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