libavformat/westwood.c

   1 /*
   2  * Westwood Studios Multimedia Formats Demuxer (VQA, AUD)
   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 westwood.c
  24  * Westwood Studios VQA & AUD file demuxers
  25  * by Mike Melanson (melanson@pcisys.net)
  26  * for more information on the Westwood file formats, visit:
  27  *   http://www.pcisys.net/~melanson/codecs/
  28  *   http://www.geocities.com/SiliconValley/8682/aud3.txt
  29  *
  30  * Implementation note: There is no definite file signature for AUD files.
  31  * The demuxer uses a probabilistic strategy for content detection. This
  32  * entails performing sanity checks on certain header values in order to
  33  * qualify a file. Refer to wsaud_probe() for the precise parameters.
  34  */
  35
  36 #include "avformat.h"
  37
  38 #define AUD_HEADER_SIZE 12
  39 #define AUD_CHUNK_PREAMBLE_SIZE 8
  40 #define AUD_CHUNK_SIGNATURE 0x0000DEAF
  41
  42 #define FORM_TAG MKBETAG('F', 'O', 'R', 'M')
  43 #define WVQA_TAG MKBETAG('W', 'V', 'Q', 'A')
  44 #define VQHD_TAG MKBETAG('V', 'Q', 'H', 'D')
  45 #define FINF_TAG MKBETAG('F', 'I', 'N', 'F')
  46 #define SND0_TAG MKBETAG('S', 'N', 'D', '0')
  47 #define SND1_TAG MKBETAG('S', 'N', 'D', '1')
  48 #define SND2_TAG MKBETAG('S', 'N', 'D', '2')
  49 #define VQFR_TAG MKBETAG('V', 'Q', 'F', 'R')
  50
  51 /* don't know what these tags are for, but acknowledge their existence */
  52 #define CINF_TAG MKBETAG('C', 'I', 'N', 'F')
  53 #define CINH_TAG MKBETAG('C', 'I', 'N', 'H')
  54 #define CIND_TAG MKBETAG('C', 'I', 'N', 'D')
  55 #define PINF_TAG MKBETAG('P', 'I', 'N', 'F')
  56 #define PINH_TAG MKBETAG('P', 'I', 'N', 'H')
  57 #define PIND_TAG MKBETAG('P', 'I', 'N', 'D')
  58 #define CMDS_TAG MKBETAG('C', 'M', 'D', 'S')
  59
  60 #define VQA_HEADER_SIZE 0x2A
  61 #define VQA_FRAMERATE 15
  62 #define VQA_VIDEO_PTS_INC (90000 / VQA_FRAMERATE)
  63 #define VQA_PREAMBLE_SIZE 8
  64
  65 typedef struct WsAudDemuxContext {
  66     int audio_samplerate;
  67     int audio_channels;
  68     int audio_bits;
  69     int audio_type;
  70     int audio_stream_index;
  71     int64_t audio_frame_counter;
  72 } WsAudDemuxContext;
  73
  74 typedef struct WsVqaDemuxContext {
  75     int audio_samplerate;
  76     int audio_channels;
  77     int audio_bits;
  78
  79     int audio_stream_index;
  80     int video_stream_index;
  81
  82     int64_t audio_frame_counter;
  83     int64_t video_pts;
  84 } WsVqaDemuxContext;
  85
  86 static int wsaud_probe(AVProbeData *p)
  87 {
  88     int field;
  89
  90     /* Probabilistic content detection strategy: There is no file signature
  91      * so perform sanity checks on various header parameters:
  92      *   8000 <= sample rate (16 bits) <= 48000  ==> 40001 acceptable numbers
  93      *   flags <= 0x03 (2 LSBs are used)         ==> 4 acceptable numbers
  94      *   compression type (8 bits) = 1 or 99     ==> 2 acceptable numbers
  95      *   first audio chunk signature (32 bits)   ==> 1 acceptable number
  96      * The number space contains 2^64 numbers. There are 40001 * 4 * 2 * 1 =
  97      * 320008 acceptable number combinations.
  98      */
  99
 100     if (p->buf_size < AUD_HEADER_SIZE + AUD_CHUNK_PREAMBLE_SIZE)
 101         return 0;
 102
 103     /* check sample rate */
 104     field = AV_RL16(&p->buf[0]);
 105     if ((field < 8000) || (field > 48000))
 106         return 0;
 107
 108     /* enforce the rule that the top 6 bits of this flags field are reserved (0);
 109      * this might not be true, but enforce it until deemed unnecessary */
 110     if (p->buf[10] & 0xFC)
 111         return 0;
 112
 113     /* note: only check for WS IMA (type 99) right now since there is no
 114      * support for type 1 */
 115     if (p->buf[11] != 99)
 116         return 0;
 117
 118     /* read ahead to the first audio chunk and validate the first header signature */
 119     if (AV_RL32(&p->buf[16]) != AUD_CHUNK_SIGNATURE)
 120         return 0;
 121
 122     /* return 1/2 certainty since this file check is a little sketchy */
 123     return AVPROBE_SCORE_MAX / 2;
 124 }
 125
 126 static int wsaud_read_header(AVFormatContext *s,
 127                              AVFormatParameters *ap)
 128 {
 129     WsAudDemuxContext *wsaud = s->priv_data;
 130     ByteIOContext *pb = s->pb;
 131     AVStream *st;
 132     unsigned char header[AUD_HEADER_SIZE];
 133
 134     if (get_buffer(pb, header, AUD_HEADER_SIZE) != AUD_HEADER_SIZE)
 135         return AVERROR(EIO);
 136     wsaud->audio_samplerate = AV_RL16(&header[0]);
 137     if (header[11] == 99)
 138         wsaud->audio_type = CODEC_ID_ADPCM_IMA_WS;
 139     else
 140         return AVERROR_INVALIDDATA;
 141
 142     /* flag 0 indicates stereo */
 143     wsaud->audio_channels = (header[10] & 0x1) + 1;
 144     /* flag 1 indicates 16 bit audio */
 145     wsaud->audio_bits = (((header[10] & 0x2) >> 1) + 1) * 8;
 146
 147     /* initialize the audio decoder stream */
 148     st = av_new_stream(s, 0);
 149     if (!st)
 150         return AVERROR(ENOMEM);
 151     av_set_pts_info(st, 33, 1, wsaud->audio_samplerate);
 152     st->codec->codec_type = CODEC_TYPE_AUDIO;
 153     st->codec->codec_id = wsaud->audio_type;
 154     st->codec->codec_tag = 0;  /* no tag */
 155     st->codec->channels = wsaud->audio_channels;
 156     st->codec->sample_rate = wsaud->audio_samplerate;
 157     st->codec->bits_per_sample = wsaud->audio_bits;
 158     st->codec->bit_rate = st->codec->channels * st->codec->sample_rate *
 159         st->codec->bits_per_sample / 4;
 160     st->codec->block_align = st->codec->channels * st->codec->bits_per_sample;
 161
 162     wsaud->audio_stream_index = st->index;
 163     wsaud->audio_frame_counter = 0;
 164
 165     return 0;
 166 }
 167
 168 static int wsaud_read_packet(AVFormatContext *s,
 169                              AVPacket *pkt)
 170 {
 171     WsAudDemuxContext *wsaud = s->priv_data;
 172     ByteIOContext *pb = s->pb;
 173     unsigned char preamble[AUD_CHUNK_PREAMBLE_SIZE];
 174     unsigned int chunk_size;
 175     int ret = 0;
 176
 177     if (get_buffer(pb, preamble, AUD_CHUNK_PREAMBLE_SIZE) !=
 178         AUD_CHUNK_PREAMBLE_SIZE)
 179         return AVERROR(EIO);
 180
 181     /* validate the chunk */
 182     if (AV_RL32(&preamble[4]) != AUD_CHUNK_SIGNATURE)
 183         return AVERROR_INVALIDDATA;
 184
 185     chunk_size = AV_RL16(&preamble[0]);
 186     ret= av_get_packet(pb, pkt, chunk_size);
 187     if (ret != chunk_size)
 188         return AVERROR(EIO);
 189     pkt->stream_index = wsaud->audio_stream_index;
 190     pkt->pts = wsaud->audio_frame_counter;
 191     pkt->pts /= wsaud->audio_samplerate;
 192
 193     /* 2 samples/byte, 1 or 2 samples per frame depending on stereo */
 194     wsaud->audio_frame_counter += (chunk_size * 2) / wsaud->audio_channels;
 195
 196     return ret;
 197 }
 198
 199 static int wsvqa_probe(AVProbeData *p)
 200 {
 201     /* need 12 bytes to qualify */
 202     if (p->buf_size < 12)
 203         return 0;
 204
 205     /* check for the VQA signatures */
 206     if ((AV_RB32(&p->buf[0]) != FORM_TAG) ||
 207         (AV_RB32(&p->buf[8]) != WVQA_TAG))
 208         return 0;
 209
 210     return AVPROBE_SCORE_MAX;
 211 }
 212
 213 static int wsvqa_read_header(AVFormatContext *s,
 214                              AVFormatParameters *ap)
 215 {
 216     WsVqaDemuxContext *wsvqa = s->priv_data;
 217     ByteIOContext *pb = s->pb;
 218     AVStream *st;
 219     unsigned char *header;
 220     unsigned char scratch[VQA_PREAMBLE_SIZE];
 221     unsigned int chunk_tag;
 222     unsigned int chunk_size;
 223
 224     /* initialize the video decoder stream */
 225     st = av_new_stream(s, 0);
 226     if (!st)
 227         return AVERROR(ENOMEM);
 228     av_set_pts_info(st, 33, 1, VQA_FRAMERATE);
 229     wsvqa->video_stream_index = st->index;
 230     st->codec->codec_type = CODEC_TYPE_VIDEO;
 231     st->codec->codec_id = CODEC_ID_WS_VQA;
 232     st->codec->codec_tag = 0;  /* no fourcc */
 233
 234     /* skip to the start of the VQA header */
 235     url_fseek(pb, 20, SEEK_SET);
 236
 237     /* the VQA header needs to go to the decoder */
 238     st->codec->extradata_size = VQA_HEADER_SIZE;
 239     st->codec->extradata = av_mallocz(VQA_HEADER_SIZE + FF_INPUT_BUFFER_PADDING_SIZE);
 240     header = (unsigned char *)st->codec->extradata;
 241     if (get_buffer(pb, st->codec->extradata, VQA_HEADER_SIZE) !=
 242         VQA_HEADER_SIZE) {
 243         av_free(st->codec->extradata);
 244         return AVERROR(EIO);
 245     }
 246     st->codec->width = AV_RL16(&header[6]);
 247     st->codec->height = AV_RL16(&header[8]);
 248
 249     /* initialize the audio decoder stream for VQA v1 or nonzero samplerate */
 250     if (AV_RL16(&header[24]) || (AV_RL16(&header[0]) == 1 && AV_RL16(&header[2]) == 1)) {
 251         st = av_new_stream(s, 0);
 252         if (!st)
 253             return AVERROR(ENOMEM);
 254         av_set_pts_info(st, 33, 1, VQA_FRAMERATE);
 255         st->codec->codec_type = CODEC_TYPE_AUDIO;
 256         if (AV_RL16(&header[0]) == 1)
 257             st->codec->codec_id = CODEC_ID_WESTWOOD_SND1;
 258         else
 259             st->codec->codec_id = CODEC_ID_ADPCM_IMA_WS;
 260         st->codec->codec_tag = 0;  /* no tag */
 261         st->codec->sample_rate = AV_RL16(&header[24]);
 262         if (!st->codec->sample_rate)
 263             st->codec->sample_rate = 22050;
 264         st->codec->channels = header[26];
 265         if (!st->codec->channels)
 266             st->codec->channels = 1;
 267         st->codec->bits_per_sample = 16;
 268         st->codec->bit_rate = st->codec->channels * st->codec->sample_rate *
 269             st->codec->bits_per_sample / 4;
 270         st->codec->block_align = st->codec->channels * st->codec->bits_per_sample;
 271
 272         wsvqa->audio_stream_index = st->index;
 273         wsvqa->audio_samplerate = st->codec->sample_rate;
 274         wsvqa->audio_channels = st->codec->channels;
 275         wsvqa->audio_frame_counter = 0;
 276     }
 277
 278     /* there are 0 or more chunks before the FINF chunk; iterate until
 279      * FINF has been skipped and the file will be ready to be demuxed */
 280     do {
 281         if (get_buffer(pb, scratch, VQA_PREAMBLE_SIZE) != VQA_PREAMBLE_SIZE) {
 282             av_free(st->codec->extradata);
 283             return AVERROR(EIO);
 284         }
 285         chunk_tag = AV_RB32(&scratch[0]);
 286         chunk_size = AV_RB32(&scratch[4]);
 287
 288         /* catch any unknown header tags, for curiousity */
 289         switch (chunk_tag) {
 290         case CINF_TAG:
 291         case CINH_TAG:
 292         case CIND_TAG:
 293         case PINF_TAG:
 294         case PINH_TAG:
 295         case PIND_TAG:
 296         case FINF_TAG:
 297         case CMDS_TAG:
 298             break;
 299
 300         default:
 301             av_log (s, AV_LOG_ERROR, " note: unknown chunk seen (%c%c%c%c)\n",
 302                 scratch[0], scratch[1],
 303                 scratch[2], scratch[3]);
 304             break;
 305         }
 306
 307         url_fseek(pb, chunk_size, SEEK_CUR);
 308     } while (chunk_tag != FINF_TAG);
 309
 310     wsvqa->video_pts = wsvqa->audio_frame_counter = 0;
 311
 312     return 0;
 313 }
 314
 315 static int wsvqa_read_packet(AVFormatContext *s,
 316                              AVPacket *pkt)
 317 {
 318     WsVqaDemuxContext *wsvqa = s->priv_data;
 319     ByteIOContext *pb = s->pb;
 320     int ret = -1;
 321     unsigned char preamble[VQA_PREAMBLE_SIZE];
 322     unsigned int chunk_type;
 323     unsigned int chunk_size;
 324     int skip_byte;
 325
 326     while (get_buffer(pb, preamble, VQA_PREAMBLE_SIZE) == VQA_PREAMBLE_SIZE) {
 327         chunk_type = AV_RB32(&preamble[0]);
 328         chunk_size = AV_RB32(&preamble[4]);
 329         skip_byte = chunk_size & 0x01;
 330
 331         if ((chunk_type == SND1_TAG) || (chunk_type == SND2_TAG) || (chunk_type == VQFR_TAG)) {
 332
 333             if (av_new_packet(pkt, chunk_size))
 334                 return AVERROR(EIO);
 335             ret = get_buffer(pb, pkt->data, chunk_size);
 336             if (ret != chunk_size) {
 337                 av_free_packet(pkt);
 338                 return AVERROR(EIO);
 339             }
 340
 341             if (chunk_type == SND2_TAG) {
 342                 pkt->stream_index = wsvqa->audio_stream_index;
 343                 /* 2 samples/byte, 1 or 2 samples per frame depending on stereo */
 344                 wsvqa->audio_frame_counter += (chunk_size * 2) / wsvqa->audio_channels;
 345             } else if(chunk_type == SND1_TAG) {
 346                 pkt->stream_index = wsvqa->audio_stream_index;
 347                 /* unpacked size is stored in header */
 348                 wsvqa->audio_frame_counter += AV_RL16(pkt->data) / wsvqa->audio_channels;
 349             } else {
 350                 pkt->stream_index = wsvqa->video_stream_index;
 351                 wsvqa->video_pts += VQA_VIDEO_PTS_INC;
 352             }
 353             /* stay on 16-bit alignment */
 354             if (skip_byte)
 355                 url_fseek(pb, 1, SEEK_CUR);
 356
 357             return ret;
 358         } else {
 359             switch(chunk_type){
 360             case CMDS_TAG:
 361             case SND0_TAG:
 362                 break;
 363             default:
 364                 av_log(s, AV_LOG_INFO, "Skipping unknown chunk 0x%08X\n", chunk_type);
 365             }
 366             url_fseek(pb, chunk_size + skip_byte, SEEK_CUR);
 367         }
 368     }
 369
 370     return ret;
 371 }
 372
 373 #ifdef CONFIG_WSAUD_DEMUXER
 374 AVInputFormat wsaud_demuxer = {
 375     "wsaud",
 376     NULL_IF_CONFIG_SMALL("Westwood Studios audio format"),
 377     sizeof(WsAudDemuxContext),
 378     wsaud_probe,
 379     wsaud_read_header,
 380     wsaud_read_packet,
 381 };
 382 #endif
 383 #ifdef CONFIG_WSVQA_DEMUXER
 384 AVInputFormat wsvqa_demuxer = {
 385     "wsvqa",
 386     NULL_IF_CONFIG_SMALL("Westwood Studios VQA format"),
 387     sizeof(WsVqaDemuxContext),
 388     wsvqa_probe,
 389     wsvqa_read_header,
 390     wsvqa_read_packet,
 391 };
 392 #endif