2 * Westwood Studios Multimedia Formats Demuxer (VQA, AUD)
3 * Copyright (c) 2003 The ffmpeg Project
5 * This file is part of FFmpeg.
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.
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.
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
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
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.
38 #define AUD_HEADER_SIZE 12
39 #define AUD_CHUNK_PREAMBLE_SIZE 8
40 #define AUD_CHUNK_SIGNATURE 0x0000DEAF
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')
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')
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
65 typedef struct WsAudDemuxContext {
70 int audio_stream_index;
71 int64_t audio_frame_counter;
74 typedef struct WsVqaDemuxContext {
79 int audio_stream_index;
80 int video_stream_index;
82 int64_t audio_frame_counter;
86 static int wsaud_probe(AVProbeData *p)
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 * compression type (8 bits) = 1 or 99 ==> 2 acceptable numbers
94 * There is a total of 24 bits. The number space contains 2^24 =
95 * 16777216 numbers. There are 40001 * 2 = 80002 acceptable combinations
96 * of numbers. There is a 80002/16777216 = 0.48% chance of a false
100 if (p->buf_size < AUD_HEADER_SIZE)
103 /* check sample rate */
104 field = AV_RL16(&p->buf[0]);
105 if ((field < 8000) || (field > 48000))
108 /* note: only check for WS IMA (type 99) right now since there is no
109 * support for type 1 */
110 if (p->buf[11] != 99)
113 /* return 1/2 certainty since this file check is a little sketchy */
114 return AVPROBE_SCORE_MAX / 2;
117 static int wsaud_read_header(AVFormatContext *s,
118 AVFormatParameters *ap)
120 WsAudDemuxContext *wsaud = s->priv_data;
121 ByteIOContext *pb = &s->pb;
123 unsigned char header[AUD_HEADER_SIZE];
125 if (get_buffer(pb, header, AUD_HEADER_SIZE) != AUD_HEADER_SIZE)
127 wsaud->audio_samplerate = AV_RL16(&header[0]);
128 if (header[11] == 99)
129 wsaud->audio_type = CODEC_ID_ADPCM_IMA_WS;
131 return AVERROR_INVALIDDATA;
133 /* flag 0 indicates stereo */
134 wsaud->audio_channels = (header[10] & 0x1) + 1;
135 /* flag 1 indicates 16 bit audio */
136 wsaud->audio_bits = (((header[10] & 0x2) >> 1) + 1) * 8;
138 /* initialize the audio decoder stream */
139 st = av_new_stream(s, 0);
141 return AVERROR_NOMEM;
142 av_set_pts_info(st, 33, 1, wsaud->audio_samplerate);
143 st->codec->codec_type = CODEC_TYPE_AUDIO;
144 st->codec->codec_id = wsaud->audio_type;
145 st->codec->codec_tag = 0; /* no tag */
146 st->codec->channels = wsaud->audio_channels;
147 st->codec->sample_rate = wsaud->audio_samplerate;
148 st->codec->bits_per_sample = wsaud->audio_bits;
149 st->codec->bit_rate = st->codec->channels * st->codec->sample_rate *
150 st->codec->bits_per_sample / 4;
151 st->codec->block_align = st->codec->channels * st->codec->bits_per_sample;
153 wsaud->audio_stream_index = st->index;
154 wsaud->audio_frame_counter = 0;
159 static int wsaud_read_packet(AVFormatContext *s,
162 WsAudDemuxContext *wsaud = s->priv_data;
163 ByteIOContext *pb = &s->pb;
164 unsigned char preamble[AUD_CHUNK_PREAMBLE_SIZE];
165 unsigned int chunk_size;
168 if (get_buffer(pb, preamble, AUD_CHUNK_PREAMBLE_SIZE) !=
169 AUD_CHUNK_PREAMBLE_SIZE)
172 /* validate the chunk */
173 if (AV_RL32(&preamble[4]) != AUD_CHUNK_SIGNATURE)
174 return AVERROR_INVALIDDATA;
176 chunk_size = AV_RL16(&preamble[0]);
177 ret= av_get_packet(pb, pkt, chunk_size);
178 if (ret != chunk_size)
180 pkt->stream_index = wsaud->audio_stream_index;
181 pkt->pts = wsaud->audio_frame_counter;
182 pkt->pts /= wsaud->audio_samplerate;
184 /* 2 samples/byte, 1 or 2 samples per frame depending on stereo */
185 wsaud->audio_frame_counter += (chunk_size * 2) / wsaud->audio_channels;
190 static int wsaud_read_close(AVFormatContext *s)
192 // WsAudDemuxContext *wsaud = s->priv_data;
198 static int wsvqa_probe(AVProbeData *p)
200 /* need 12 bytes to qualify */
201 if (p->buf_size < 12)
204 /* check for the VQA signatures */
205 if ((AV_RB32(&p->buf[0]) != FORM_TAG) ||
206 (AV_RB32(&p->buf[8]) != WVQA_TAG))
209 return AVPROBE_SCORE_MAX;
212 static int wsvqa_read_header(AVFormatContext *s,
213 AVFormatParameters *ap)
215 WsVqaDemuxContext *wsvqa = s->priv_data;
216 ByteIOContext *pb = &s->pb;
218 unsigned char *header;
219 unsigned char scratch[VQA_PREAMBLE_SIZE];
220 unsigned int chunk_tag;
221 unsigned int chunk_size;
223 /* initialize the video decoder stream */
224 st = av_new_stream(s, 0);
226 return AVERROR_NOMEM;
227 av_set_pts_info(st, 33, 1, VQA_FRAMERATE);
228 wsvqa->video_stream_index = st->index;
229 st->codec->codec_type = CODEC_TYPE_VIDEO;
230 st->codec->codec_id = CODEC_ID_WS_VQA;
231 st->codec->codec_tag = 0; /* no fourcc */
233 /* skip to the start of the VQA header */
234 url_fseek(pb, 20, SEEK_SET);
236 /* the VQA header needs to go to the decoder */
237 st->codec->extradata_size = VQA_HEADER_SIZE;
238 st->codec->extradata = av_mallocz(VQA_HEADER_SIZE + FF_INPUT_BUFFER_PADDING_SIZE);
239 header = (unsigned char *)st->codec->extradata;
240 if (get_buffer(pb, st->codec->extradata, VQA_HEADER_SIZE) !=
242 av_free(st->codec->extradata);
245 st->codec->width = AV_RL16(&header[6]);
246 st->codec->height = AV_RL16(&header[8]);
248 /* initialize the audio decoder stream for VQA v1 or nonzero samplerate */
249 if (AV_RL16(&header[24]) || (AV_RL16(&header[0]) == 1 && AV_RL16(&header[2]) == 1)) {
250 st = av_new_stream(s, 0);
252 return AVERROR_NOMEM;
253 av_set_pts_info(st, 33, 1, VQA_FRAMERATE);
254 st->codec->codec_type = CODEC_TYPE_AUDIO;
255 if (AV_RL16(&header[0]) == 1)
256 st->codec->codec_id = CODEC_ID_WESTWOOD_SND1;
258 st->codec->codec_id = CODEC_ID_ADPCM_IMA_WS;
259 st->codec->codec_tag = 0; /* no tag */
260 st->codec->sample_rate = AV_RL16(&header[24]);
261 if (!st->codec->sample_rate)
262 st->codec->sample_rate = 22050;
263 st->codec->channels = header[26];
264 if (!st->codec->channels)
265 st->codec->channels = 1;
266 st->codec->bits_per_sample = 16;
267 st->codec->bit_rate = st->codec->channels * st->codec->sample_rate *
268 st->codec->bits_per_sample / 4;
269 st->codec->block_align = st->codec->channels * st->codec->bits_per_sample;
271 wsvqa->audio_stream_index = st->index;
272 wsvqa->audio_samplerate = st->codec->sample_rate;
273 wsvqa->audio_channels = st->codec->channels;
274 wsvqa->audio_frame_counter = 0;
277 /* there are 0 or more chunks before the FINF chunk; iterate until
278 * FINF has been skipped and the file will be ready to be demuxed */
280 if (get_buffer(pb, scratch, VQA_PREAMBLE_SIZE) != VQA_PREAMBLE_SIZE) {
281 av_free(st->codec->extradata);
284 chunk_tag = AV_RB32(&scratch[0]);
285 chunk_size = AV_RB32(&scratch[4]);
287 /* catch any unknown header tags, for curiousity */
300 av_log (s, AV_LOG_ERROR, " note: unknown chunk seen (%c%c%c%c)\n",
301 scratch[0], scratch[1],
302 scratch[2], scratch[3]);
306 url_fseek(pb, chunk_size, SEEK_CUR);
307 } while (chunk_tag != FINF_TAG);
309 wsvqa->video_pts = wsvqa->audio_frame_counter = 0;
314 static int wsvqa_read_packet(AVFormatContext *s,
317 WsVqaDemuxContext *wsvqa = s->priv_data;
318 ByteIOContext *pb = &s->pb;
320 unsigned char preamble[VQA_PREAMBLE_SIZE];
321 unsigned int chunk_type;
322 unsigned int chunk_size;
325 while (get_buffer(pb, preamble, VQA_PREAMBLE_SIZE) == VQA_PREAMBLE_SIZE) {
326 chunk_type = AV_RB32(&preamble[0]);
327 chunk_size = AV_RB32(&preamble[4]);
328 skip_byte = chunk_size & 0x01;
330 if ((chunk_type == SND1_TAG) || (chunk_type == SND2_TAG) || (chunk_type == VQFR_TAG)) {
332 if (av_new_packet(pkt, chunk_size))
334 ret = get_buffer(pb, pkt->data, chunk_size);
335 if (ret != chunk_size) {
340 if (chunk_type == SND2_TAG) {
341 pkt->stream_index = wsvqa->audio_stream_index;
342 /* 2 samples/byte, 1 or 2 samples per frame depending on stereo */
343 wsvqa->audio_frame_counter += (chunk_size * 2) / wsvqa->audio_channels;
344 } else if(chunk_type == SND1_TAG) {
345 pkt->stream_index = wsvqa->audio_stream_index;
346 /* unpacked size is stored in header */
347 wsvqa->audio_frame_counter += AV_RL16(pkt->data) / wsvqa->audio_channels;
349 pkt->stream_index = wsvqa->video_stream_index;
350 wsvqa->video_pts += VQA_VIDEO_PTS_INC;
352 /* stay on 16-bit alignment */
354 url_fseek(pb, 1, SEEK_CUR);
363 av_log(s, AV_LOG_INFO, "Skipping unknown chunk 0x%08X\n", chunk_type);
365 url_fseek(pb, chunk_size + skip_byte, SEEK_CUR);
372 static int wsvqa_read_close(AVFormatContext *s)
374 // WsVqaDemuxContext *wsvqa = s->priv_data;
379 #ifdef CONFIG_WSAUD_DEMUXER
380 AVInputFormat wsaud_demuxer = {
382 "Westwood Studios audio format",
383 sizeof(WsAudDemuxContext),
390 #ifdef CONFIG_WSVQA_DEMUXER
391 AVInputFormat wsvqa_demuxer = {
393 "Westwood Studios VQA format",
394 sizeof(WsVqaDemuxContext),