ffmpeg/libavformat/rtpdec_vp8.c
Martin Storsjö c7d4de3d73 rtpdec_vp8: Don't return known-broken packets
This is built on the assumption that the first partition of each
VP8 packet is essential for decoding any later packet - if this
partition is broken/missed, the arithmetic coder gets out of sync
and decoding the bitstream in further packet ends up with total
garbage. If packets of a frame are lost, make sure the first
partition is intact (return only this part of the packet, nothing
else), otherwise stop returning data until the next keyframe is
received.

Alternatively, one would simply not return any packets at all
until the next keyframe, if packet loss is detected.

Signed-off-by: Martin Storsjö <martin@martin.st>
2012-12-21 14:14:47 +02:00

271 lines
8.8 KiB
C

/*
* RTP VP8 Depacketizer
* Copyright (c) 2010 Josh Allmann
* Copyright (c) 2012 Martin Storsjo
*
* This file is part of Libav.
*
* Libav is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
* version 2.1 of the License, or (at your option) any later version.
*
* Libav is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with Libav; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
*/
/**
* @file
* @brief RTP support for the VP8 payload
* @author Josh Allmann <joshua.allmann@gmail.com>
* @see http://tools.ietf.org/html/draft-ietf-payload-vp8-05
*/
#include "libavcodec/bytestream.h"
#include "rtpdec_formats.h"
struct PayloadContext {
AVIOContext *data;
uint32_t timestamp;
int is_keyframe;
int sequence_ok;
int first_part_size;
uint16_t prev_seq;
int prev_pictureid;
int broken_frame;
};
static void vp8_free_buffer(PayloadContext *vp8)
{
uint8_t *tmp;
if (!vp8->data)
return;
avio_close_dyn_buf(vp8->data, &tmp);
av_free(tmp);
vp8->data = NULL;
}
static int vp8_broken_sequence(AVFormatContext *ctx, PayloadContext *vp8,
const char *msg)
{
vp8->sequence_ok = 0;
av_log(ctx, AV_LOG_WARNING, "%s", msg);
vp8_free_buffer(vp8);
return AVERROR(EAGAIN);
}
static int vp8_handle_packet(AVFormatContext *ctx, PayloadContext *vp8,
AVStream *st, AVPacket *pkt, uint32_t *timestamp,
const uint8_t *buf, int len, uint16_t seq,
int flags)
{
int start_partition, end_packet;
int extended_bits, part_id;
int pictureid_present = 0, tl0picidx_present = 0, tid_present = 0,
keyidx_present = 0;
int pictureid = -1, pictureid_mask;
int returned_old_frame = 0;
uint32_t old_timestamp;
if (!buf) {
if (vp8->data) {
int ret = ff_rtp_finalize_packet(pkt, &vp8->data, st->index);
if (ret < 0)
return ret;
return 0;
}
return AVERROR(EAGAIN);
}
if (len < 1)
return AVERROR_INVALIDDATA;
extended_bits = buf[0] & 0x80;
start_partition = buf[0] & 0x10;
part_id = buf[0] & 0x0f;
end_packet = flags & RTP_FLAG_MARKER;
buf++;
len--;
if (extended_bits) {
if (len < 1)
return AVERROR_INVALIDDATA;
pictureid_present = buf[0] & 0x80;
tl0picidx_present = buf[0] & 0x40;
tid_present = buf[0] & 0x20;
keyidx_present = buf[0] & 0x10;
buf++;
len--;
}
if (pictureid_present) {
if (len < 1)
return AVERROR_INVALIDDATA;
if (buf[0] & 0x80) {
if (len < 2)
return AVERROR_INVALIDDATA;
pictureid = AV_RB16(buf) & 0x7fff;
pictureid_mask = 0x7fff;
buf += 2;
len -= 2;
} else {
pictureid = buf[0] & 0x7f;
pictureid_mask = 0x7f;
buf++;
len--;
}
}
if (tl0picidx_present) {
// Ignoring temporal level zero index
buf++;
len--;
}
if (tid_present || keyidx_present) {
// Ignoring temporal layer index, layer sync bit and keyframe index
buf++;
len--;
}
if (len < 1)
return AVERROR_INVALIDDATA;
if (start_partition && part_id == 0 && len >= 3) {
int res;
int non_key = buf[0] & 0x01;
if (!non_key) {
vp8_free_buffer(vp8);
// Keyframe, decoding ok again
vp8->sequence_ok = 1;
} else {
int can_continue = vp8->data && !vp8->is_keyframe &&
avio_tell(vp8->data) >= vp8->first_part_size;
if (!vp8->sequence_ok)
return AVERROR(EAGAIN);
if (pictureid >= 0) {
if (pictureid != ((vp8->prev_pictureid + 1) & pictureid_mask)) {
return vp8_broken_sequence(ctx, vp8,
"Missed a picture, sequence broken\n");
} else {
if (vp8->data && !can_continue)
return vp8_broken_sequence(ctx, vp8,
"Missed a picture, sequence broken\n");
}
} else {
uint16_t expected_seq = vp8->prev_seq + 1;
int16_t diff = seq - expected_seq;
if (vp8->data) {
// No picture id, so we can't know if missed packets
// contained any new frames. If diff == 0, we did get
// later packets from the same frame (matching timestamp),
// so we know we didn't miss any frame. If diff == 1 and
// we still have data (not flushed by the end of frame
// marker), the single missed packet must have been part
// of the same frame.
if ((diff == 0 || diff == 1) && can_continue) {
// Proceed with what we have
} else {
return vp8_broken_sequence(ctx, vp8,
"Missed too much, sequence broken\n");
}
} else {
if (diff != 0)
return vp8_broken_sequence(ctx, vp8,
"Missed unknown data, sequence broken\n");
}
}
if (vp8->data) {
if (avio_tell(vp8->data) >= vp8->first_part_size) {
int ret = ff_rtp_finalize_packet(pkt, &vp8->data, st->index);
if (ret < 0)
return ret;
pkt->size = vp8->first_part_size;
returned_old_frame = 1;
old_timestamp = vp8->timestamp;
} else {
// Shouldn't happen
vp8_free_buffer(vp8);
}
}
}
vp8->first_part_size = (AV_RL16(&buf[1]) << 3 | buf[0] >> 5) + 3;
if ((res = avio_open_dyn_buf(&vp8->data)) < 0)
return res;
vp8->timestamp = *timestamp;
vp8->broken_frame = 0;
vp8->prev_pictureid = pictureid;
vp8->is_keyframe = !non_key;
} else {
uint16_t expected_seq = vp8->prev_seq + 1;
if (!vp8->sequence_ok)
return AVERROR(EAGAIN);
if (vp8->timestamp != *timestamp) {
// Missed the start of the new frame, sequence broken
vp8->sequence_ok = 0;
av_log(ctx, AV_LOG_WARNING,
"Received no start marker; dropping frame\n");
vp8_free_buffer(vp8);
return AVERROR(EAGAIN);
}
if (seq != expected_seq) {
if (vp8->is_keyframe) {
return vp8_broken_sequence(ctx, vp8,
"Missed part of a keyframe, sequence broken\n");
} else if (vp8->data && avio_tell(vp8->data) >= vp8->first_part_size) {
vp8->broken_frame = 1;
} else {
return vp8_broken_sequence(ctx, vp8,
"Missed part of the first partition, sequence broken\n");
}
}
}
if (!vp8->data)
return vp8_broken_sequence(ctx, vp8, "Received no start marker\n");
vp8->prev_seq = seq;
avio_write(vp8->data, buf, len);
if (end_packet) {
int ret;
if (returned_old_frame) {
*timestamp = old_timestamp;
return 1;
}
ret = ff_rtp_finalize_packet(pkt, &vp8->data, st->index);
if (ret < 0)
return ret;
if (vp8->broken_frame)
pkt->size = vp8->first_part_size;
return 0;
}
return AVERROR(EAGAIN);
}
static PayloadContext *vp8_new_context(void)
{
return av_mallocz(sizeof(PayloadContext));
}
static void vp8_free_context(PayloadContext *vp8)
{
vp8_free_buffer(vp8);
av_free(vp8);
}
RTPDynamicProtocolHandler ff_vp8_dynamic_handler = {
.enc_name = "VP8",
.codec_type = AVMEDIA_TYPE_VIDEO,
.codec_id = AV_CODEC_ID_VP8,
.alloc = vp8_new_context,
.free = vp8_free_context,
.parse_packet = vp8_handle_packet,
};