| 1 | /* |
| 2 | * H.264 MP4 to Annex B byte stream format filter |
| 3 | * Copyright (c) 2007 Benoit Fouet <benoit.fouet@free.fr> |
| 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 | #include <string.h> |
| 23 | |
| 24 | #include "libavutil/intreadwrite.h" |
| 25 | #include "libavutil/mem.h" |
| 26 | #include "avcodec.h" |
| 27 | |
| 28 | typedef struct H264BSFContext { |
| 29 | int32_t sps_offset; |
| 30 | int32_t pps_offset; |
| 31 | uint8_t length_size; |
| 32 | uint8_t new_idr; |
| 33 | uint8_t idr_sps_seen; |
| 34 | uint8_t idr_pps_seen; |
| 35 | int extradata_parsed; |
| 36 | } H264BSFContext; |
| 37 | |
| 38 | static int alloc_and_copy(uint8_t **poutbuf, int *poutbuf_size, |
| 39 | const uint8_t *sps_pps, uint32_t sps_pps_size, |
| 40 | const uint8_t *in, uint32_t in_size) |
| 41 | { |
| 42 | uint32_t offset = *poutbuf_size; |
| 43 | uint8_t nal_header_size = offset ? 3 : 4; |
| 44 | int err; |
| 45 | |
| 46 | *poutbuf_size += sps_pps_size + in_size + nal_header_size; |
| 47 | if ((err = av_reallocp(poutbuf, |
| 48 | *poutbuf_size + FF_INPUT_BUFFER_PADDING_SIZE)) < 0) { |
| 49 | *poutbuf_size = 0; |
| 50 | return err; |
| 51 | } |
| 52 | if (sps_pps) |
| 53 | memcpy(*poutbuf + offset, sps_pps, sps_pps_size); |
| 54 | memcpy(*poutbuf + sps_pps_size + nal_header_size + offset, in, in_size); |
| 55 | if (!offset) { |
| 56 | AV_WB32(*poutbuf + sps_pps_size, 1); |
| 57 | } else { |
| 58 | (*poutbuf + offset + sps_pps_size)[0] = |
| 59 | (*poutbuf + offset + sps_pps_size)[1] = 0; |
| 60 | (*poutbuf + offset + sps_pps_size)[2] = 1; |
| 61 | } |
| 62 | |
| 63 | return 0; |
| 64 | } |
| 65 | |
| 66 | static int h264_extradata_to_annexb(H264BSFContext *ctx, AVCodecContext *avctx, const int padding) |
| 67 | { |
| 68 | uint16_t unit_size; |
| 69 | uint64_t total_size = 0; |
| 70 | uint8_t *out = NULL, unit_nb, sps_done = 0, |
| 71 | sps_seen = 0, pps_seen = 0; |
| 72 | const uint8_t *extradata = avctx->extradata + 4; |
| 73 | static const uint8_t nalu_header[4] = { 0, 0, 0, 1 }; |
| 74 | int length_size = (*extradata++ & 0x3) + 1; // retrieve length coded size |
| 75 | |
| 76 | ctx->sps_offset = ctx->pps_offset = -1; |
| 77 | |
| 78 | /* retrieve sps and pps unit(s) */ |
| 79 | unit_nb = *extradata++ & 0x1f; /* number of sps unit(s) */ |
| 80 | if (!unit_nb) { |
| 81 | goto pps; |
| 82 | } else { |
| 83 | ctx->sps_offset = 0; |
| 84 | sps_seen = 1; |
| 85 | } |
| 86 | |
| 87 | while (unit_nb--) { |
| 88 | int err; |
| 89 | |
| 90 | unit_size = AV_RB16(extradata); |
| 91 | total_size += unit_size + 4; |
| 92 | if (total_size > INT_MAX - padding) { |
| 93 | av_log(avctx, AV_LOG_ERROR, |
| 94 | "Too big extradata size, corrupted stream or invalid MP4/AVCC bitstream\n"); |
| 95 | av_free(out); |
| 96 | return AVERROR(EINVAL); |
| 97 | } |
| 98 | if (extradata + 2 + unit_size > avctx->extradata + avctx->extradata_size) { |
| 99 | av_log(avctx, AV_LOG_ERROR, "Packet header is not contained in global extradata, " |
| 100 | "corrupted stream or invalid MP4/AVCC bitstream\n"); |
| 101 | av_free(out); |
| 102 | return AVERROR(EINVAL); |
| 103 | } |
| 104 | if ((err = av_reallocp(&out, total_size + padding)) < 0) |
| 105 | return err; |
| 106 | memcpy(out + total_size - unit_size - 4, nalu_header, 4); |
| 107 | memcpy(out + total_size - unit_size, extradata + 2, unit_size); |
| 108 | extradata += 2 + unit_size; |
| 109 | pps: |
| 110 | if (!unit_nb && !sps_done++) { |
| 111 | unit_nb = *extradata++; /* number of pps unit(s) */ |
| 112 | if (unit_nb) { |
| 113 | ctx->pps_offset = (extradata - 1) - (avctx->extradata + 4); |
| 114 | pps_seen = 1; |
| 115 | } |
| 116 | } |
| 117 | } |
| 118 | |
| 119 | if (out) |
| 120 | memset(out + total_size, 0, padding); |
| 121 | |
| 122 | if (!sps_seen) |
| 123 | av_log(avctx, AV_LOG_WARNING, |
| 124 | "Warning: SPS NALU missing or invalid. " |
| 125 | "The resulting stream may not play.\n"); |
| 126 | |
| 127 | if (!pps_seen) |
| 128 | av_log(avctx, AV_LOG_WARNING, |
| 129 | "Warning: PPS NALU missing or invalid. " |
| 130 | "The resulting stream may not play.\n"); |
| 131 | |
| 132 | av_free(avctx->extradata); |
| 133 | avctx->extradata = out; |
| 134 | avctx->extradata_size = total_size; |
| 135 | |
| 136 | return length_size; |
| 137 | } |
| 138 | |
| 139 | static int h264_mp4toannexb_filter(AVBitStreamFilterContext *bsfc, |
| 140 | AVCodecContext *avctx, const char *args, |
| 141 | uint8_t **poutbuf, int *poutbuf_size, |
| 142 | const uint8_t *buf, int buf_size, |
| 143 | int keyframe) |
| 144 | { |
| 145 | H264BSFContext *ctx = bsfc->priv_data; |
| 146 | int i; |
| 147 | uint8_t unit_type; |
| 148 | int32_t nal_size; |
| 149 | uint32_t cumul_size = 0; |
| 150 | const uint8_t *buf_end = buf + buf_size; |
| 151 | int ret = 0; |
| 152 | |
| 153 | /* nothing to filter */ |
| 154 | if (!avctx->extradata || avctx->extradata_size < 6) { |
| 155 | *poutbuf = (uint8_t *)buf; |
| 156 | *poutbuf_size = buf_size; |
| 157 | return 0; |
| 158 | } |
| 159 | |
| 160 | /* retrieve sps and pps NAL units from extradata */ |
| 161 | if (!ctx->extradata_parsed) { |
| 162 | ret = h264_extradata_to_annexb(ctx, avctx, FF_INPUT_BUFFER_PADDING_SIZE); |
| 163 | if (ret < 0) |
| 164 | return ret; |
| 165 | ctx->length_size = ret; |
| 166 | ctx->new_idr = 1; |
| 167 | ctx->idr_sps_seen = 0; |
| 168 | ctx->idr_pps_seen = 0; |
| 169 | ctx->extradata_parsed = 1; |
| 170 | } |
| 171 | |
| 172 | *poutbuf_size = 0; |
| 173 | *poutbuf = NULL; |
| 174 | do { |
| 175 | ret= AVERROR(EINVAL); |
| 176 | if (buf + ctx->length_size > buf_end) |
| 177 | goto fail; |
| 178 | |
| 179 | for (nal_size = 0, i = 0; i<ctx->length_size; i++) |
| 180 | nal_size = (nal_size << 8) | buf[i]; |
| 181 | |
| 182 | buf += ctx->length_size; |
| 183 | unit_type = *buf & 0x1f; |
| 184 | |
| 185 | if (buf + nal_size > buf_end || nal_size < 0) |
| 186 | goto fail; |
| 187 | |
| 188 | if (unit_type == 7) |
| 189 | ctx->idr_sps_seen = ctx->new_idr = 1; |
| 190 | else if (unit_type == 8) { |
| 191 | ctx->idr_pps_seen = ctx->new_idr = 1; |
| 192 | /* if SPS has not been seen yet, prepend the AVCC one to PPS */ |
| 193 | if (!ctx->idr_sps_seen) { |
| 194 | if (ctx->sps_offset == -1) |
| 195 | av_log(avctx, AV_LOG_WARNING, "SPS not present in the stream, nor in AVCC, stream may be unreadable\n"); |
| 196 | else { |
| 197 | if ((ret = alloc_and_copy(poutbuf, poutbuf_size, |
| 198 | avctx->extradata + ctx->sps_offset, |
| 199 | ctx->pps_offset != -1 ? ctx->pps_offset : avctx->extradata_size - ctx->sps_offset, |
| 200 | buf, nal_size)) < 0) |
| 201 | goto fail; |
| 202 | ctx->idr_sps_seen = 1; |
| 203 | goto next_nal; |
| 204 | } |
| 205 | } |
| 206 | } |
| 207 | |
| 208 | /* if this is a new IDR picture following an IDR picture, reset the idr flag. |
| 209 | * Just check first_mb_in_slice to be 0 as this is the simplest solution. |
| 210 | * This could be checking idr_pic_id instead, but would complexify the parsing. */ |
| 211 | if (!ctx->new_idr && unit_type == 5 && (buf[1] & 0x80)) |
| 212 | ctx->new_idr = 1; |
| 213 | |
| 214 | /* prepend only to the first type 5 NAL unit of an IDR picture, if no sps/pps are already present */ |
| 215 | if (ctx->new_idr && unit_type == 5 && !ctx->idr_sps_seen && !ctx->idr_pps_seen) { |
| 216 | if ((ret=alloc_and_copy(poutbuf, poutbuf_size, |
| 217 | avctx->extradata, avctx->extradata_size, |
| 218 | buf, nal_size)) < 0) |
| 219 | goto fail; |
| 220 | ctx->new_idr = 0; |
| 221 | /* if only SPS has been seen, also insert PPS */ |
| 222 | } else if (ctx->new_idr && unit_type == 5 && ctx->idr_sps_seen && !ctx->idr_pps_seen) { |
| 223 | if (ctx->pps_offset == -1) { |
| 224 | av_log(avctx, AV_LOG_WARNING, "PPS not present in the stream, nor in AVCC, stream may be unreadable\n"); |
| 225 | if ((ret = alloc_and_copy(poutbuf, poutbuf_size, |
| 226 | NULL, 0, buf, nal_size)) < 0) |
| 227 | goto fail; |
| 228 | } else if ((ret = alloc_and_copy(poutbuf, poutbuf_size, |
| 229 | avctx->extradata + ctx->pps_offset, avctx->extradata_size - ctx->pps_offset, |
| 230 | buf, nal_size)) < 0) |
| 231 | goto fail; |
| 232 | } else { |
| 233 | if ((ret=alloc_and_copy(poutbuf, poutbuf_size, |
| 234 | NULL, 0, buf, nal_size)) < 0) |
| 235 | goto fail; |
| 236 | if (!ctx->new_idr && unit_type == 1) { |
| 237 | ctx->new_idr = 1; |
| 238 | ctx->idr_sps_seen = 0; |
| 239 | ctx->idr_pps_seen = 0; |
| 240 | } |
| 241 | } |
| 242 | |
| 243 | next_nal: |
| 244 | buf += nal_size; |
| 245 | cumul_size += nal_size + ctx->length_size; |
| 246 | } while (cumul_size < buf_size); |
| 247 | |
| 248 | return 1; |
| 249 | |
| 250 | fail: |
| 251 | av_freep(poutbuf); |
| 252 | *poutbuf_size = 0; |
| 253 | return ret; |
| 254 | } |
| 255 | |
| 256 | AVBitStreamFilter ff_h264_mp4toannexb_bsf = { |
| 257 | .name = "h264_mp4toannexb", |
| 258 | .priv_data_size = sizeof(H264BSFContext), |
| 259 | .filter = h264_mp4toannexb_filter, |
| 260 | }; |