mirror of
https://github.com/xenia-project/FFmpeg.git
synced 2024-11-27 13:30:45 +00:00
3804d73e46
* commit '5d8cae45737bed6239bd6b6e0698802dbe1463c8': rtpdec: Get rid of all trivial .alloc/.free functions Merged-by: Michael Niedermayer <michaelni@gmx.at>
220 lines
7.5 KiB
C
220 lines
7.5 KiB
C
/*
|
|
* RTP Depacketization of QCELP/PureVoice, RFC 2658
|
|
* Copyright (c) 2010 Martin Storsjo
|
|
*
|
|
* This file is part of FFmpeg.
|
|
*
|
|
* FFmpeg 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.
|
|
*
|
|
* FFmpeg 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 FFmpeg; if not, write to the Free Software
|
|
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
|
|
*/
|
|
|
|
#include "rtpdec_formats.h"
|
|
|
|
static const uint8_t frame_sizes[] = {
|
|
1, 4, 8, 17, 35
|
|
};
|
|
|
|
typedef struct InterleavePacket {
|
|
int pos;
|
|
int size;
|
|
/* The largest frame is 35 bytes, only 10 frames are allowed per
|
|
* packet, and we return the first one immediately, so allocate
|
|
* space for 9 frames */
|
|
uint8_t data[35*9];
|
|
} InterleavePacket;
|
|
|
|
struct PayloadContext {
|
|
int interleave_size;
|
|
int interleave_index;
|
|
InterleavePacket group[6];
|
|
int group_finished;
|
|
|
|
/* The maximum packet size, 10 frames of 35 bytes each, and one
|
|
* packet header byte. */
|
|
uint8_t next_data[1 + 35*10];
|
|
int next_size;
|
|
uint32_t next_timestamp;
|
|
};
|
|
|
|
static int return_stored_frame(AVFormatContext *ctx, PayloadContext *data,
|
|
AVStream *st, AVPacket *pkt, uint32_t *timestamp,
|
|
const uint8_t *buf, int len);
|
|
|
|
static int store_packet(AVFormatContext *ctx, PayloadContext *data,
|
|
AVStream *st, AVPacket *pkt, uint32_t *timestamp,
|
|
const uint8_t *buf, int len)
|
|
{
|
|
int interleave_size, interleave_index;
|
|
int frame_size, ret;
|
|
InterleavePacket* ip;
|
|
|
|
if (len < 2)
|
|
return AVERROR_INVALIDDATA;
|
|
|
|
interleave_size = buf[0] >> 3 & 7;
|
|
interleave_index = buf[0] & 7;
|
|
|
|
if (interleave_size > 5) {
|
|
av_log(ctx, AV_LOG_ERROR, "Invalid interleave size %d\n",
|
|
interleave_size);
|
|
return AVERROR_INVALIDDATA;
|
|
}
|
|
if (interleave_index > interleave_size) {
|
|
av_log(ctx, AV_LOG_ERROR, "Invalid interleave index %d/%d\n",
|
|
interleave_index, interleave_size);
|
|
return AVERROR_INVALIDDATA;
|
|
}
|
|
if (interleave_size != data->interleave_size) {
|
|
int i;
|
|
/* First packet, or changed interleave size */
|
|
data->interleave_size = interleave_size;
|
|
data->interleave_index = 0;
|
|
for (i = 0; i < 6; i++)
|
|
data->group[i].size = 0;
|
|
}
|
|
|
|
if (interleave_index < data->interleave_index) {
|
|
/* Wrapped around - missed the last packet of the previous group. */
|
|
if (data->group_finished) {
|
|
/* No more data in the packets in this interleaving group, just
|
|
* start processing the next one */
|
|
data->interleave_index = 0;
|
|
} else {
|
|
/* Stash away the current packet, emit everything we have of the
|
|
* previous group. */
|
|
for (; data->interleave_index <= interleave_size;
|
|
data->interleave_index++)
|
|
data->group[data->interleave_index].size = 0;
|
|
|
|
if (len > sizeof(data->next_data))
|
|
return AVERROR_INVALIDDATA;
|
|
memcpy(data->next_data, buf, len);
|
|
data->next_size = len;
|
|
data->next_timestamp = *timestamp;
|
|
*timestamp = RTP_NOTS_VALUE;
|
|
|
|
data->interleave_index = 0;
|
|
return return_stored_frame(ctx, data, st, pkt, timestamp, buf, len);
|
|
}
|
|
}
|
|
if (interleave_index > data->interleave_index) {
|
|
/* We missed a packet */
|
|
for (; data->interleave_index < interleave_index;
|
|
data->interleave_index++)
|
|
data->group[data->interleave_index].size = 0;
|
|
}
|
|
data->interleave_index = interleave_index;
|
|
|
|
if (buf[1] >= FF_ARRAY_ELEMS(frame_sizes))
|
|
return AVERROR_INVALIDDATA;
|
|
frame_size = frame_sizes[buf[1]];
|
|
if (1 + frame_size > len)
|
|
return AVERROR_INVALIDDATA;
|
|
|
|
if (len - 1 - frame_size > sizeof(data->group[0].data))
|
|
return AVERROR_INVALIDDATA;
|
|
|
|
if ((ret = av_new_packet(pkt, frame_size)) < 0)
|
|
return ret;
|
|
memcpy(pkt->data, &buf[1], frame_size);
|
|
pkt->stream_index = st->index;
|
|
|
|
ip = &data->group[data->interleave_index];
|
|
ip->size = len - 1 - frame_size;
|
|
ip->pos = 0;
|
|
memcpy(ip->data, &buf[1 + frame_size], ip->size);
|
|
/* Each packet must contain the same number of frames according to the
|
|
* RFC. If there's no data left in this packet, there shouldn't be any
|
|
* in any of the other frames in the interleaving group either. */
|
|
data->group_finished = ip->size == 0;
|
|
|
|
if (interleave_index == interleave_size) {
|
|
data->interleave_index = 0;
|
|
return !data->group_finished;
|
|
} else {
|
|
data->interleave_index++;
|
|
return 0;
|
|
}
|
|
}
|
|
|
|
static int return_stored_frame(AVFormatContext *ctx, PayloadContext *data,
|
|
AVStream *st, AVPacket *pkt, uint32_t *timestamp,
|
|
const uint8_t *buf, int len)
|
|
{
|
|
InterleavePacket* ip = &data->group[data->interleave_index];
|
|
int frame_size, ret;
|
|
|
|
if (data->group_finished && data->interleave_index == 0) {
|
|
*timestamp = data->next_timestamp;
|
|
ret = store_packet(ctx, data, st, pkt, timestamp, data->next_data,
|
|
data->next_size);
|
|
data->next_size = 0;
|
|
return ret;
|
|
}
|
|
|
|
if (ip->size == 0) {
|
|
/* No stored data for this interleave block, output an empty packet */
|
|
if ((ret = av_new_packet(pkt, 1)) < 0)
|
|
return ret;
|
|
pkt->data[0] = 0; // Blank - could also be 14, Erasure
|
|
} else {
|
|
if (ip->pos >= ip->size)
|
|
return AVERROR_INVALIDDATA;
|
|
if (ip->data[ip->pos] >= FF_ARRAY_ELEMS(frame_sizes))
|
|
return AVERROR_INVALIDDATA;
|
|
frame_size = frame_sizes[ip->data[ip->pos]];
|
|
if (ip->pos + frame_size > ip->size)
|
|
return AVERROR_INVALIDDATA;
|
|
|
|
if ((ret = av_new_packet(pkt, frame_size)) < 0)
|
|
return ret;
|
|
memcpy(pkt->data, &ip->data[ip->pos], frame_size);
|
|
|
|
ip->pos += frame_size;
|
|
data->group_finished = ip->pos >= ip->size;
|
|
}
|
|
pkt->stream_index = st->index;
|
|
|
|
if (data->interleave_index == data->interleave_size) {
|
|
data->interleave_index = 0;
|
|
if (!data->group_finished)
|
|
return 1;
|
|
else
|
|
return data->next_size > 0;
|
|
} else {
|
|
data->interleave_index++;
|
|
return 1;
|
|
}
|
|
}
|
|
|
|
static int qcelp_parse_packet(AVFormatContext *ctx, PayloadContext *data,
|
|
AVStream *st, AVPacket *pkt, uint32_t *timestamp,
|
|
const uint8_t *buf, int len, uint16_t seq,
|
|
int flags)
|
|
{
|
|
if (buf)
|
|
return store_packet(ctx, data, st, pkt, timestamp, buf, len);
|
|
else
|
|
return return_stored_frame(ctx, data, st, pkt, timestamp, buf, len);
|
|
}
|
|
|
|
RTPDynamicProtocolHandler ff_qcelp_dynamic_handler = {
|
|
.enc_name = "x-Purevoice",
|
|
.codec_type = AVMEDIA_TYPE_AUDIO,
|
|
.codec_id = AV_CODEC_ID_QCELP,
|
|
.priv_data_size = sizeof(PayloadContext),
|
|
.static_payload_id = 12,
|
|
.parse_packet = qcelp_parse_packet,
|
|
};
|