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.

393 lines
14KB

  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 "bswapdsp.h"
  27. #include "get_bits.h"
  28. #include "internal.h"
  29. #include "raw.h"
  30. #include "libavutil/avassert.h"
  31. #include "libavutil/buffer.h"
  32. #include "libavutil/common.h"
  33. #include "libavutil/intreadwrite.h"
  34. #include "libavutil/imgutils.h"
  35. #include "libavutil/opt.h"
  36. typedef struct RawVideoContext {
  37. AVClass *av_class;
  38. AVBufferRef *palette;
  39. int frame_size; /* size of the frame in bytes */
  40. int flip;
  41. int is_1_2_4_bpp; // 1 bpp raw in mov, and 2 or 4 bpp raw in avi/mov
  42. int is_yuv2;
  43. int is_lt_16bpp; // 16bpp pixfmt and bits_per_coded_sample < 16
  44. int tff;
  45. BswapDSPContext bbdsp;
  46. void *bitstream_buf;
  47. unsigned int bitstream_buf_size;
  48. } RawVideoContext;
  49. static const AVOption options[]={
  50. {"top", "top field first", offsetof(RawVideoContext, tff), AV_OPT_TYPE_BOOL, {.i64 = -1}, -1, 1, AV_OPT_FLAG_DECODING_PARAM|AV_OPT_FLAG_VIDEO_PARAM},
  51. {NULL}
  52. };
  53. static const AVClass rawdec_class = {
  54. .class_name = "rawdec",
  55. .option = options,
  56. .version = LIBAVUTIL_VERSION_INT,
  57. };
  58. static av_cold int raw_init_decoder(AVCodecContext *avctx)
  59. {
  60. RawVideoContext *context = avctx->priv_data;
  61. const AVPixFmtDescriptor *desc;
  62. ff_bswapdsp_init(&context->bbdsp);
  63. if ( avctx->codec_tag == MKTAG('r','a','w',' ')
  64. || avctx->codec_tag == MKTAG('N','O','1','6'))
  65. avctx->pix_fmt = avpriv_find_pix_fmt(avpriv_pix_fmt_bps_mov,
  66. avctx->bits_per_coded_sample);
  67. else if (avctx->codec_tag == MKTAG('W', 'R', 'A', 'W'))
  68. avctx->pix_fmt = avpriv_find_pix_fmt(avpriv_pix_fmt_bps_avi,
  69. avctx->bits_per_coded_sample);
  70. else if (avctx->codec_tag && (avctx->codec_tag & 0xFFFFFF) != MKTAG('B','I','T', 0))
  71. avctx->pix_fmt = avpriv_find_pix_fmt(ff_raw_pix_fmt_tags, avctx->codec_tag);
  72. else if (avctx->pix_fmt == AV_PIX_FMT_NONE && avctx->bits_per_coded_sample)
  73. avctx->pix_fmt = avpriv_find_pix_fmt(avpriv_pix_fmt_bps_avi,
  74. avctx->bits_per_coded_sample);
  75. desc = av_pix_fmt_desc_get(avctx->pix_fmt);
  76. if (!desc) {
  77. av_log(avctx, AV_LOG_ERROR, "Invalid pixel format.\n");
  78. return AVERROR(EINVAL);
  79. }
  80. if (desc->flags & (AV_PIX_FMT_FLAG_PAL | AV_PIX_FMT_FLAG_PSEUDOPAL)) {
  81. context->palette = av_buffer_alloc(AVPALETTE_SIZE);
  82. if (!context->palette)
  83. return AVERROR(ENOMEM);
  84. if (desc->flags & AV_PIX_FMT_FLAG_PSEUDOPAL)
  85. avpriv_set_systematic_pal2((uint32_t*)context->palette->data, avctx->pix_fmt);
  86. else {
  87. memset(context->palette->data, 0, AVPALETTE_SIZE);
  88. if (avctx->bits_per_coded_sample == 1)
  89. memset(context->palette->data, 0xff, 4);
  90. }
  91. }
  92. if ((avctx->extradata_size >= 9 &&
  93. !memcmp(avctx->extradata + avctx->extradata_size - 9, "BottomUp", 9)) ||
  94. avctx->codec_tag == MKTAG('c','y','u','v') ||
  95. avctx->codec_tag == MKTAG(3, 0, 0, 0) ||
  96. avctx->codec_tag == MKTAG('W','R','A','W'))
  97. context->flip = 1;
  98. if (avctx->codec_tag == AV_RL32("yuv2") &&
  99. avctx->pix_fmt == AV_PIX_FMT_YUYV422)
  100. context->is_yuv2 = 1;
  101. return 0;
  102. }
  103. static void flip(AVCodecContext *avctx, AVFrame *frame)
  104. {
  105. frame->data[0] += frame->linesize[0] * (avctx->height - 1);
  106. frame->linesize[0] *= -1;
  107. }
  108. /*
  109. * Scale sample to 16-bit resolution
  110. */
  111. #define SCALE16(x, bits) (((x) << (16 - (bits))) | ((x) >> (2 * (bits) - 16)))
  112. /**
  113. * Scale buffer to 16 bits per coded sample resolution
  114. */
  115. #define MKSCALE16(name, r16, w16) \
  116. static void name(AVCodecContext *avctx, uint8_t * dst, const uint8_t *buf, int buf_size, int packed) \
  117. { \
  118. int i; \
  119. if (!packed) { \
  120. for (i = 0; i + 1 < buf_size; i += 2) \
  121. w16(dst + i, SCALE16(r16(buf + i), avctx->bits_per_coded_sample)); \
  122. } else { \
  123. GetBitContext gb; \
  124. init_get_bits(&gb, buf, buf_size * 8); \
  125. for (i = 0; i < avctx->width * avctx->height; i++) { \
  126. int sample = get_bits(&gb, avctx->bits_per_coded_sample); \
  127. w16(dst + i*2, SCALE16(sample, avctx->bits_per_coded_sample)); \
  128. } \
  129. } \
  130. }
  131. MKSCALE16(scale16be, AV_RB16, AV_WB16)
  132. MKSCALE16(scale16le, AV_RL16, AV_WL16)
  133. static int raw_decode(AVCodecContext *avctx, void *data, int *got_frame,
  134. AVPacket *avpkt)
  135. {
  136. const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(avctx->pix_fmt);
  137. RawVideoContext *context = avctx->priv_data;
  138. const uint8_t *buf = avpkt->data;
  139. int buf_size = avpkt->size;
  140. int linesize_align = 4;
  141. int res, len;
  142. int need_copy;
  143. AVFrame *frame = data;
  144. if ((avctx->bits_per_coded_sample == 4 || avctx->bits_per_coded_sample == 2
  145. || avctx->bits_per_coded_sample == 1) &&
  146. avctx->pix_fmt == AV_PIX_FMT_PAL8 &&
  147. (!avctx->codec_tag || avctx->codec_tag == MKTAG('r','a','w',' '))) {
  148. context->is_1_2_4_bpp = 1;
  149. context->frame_size = av_image_get_buffer_size(avctx->pix_fmt,
  150. FFALIGN(avctx->width, 32),
  151. avctx->height, 1);
  152. } else {
  153. context->is_lt_16bpp = av_get_bits_per_pixel(desc) == 16 && avctx->bits_per_coded_sample && avctx->bits_per_coded_sample < 16;
  154. context->frame_size = av_image_get_buffer_size(avctx->pix_fmt, avctx->width,
  155. avctx->height, 1);
  156. }
  157. if (context->frame_size < 0)
  158. return context->frame_size;
  159. need_copy = !avpkt->buf || context->is_1_2_4_bpp || context->is_yuv2 || context->is_lt_16bpp;
  160. frame->pict_type = AV_PICTURE_TYPE_I;
  161. frame->key_frame = 1;
  162. res = ff_decode_frame_props(avctx, frame);
  163. if (res < 0)
  164. return res;
  165. av_frame_set_pkt_pos (frame, avctx->internal->pkt->pos);
  166. av_frame_set_pkt_duration(frame, avctx->internal->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. // 1 bpp raw in mov, and 2 or 4 bpp raw in avi/mov
  180. if (context->is_1_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 if (avctx->bits_per_coded_sample == 2) {
  191. for (i = 0; 4 * i + 3 < buf_size && i<avpkt->size; i++) {
  192. dst[4 * i + 0] = buf[i] >> 6;
  193. dst[4 * i + 1] = buf[i] >> 4 & 3;
  194. dst[4 * i + 2] = buf[i] >> 2 & 3;
  195. dst[4 * i + 3] = buf[i] & 3;
  196. }
  197. linesize_align = 16;
  198. } else {
  199. av_assert0(avctx->bits_per_coded_sample == 1);
  200. for (i = 0; 8 * i + 7 < buf_size && i<avpkt->size; i++) {
  201. dst[8 * i + 0] = buf[i] >> 7 & 1;
  202. dst[8 * i + 1] = buf[i] >> 6 & 1;
  203. dst[8 * i + 2] = buf[i] >> 5 & 1;
  204. dst[8 * i + 3] = buf[i] >> 4 & 1;
  205. dst[8 * i + 4] = buf[i] >> 3 & 1;
  206. dst[8 * i + 5] = buf[i] >> 2 & 1;
  207. dst[8 * i + 6] = buf[i] >> 1 & 1;
  208. dst[8 * i + 7] = buf[i] & 1;
  209. }
  210. linesize_align = 32;
  211. }
  212. buf = dst;
  213. } else if (context->is_lt_16bpp) {
  214. uint8_t *dst = frame->buf[0]->data;
  215. int packed = (avctx->codec_tag & 0xFFFFFF) == MKTAG('B','I','T', 0);
  216. int swap = avctx->codec_tag >> 24;
  217. if (packed && swap) {
  218. av_fast_padded_malloc(&context->bitstream_buf, &context->bitstream_buf_size, buf_size);
  219. if (!context->bitstream_buf)
  220. return AVERROR(ENOMEM);
  221. if (swap == 16)
  222. context->bbdsp.bswap16_buf(context->bitstream_buf, (const uint16_t*)buf, buf_size / 2);
  223. else if (swap == 32)
  224. context->bbdsp.bswap_buf(context->bitstream_buf, (const uint32_t*)buf, buf_size / 4);
  225. else
  226. return AVERROR_INVALIDDATA;
  227. buf = context->bitstream_buf;
  228. }
  229. if (desc->flags & AV_PIX_FMT_FLAG_BE)
  230. scale16be(avctx, dst, buf, buf_size, packed);
  231. else
  232. scale16le(avctx, dst, buf, buf_size, packed);
  233. buf = dst;
  234. } else if (need_copy) {
  235. memcpy(frame->buf[0]->data, buf, buf_size);
  236. buf = frame->buf[0]->data;
  237. }
  238. if (avctx->codec_tag == MKTAG('A', 'V', '1', 'x') ||
  239. avctx->codec_tag == MKTAG('A', 'V', 'u', 'p'))
  240. buf += buf_size - context->frame_size;
  241. len = context->frame_size - (avctx->pix_fmt==AV_PIX_FMT_PAL8 ? AVPALETTE_SIZE : 0);
  242. if (buf_size < len && ((avctx->codec_tag & 0xFFFFFF) != MKTAG('B','I','T', 0) || !need_copy)) {
  243. av_log(avctx, AV_LOG_ERROR, "Invalid buffer size, packet size %d < expected frame_size %d\n", buf_size, len);
  244. av_buffer_unref(&frame->buf[0]);
  245. return AVERROR(EINVAL);
  246. }
  247. if ((res = av_image_fill_arrays(frame->data, frame->linesize,
  248. buf, avctx->pix_fmt,
  249. avctx->width, avctx->height, 1)) < 0) {
  250. av_buffer_unref(&frame->buf[0]);
  251. return res;
  252. }
  253. if (avctx->pix_fmt == AV_PIX_FMT_PAL8) {
  254. const uint8_t *pal = av_packet_get_side_data(avpkt, AV_PKT_DATA_PALETTE,
  255. NULL);
  256. if (pal) {
  257. av_buffer_unref(&context->palette);
  258. context->palette = av_buffer_alloc(AVPALETTE_SIZE);
  259. if (!context->palette) {
  260. av_buffer_unref(&frame->buf[0]);
  261. return AVERROR(ENOMEM);
  262. }
  263. memcpy(context->palette->data, pal, AVPALETTE_SIZE);
  264. frame->palette_has_changed = 1;
  265. }
  266. }
  267. if ((avctx->pix_fmt==AV_PIX_FMT_BGR24 ||
  268. avctx->pix_fmt==AV_PIX_FMT_GRAY8 ||
  269. avctx->pix_fmt==AV_PIX_FMT_RGB555LE ||
  270. avctx->pix_fmt==AV_PIX_FMT_RGB555BE ||
  271. avctx->pix_fmt==AV_PIX_FMT_RGB565LE ||
  272. avctx->pix_fmt==AV_PIX_FMT_MONOWHITE ||
  273. avctx->pix_fmt==AV_PIX_FMT_PAL8) &&
  274. FFALIGN(frame->linesize[0], linesize_align) * avctx->height <= buf_size)
  275. frame->linesize[0] = FFALIGN(frame->linesize[0], linesize_align);
  276. if (avctx->pix_fmt == AV_PIX_FMT_NV12 && avctx->codec_tag == MKTAG('N', 'V', '1', '2') &&
  277. FFALIGN(frame->linesize[0], linesize_align) * avctx->height +
  278. FFALIGN(frame->linesize[1], linesize_align) * ((avctx->height + 1) / 2) <= buf_size) {
  279. int la0 = FFALIGN(frame->linesize[0], linesize_align);
  280. frame->data[1] += (la0 - frame->linesize[0]) * avctx->height;
  281. frame->linesize[0] = la0;
  282. frame->linesize[1] = FFALIGN(frame->linesize[1], linesize_align);
  283. }
  284. if ((avctx->pix_fmt == AV_PIX_FMT_PAL8 && buf_size < context->frame_size) ||
  285. (desc->flags & AV_PIX_FMT_FLAG_PSEUDOPAL)) {
  286. frame->buf[1] = av_buffer_ref(context->palette);
  287. if (!frame->buf[1]) {
  288. av_buffer_unref(&frame->buf[0]);
  289. return AVERROR(ENOMEM);
  290. }
  291. frame->data[1] = frame->buf[1]->data;
  292. }
  293. if (avctx->pix_fmt == AV_PIX_FMT_BGR24 &&
  294. ((frame->linesize[0] + 3) & ~3) * avctx->height <= buf_size)
  295. frame->linesize[0] = (frame->linesize[0] + 3) & ~3;
  296. if (context->flip)
  297. flip(avctx, frame);
  298. if (avctx->codec_tag == MKTAG('Y', 'V', '1', '2') ||
  299. avctx->codec_tag == MKTAG('Y', 'V', '1', '6') ||
  300. avctx->codec_tag == MKTAG('Y', 'V', '2', '4') ||
  301. avctx->codec_tag == MKTAG('Y', 'V', 'U', '9'))
  302. FFSWAP(uint8_t *, frame->data[1], frame->data[2]);
  303. if (avctx->codec_tag == AV_RL32("I420") && (avctx->width+1)*(avctx->height+1) * 3/2 == buf_size) {
  304. frame->data[1] = frame->data[1] + (avctx->width+1)*(avctx->height+1) -avctx->width*avctx->height;
  305. frame->data[2] = frame->data[2] + ((avctx->width+1)*(avctx->height+1) -avctx->width*avctx->height)*5/4;
  306. }
  307. if (avctx->codec_tag == AV_RL32("yuv2") &&
  308. avctx->pix_fmt == AV_PIX_FMT_YUYV422) {
  309. int x, y;
  310. uint8_t *line = frame->data[0];
  311. for (y = 0; y < avctx->height; y++) {
  312. for (x = 0; x < avctx->width; x++)
  313. line[2 * x + 1] ^= 0x80;
  314. line += frame->linesize[0];
  315. }
  316. }
  317. if (avctx->field_order > AV_FIELD_PROGRESSIVE) { /* we have interlaced material flagged in container */
  318. frame->interlaced_frame = 1;
  319. if (avctx->field_order == AV_FIELD_TT || avctx->field_order == AV_FIELD_TB)
  320. frame->top_field_first = 1;
  321. }
  322. *got_frame = 1;
  323. return buf_size;
  324. }
  325. static av_cold int raw_close_decoder(AVCodecContext *avctx)
  326. {
  327. RawVideoContext *context = avctx->priv_data;
  328. av_buffer_unref(&context->palette);
  329. return 0;
  330. }
  331. AVCodec ff_rawvideo_decoder = {
  332. .name = "rawvideo",
  333. .long_name = NULL_IF_CONFIG_SMALL("raw video"),
  334. .type = AVMEDIA_TYPE_VIDEO,
  335. .id = AV_CODEC_ID_RAWVIDEO,
  336. .priv_data_size = sizeof(RawVideoContext),
  337. .init = raw_init_decoder,
  338. .close = raw_close_decoder,
  339. .decode = raw_decode,
  340. .priv_class = &rawdec_class,
  341. .capabilities = AV_CODEC_CAP_PARAM_CHANGE,
  342. };