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.

361 lines
13KB

  1. /*
  2. * Raw Video Decoder
  3. * Copyright (c) 2001 Fabrice Bellard
  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. * Raw Video Decoder
  24. */
  25. #include "avcodec.h"
  26. #include "raw.h"
  27. #include "libavutil/avassert.h"
  28. #include "libavutil/buffer.h"
  29. #include "libavutil/common.h"
  30. #include "libavutil/intreadwrite.h"
  31. #include "libavutil/imgutils.h"
  32. #include "libavutil/opt.h"
  33. typedef struct RawVideoContext {
  34. AVClass *av_class;
  35. AVBufferRef *palette;
  36. int frame_size; /* size of the frame in bytes */
  37. int flip;
  38. int is_2_4_bpp; // 2 or 4 bpp raw in avi/mov
  39. int is_yuv2;
  40. int tff;
  41. } RawVideoContext;
  42. static const AVOption options[]={
  43. {"top", "top field first", offsetof(RawVideoContext, tff), AV_OPT_TYPE_INT, {.i64 = -1}, -1, 1, AV_OPT_FLAG_DECODING_PARAM|AV_OPT_FLAG_VIDEO_PARAM},
  44. {NULL}
  45. };
  46. static const AVClass rawdec_class = {
  47. .class_name = "rawdec",
  48. .option = options,
  49. .version = LIBAVUTIL_VERSION_INT,
  50. };
  51. static const PixelFormatTag pix_fmt_bps_avi[] = {
  52. { AV_PIX_FMT_MONOWHITE, 1 },
  53. { AV_PIX_FMT_PAL8, 2 },
  54. { AV_PIX_FMT_PAL8, 4 },
  55. { AV_PIX_FMT_PAL8, 8 },
  56. { AV_PIX_FMT_RGB444LE, 12 },
  57. { AV_PIX_FMT_RGB555LE, 15 },
  58. { AV_PIX_FMT_RGB555LE, 16 },
  59. { AV_PIX_FMT_BGR24, 24 },
  60. { AV_PIX_FMT_BGRA, 32 },
  61. { AV_PIX_FMT_NONE, 0 },
  62. };
  63. static const PixelFormatTag pix_fmt_bps_mov[] = {
  64. { AV_PIX_FMT_MONOWHITE, 1 },
  65. { AV_PIX_FMT_PAL8, 2 },
  66. { AV_PIX_FMT_PAL8, 4 },
  67. { AV_PIX_FMT_PAL8, 8 },
  68. // FIXME swscale does not support 16 bit in .mov, sample 16bit.mov
  69. // http://developer.apple.com/documentation/QuickTime/QTFF/QTFFChap3/qtff3.html
  70. { AV_PIX_FMT_RGB555BE, 16 },
  71. { AV_PIX_FMT_RGB24, 24 },
  72. { AV_PIX_FMT_ARGB, 32 },
  73. { AV_PIX_FMT_MONOWHITE,33 },
  74. { AV_PIX_FMT_NONE, 0 },
  75. };
  76. enum AVPixelFormat avpriv_find_pix_fmt(const PixelFormatTag *tags,
  77. unsigned int fourcc)
  78. {
  79. while (tags->pix_fmt >= 0) {
  80. if (tags->fourcc == fourcc)
  81. return tags->pix_fmt;
  82. tags++;
  83. }
  84. return AV_PIX_FMT_NONE;
  85. }
  86. #if LIBAVCODEC_VERSION_MAJOR < 55
  87. enum AVPixelFormat ff_find_pix_fmt(const PixelFormatTag *tags, unsigned int fourcc)
  88. {
  89. return avpriv_find_pix_fmt(tags, fourcc);
  90. }
  91. #endif
  92. static av_cold int raw_init_decoder(AVCodecContext *avctx)
  93. {
  94. RawVideoContext *context = avctx->priv_data;
  95. const AVPixFmtDescriptor *desc;
  96. if ( avctx->codec_tag == MKTAG('r','a','w',' ')
  97. || avctx->codec_tag == MKTAG('N','O','1','6'))
  98. avctx->pix_fmt = avpriv_find_pix_fmt(pix_fmt_bps_mov,
  99. avctx->bits_per_coded_sample);
  100. else if (avctx->codec_tag == MKTAG('W', 'R', 'A', 'W'))
  101. avctx->pix_fmt = avpriv_find_pix_fmt(pix_fmt_bps_avi,
  102. avctx->bits_per_coded_sample);
  103. else if (avctx->codec_tag)
  104. avctx->pix_fmt = avpriv_find_pix_fmt(ff_raw_pix_fmt_tags, avctx->codec_tag);
  105. else if (avctx->pix_fmt == AV_PIX_FMT_NONE && avctx->bits_per_coded_sample)
  106. avctx->pix_fmt = avpriv_find_pix_fmt(pix_fmt_bps_avi,
  107. avctx->bits_per_coded_sample);
  108. desc = av_pix_fmt_desc_get(avctx->pix_fmt);
  109. if (!desc) {
  110. av_log(avctx, AV_LOG_ERROR, "Invalid pixel format.\n");
  111. return AVERROR(EINVAL);
  112. }
  113. if (desc->flags & (AV_PIX_FMT_FLAG_PAL | AV_PIX_FMT_FLAG_PSEUDOPAL)) {
  114. context->palette = av_buffer_alloc(AVPALETTE_SIZE);
  115. if (!context->palette)
  116. return AVERROR(ENOMEM);
  117. if (desc->flags & AV_PIX_FMT_FLAG_PSEUDOPAL)
  118. avpriv_set_systematic_pal2((uint32_t*)context->palette->data, avctx->pix_fmt);
  119. else
  120. memset(context->palette->data, 0, AVPALETTE_SIZE);
  121. }
  122. if ((avctx->bits_per_coded_sample == 4 || avctx->bits_per_coded_sample == 2) &&
  123. avctx->pix_fmt == AV_PIX_FMT_PAL8 &&
  124. (!avctx->codec_tag || avctx->codec_tag == MKTAG('r','a','w',' '))) {
  125. context->is_2_4_bpp = 1;
  126. context->frame_size = avpicture_get_size(avctx->pix_fmt,
  127. FFALIGN(avctx->width, 16),
  128. avctx->height);
  129. } else {
  130. context->frame_size = avpicture_get_size(avctx->pix_fmt, avctx->width,
  131. avctx->height);
  132. }
  133. if ((avctx->extradata_size >= 9 &&
  134. !memcmp(avctx->extradata + avctx->extradata_size - 9, "BottomUp", 9)) ||
  135. avctx->codec_tag == MKTAG('c','y','u','v') ||
  136. avctx->codec_tag == MKTAG(3, 0, 0, 0) ||
  137. avctx->codec_tag == MKTAG('W','R','A','W'))
  138. context->flip = 1;
  139. if (avctx->codec_tag == AV_RL32("yuv2") &&
  140. avctx->pix_fmt == AV_PIX_FMT_YUYV422)
  141. context->is_yuv2 = 1;
  142. return 0;
  143. }
  144. static void flip(AVCodecContext *avctx, AVPicture *picture)
  145. {
  146. picture->data[0] += picture->linesize[0] * (avctx->height - 1);
  147. picture->linesize[0] *= -1;
  148. }
  149. static int raw_decode(AVCodecContext *avctx, void *data, int *got_frame,
  150. AVPacket *avpkt)
  151. {
  152. const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(avctx->pix_fmt);
  153. RawVideoContext *context = avctx->priv_data;
  154. const uint8_t *buf = avpkt->data;
  155. int buf_size = avpkt->size;
  156. int linesize_align = 4;
  157. int res, len;
  158. int need_copy = !avpkt->buf || context->is_2_4_bpp || context->is_yuv2;
  159. AVFrame *frame = data;
  160. AVPicture *picture = data;
  161. frame->pict_type = AV_PICTURE_TYPE_I;
  162. frame->key_frame = 1;
  163. frame->reordered_opaque = avctx->reordered_opaque;
  164. frame->pkt_pts = avctx->pkt->pts;
  165. av_frame_set_pkt_pos (frame, avctx->pkt->pos);
  166. av_frame_set_pkt_duration(frame, avctx->pkt->duration);
  167. if (context->tff >= 0) {
  168. frame->interlaced_frame = 1;
  169. frame->top_field_first = context->tff;
  170. }
  171. if ((res = av_image_check_size(avctx->width, avctx->height, 0, avctx)) < 0)
  172. return res;
  173. if (need_copy)
  174. frame->buf[0] = av_buffer_alloc(FFMAX(context->frame_size, buf_size));
  175. else
  176. frame->buf[0] = av_buffer_ref(avpkt->buf);
  177. if (!frame->buf[0])
  178. return AVERROR(ENOMEM);
  179. //2bpp and 4bpp raw in avi and mov (yes this is ugly ...)
  180. if (context->is_2_4_bpp) {
  181. int i;
  182. uint8_t *dst = frame->buf[0]->data;
  183. buf_size = context->frame_size - AVPALETTE_SIZE;
  184. if (avctx->bits_per_coded_sample == 4) {
  185. for (i = 0; 2 * i + 1 < buf_size && i<avpkt->size; i++) {
  186. dst[2 * i + 0] = buf[i] >> 4;
  187. dst[2 * i + 1] = buf[i] & 15;
  188. }
  189. linesize_align = 8;
  190. } else {
  191. av_assert0(avctx->bits_per_coded_sample == 2);
  192. for (i = 0; 4 * i + 3 < buf_size && i<avpkt->size; i++) {
  193. dst[4 * i + 0] = buf[i] >> 6;
  194. dst[4 * i + 1] = buf[i] >> 4 & 3;
  195. dst[4 * i + 2] = buf[i] >> 2 & 3;
  196. dst[4 * i + 3] = buf[i] & 3;
  197. }
  198. linesize_align = 16;
  199. }
  200. buf = dst;
  201. } else if (need_copy) {
  202. memcpy(frame->buf[0]->data, buf, buf_size);
  203. buf = frame->buf[0]->data;
  204. }
  205. if (avctx->codec_tag == MKTAG('A', 'V', '1', 'x') ||
  206. avctx->codec_tag == MKTAG('A', 'V', 'u', 'p'))
  207. buf += buf_size - context->frame_size;
  208. len = context->frame_size - (avctx->pix_fmt==AV_PIX_FMT_PAL8 ? AVPALETTE_SIZE : 0);
  209. if (buf_size < len) {
  210. av_log(avctx, AV_LOG_ERROR, "Invalid buffer size, packet size %d < expected frame_size %d\n", buf_size, len);
  211. av_buffer_unref(&frame->buf[0]);
  212. return AVERROR(EINVAL);
  213. }
  214. if ((res = avpicture_fill(picture, buf, avctx->pix_fmt,
  215. avctx->width, avctx->height)) < 0) {
  216. av_buffer_unref(&frame->buf[0]);
  217. return res;
  218. }
  219. if (avctx->pix_fmt == AV_PIX_FMT_PAL8) {
  220. const uint8_t *pal = av_packet_get_side_data(avpkt, AV_PKT_DATA_PALETTE,
  221. NULL);
  222. if (pal) {
  223. av_buffer_unref(&context->palette);
  224. context->palette = av_buffer_alloc(AVPALETTE_SIZE);
  225. if (!context->palette) {
  226. av_buffer_unref(&frame->buf[0]);
  227. return AVERROR(ENOMEM);
  228. }
  229. memcpy(context->palette->data, pal, AVPALETTE_SIZE);
  230. frame->palette_has_changed = 1;
  231. }
  232. }
  233. if ((avctx->pix_fmt==AV_PIX_FMT_BGR24 ||
  234. avctx->pix_fmt==AV_PIX_FMT_GRAY8 ||
  235. avctx->pix_fmt==AV_PIX_FMT_RGB555LE ||
  236. avctx->pix_fmt==AV_PIX_FMT_RGB555BE ||
  237. avctx->pix_fmt==AV_PIX_FMT_RGB565LE ||
  238. avctx->pix_fmt==AV_PIX_FMT_MONOWHITE ||
  239. avctx->pix_fmt==AV_PIX_FMT_PAL8) &&
  240. FFALIGN(frame->linesize[0], linesize_align) * avctx->height <= buf_size)
  241. frame->linesize[0] = FFALIGN(frame->linesize[0], linesize_align);
  242. if (avctx->pix_fmt == AV_PIX_FMT_NV12 && avctx->codec_tag == MKTAG('N', 'V', '1', '2') &&
  243. FFALIGN(frame->linesize[0], linesize_align) * avctx->height +
  244. FFALIGN(frame->linesize[1], linesize_align) * ((avctx->height + 1) / 2) <= buf_size) {
  245. int la0 = FFALIGN(frame->linesize[0], linesize_align);
  246. frame->data[1] += (la0 - frame->linesize[0]) * avctx->height;
  247. frame->linesize[0] = la0;
  248. frame->linesize[1] = FFALIGN(frame->linesize[1], linesize_align);
  249. }
  250. if ((avctx->pix_fmt == AV_PIX_FMT_PAL8 && buf_size < context->frame_size) ||
  251. (desc->flags & AV_PIX_FMT_FLAG_PSEUDOPAL)) {
  252. frame->buf[1] = av_buffer_ref(context->palette);
  253. if (!frame->buf[1]) {
  254. av_buffer_unref(&frame->buf[0]);
  255. return AVERROR(ENOMEM);
  256. }
  257. frame->data[1] = frame->buf[1]->data;
  258. }
  259. if (avctx->pix_fmt == AV_PIX_FMT_BGR24 &&
  260. ((frame->linesize[0] + 3) & ~3) * avctx->height <= buf_size)
  261. frame->linesize[0] = (frame->linesize[0] + 3) & ~3;
  262. if (context->flip)
  263. flip(avctx, picture);
  264. if (avctx->codec_tag == MKTAG('Y', 'V', '1', '2') ||
  265. avctx->codec_tag == MKTAG('Y', 'V', '1', '6') ||
  266. avctx->codec_tag == MKTAG('Y', 'V', '2', '4') ||
  267. avctx->codec_tag == MKTAG('Y', 'V', 'U', '9'))
  268. FFSWAP(uint8_t *, picture->data[1], picture->data[2]);
  269. if (avctx->codec_tag == AV_RL32("I420") && (avctx->width+1)*(avctx->height+1) * 3/2 == buf_size) {
  270. picture->data[1] = picture->data[1] + (avctx->width+1)*(avctx->height+1) -avctx->width*avctx->height;
  271. picture->data[2] = picture->data[2] + ((avctx->width+1)*(avctx->height+1) -avctx->width*avctx->height)*5/4;
  272. }
  273. if (avctx->codec_tag == AV_RL32("yuv2") &&
  274. avctx->pix_fmt == AV_PIX_FMT_YUYV422) {
  275. int x, y;
  276. uint8_t *line = picture->data[0];
  277. for (y = 0; y < avctx->height; y++) {
  278. for (x = 0; x < avctx->width; x++)
  279. line[2 * x + 1] ^= 0x80;
  280. line += picture->linesize[0];
  281. }
  282. }
  283. if (avctx->codec_tag == AV_RL32("YVYU") &&
  284. avctx->pix_fmt == AV_PIX_FMT_YUYV422) {
  285. int x, y;
  286. uint8_t *line = picture->data[0];
  287. for(y = 0; y < avctx->height; y++) {
  288. for(x = 0; x < avctx->width - 1; x += 2)
  289. FFSWAP(uint8_t, line[2*x + 1], line[2*x + 3]);
  290. line += picture->linesize[0];
  291. }
  292. }
  293. if (avctx->field_order > AV_FIELD_PROGRESSIVE) { /* we have interlaced material flagged in container */
  294. frame->interlaced_frame = 1;
  295. if (avctx->field_order == AV_FIELD_TT || avctx->field_order == AV_FIELD_TB)
  296. frame->top_field_first = 1;
  297. }
  298. *got_frame = 1;
  299. return buf_size;
  300. }
  301. static av_cold int raw_close_decoder(AVCodecContext *avctx)
  302. {
  303. RawVideoContext *context = avctx->priv_data;
  304. av_buffer_unref(&context->palette);
  305. return 0;
  306. }
  307. AVCodec ff_rawvideo_decoder = {
  308. .name = "rawvideo",
  309. .type = AVMEDIA_TYPE_VIDEO,
  310. .id = AV_CODEC_ID_RAWVIDEO,
  311. .priv_data_size = sizeof(RawVideoContext),
  312. .init = raw_init_decoder,
  313. .close = raw_close_decoder,
  314. .decode = raw_decode,
  315. .long_name = NULL_IF_CONFIG_SMALL("raw video"),
  316. .priv_class = &rawdec_class,
  317. };