rtpdec: experimental VP9 depacketizer (draft 0)
[libav.git] / libavformat / rtpdec_vp9.c
1 /*
2 * RTP parser for VP9 payload format (draft version 0) - experimental
3 * Copyright (c) 2015 Thomas Volkert <thomas@homer-conferencing.com>
4 *
5 * This file is part of Libav.
6 *
7 * Libav 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 * Libav 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 Libav; 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/intreadwrite.h"
23
24 #include "avio_internal.h"
25 #include "rtpdec_formats.h"
26
27 #define RTP_VP9_DESC_REQUIRED_SIZE 1
28
29 struct PayloadContext {
30 AVIOContext *buf;
31 uint32_t timestamp;
32 };
33
34 static av_cold int vp9_init(AVFormatContext *ctx, int st_index,
35 PayloadContext *data)
36 {
37 av_log(ctx, AV_LOG_WARNING,
38 "RTP/VP9 support is still experimental\n");
39
40 return 0;
41 }
42
43 static int vp9_handle_packet(AVFormatContext *ctx, PayloadContext *rtp_vp9_ctx,
44 AVStream *st, AVPacket *pkt, uint32_t *timestamp,
45 const uint8_t *buf, int len, uint16_t seq,
46 int flags)
47 {
48 int has_pic_id, has_layer_idc, has_ref_idc, has_ss_data, has_su_data;
49 av_unused int pic_id = 0, non_key_frame = 0;
50 av_unused int layer_temporal = -1, layer_spatial = -1, layer_quality = -1;
51 int ref_fields = 0, has_ref_field_ext_pic_id = 0;
52 int first_fragment, last_fragment;
53 int rtp_m;
54 int res = 0;
55
56 /* drop data of previous packets in case of non-continuous (lossy) packet stream */
57 if (rtp_vp9_ctx->buf && rtp_vp9_ctx->timestamp != *timestamp)
58 ffio_free_dyn_buf(&rtp_vp9_ctx->buf);
59
60 /* sanity check for size of input packet: 1 byte payload at least */
61 if (len < RTP_VP9_DESC_REQUIRED_SIZE + 1) {
62 av_log(ctx, AV_LOG_ERROR, "Too short RTP/VP9 packet, got %d bytes\n", len);
63 return AVERROR_INVALIDDATA;
64 }
65
66 /*
67 * decode the required VP9 payload descriptor according to section 4.2 of the spec.:
68 *
69 * 0 1 2 3 4 5 6 7
70 * +-+-+-+-+-+-+-+-+
71 * |I|L|F|B|E|V|U|-| (REQUIRED)
72 * +-+-+-+-+-+-+-+-+
73 *
74 * I: PictureID present
75 * L: Layer indices present
76 * F: Reference indices present
77 * B: Start of VP9 frame
78 * E: End of picture
79 * V: Scalability Structure (SS) present
80 * U: Scalability Structure Update (SU) present
81 */
82 has_pic_id = !!(buf[0] & 0x80);
83 has_layer_idc = !!(buf[0] & 0x40);
84 has_ref_idc = !!(buf[0] & 0x20);
85 first_fragment = !!(buf[0] & 0x10);
86 last_fragment = !!(buf[0] & 0x08);
87 has_ss_data = !!(buf[0] & 0x04);
88 has_su_data = !!(buf[0] & 0x02);
89
90 rtp_m = !!(flags & RTP_FLAG_MARKER);
91
92 /* sanity check for markers: B should always be equal to the RTP M marker */
93 if (last_fragment != rtp_m) {
94 av_log(ctx, AV_LOG_ERROR, "Invalid combination of B and M marker (%d != %d)\n", last_fragment, rtp_m);
95 return AVERROR_INVALIDDATA;
96 }
97
98 /* pass the extensions field */
99 buf += RTP_VP9_DESC_REQUIRED_SIZE;
100 len -= RTP_VP9_DESC_REQUIRED_SIZE;
101
102 /*
103 * decode the 1-byte/2-byte picture ID:
104 *
105 * 0 1 2 3 4 5 6 7
106 * +-+-+-+-+-+-+-+-+
107 * I: |M|PICTURE ID | (RECOMMENDED)
108 * +-+-+-+-+-+-+-+-+
109 * M: | EXTENDED PID | (RECOMMENDED)
110 * +-+-+-+-+-+-+-+-+
111 *
112 * M: The most significant bit of the first octet is an extension flag.
113 * PictureID: 8 or 16 bits including the M bit.
114 */
115 if (has_pic_id) {
116 if (len < 1) {
117 av_log(ctx, AV_LOG_ERROR, "Too short RTP/VP9 packet\n");
118 return AVERROR_INVALIDDATA;
119 }
120
121 /* check for 1-byte or 2-byte picture index */
122 if (buf[0] & 0x80) {
123 if (len < 2) {
124 av_log(ctx, AV_LOG_ERROR, "Too short RTP/VP9 packet\n");
125 return AVERROR_INVALIDDATA;
126 }
127 pic_id = AV_RB16(buf) & 0x7fff;
128 buf += 2;
129 len -= 2;
130 } else {
131 pic_id = buf[0] & 0x7f;
132 buf++;
133 len--;
134 }
135 }
136
137 /*
138 * decode layer indices
139 *
140 * 0 1 2 3 4 5 6 7
141 * +-+-+-+-+-+-+-+-+
142 * L: | T | S | Q | R | (CONDITIONALLY RECOMMENDED)
143 * +-+-+-+-+-+-+-+-+
144 *
145 * T, S and Q are 2-bit indices for temporal, spatial, and quality layers.
146 * If "F" is set in the initial octet, R is 2 bits representing the number
147 * of reference fields this frame refers to.
148 */
149 if (has_layer_idc) {
150 if (len < 1) {
151 av_log(ctx, AV_LOG_ERROR, "Too short RTP/VP9 packet\n");
152 return AVERROR_INVALIDDATA;
153 }
154 layer_temporal = buf[0] & 0xC0;
155 layer_spatial = buf[0] & 0x30;
156 layer_quality = buf[0] & 0x0C;
157 if (has_ref_idc) {
158 ref_fields = buf[0] & 0x03;
159 if (ref_fields)
160 non_key_frame = 1;
161 }
162 buf++;
163 len--;
164 }
165
166 /*
167 * decode the reference fields
168 *
169 * 0 1 2 3 4 5 6 7
170 * +-+-+-+-+-+-+-+-+ -\
171 * F: | PID |X| RS| RQ| (OPTIONAL) .
172 * +-+-+-+-+-+-+-+-+ . - R times
173 * X: | EXTENDED PID | (OPTIONAL) .
174 * +-+-+-+-+-+-+-+-+ -/
175 *
176 * PID: The relative Picture ID referred to by this frame.
177 * RS and RQ: The spatial and quality layer IDs.
178 * X: 1 if this layer index has an extended relative Picture ID.
179 */
180 if (has_ref_idc) {
181 while (ref_fields) {
182 if (len < 1) {
183 av_log(ctx, AV_LOG_ERROR, "Too short RTP/VP9 packet\n");
184 return AVERROR_INVALIDDATA;
185 }
186
187 has_ref_field_ext_pic_id = buf[0] & 0x10;
188
189 /* pass ref. field */
190 if (has_ref_field_ext_pic_id) {
191 if (len < 2) {
192 av_log(ctx, AV_LOG_ERROR, "Too short RTP/VP9 packet\n");
193 return AVERROR_INVALIDDATA;
194 }
195
196 /* ignore ref. data */
197
198 buf += 2;
199 len -= 2;
200 } else {
201
202 /* ignore ref. data */
203
204 buf++;
205 len--;
206 }
207 ref_fields--;
208 }
209 }
210
211 /*
212 * decode the scalability structure (SS)
213 *
214 * 0 1 2 3 4 5 6 7
215 * +-+-+-+-+-+-+-+-+
216 * V: | PATTERN LENGTH|
217 * +-+-+-+-+-+-+-+-+ -\
218 * | T | S | Q | R | (OPTIONAL) .
219 * +-+-+-+-+-+-+-+-+ -\ .
220 * | PID |X| RS| RQ| (OPTIONAL) . . - PAT. LEN. times
221 * +-+-+-+-+-+-+-+-+ . - R times .
222 * X: | EXTENDED PID | (OPTIONAL) . .
223 * +-+-+-+-+-+-+-+-+ -/ -/
224 *
225 * PID: The relative Picture ID referred to by this frame.
226 * RS and RQ: The spatial and quality layer IDs.
227 * X: 1 if this layer index has an extended relative Picture ID.
228 */
229 if (has_ss_data) {
230 avpriv_report_missing_feature(ctx, "VP9 scalability structure data");
231 return AVERROR(ENOSYS);
232 }
233
234 /*
235 * decode the scalability update structure (SU)
236 *
237 * spec. is tbd
238 */
239 if (has_su_data) {
240 avpriv_report_missing_feature(ctx, "VP9 scalability update structure data");
241 return AVERROR(ENOSYS);
242 }
243
244 /*
245 * decode the VP9 payload header
246 *
247 * spec. is tbd
248 */
249 //XXX: implement when specified
250
251 /* sanity check: 1 byte payload as minimum */
252 if (len < 1) {
253 av_log(ctx, AV_LOG_ERROR, "Too short RTP/VP9 packet\n");
254 return AVERROR_INVALIDDATA;
255 }
256
257 /* start frame buffering with new dynamic buffer */
258 if (!rtp_vp9_ctx->buf) {
259 /* sanity check: a new frame should have started */
260 if (first_fragment) {
261 res = avio_open_dyn_buf(&rtp_vp9_ctx->buf);
262 if (res < 0)
263 return res;
264 /* update the timestamp in the frame packet with the one from the RTP packet */
265 rtp_vp9_ctx->timestamp = *timestamp;
266 } else {
267 /* frame not started yet, need more packets */
268 return AVERROR(EAGAIN);
269 }
270 }
271
272 /* write the fragment to the dyn. buffer */
273 avio_write(rtp_vp9_ctx->buf, buf, len);
274
275 /* do we need more fragments? */
276 if (!last_fragment)
277 return AVERROR(EAGAIN);
278
279 /* close frame buffering and create resulting A/V packet */
280 res = ff_rtp_finalize_packet(pkt, &rtp_vp9_ctx->buf, st->index);
281 if (res < 0)
282 return res;
283
284 return 0;
285 }
286
287 RTPDynamicProtocolHandler ff_vp9_dynamic_handler = {
288 .enc_name = "VP9",
289 .codec_type = AVMEDIA_TYPE_VIDEO,
290 .codec_id = AV_CODEC_ID_VP9,
291 .priv_data_size = sizeof(PayloadContext),
292 .init = vp9_init,
293 .parse_packet = vp9_handle_packet
294 };