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.
36 #include "libavutil/intreadwrite.h"
40 #define AUD_HEADER_SIZE 12
41 #define AUD_CHUNK_PREAMBLE_SIZE 8
42 #define AUD_CHUNK_SIGNATURE 0x0000DEAF
44 #define FORM_TAG MKBETAG('F', 'O', 'R', 'M')
45 #define WVQA_TAG MKBETAG('W', 'V', 'Q', 'A')
46 #define VQHD_TAG MKBETAG('V', 'Q', 'H', 'D')
47 #define FINF_TAG MKBETAG('F', 'I', 'N', 'F')
48 #define SND0_TAG MKBETAG('S', 'N', 'D', '0')
49 #define SND1_TAG MKBETAG('S', 'N', 'D', '1')
50 #define SND2_TAG MKBETAG('S', 'N', 'D', '2')
51 #define VQFR_TAG MKBETAG('V', 'Q', 'F', 'R')
53 /* don't know what these tags are for, but acknowledge their existence */
54 #define CINF_TAG MKBETAG('C', 'I', 'N', 'F')
55 #define CINH_TAG MKBETAG('C', 'I', 'N', 'H')
56 #define CIND_TAG MKBETAG('C', 'I', 'N', 'D')
57 #define PINF_TAG MKBETAG('P', 'I', 'N', 'F')
58 #define PINH_TAG MKBETAG('P', 'I', 'N', 'H')
59 #define PIND_TAG MKBETAG('P', 'I', 'N', 'D')
60 #define CMDS_TAG MKBETAG('C', 'M', 'D', 'S')
62 #define VQA_HEADER_SIZE 0x2A
63 #define VQA_FRAMERATE 15
64 #define VQA_PREAMBLE_SIZE 8
66 typedef struct WsAudDemuxContext {
70 enum CodecID audio_type;
71 int audio_stream_index;
72 int64_t audio_frame_counter;
75 typedef struct WsVqaDemuxContext {
80 int audio_stream_index;
81 int video_stream_index;
83 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 * 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.
100 if (p->buf_size < AUD_HEADER_SIZE + AUD_CHUNK_PREAMBLE_SIZE)
103 /* check sample rate */
104 field = AV_RL16(&p->buf[0]);
105 if ((field < 8000) || (field > 48000))
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)
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)
118 /* read ahead to the first audio chunk and validate the first header signature */
119 if (AV_RL32(&p->buf[16]) != AUD_CHUNK_SIGNATURE)
122 /* return 1/2 certainty since this file check is a little sketchy */
123 return AVPROBE_SCORE_MAX / 2;
126 static int wsaud_read_header(AVFormatContext *s,
127 AVFormatParameters *ap)
129 WsAudDemuxContext *wsaud = s->priv_data;
130 AVIOContext *pb = s->pb;
132 unsigned char header[AUD_HEADER_SIZE];
134 if (avio_read(pb, header, AUD_HEADER_SIZE) != AUD_HEADER_SIZE)
136 wsaud->audio_samplerate = AV_RL16(&header[0]);
137 if (header[11] == 99)
138 wsaud->audio_type = CODEC_ID_ADPCM_IMA_WS;
140 return AVERROR_INVALIDDATA;
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;
147 /* initialize the audio decoder stream */
148 st = avformat_new_stream(s, NULL);
150 return AVERROR(ENOMEM);
151 avpriv_set_pts_info(st, 33, 1, wsaud->audio_samplerate);
152 st->codec->codec_type = AVMEDIA_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_coded_sample = wsaud->audio_bits;
158 st->codec->bit_rate = st->codec->channels * st->codec->sample_rate *
159 st->codec->bits_per_coded_sample / 4;
160 st->codec->block_align = st->codec->channels * st->codec->bits_per_coded_sample;
162 wsaud->audio_stream_index = st->index;
163 wsaud->audio_frame_counter = 0;
168 static int wsaud_read_packet(AVFormatContext *s,
171 WsAudDemuxContext *wsaud = s->priv_data;
172 AVIOContext *pb = s->pb;
173 unsigned char preamble[AUD_CHUNK_PREAMBLE_SIZE];
174 unsigned int chunk_size;
177 if (avio_read(pb, preamble, AUD_CHUNK_PREAMBLE_SIZE) !=
178 AUD_CHUNK_PREAMBLE_SIZE)
181 /* validate the chunk */
182 if (AV_RL32(&preamble[4]) != AUD_CHUNK_SIGNATURE)
183 return AVERROR_INVALIDDATA;
185 chunk_size = AV_RL16(&preamble[0]);
186 ret= av_get_packet(pb, pkt, chunk_size);
187 if (ret != chunk_size)
189 pkt->stream_index = wsaud->audio_stream_index;
190 pkt->pts = wsaud->audio_frame_counter;
191 pkt->pts /= wsaud->audio_samplerate;
193 /* 2 samples/byte, 1 or 2 samples per frame depending on stereo */
194 wsaud->audio_frame_counter += (chunk_size * 2) / wsaud->audio_channels;
199 static int wsvqa_probe(AVProbeData *p)
201 /* need 12 bytes to qualify */
202 if (p->buf_size < 12)
205 /* check for the VQA signatures */
206 if ((AV_RB32(&p->buf[0]) != FORM_TAG) ||
207 (AV_RB32(&p->buf[8]) != WVQA_TAG))
210 return AVPROBE_SCORE_MAX;
213 static int wsvqa_read_header(AVFormatContext *s,
214 AVFormatParameters *ap)
216 WsVqaDemuxContext *wsvqa = s->priv_data;
217 AVIOContext *pb = s->pb;
219 unsigned char *header;
220 unsigned char scratch[VQA_PREAMBLE_SIZE];
221 unsigned int chunk_tag;
222 unsigned int chunk_size;
224 /* initialize the video decoder stream */
225 st = avformat_new_stream(s, NULL);
227 return AVERROR(ENOMEM);
228 avpriv_set_pts_info(st, 33, 1, VQA_FRAMERATE);
229 wsvqa->video_stream_index = st->index;
230 st->codec->codec_type = AVMEDIA_TYPE_VIDEO;
231 st->codec->codec_id = CODEC_ID_WS_VQA;
232 st->codec->codec_tag = 0; /* no fourcc */
234 /* skip to the start of the VQA header */
235 avio_seek(pb, 20, SEEK_SET);
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 (avio_read(pb, st->codec->extradata, VQA_HEADER_SIZE) !=
243 av_free(st->codec->extradata);
246 st->codec->width = AV_RL16(&header[6]);
247 st->codec->height = AV_RL16(&header[8]);
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 = avformat_new_stream(s, NULL);
253 return AVERROR(ENOMEM);
254 avpriv_set_pts_info(st, 33, 1, VQA_FRAMERATE);
255 st->codec->codec_type = AVMEDIA_TYPE_AUDIO;
256 if (AV_RL16(&header[0]) == 1)
257 st->codec->codec_id = CODEC_ID_WESTWOOD_SND1;
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_coded_sample = 16;
268 st->codec->bit_rate = st->codec->channels * st->codec->sample_rate *
269 st->codec->bits_per_coded_sample / 4;
270 st->codec->block_align = st->codec->channels * st->codec->bits_per_coded_sample;
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;
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 */
281 if (avio_read(pb, scratch, VQA_PREAMBLE_SIZE) != VQA_PREAMBLE_SIZE)
283 chunk_tag = AV_RB32(&scratch[0]);
284 chunk_size = AV_RB32(&scratch[4]);
286 /* catch any unknown header tags, for curiousity */
299 av_log (s, AV_LOG_ERROR, " note: unknown chunk seen (%c%c%c%c)\n",
300 scratch[0], scratch[1],
301 scratch[2], scratch[3]);
305 avio_skip(pb, chunk_size);
306 } while (chunk_tag != FINF_TAG);
311 static int wsvqa_read_packet(AVFormatContext *s,
314 WsVqaDemuxContext *wsvqa = s->priv_data;
315 AVIOContext *pb = s->pb;
317 unsigned char preamble[VQA_PREAMBLE_SIZE];
318 unsigned int chunk_type;
319 unsigned int chunk_size;
322 while (avio_read(pb, preamble, VQA_PREAMBLE_SIZE) == VQA_PREAMBLE_SIZE) {
323 chunk_type = AV_RB32(&preamble[0]);
324 chunk_size = AV_RB32(&preamble[4]);
326 skip_byte = chunk_size & 0x01;
328 if ((chunk_type == SND2_TAG || chunk_type == SND1_TAG) && wsvqa->audio_channels == 0) {
329 av_log(s, AV_LOG_ERROR, "audio chunk without any audio header information found\n");
330 return AVERROR_INVALIDDATA;
333 if ((chunk_type == SND1_TAG) || (chunk_type == SND2_TAG) || (chunk_type == VQFR_TAG)) {
335 ret= av_get_packet(pb, pkt, chunk_size);
339 if (chunk_type == SND2_TAG) {
340 pkt->stream_index = wsvqa->audio_stream_index;
341 /* 2 samples/byte, 1 or 2 samples per frame depending on stereo */
342 wsvqa->audio_frame_counter += (chunk_size * 2) / wsvqa->audio_channels;
343 } else if(chunk_type == SND1_TAG) {
344 pkt->stream_index = wsvqa->audio_stream_index;
345 /* unpacked size is stored in header */
346 wsvqa->audio_frame_counter += AV_RL16(pkt->data) / wsvqa->audio_channels;
348 pkt->stream_index = wsvqa->video_stream_index;
350 /* stay on 16-bit alignment */
361 av_log(s, AV_LOG_INFO, "Skipping unknown chunk 0x%08X\n", chunk_type);
363 avio_skip(pb, chunk_size + skip_byte);
370 #if CONFIG_WSAUD_DEMUXER
371 AVInputFormat ff_wsaud_demuxer = {
373 .long_name = NULL_IF_CONFIG_SMALL("Westwood Studios audio format"),
374 .priv_data_size = sizeof(WsAudDemuxContext),
375 .read_probe = wsaud_probe,
376 .read_header = wsaud_read_header,
377 .read_packet = wsaud_read_packet,
380 #if CONFIG_WSVQA_DEMUXER
381 AVInputFormat ff_wsvqa_demuxer = {
383 .long_name = NULL_IF_CONFIG_SMALL("Westwood Studios VQA format"),
384 .priv_data_size = sizeof(WsVqaDemuxContext),
385 .read_probe = wsvqa_probe,
386 .read_header = wsvqa_read_header,
387 .read_packet = wsvqa_read_packet,