You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.

351 lines
12KB

  1. /*
  2. * id Quake II CIN File Demuxer
  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. * @file
  23. * id Quake II CIN file demuxer by Mike Melanson (melanson@pcisys.net)
  24. * For more information about the id CIN format, visit:
  25. * http://www.csse.monash.edu.au/~timf/
  26. *
  27. * CIN is a somewhat quirky and ill-defined format. Here are some notes
  28. * for anyone trying to understand the technical details of this format:
  29. *
  30. * The format has no definite file signature. This is problematic for a
  31. * general-purpose media player that wants to automatically detect file
  32. * types. However, a CIN file does start with 5 32-bit numbers that
  33. * specify audio and video parameters. This demuxer gets around the lack
  34. * of file signature by performing sanity checks on those parameters.
  35. * Probabalistically, this is a reasonable solution since the number of
  36. * valid combinations of the 5 parameters is a very small subset of the
  37. * total 160-bit number space.
  38. *
  39. * Refer to the function idcin_probe() for the precise A/V parameters
  40. * that this demuxer allows.
  41. *
  42. * Next, each audio and video frame has a duration of 1/14 sec. If the
  43. * audio sample rate is a multiple of the common frequency 22050 Hz it will
  44. * divide evenly by 14. However, if the sample rate is 11025 Hz:
  45. * 11025 (samples/sec) / 14 (frames/sec) = 787.5 (samples/frame)
  46. * The way the CIN stores audio in this case is by storing 787 sample
  47. * frames in the first audio frame and 788 sample frames in the second
  48. * audio frame. Therefore, the total number of bytes in an audio frame
  49. * is given as:
  50. * audio frame #0: 787 * (bytes/sample) * (# channels) bytes in frame
  51. * audio frame #1: 788 * (bytes/sample) * (# channels) bytes in frame
  52. * audio frame #2: 787 * (bytes/sample) * (# channels) bytes in frame
  53. * audio frame #3: 788 * (bytes/sample) * (# channels) bytes in frame
  54. *
  55. * Finally, not all id CIN creation tools agree on the resolution of the
  56. * color palette, apparently. Some creation tools specify red, green, and
  57. * blue palette components in terms of 6-bit VGA color DAC values which
  58. * range from 0..63. Other tools specify the RGB components as full 8-bit
  59. * values that range from 0..255. Since there are no markers in the file to
  60. * differentiate between the two variants, this demuxer uses the following
  61. * heuristic:
  62. * - load the 768 palette bytes from disk
  63. * - assume that they will need to be shifted left by 2 bits to
  64. * transform them from 6-bit values to 8-bit values
  65. * - scan through all 768 palette bytes
  66. * - if any bytes exceed 63, do not shift the bytes at all before
  67. * transmitting them to the video decoder
  68. */
  69. #include "libavutil/audioconvert.h"
  70. #include "libavutil/imgutils.h"
  71. #include "libavutil/intreadwrite.h"
  72. #include "avformat.h"
  73. #include "internal.h"
  74. #define HUFFMAN_TABLE_SIZE (64 * 1024)
  75. #define IDCIN_FPS 14
  76. typedef struct IdcinDemuxContext {
  77. int video_stream_index;
  78. int audio_stream_index;
  79. int audio_chunk_size1;
  80. int audio_chunk_size2;
  81. int block_align;
  82. /* demux state variables */
  83. int current_audio_chunk;
  84. int next_chunk_is_video;
  85. int audio_present;
  86. int64_t first_pkt_pos;
  87. } IdcinDemuxContext;
  88. static int idcin_probe(AVProbeData *p)
  89. {
  90. unsigned int number;
  91. /*
  92. * This is what you could call a "probabilistic" file check: id CIN
  93. * files don't have a definite file signature. In lieu of such a marker,
  94. * perform sanity checks on the 5 32-bit header fields:
  95. * width, height: greater than 0, less than or equal to 1024
  96. * audio sample rate: greater than or equal to 8000, less than or
  97. * equal to 48000, or 0 for no audio
  98. * audio sample width (bytes/sample): 0 for no audio, or 1 or 2
  99. * audio channels: 0 for no audio, or 1 or 2
  100. */
  101. /* check we have enough data to do all checks, otherwise the
  102. 0-padding may cause a wrong recognition */
  103. if (p->buf_size < 20)
  104. return 0;
  105. /* check the video width */
  106. number = AV_RL32(&p->buf[0]);
  107. if ((number == 0) || (number > 1024))
  108. return 0;
  109. /* check the video height */
  110. number = AV_RL32(&p->buf[4]);
  111. if ((number == 0) || (number > 1024))
  112. return 0;
  113. /* check the audio sample rate */
  114. number = AV_RL32(&p->buf[8]);
  115. if ((number != 0) && ((number < 8000) | (number > 48000)))
  116. return 0;
  117. /* check the audio bytes/sample */
  118. number = AV_RL32(&p->buf[12]);
  119. if (number > 2)
  120. return 0;
  121. /* check the audio channels */
  122. number = AV_RL32(&p->buf[16]);
  123. if (number > 2)
  124. return 0;
  125. /* return half certainly since this check is a bit sketchy */
  126. return AVPROBE_SCORE_MAX / 2;
  127. }
  128. static int idcin_read_header(AVFormatContext *s)
  129. {
  130. AVIOContext *pb = s->pb;
  131. IdcinDemuxContext *idcin = s->priv_data;
  132. AVStream *st;
  133. unsigned int width, height;
  134. unsigned int sample_rate, bytes_per_sample, channels;
  135. /* get the 5 header parameters */
  136. width = avio_rl32(pb);
  137. height = avio_rl32(pb);
  138. sample_rate = avio_rl32(pb);
  139. bytes_per_sample = avio_rl32(pb);
  140. channels = avio_rl32(pb);
  141. if (av_image_check_size(width, height, 0, s) < 0)
  142. return AVERROR_INVALIDDATA;
  143. if (sample_rate > 0) {
  144. if (sample_rate < 14 || sample_rate > INT_MAX) {
  145. av_log(s, AV_LOG_ERROR, "invalid sample rate: %u\n", sample_rate);
  146. return AVERROR_INVALIDDATA;
  147. }
  148. if (bytes_per_sample < 1 || bytes_per_sample > 2) {
  149. av_log(s, AV_LOG_ERROR, "invalid bytes per sample: %u\n",
  150. bytes_per_sample);
  151. return AVERROR_INVALIDDATA;
  152. }
  153. if (channels < 1 || channels > 2) {
  154. av_log(s, AV_LOG_ERROR, "invalid channels: %u\n", channels);
  155. return AVERROR_INVALIDDATA;
  156. }
  157. idcin->audio_present = 1;
  158. } else {
  159. /* if sample rate is 0, assume no audio */
  160. idcin->audio_present = 0;
  161. }
  162. st = avformat_new_stream(s, NULL);
  163. if (!st)
  164. return AVERROR(ENOMEM);
  165. avpriv_set_pts_info(st, 33, 1, IDCIN_FPS);
  166. st->start_time = 0;
  167. idcin->video_stream_index = st->index;
  168. st->codec->codec_type = AVMEDIA_TYPE_VIDEO;
  169. st->codec->codec_id = AV_CODEC_ID_IDCIN;
  170. st->codec->codec_tag = 0; /* no fourcc */
  171. st->codec->width = width;
  172. st->codec->height = height;
  173. /* load up the Huffman tables into extradata */
  174. st->codec->extradata_size = HUFFMAN_TABLE_SIZE;
  175. st->codec->extradata = av_malloc(HUFFMAN_TABLE_SIZE);
  176. if (avio_read(pb, st->codec->extradata, HUFFMAN_TABLE_SIZE) !=
  177. HUFFMAN_TABLE_SIZE)
  178. return AVERROR(EIO);
  179. if (idcin->audio_present) {
  180. idcin->audio_present = 1;
  181. st = avformat_new_stream(s, NULL);
  182. if (!st)
  183. return AVERROR(ENOMEM);
  184. avpriv_set_pts_info(st, 63, 1, sample_rate);
  185. st->start_time = 0;
  186. idcin->audio_stream_index = st->index;
  187. st->codec->codec_type = AVMEDIA_TYPE_AUDIO;
  188. st->codec->codec_tag = 1;
  189. st->codec->channels = channels;
  190. st->codec->channel_layout = channels > 1 ? AV_CH_LAYOUT_STEREO :
  191. AV_CH_LAYOUT_MONO;
  192. st->codec->sample_rate = sample_rate;
  193. st->codec->bits_per_coded_sample = bytes_per_sample * 8;
  194. st->codec->bit_rate = sample_rate * bytes_per_sample * 8 * channels;
  195. st->codec->block_align = idcin->block_align = bytes_per_sample * channels;
  196. if (bytes_per_sample == 1)
  197. st->codec->codec_id = AV_CODEC_ID_PCM_U8;
  198. else
  199. st->codec->codec_id = AV_CODEC_ID_PCM_S16LE;
  200. if (sample_rate % 14 != 0) {
  201. idcin->audio_chunk_size1 = (sample_rate / 14) *
  202. bytes_per_sample * channels;
  203. idcin->audio_chunk_size2 = (sample_rate / 14 + 1) *
  204. bytes_per_sample * channels;
  205. } else {
  206. idcin->audio_chunk_size1 = idcin->audio_chunk_size2 =
  207. (sample_rate / 14) * bytes_per_sample * channels;
  208. }
  209. idcin->current_audio_chunk = 0;
  210. }
  211. idcin->next_chunk_is_video = 1;
  212. idcin->first_pkt_pos = avio_tell(s->pb);
  213. return 0;
  214. }
  215. static int idcin_read_packet(AVFormatContext *s,
  216. AVPacket *pkt)
  217. {
  218. int ret;
  219. unsigned int command;
  220. unsigned int chunk_size;
  221. IdcinDemuxContext *idcin = s->priv_data;
  222. AVIOContext *pb = s->pb;
  223. int i;
  224. int palette_scale;
  225. unsigned char r, g, b;
  226. unsigned char palette_buffer[768];
  227. uint32_t palette[256];
  228. if (url_feof(s->pb))
  229. return AVERROR(EIO);
  230. if (idcin->next_chunk_is_video) {
  231. command = avio_rl32(pb);
  232. if (command == 2) {
  233. return AVERROR(EIO);
  234. } else if (command == 1) {
  235. /* trigger a palette change */
  236. if (avio_read(pb, palette_buffer, 768) != 768)
  237. return AVERROR(EIO);
  238. /* scale the palette as necessary */
  239. palette_scale = 2;
  240. for (i = 0; i < 768; i++)
  241. if (palette_buffer[i] > 63) {
  242. palette_scale = 0;
  243. break;
  244. }
  245. for (i = 0; i < 256; i++) {
  246. r = palette_buffer[i * 3 ] << palette_scale;
  247. g = palette_buffer[i * 3 + 1] << palette_scale;
  248. b = palette_buffer[i * 3 + 2] << palette_scale;
  249. palette[i] = (0xFFU << 24) | (r << 16) | (g << 8) | (b);
  250. if (palette_scale == 2)
  251. palette[i] |= palette[i] >> 6 & 0x30303;
  252. }
  253. }
  254. chunk_size = avio_rl32(pb);
  255. /* skip the number of decoded bytes (always equal to width * height) */
  256. avio_skip(pb, 4);
  257. if (chunk_size < 4)
  258. return AVERROR_INVALIDDATA;
  259. chunk_size -= 4;
  260. ret= av_get_packet(pb, pkt, chunk_size);
  261. if (ret < 0)
  262. return ret;
  263. if (command == 1) {
  264. uint8_t *pal;
  265. pal = av_packet_new_side_data(pkt, AV_PKT_DATA_PALETTE,
  266. AVPALETTE_SIZE);
  267. if (!pal)
  268. return AVERROR(ENOMEM);
  269. memcpy(pal, palette, AVPALETTE_SIZE);
  270. pkt->flags |= AV_PKT_FLAG_KEY;
  271. }
  272. pkt->stream_index = idcin->video_stream_index;
  273. pkt->duration = 1;
  274. } else {
  275. /* send out the audio chunk */
  276. if (idcin->current_audio_chunk)
  277. chunk_size = idcin->audio_chunk_size2;
  278. else
  279. chunk_size = idcin->audio_chunk_size1;
  280. ret= av_get_packet(pb, pkt, chunk_size);
  281. if (ret < 0)
  282. return ret;
  283. pkt->stream_index = idcin->audio_stream_index;
  284. pkt->duration = chunk_size / idcin->block_align;
  285. idcin->current_audio_chunk ^= 1;
  286. }
  287. if (idcin->audio_present)
  288. idcin->next_chunk_is_video ^= 1;
  289. return ret;
  290. }
  291. static int idcin_read_seek(AVFormatContext *s, int stream_index,
  292. int64_t timestamp, int flags)
  293. {
  294. IdcinDemuxContext *idcin = s->priv_data;
  295. if (idcin->first_pkt_pos > 0) {
  296. int ret = avio_seek(s->pb, idcin->first_pkt_pos, SEEK_SET);
  297. if (ret < 0)
  298. return ret;
  299. ff_update_cur_dts(s, s->streams[idcin->video_stream_index], 0);
  300. idcin->next_chunk_is_video = 1;
  301. idcin->current_audio_chunk = 0;
  302. return 0;
  303. }
  304. return -1;
  305. }
  306. AVInputFormat ff_idcin_demuxer = {
  307. .name = "idcin",
  308. .long_name = NULL_IF_CONFIG_SMALL("id Cinematic"),
  309. .priv_data_size = sizeof(IdcinDemuxContext),
  310. .read_probe = idcin_probe,
  311. .read_header = idcin_read_header,
  312. .read_packet = idcin_read_packet,
  313. .read_seek = idcin_read_seek,
  314. .flags = AVFMT_NO_BYTE_SEEK,
  315. };