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 #define VQA_HEADER_SIZE 0x2A
  65 #define VQA_FRAMERATE 15
  66 #define VQA_VIDEO_PTS_INC (90000 / VQA_FRAMERATE)
  67 #define VQA_PREAMBLE_SIZE 8
  68
  69 typedef struct WsAudDemuxContext {
  70     int audio_samplerate;
  71     int audio_channels;
  72     int audio_bits;
  73     int audio_type;
  74     int audio_stream_index;
  75     int64_t audio_frame_counter;
  76 } WsAudDemuxContext;
  77
  78 typedef struct WsVqaDemuxContext {
  79     int audio_samplerate;
  80     int audio_channels;
  81     int audio_bits;
  82
  83     int audio_stream_index;
  84     int video_stream_index;
  85
  86     int64_t audio_frame_counter;
  87     int64_t video_pts;
  88 } WsVqaDemuxContext;
  89
  90 static int wsaud_probe(AVProbeData *p)
  91 {
  92     int field;
  93
  94     /* Probabilistic content detection strategy: There is no file signature
  95      * so perform sanity checks on various header parameters:
  96      *   8000 <= sample rate (16 bits) <= 48000  ==> 40001 acceptable numbers
  97      *   compression type (8 bits) = 1 or 99     ==> 2 acceptable numbers
  98      * There is a total of 24 bits. The number space contains 2^24 =
  99      * 16777216 numbers. There are 40001 * 2 = 80002 acceptable combinations
 100      * of numbers. There is a 80002/16777216 = 0.48% chance of a false
 101      * positive.
 102      */
 103
 104     if (p->buf_size < AUD_HEADER_SIZE)
 105         return 0;
 106
 107     /* check sample rate */
 108     field = LE_16(&p->buf[0]);
 109     if ((field < 8000) || (field > 48000))
 110         return 0;
 111
 112     /* note: only check for WS IMA (type 99) right now since there is no
 113      * support for type 1 */
 114     if (p->buf[11] != 99)
 115         return 0;
 116
 117     /* return 1/2 certainty since this file check is a little sketchy */
 118     return AVPROBE_SCORE_MAX / 2;
 119 }
 120
 121 static int wsaud_read_header(AVFormatContext *s,
 122                              AVFormatParameters *ap)
 123 {
 124     WsAudDemuxContext *wsaud = (WsAudDemuxContext *)s->priv_data;
 125     ByteIOContext *pb = &s->pb;
 126     AVStream *st;
 127     unsigned char header[AUD_HEADER_SIZE];
 128
 129     if (get_buffer(pb, header, AUD_HEADER_SIZE) != AUD_HEADER_SIZE)
 130         return -EIO;
 131     wsaud->audio_samplerate = LE_16(&header[0]);
 132     if (header[11] == 99)
 133         wsaud->audio_type = CODEC_ID_ADPCM_IMA_WS;
 134     else
 135         return AVERROR_INVALIDDATA;
 136
 137     /* flag 0 indicates stereo */
 138     wsaud->audio_channels = (header[10] & 0x1) + 1;
 139     /* flag 1 indicates 16 bit audio */
 140     wsaud->audio_bits = (((header[10] & 0x2) >> 1) + 1) * 8;
 141
 142     /* set the pts reference the same as the sample rate */
 143     s->pts_num = 1;
 144     s->pts_den = wsaud->audio_samplerate;
 145
 146     /* initialize the audio decoder stream */
 147     st = av_new_stream(s, 0);
 148     if (!st)
 149         return AVERROR_NOMEM;
 150     st->codec.codec_type = CODEC_TYPE_AUDIO;
 151     st->codec.codec_id = wsaud->audio_type;
 152     st->codec.codec_tag = 0;  /* no tag */
 153     st->codec.channels = wsaud->audio_channels;
 154     st->codec.sample_rate = wsaud->audio_samplerate;
 155     st->codec.bits_per_sample = wsaud->audio_bits;
 156     st->codec.bit_rate = st->codec.channels * st->codec.sample_rate *
 157         st->codec.bits_per_sample / 4;
 158     st->codec.block_align = st->codec.channels * st->codec.bits_per_sample;
 159
 160     wsaud->audio_stream_index = st->index;
 161     wsaud->audio_frame_counter = 0;
 162
 163     return 0;
 164 }
 165
 166 static int wsaud_read_packet(AVFormatContext *s,
 167                              AVPacket *pkt)
 168 {
 169     WsAudDemuxContext *wsaud = (WsAudDemuxContext *)s->priv_data;
 170     ByteIOContext *pb = &s->pb;
 171     unsigned char preamble[AUD_CHUNK_PREAMBLE_SIZE];
 172     unsigned int chunk_size;
 173     int ret = 0;
 174
 175     if (get_buffer(pb, preamble, AUD_CHUNK_PREAMBLE_SIZE) !=
 176         AUD_CHUNK_PREAMBLE_SIZE)
 177         return -EIO;
 178
 179     /* validate the chunk */
 180     if (LE_32(&preamble[4]) != AUD_CHUNK_SIGNATURE)
 181         return AVERROR_INVALIDDATA;
 182
 183     chunk_size = LE_16(&preamble[0]);
 184     if (av_new_packet(pkt, chunk_size))
 185         return -EIO;
 186     pkt->stream_index = wsaud->audio_stream_index;
 187     pkt->pts = wsaud->audio_frame_counter;
 188     pkt->pts /= wsaud->audio_samplerate;
 189     if ((ret = get_buffer(pb, pkt->data, chunk_size)) != chunk_size) {
 190         av_free_packet(pkt);
 191         ret = -EIO;
 192     }
 193
 194     /* 2 samples/byte, 1 or 2 samples per frame depending on stereo */
 195     wsaud->audio_frame_counter += (chunk_size * 2) / wsaud->audio_channels;
 196
 197     return ret;
 198 }
 199
 200 static int wsaud_read_close(AVFormatContext *s)
 201 {
 202 //    WsAudDemuxContext *wsaud = (WsAudDemuxContext *)s->priv_data;
 203
 204     return 0;
 205 }
 206
 207
 208 static int wsvqa_probe(AVProbeData *p)
 209 {
 210     /* need 12 bytes to qualify */
 211     if (p->buf_size < 12)
 212         return 0;
 213
 214     /* check for the VQA signatures */
 215     if ((BE_32(&p->buf[0]) != FORM_TAG) ||
 216         (BE_32(&p->buf[8]) != WVQA_TAG))
 217         return 0;
 218
 219     return AVPROBE_SCORE_MAX;
 220 }
 221
 222 static int wsvqa_read_header(AVFormatContext *s,
 223                              AVFormatParameters *ap)
 224 {
 225     WsVqaDemuxContext *wsvqa = (WsVqaDemuxContext *)s->priv_data;
 226     ByteIOContext *pb = &s->pb;
 227     AVStream *st;
 228     unsigned char *header;
 229     unsigned char scratch[VQA_PREAMBLE_SIZE];
 230
 231     /* set the pts reference (1 pts = 1/90000) */
 232     s->pts_num = 1;
 233     s->pts_den = 90000;
 234
 235     /* initialize the video decoder stream */
 236     st = av_new_stream(s, 0);
 237     if (!st)
 238         return AVERROR_NOMEM;
 239     wsvqa->video_stream_index = st->index;
 240     st->codec.codec_type = CODEC_TYPE_VIDEO;
 241     st->codec.codec_id = CODEC_ID_WS_VQA;
 242     st->codec.codec_tag = 0;  /* no fourcc */
 243
 244     /* skip to the start of the VQA header */
 245     url_fseek(pb, 20, SEEK_SET);
 246
 247     /* the VQA header needs to go to the decoder */
 248     st->codec.extradata_size = VQA_HEADER_SIZE;
 249     st->codec.extradata = av_malloc(VQA_HEADER_SIZE);
 250     header = (unsigned char *)st->codec.extradata;
 251     if (get_buffer(pb, st->codec.extradata, VQA_HEADER_SIZE) !=
 252         VQA_HEADER_SIZE) {
 253         av_free(st->codec.extradata);
 254         return -EIO;
 255     }
 256     st->codec.width = LE_16(&header[6]);
 257     st->codec.height = LE_16(&header[8]);
 258
 259     /* initialize the audio decoder stream */
 260     st = av_new_stream(s, 0);
 261     if (!st)
 262         return AVERROR_NOMEM;
 263     st->codec.codec_type = CODEC_TYPE_AUDIO;
 264     st->codec.codec_id = CODEC_ID_ADPCM_IMA_WS;
 265     st->codec.codec_tag = 0;  /* no tag */
 266     st->codec.sample_rate = LE_16(&header[24]);
 267     st->codec.channels = header[26];
 268     st->codec.bits_per_sample = 16;
 269     st->codec.bit_rate = st->codec.channels * st->codec.sample_rate *
 270         st->codec.bits_per_sample / 4;
 271     st->codec.block_align = st->codec.channels * st->codec.bits_per_sample;
 272
 273     wsvqa->audio_stream_index = st->index;
 274     wsvqa->audio_samplerate = st->codec.sample_rate;
 275     wsvqa->audio_channels = st->codec.channels;
 276     wsvqa->audio_frame_counter = 0;
 277
 278     /* skip the useless FINF chunk index */
 279     if (get_buffer(pb, scratch, VQA_PREAMBLE_SIZE) != VQA_PREAMBLE_SIZE) {
 280         av_free(st->codec.extradata);
 281         return -EIO;
 282     }
 283     url_fseek(pb, BE_32(&scratch[4]), SEEK_CUR);
 284     wsvqa->video_pts = wsvqa->audio_frame_counter = 0;
 285
 286     return 0;
 287 }
 288
 289 static int wsvqa_read_packet(AVFormatContext *s,
 290                              AVPacket *pkt)
 291 {
 292     WsVqaDemuxContext *wsvqa = (WsVqaDemuxContext *)s->priv_data;
 293     ByteIOContext *pb = &s->pb;
 294     int ret = 0;
 295     unsigned char preamble[VQA_PREAMBLE_SIZE];
 296     unsigned int chunk_type;
 297     unsigned int chunk_size;
 298     int skip_byte;
 299
 300     if (get_buffer(pb, preamble, VQA_PREAMBLE_SIZE) != VQA_PREAMBLE_SIZE)
 301         return -EIO;
 302
 303     chunk_type = BE_32(&preamble[0]);
 304     chunk_size = BE_32(&preamble[4]);
 305     skip_byte = chunk_size & 0x01;
 306
 307     if ((chunk_type == SND2_TAG) || (chunk_type == VQFR_TAG)) {
 308
 309         if (av_new_packet(pkt, chunk_size))
 310             return -EIO;
 311         ret = get_buffer(pb, pkt->data, chunk_size);
 312         if (ret != chunk_size) {
 313             av_free_packet(pkt);
 314             ret = -EIO;
 315         }
 316
 317         if (chunk_type == SND2_TAG) {
 318             pkt->stream_index = wsvqa->audio_stream_index;
 319
 320             pkt->pts = 90000;
 321             pkt->pts *= wsvqa->audio_frame_counter;
 322             pkt->pts /= wsvqa->audio_samplerate;
 323
 324             /* 2 samples/byte, 1 or 2 samples per frame depending on stereo */
 325             wsvqa->audio_frame_counter += (chunk_size * 2) /
 326                 wsvqa->audio_channels;
 327         } else {
 328             pkt->stream_index = wsvqa->video_stream_index;
 329             pkt->pts = wsvqa->video_pts;
 330             wsvqa->video_pts += VQA_VIDEO_PTS_INC;
 331         }
 332
 333     } else
 334         return AVERROR_INVALIDDATA;
 335
 336     /* stay on 16-bit alignment */
 337     if (skip_byte)
 338         url_fseek(pb, 1, SEEK_CUR);
 339
 340     return ret;
 341 }
 342
 343 static int wsvqa_read_close(AVFormatContext *s)
 344 {
 345 //    WsVqaDemuxContext *wsvqa = (WsVqaDemuxContext *)s->priv_data;
 346
 347     return 0;
 348 }
 349
 350 static AVInputFormat wsaud_iformat = {
 351     "wsaud",
 352     "Westwood Studios audio format",
 353     sizeof(WsAudDemuxContext),
 354     wsaud_probe,
 355     wsaud_read_header,
 356     wsaud_read_packet,
 357     wsaud_read_close,
 358 };
 359
 360 static AVInputFormat wsvqa_iformat = {
 361     "wsvqa",
 362     "Westwood Studios VQA format",
 363     sizeof(WsVqaDemuxContext),
 364     wsvqa_probe,
 365     wsvqa_read_header,
 366     wsvqa_read_packet,
 367     wsvqa_read_close,
 368 };
 369
 370 int westwood_init(void)
 371 {
 372     av_register_input_format(&wsaud_iformat);
 373     av_register_input_format(&wsvqa_iformat);
 374     return 0;
 375 }