VQFR_TAG)) {
 345
 346         if (av_new_packet(pkt, chunk_size))
 347             return -EIO;
 348         ret = get_buffer(pb, pkt->data, chunk_size);
 349         if (ret != chunk_size) {
 350             av_free_packet(pkt);
 351             ret = -EIO;
 352         }
 353
 354         if (chunk_type == SND2_TAG) {
 355             pkt->stream_index = wsvqa->audio_stream_index;
 356
 357             pkt->pts = 90000;
 358             pkt->pts *= wsvqa->audio_frame_counter;
 359             pkt->pts /= wsvqa->audio_samplerate;
 360
 361             /* 2 samples/byte, 1 or 2 samples per frame depending on stereo */
 362             wsvqa->audio_frame_counter += (chunk_size * 2) /
 363                 wsvqa->audio_channels;
 364         } else {
 365             pkt->stream_index = wsvqa->video_stream_index;
 366             pkt->pts = wsvqa->video_pts;
 367             wsvqa->video_pts += VQA_VIDEO_PTS_INC;
 368         }
 369
 370     } else
 371         return AVERROR_INVALIDDATA;
 372
 373     /* stay on 16-bit alignment */
 374     if (skip_byte)
 375         url_fseek(pb, 1, SEEK_CUR);
 376
 377     return ret;
 378 }
 379
 380 static int wsvqa_read_close(AVFormatContext *s)
 381 {
 382 //    WsVqaDemuxContext *wsvqa = (WsVqaDemuxContext *)s->priv_data;
 383
 384     return 0;
 385 }
 386
 387 static AVInputFormat wsaud_iformat = {
 388     "wsaud",
 389     "Westwood Studios audio format",
 390     sizeof(WsAudDemuxContext),
 391     wsaud_probe,
 392     wsaud_read_header,
 393     wsaud_read_packet,
 394     wsaud_read_close,
 395 };
 396
 397 static AVInputFormat wsvqa_iformat = {
 398     "wsvqa",
 399     "Westwood Studios VQA format",
 400     sizeof(WsVqaDemuxContext),
 401     wsvqa_probe,
 402     wsvqa_read_header,
 403     wsvqa_read_packet,
 404     wsvqa_read_close,
 405 };
 406
 407 int westwood_init(void)
 408 {
 409     av_register_input_format(&wsaud_iformat);
 410     av_register_input_format(&wsvqa_iformat);
 411     return 0;
 412 }