blob: 9900e7ab1f796d3171ad2688c2662b081f61a6ac
| 1 | /* |
| 2 | * VP9 compatible video decoder |
| 3 | * |
| 4 | * Copyright (C) 2013 Ronald S. Bultje <rsbultje gmail com> |
| 5 | * Copyright (C) 2013 Clément Bœsch <u pkh me> |
| 6 | * |
| 7 | * This file is part of FFmpeg. |
| 8 | * |
| 9 | * FFmpeg is free software; you can redistribute it and/or |
| 10 | * modify it under the terms of the GNU Lesser General Public |
| 11 | * License as published by the Free Software Foundation; either |
| 12 | * version 2.1 of the License, or (at your option) any later version. |
| 13 | * |
| 14 | * FFmpeg is distributed in the hope that it will be useful, |
| 15 | * but WITHOUT ANY WARRANTY; without even the implied warranty of |
| 16 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU |
| 17 | * Lesser General Public License for more details. |
| 18 | * |
| 19 | * You should have received a copy of the GNU Lesser General Public |
| 20 | * License along with FFmpeg; if not, write to the Free Software |
| 21 | * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA |
| 22 | */ |
| 23 | |
| 24 | #include "libavutil/intreadwrite.h" |
| 25 | #include "libavcodec/get_bits.h" |
| 26 | #include "parser.h" |
| 27 | |
| 28 | typedef struct VP9ParseContext { |
| 29 | int n_frames; // 1-8 |
| 30 | int size[8]; |
| 31 | int marker_size; |
| 32 | int64_t pts; |
| 33 | } VP9ParseContext; |
| 34 | |
| 35 | static int parse_frame(AVCodecParserContext *ctx, const uint8_t *buf, int size) |
| 36 | { |
| 37 | VP9ParseContext *s = ctx->priv_data; |
| 38 | GetBitContext gb; |
| 39 | int res, profile, keyframe, invisible; |
| 40 | |
| 41 | if ((res = init_get_bits8(&gb, buf, size)) < 0) |
| 42 | return res; |
| 43 | get_bits(&gb, 2); // frame marker |
| 44 | profile = get_bits1(&gb); |
| 45 | profile |= get_bits1(&gb) << 1; |
| 46 | if (profile == 3) profile += get_bits1(&gb); |
| 47 | |
| 48 | if (get_bits1(&gb)) { |
| 49 | keyframe = 0; |
| 50 | invisible = 0; |
| 51 | } else { |
| 52 | keyframe = !get_bits1(&gb); |
| 53 | invisible = !get_bits1(&gb); |
| 54 | } |
| 55 | |
| 56 | if (!keyframe) { |
| 57 | ctx->pict_type = AV_PICTURE_TYPE_P; |
| 58 | ctx->key_frame = 0; |
| 59 | } else { |
| 60 | ctx->pict_type = AV_PICTURE_TYPE_I; |
| 61 | ctx->key_frame = 1; |
| 62 | } |
| 63 | |
| 64 | if (!invisible) { |
| 65 | if (ctx->pts == AV_NOPTS_VALUE) |
| 66 | ctx->pts = s->pts; |
| 67 | s->pts = AV_NOPTS_VALUE; |
| 68 | } else if (ctx->pts != AV_NOPTS_VALUE) { |
| 69 | s->pts = ctx->pts; |
| 70 | ctx->pts = AV_NOPTS_VALUE; |
| 71 | } |
| 72 | |
| 73 | return 0; |
| 74 | } |
| 75 | |
| 76 | static int parse(AVCodecParserContext *ctx, |
| 77 | AVCodecContext *avctx, |
| 78 | const uint8_t **out_data, int *out_size, |
| 79 | const uint8_t *data, int size) |
| 80 | { |
| 81 | VP9ParseContext *s = ctx->priv_data; |
| 82 | int full_size = size; |
| 83 | int marker; |
| 84 | |
| 85 | if (size <= 0) { |
| 86 | *out_size = 0; |
| 87 | *out_data = data; |
| 88 | |
| 89 | return 0; |
| 90 | } |
| 91 | |
| 92 | if (s->n_frames > 0) { |
| 93 | int i; |
| 94 | int size_sum = 0; |
| 95 | |
| 96 | for (i = 0; i < s->n_frames ;i++) |
| 97 | size_sum += s->size[i]; |
| 98 | size_sum += s->marker_size; |
| 99 | |
| 100 | if (size_sum != size) { |
| 101 | av_log(avctx, AV_LOG_ERROR, "Inconsistent input frame sizes %d %d\n", |
| 102 | size_sum, size); |
| 103 | s->n_frames = 0; |
| 104 | } |
| 105 | } |
| 106 | |
| 107 | if (s->n_frames > 0) { |
| 108 | *out_data = data; |
| 109 | *out_size = s->size[--s->n_frames]; |
| 110 | parse_frame(ctx, *out_data, *out_size); |
| 111 | |
| 112 | return s->n_frames > 0 ? *out_size : size /* i.e. include idx tail */; |
| 113 | } |
| 114 | |
| 115 | marker = data[size - 1]; |
| 116 | if ((marker & 0xe0) == 0xc0) { |
| 117 | int nbytes = 1 + ((marker >> 3) & 0x3); |
| 118 | int n_frames = 1 + (marker & 0x7), idx_sz = 2 + n_frames * nbytes; |
| 119 | |
| 120 | if (size >= idx_sz && data[size - idx_sz] == marker) { |
| 121 | const uint8_t *idx = data + size + 1 - idx_sz; |
| 122 | int first = 1; |
| 123 | |
| 124 | switch (nbytes) { |
| 125 | #define case_n(a, rd) \ |
| 126 | case a: \ |
| 127 | while (n_frames--) { \ |
| 128 | unsigned sz = rd; \ |
| 129 | idx += a; \ |
| 130 | if (sz == 0 || sz > size) { \ |
| 131 | s->n_frames = 0; \ |
| 132 | *out_size = size; \ |
| 133 | *out_data = data; \ |
| 134 | av_log(avctx, AV_LOG_ERROR, \ |
| 135 | "Invalid superframe packet size: %u frame size: %d\n", \ |
| 136 | sz, size); \ |
| 137 | return full_size; \ |
| 138 | } \ |
| 139 | if (first) { \ |
| 140 | first = 0; \ |
| 141 | *out_data = data; \ |
| 142 | *out_size = sz; \ |
| 143 | s->n_frames = n_frames; \ |
| 144 | } else { \ |
| 145 | s->size[n_frames] = sz; \ |
| 146 | } \ |
| 147 | data += sz; \ |
| 148 | size -= sz; \ |
| 149 | } \ |
| 150 | s->marker_size = size; \ |
| 151 | parse_frame(ctx, *out_data, *out_size); \ |
| 152 | return s->n_frames > 0 ? *out_size : full_size |
| 153 | |
| 154 | case_n(1, *idx); |
| 155 | case_n(2, AV_RL16(idx)); |
| 156 | case_n(3, AV_RL24(idx)); |
| 157 | case_n(4, AV_RL32(idx)); |
| 158 | } |
| 159 | } |
| 160 | } |
| 161 | |
| 162 | *out_data = data; |
| 163 | *out_size = size; |
| 164 | parse_frame(ctx, data, size); |
| 165 | |
| 166 | return size; |
| 167 | } |
| 168 | |
| 169 | AVCodecParser ff_vp9_parser = { |
| 170 | .codec_ids = { AV_CODEC_ID_VP9 }, |
| 171 | .priv_data_size = sizeof(VP9ParseContext), |
| 172 | .parser_parse = parse, |
| 173 | }; |
| 174 |