blob: 8687e654a3c5268a1c9e4a3d353db87662f0a6c3
| 1 | /* |
| 2 | * RTP AMR Depacketizer, RFC 3267 |
| 3 | * Copyright (c) 2010 Martin Storsjo |
| 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 "libavutil/channel_layout.h" |
| 23 | #include "avformat.h" |
| 24 | #include "rtpdec_formats.h" |
| 25 | #include "libavutil/avstring.h" |
| 26 | |
| 27 | static const uint8_t frame_sizes_nb[16] = { |
| 28 | 12, 13, 15, 17, 19, 20, 26, 31, 5, 0, 0, 0, 0, 0, 0, 0 |
| 29 | }; |
| 30 | static const uint8_t frame_sizes_wb[16] = { |
| 31 | 17, 23, 32, 36, 40, 46, 50, 58, 60, 5, 5, 0, 0, 0, 0, 0 |
| 32 | }; |
| 33 | |
| 34 | struct PayloadContext { |
| 35 | int octet_align; |
| 36 | int crc; |
| 37 | int interleaving; |
| 38 | int channels; |
| 39 | }; |
| 40 | |
| 41 | static av_cold int amr_init(AVFormatContext *s, int st_index, PayloadContext *data) |
| 42 | { |
| 43 | data->channels = 1; |
| 44 | return 0; |
| 45 | } |
| 46 | |
| 47 | static int amr_handle_packet(AVFormatContext *ctx, PayloadContext *data, |
| 48 | AVStream *st, AVPacket *pkt, uint32_t *timestamp, |
| 49 | const uint8_t *buf, int len, uint16_t seq, |
| 50 | int flags) |
| 51 | { |
| 52 | const uint8_t *frame_sizes = NULL; |
| 53 | int frames; |
| 54 | int i; |
| 55 | const uint8_t *speech_data; |
| 56 | uint8_t *ptr; |
| 57 | |
| 58 | if (st->codecpar->codec_id == AV_CODEC_ID_AMR_NB) { |
| 59 | frame_sizes = frame_sizes_nb; |
| 60 | } else if (st->codecpar->codec_id == AV_CODEC_ID_AMR_WB) { |
| 61 | frame_sizes = frame_sizes_wb; |
| 62 | } else { |
| 63 | av_log(ctx, AV_LOG_ERROR, "Bad codec ID\n"); |
| 64 | return AVERROR_INVALIDDATA; |
| 65 | } |
| 66 | |
| 67 | if (st->codecpar->channels != 1) { |
| 68 | av_log(ctx, AV_LOG_ERROR, "Only mono AMR is supported\n"); |
| 69 | return AVERROR_INVALIDDATA; |
| 70 | } |
| 71 | st->codecpar->channel_layout = AV_CH_LAYOUT_MONO; |
| 72 | |
| 73 | /* The AMR RTP packet consists of one header byte, followed |
| 74 | * by one TOC byte for each AMR frame in the packet, followed |
| 75 | * by the speech data for all the AMR frames. |
| 76 | * |
| 77 | * The header byte contains only a codec mode request, for |
| 78 | * requesting what kind of AMR data the sender wants to |
| 79 | * receive. Not used at the moment. |
| 80 | */ |
| 81 | |
| 82 | /* Count the number of frames in the packet. The highest bit |
| 83 | * is set in a TOC byte if there are more frames following. |
| 84 | */ |
| 85 | for (frames = 1; frames < len && (buf[frames] & 0x80); frames++) ; |
| 86 | |
| 87 | if (1 + frames >= len) { |
| 88 | /* We hit the end of the packet while counting frames. */ |
| 89 | av_log(ctx, AV_LOG_ERROR, "No speech data found\n"); |
| 90 | return AVERROR_INVALIDDATA; |
| 91 | } |
| 92 | |
| 93 | speech_data = buf + 1 + frames; |
| 94 | |
| 95 | /* Everything except the codec mode request byte should be output. */ |
| 96 | if (av_new_packet(pkt, len - 1)) { |
| 97 | av_log(ctx, AV_LOG_ERROR, "Out of memory\n"); |
| 98 | return AVERROR(ENOMEM); |
| 99 | } |
| 100 | pkt->stream_index = st->index; |
| 101 | ptr = pkt->data; |
| 102 | |
| 103 | for (i = 0; i < frames; i++) { |
| 104 | uint8_t toc = buf[1 + i]; |
| 105 | int frame_size = frame_sizes[(toc >> 3) & 0x0f]; |
| 106 | |
| 107 | if (speech_data + frame_size > buf + len) { |
| 108 | /* Too little speech data */ |
| 109 | av_log(ctx, AV_LOG_WARNING, "Too little speech data in the RTP packet\n"); |
| 110 | /* Set the unwritten part of the packet to zero. */ |
| 111 | memset(ptr, 0, pkt->data + pkt->size - ptr); |
| 112 | pkt->size = ptr - pkt->data; |
| 113 | return 0; |
| 114 | } |
| 115 | |
| 116 | /* Extract the AMR frame mode from the TOC byte */ |
| 117 | *ptr++ = toc & 0x7C; |
| 118 | |
| 119 | /* Copy the speech data */ |
| 120 | memcpy(ptr, speech_data, frame_size); |
| 121 | speech_data += frame_size; |
| 122 | ptr += frame_size; |
| 123 | } |
| 124 | |
| 125 | if (speech_data < buf + len) { |
| 126 | av_log(ctx, AV_LOG_WARNING, "Too much speech data in the RTP packet?\n"); |
| 127 | /* Set the unwritten part of the packet to zero. */ |
| 128 | memset(ptr, 0, pkt->data + pkt->size - ptr); |
| 129 | pkt->size = ptr - pkt->data; |
| 130 | } |
| 131 | |
| 132 | return 0; |
| 133 | } |
| 134 | |
| 135 | static int amr_parse_fmtp(AVFormatContext *s, |
| 136 | AVStream *stream, PayloadContext *data, |
| 137 | const char *attr, const char *value) |
| 138 | { |
| 139 | /* Some AMR SDP configurations contain "octet-align", without |
| 140 | * the trailing =1. Therefore, if the value is empty, |
| 141 | * interpret it as "1". |
| 142 | */ |
| 143 | if (!strcmp(value, "")) { |
| 144 | av_log(s, AV_LOG_WARNING, "AMR fmtp attribute %s had " |
| 145 | "nonstandard empty value\n", attr); |
| 146 | value = "1"; |
| 147 | } |
| 148 | if (!strcmp(attr, "octet-align")) |
| 149 | data->octet_align = atoi(value); |
| 150 | else if (!strcmp(attr, "crc")) |
| 151 | data->crc = atoi(value); |
| 152 | else if (!strcmp(attr, "interleaving")) |
| 153 | data->interleaving = atoi(value); |
| 154 | else if (!strcmp(attr, "channels")) |
| 155 | data->channels = atoi(value); |
| 156 | return 0; |
| 157 | } |
| 158 | |
| 159 | static int amr_parse_sdp_line(AVFormatContext *s, int st_index, |
| 160 | PayloadContext *data, const char *line) |
| 161 | { |
| 162 | const char *p; |
| 163 | int ret; |
| 164 | |
| 165 | if (st_index < 0) |
| 166 | return 0; |
| 167 | |
| 168 | /* Parse an fmtp line this one: |
| 169 | * a=fmtp:97 octet-align=1; interleaving=0 |
| 170 | * That is, a normal fmtp: line followed by semicolon & space |
| 171 | * separated key/value pairs. |
| 172 | */ |
| 173 | if (av_strstart(line, "fmtp:", &p)) { |
| 174 | ret = ff_parse_fmtp(s, s->streams[st_index], data, p, amr_parse_fmtp); |
| 175 | if (!data->octet_align || data->crc || |
| 176 | data->interleaving || data->channels != 1) { |
| 177 | av_log(s, AV_LOG_ERROR, "Unsupported RTP/AMR configuration!\n"); |
| 178 | return -1; |
| 179 | } |
| 180 | return ret; |
| 181 | } |
| 182 | return 0; |
| 183 | } |
| 184 | |
| 185 | RTPDynamicProtocolHandler ff_amr_nb_dynamic_handler = { |
| 186 | .enc_name = "AMR", |
| 187 | .codec_type = AVMEDIA_TYPE_AUDIO, |
| 188 | .codec_id = AV_CODEC_ID_AMR_NB, |
| 189 | .priv_data_size = sizeof(PayloadContext), |
| 190 | .init = amr_init, |
| 191 | .parse_sdp_a_line = amr_parse_sdp_line, |
| 192 | .parse_packet = amr_handle_packet, |
| 193 | }; |
| 194 | |
| 195 | RTPDynamicProtocolHandler ff_amr_wb_dynamic_handler = { |
| 196 | .enc_name = "AMR-WB", |
| 197 | .codec_type = AVMEDIA_TYPE_AUDIO, |
| 198 | .codec_id = AV_CODEC_ID_AMR_WB, |
| 199 | .priv_data_size = sizeof(PayloadContext), |
| 200 | .init = amr_init, |
| 201 | .parse_sdp_a_line = amr_parse_sdp_line, |
| 202 | .parse_packet = amr_handle_packet, |
| 203 | }; |
| 204 |