2 * RTP parser for VP9 payload format (draft version 0) - experimental
3 * Copyright (c) 2015 Thomas Volkert <thomas@homer-conferencing.com>
5 * This file is part of FFmpeg.
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.
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.
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
23 #include "libavcodec/bytestream.h"
25 #include "rtpdec_formats.h"
27 #define RTP_VP9_DESC_REQUIRED_SIZE 1
29 struct PayloadContext {
34 static void vp9_free_dyn_buffer(AVIOContext **dyn_buf)
36 uint8_t *ptr_dyn_buffer;
37 avio_close_dyn_buf(*dyn_buf, &ptr_dyn_buffer);
38 av_free(ptr_dyn_buffer);
42 static av_cold int vp9_init(AVFormatContext *ctx, int st_index,
45 av_dlog(ctx, "vp9_init() for stream %d\n", st_index);
46 av_log(ctx, AV_LOG_WARNING,
47 "RTP/VP9 support is still experimental\n");
52 ctx->streams[st_index]->need_parsing = AVSTREAM_PARSE_FULL;
57 static int vp9_handle_packet(AVFormatContext *ctx, PayloadContext *rtp_vp9_ctx,
58 AVStream *st, AVPacket *pkt, uint32_t *timestamp,
59 const uint8_t *buf, int len, uint16_t seq,
62 int has_pic_id, has_layer_idc, has_ref_idc, has_ss_data, has_su_data;
63 av_unused int pic_id = 0, non_key_frame = 0;
64 av_unused int layer_temporal = -1, layer_spatial = -1, layer_quality = -1;
65 int ref_fields = 0, has_ref_field_ext_pic_id = 0;
66 int first_fragment, last_fragment;
69 /* drop data of previous packets in case of non-continuous (lossy) packet stream */
70 if (rtp_vp9_ctx->buf && rtp_vp9_ctx->timestamp != *timestamp) {
71 vp9_free_dyn_buffer(&rtp_vp9_ctx->buf);
74 /* sanity check for size of input packet: 1 byte payload at least */
75 if (len < RTP_VP9_DESC_REQUIRED_SIZE + 1) {
76 av_log(ctx, AV_LOG_ERROR, "Too short RTP/VP9 packet, got %d bytes\n", len);
77 return AVERROR_INVALIDDATA;
81 decode the required VP9 payload descriptor according to section 4.2 of the spec.:
85 |I|L|F|B|E|V|U|-| (REQUIRED)
89 L: Layer indices present
90 F: Reference indices present
93 V: Scalability Structure (SS) present
94 U: Scalability Structure Update (SU) present
96 has_pic_id = buf[0] & 0x80;
97 has_layer_idc = buf[0] & 0x40;
98 has_ref_idc = buf[0] & 0x20;
99 first_fragment = buf[0] & 0x10;
100 last_fragment = buf[0] & 0x08;
101 has_ss_data = buf[0] & 0x04;
102 has_su_data = buf[0] & 0x02;
104 /* sanity check for markers: B should always be equal to the RTP M marker */
105 if (last_fragment >> 2 != flags & RTP_FLAG_MARKER) {
106 av_log(ctx, AV_LOG_ERROR, "Invalid combination of B and M marker\n");
107 return AVERROR_INVALIDDATA;
110 /* pass the extensions field */
111 buf += RTP_VP9_DESC_REQUIRED_SIZE;
112 len -= RTP_VP9_DESC_REQUIRED_SIZE;
115 decode the 1-byte/2-byte picture ID:
119 I: |M|PICTURE ID | (RECOMMENDED)
121 M: | EXTENDED PID | (RECOMMENDED)
124 M: The most significant bit of the first octet is an extension flag.
125 PictureID: 8 or 16 bits including the M bit.
129 av_log(ctx, AV_LOG_ERROR, "Too short RTP/VP9 packet\n");
130 return AVERROR_INVALIDDATA;
133 /* check for 1-byte or 2-byte picture index */
136 av_log(ctx, AV_LOG_ERROR, "Too short RTP/VP9 packet\n");
137 return AVERROR_INVALIDDATA;
139 pic_id = AV_RB16(buf) & 0x7fff;
143 pic_id = buf[0] & 0x7f;
154 L: | T | S | Q | R | (CONDITIONALLY RECOMMENDED)
157 T, S and Q are 2-bit indices for temporal, spatial, and quality layers.
158 If "F" is set in the initial octet, R is 2 bits representing the number
159 of reference fields this frame refers to.
163 av_log(ctx, AV_LOG_ERROR, "Too short RTP/VP9 packet");
164 return AVERROR_INVALIDDATA;
166 layer_temporal = buf[0] & 0xC0;
167 layer_spatial = buf[0] & 0x30;
168 layer_quality = buf[0] & 0x0C;
170 ref_fields = buf[0] & 0x03;
179 decode the reference fields
183 F: | PID |X| RS| RQ| (OPTIONAL) .
184 +-+-+-+-+-+-+-+-+ . - R times
185 X: | EXTENDED PID | (OPTIONAL) .
188 PID: The relative Picture ID referred to by this frame.
189 RS and RQ: The spatial and quality layer IDs.
190 X: 1 if this layer index has an extended relative Picture ID.
195 av_log(ctx, AV_LOG_ERROR, "Too short RTP/VP9 packet\n");
196 return AVERROR_INVALIDDATA;
199 has_ref_field_ext_pic_id = buf[0] & 0x10;
201 /* pass ref. field */
202 if (has_ref_field_ext_pic_id) {
204 av_log(ctx, AV_LOG_ERROR, "Too short RTP/VP9 packet\n");
205 return AVERROR_INVALIDDATA;
208 /* ignore ref. data */
214 /* ignore ref. data */
224 decode the scalability structure (SS)
230 | T | S | Q | R | (OPTIONAL) .
231 +-+-+-+-+-+-+-+-+ -\ .
232 | PID |X| RS| RQ| (OPTIONAL) . . - PAT. LEN. times
233 +-+-+-+-+-+-+-+-+ . - R times .
234 X: | EXTENDED PID | (OPTIONAL) . .
235 +-+-+-+-+-+-+-+-+ -/ -/
237 PID: The relative Picture ID referred to by this frame.
238 RS and RQ: The spatial and quality layer IDs.
239 X: 1 if this layer index has an extended relative Picture ID.
242 avpriv_report_missing_feature(ctx, "VP9 scalability structure data\n");
243 return AVERROR_PATCHWELCOME;
247 decode the scalability update structure (SU)
252 avpriv_report_missing_feature(ctx, "VP9 scalability update structure data\n");
253 return AVERROR_PATCHWELCOME;
257 decode the VP9 payload header
261 //XXX: implement when specified
263 /* sanity check: 1 byte payload as minimum */
265 av_log(ctx, AV_LOG_ERROR, "Too short RTP/VP9 packet\n");
266 return AVERROR_INVALIDDATA;
269 /* start frame buffering with new dynamic buffer */
270 if (!rtp_vp9_ctx->buf) {
271 /* sanity check: a new frame should have started */
272 if (first_fragment) {
273 res = avio_open_dyn_buf(&rtp_vp9_ctx->buf);
276 /* update the timestamp in the frame packet with the one from the RTP packet */
277 rtp_vp9_ctx->timestamp = *timestamp;
279 /* frame not started yet, need more packets */
280 return AVERROR(EAGAIN);
284 /* write the fragment to the dyn. buffer */
285 avio_write(rtp_vp9_ctx->buf, buf, len);
287 /* do we need more fragments? */
289 return AVERROR(EAGAIN);
291 /* close frame buffering and create resulting A/V packet */
292 res = ff_rtp_finalize_packet(pkt, &rtp_vp9_ctx->buf, st->index);
299 RTPDynamicProtocolHandler ff_vp9_dynamic_handler = {
301 .codec_type = AVMEDIA_TYPE_VIDEO,
302 .codec_id = AV_CODEC_ID_VP9,
304 .priv_data_size = sizeof(PayloadContext),
305 .parse_packet = vp9_handle_packet