Add RTP packetization of Theora and Vorbis

Patch by Josh Allmann, joshua dot allmann at gmail

Originally committed as revision 24735 to svn://svn.ffmpeg.org/ffmpeg/trunk
oldabi
Josh Allmann 15 years ago committed by Martin Storsjö
parent a63dbf3970
commit 91af5601c1
  1. 1
      Changelog
  2. 1
      libavformat/Makefile
  3. 2
      libavformat/avformat.h
  4. 14
      libavformat/rtpenc.c
  5. 1
      libavformat/rtpenc.h
  6. 125
      libavformat/rtpenc_xiph.c
  7. 2
      libavformat/rtsp.c
  8. 115
      libavformat/sdp.c

@ -27,6 +27,7 @@ version <next>:
- SubRip subtitle file muxer and demuxer - SubRip subtitle file muxer and demuxer
- Chinese AVS encoding via libxavs - Chinese AVS encoding via libxavs
- ffprobe -show_packets option added - ffprobe -show_packets option added
- RTP packetization of Theora and Vorbis

@ -219,6 +219,7 @@ OBJS-$(CONFIG_RTP_MUXER) += rtp.o \
rtpenc_mpv.o \ rtpenc_mpv.o \
rtpenc.o \ rtpenc.o \
rtpenc_h264.o \ rtpenc_h264.o \
rtpenc_xiph.o \
avc.o avc.o
OBJS-$(CONFIG_RTSP_DEMUXER) += rtsp.o httpauth.o OBJS-$(CONFIG_RTSP_DEMUXER) += rtsp.o httpauth.o
OBJS-$(CONFIG_RTSP_MUXER) += rtsp.o rtspenc.o httpauth.o OBJS-$(CONFIG_RTSP_MUXER) += rtsp.o rtspenc.o httpauth.o

@ -23,7 +23,7 @@
#define LIBAVFORMAT_VERSION_MAJOR 52 #define LIBAVFORMAT_VERSION_MAJOR 52
#define LIBAVFORMAT_VERSION_MINOR 78 #define LIBAVFORMAT_VERSION_MINOR 78
#define LIBAVFORMAT_VERSION_MICRO 0 #define LIBAVFORMAT_VERSION_MICRO 1
#define LIBAVFORMAT_VERSION_INT AV_VERSION_INT(LIBAVFORMAT_VERSION_MAJOR, \ #define LIBAVFORMAT_VERSION_INT AV_VERSION_INT(LIBAVFORMAT_VERSION_MAJOR, \
LIBAVFORMAT_VERSION_MINOR, \ LIBAVFORMAT_VERSION_MINOR, \

@ -53,6 +53,8 @@ static int is_supported(enum CodecID id)
case CODEC_ID_MPEG2TS: case CODEC_ID_MPEG2TS:
case CODEC_ID_AMR_NB: case CODEC_ID_AMR_NB:
case CODEC_ID_AMR_WB: case CODEC_ID_AMR_WB:
case CODEC_ID_VORBIS:
case CODEC_ID_THEORA:
return 1; return 1;
default: default:
return 0; return 0;
@ -135,6 +137,13 @@ static int rtp_write_header(AVFormatContext *s1)
s->nal_length_size = (st->codec->extradata[4] & 0x03) + 1; s->nal_length_size = (st->codec->extradata[4] & 0x03) + 1;
} }
break; break;
case CODEC_ID_VORBIS:
case CODEC_ID_THEORA:
if (!s->max_frames_per_packet) s->max_frames_per_packet = 15;
s->max_frames_per_packet = av_clip(s->max_frames_per_packet, 1, 15);
s->max_payload_size -= 6; // ident+frag+tdt/vdt+pkt_num+pkt_length
s->num_frames = 0;
goto defaultcase;
case CODEC_ID_AMR_NB: case CODEC_ID_AMR_NB:
case CODEC_ID_AMR_WB: case CODEC_ID_AMR_WB:
if (!s->max_frames_per_packet) if (!s->max_frames_per_packet)
@ -155,6 +164,7 @@ static int rtp_write_header(AVFormatContext *s1)
case CODEC_ID_AAC: case CODEC_ID_AAC:
s->num_frames = 0; s->num_frames = 0;
default: default:
defaultcase:
if (st->codec->codec_type == AVMEDIA_TYPE_AUDIO) { if (st->codec->codec_type == AVMEDIA_TYPE_AUDIO) {
av_set_pts_info(st, 32, 1, st->codec->sample_rate); av_set_pts_info(st, 32, 1, st->codec->sample_rate);
} }
@ -393,6 +403,10 @@ static int rtp_write_packet(AVFormatContext *s1, AVPacket *pkt)
case CODEC_ID_H263P: case CODEC_ID_H263P:
ff_rtp_send_h263(s1, pkt->data, size); ff_rtp_send_h263(s1, pkt->data, size);
break; break;
case CODEC_ID_VORBIS:
case CODEC_ID_THEORA:
ff_rtp_send_xiph(s1, pkt->data, size);
break;
default: default:
/* better than nothing : send the codec raw data */ /* better than nothing : send the codec raw data */
rtp_send_raw(s1, pkt->data, size); rtp_send_raw(s1, pkt->data, size);

@ -67,5 +67,6 @@ void ff_rtp_send_h263(AVFormatContext *s1, const uint8_t *buf1, int size);
void ff_rtp_send_aac(AVFormatContext *s1, const uint8_t *buff, int size); void ff_rtp_send_aac(AVFormatContext *s1, const uint8_t *buff, int size);
void ff_rtp_send_amr(AVFormatContext *s1, const uint8_t *buff, int size); void ff_rtp_send_amr(AVFormatContext *s1, const uint8_t *buff, int size);
void ff_rtp_send_mpegvideo(AVFormatContext *s1, const uint8_t *buf1, int size); void ff_rtp_send_mpegvideo(AVFormatContext *s1, const uint8_t *buf1, int size);
void ff_rtp_send_xiph(AVFormatContext *s1, const uint8_t *buff, int size);
#endif /* AVFORMAT_RTPENC_H */ #endif /* AVFORMAT_RTPENC_H */

@ -0,0 +1,125 @@
/*
* RTP packetization for Xiph audio and video
* Copyright (c) 2010 Josh Allmann
*
* 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 "avformat.h"
#include "rtpenc.h"
/**
* Packetize Xiph frames into RTP according to
* RFC 5215 (Vorbis) and the Theora RFC draft.
* (http://svn.xiph.org/trunk/theora/doc/draft-ietf-avt-rtp-theora-00.txt)
*/
void ff_rtp_send_xiph(AVFormatContext *s1, const uint8_t *buff, int size)
{
RTPMuxContext *s = s1->priv_data;
int max_pkt_size, xdt, frag;
uint8_t *q;
max_pkt_size = s->max_payload_size;
// set xiph data type
switch (*buff) {
case 0x01: // vorbis id
case 0x05: // vorbis setup
case 0x80: // theora header
case 0x82: // theora tables
xdt = 1; // packed config payload
break;
case 0x03: // vorbis comments
case 0x81: // theora comments
xdt = 2; // comment payload
break;
default:
xdt = 0; // raw data payload
break;
}
// Set ident. Must match the one in sdp.c
// Probably need a non-fixed way of generating
// this, but it has to be done in SDP and passed in from there.
q = s->buf;
*q++ = 0xfe;
*q++ = 0xcd;
*q++ = 0xba;
// set fragment
// 0 - whole frame (possibly multiple frames)
// 1 - first fragment
// 2 - fragment continuation
// 3 - last fragmement
frag = size <= max_pkt_size ? 0 : 1;
if (!frag && !xdt) { // do we have a whole frame of raw data?
unsigned end_ptr = (unsigned)s->buf + 6 + max_pkt_size; // what we're allowed to write
unsigned ptr = (unsigned)s->buf_ptr + 2 + size; // what we're going to write
int remaining = end_ptr - ptr;
if ((s->num_frames > 0 && remaining < 0) ||
s->num_frames >= s->max_frames_per_packet) {
// send previous packets now; no room for new data
ff_rtp_send_data(s1, s->buf, s->buf_ptr - s->buf, 0);
s->num_frames = 0;
}
// buffer current frame to send later
if (0 == s->num_frames) s->timestamp = s->cur_timestamp;
s->num_frames++;
// Set packet header. Normally, this is OR'd with frag and xdt,
// but those are zero, so omitted here
*q++ = s->num_frames;
if (s->num_frames > 1) q = s->buf_ptr; // jump ahead if needed
*q++ = (size >> 8) & 0xff;
*q++ = size & 0xff;
memcpy(q, buff, size);
q += size;
s->buf_ptr = q;
return;
} else if (s->num_frames) {
// immediately send buffered frames if buffer is not raw data,
// or if current frame is fragmented.
ff_rtp_send_data(s1, s->buf, s->buf_ptr - s->buf, 0);
}
s->timestamp = s->cur_timestamp;
s->num_frames = 0;
s->buf_ptr = q;
while (size > 0) {
int len = (!frag || frag == 3) ? size : max_pkt_size;
q = s->buf_ptr;
// set packet headers
*q++ = (frag << 6) | (xdt << 4); // num_frames = 0
*q++ = (len >> 8) & 0xff;
*q++ = len & 0xff;
// set packet body
memcpy(q, buff, len);
q += len;
buff += len;
size -= len;
ff_rtp_send_data(s1, s->buf, q - s->buf, 0);
frag = size <= max_pkt_size ? 3 : 2;
}
}

@ -52,7 +52,7 @@ int rtsp_default_protocols = (1 << RTSP_LOWER_TRANSPORT_UDP);
#define SELECT_TIMEOUT_MS 100 #define SELECT_TIMEOUT_MS 100
#define READ_PACKET_TIMEOUT_S 10 #define READ_PACKET_TIMEOUT_S 10
#define MAX_TIMEOUTS READ_PACKET_TIMEOUT_S * 1000 / SELECT_TIMEOUT_MS #define MAX_TIMEOUTS READ_PACKET_TIMEOUT_S * 1000 / SELECT_TIMEOUT_MS
#define SDP_MAX_SIZE 8192 #define SDP_MAX_SIZE 16384
static void get_word_until_chars(char *buf, int buf_size, static void get_word_until_chars(char *buf, int buf_size,
const char *sep, const char **pp) const char *sep, const char **pp)

@ -21,6 +21,7 @@
#include <string.h> #include <string.h>
#include "libavutil/avstring.h" #include "libavutil/avstring.h"
#include "libavutil/base64.h" #include "libavutil/base64.h"
#include "libavcodec/xiph.h"
#include "avformat.h" #include "avformat.h"
#include "internal.h" #include "internal.h"
#include "avc.h" #include "avc.h"
@ -220,6 +221,75 @@ static char *extradata2config(AVCodecContext *c)
return config; return config;
} }
static char *xiph_extradata2config(AVCodecContext *c)
{
char *config, *encoded_config;
uint8_t *header_start[3];
int headers_len, header_len[3], config_len;
int first_header_size;
switch (c->codec_id) {
case CODEC_ID_THEORA:
first_header_size = 42;
break;
case CODEC_ID_VORBIS:
first_header_size = 30;
break;
default:
av_log(c, AV_LOG_ERROR, "Unsupported Xiph codec ID\n");
return NULL;
}
if (ff_split_xiph_headers(c->extradata, c->extradata_size,
first_header_size, header_start,
header_len) < 0) {
av_log(c, AV_LOG_ERROR, "Extradata corrupt.\n");
return NULL;
}
headers_len = header_len[0] + header_len[2];
config_len = 4 + // count
3 + // ident
2 + // packet size
1 + // header count
2 + // header size
headers_len; // and the rest
config = av_malloc(config_len);
if (!config)
goto xiph_fail;
encoded_config = av_malloc(AV_BASE64_SIZE(config_len));
if (!encoded_config) {
av_free(config);
goto xiph_fail;
}
config[0] = config[1] = config[2] = 0;
config[3] = 1;
config[4] = 0xfe; // ident must match the one in rtpenc_xiph.c
config[5] = 0xcd;
config[6] = 0xba;
config[7] = (headers_len >> 8) & 0xff;
config[8] = headers_len & 0xff;
config[9] = 2;
config[10] = header_len[0];
config[11] = 0; // size of comment header; nonexistent
memcpy(config + 12, header_start[0], header_len[0]);
memcpy(config + 12 + header_len[0], header_start[2], header_len[2]);
av_base64_encode(encoded_config, AV_BASE64_SIZE(config_len),
config, config_len);
av_free(config);
return encoded_config;
xiph_fail:
av_log(c, AV_LOG_ERROR,
"Not enough memory for configuration string\n");
return NULL;
}
static char *sdp_write_media_attributes(char *buff, int size, AVCodecContext *c, int payload_type) static char *sdp_write_media_attributes(char *buff, int size, AVCodecContext *c, int payload_type)
{ {
char *config = NULL; char *config = NULL;
@ -297,6 +367,51 @@ static char *sdp_write_media_attributes(char *buff, int size, AVCodecContext *c,
payload_type, c->sample_rate, c->channels, payload_type, c->sample_rate, c->channels,
payload_type); payload_type);
break; break;
case CODEC_ID_VORBIS:
if (c->extradata_size)
config = xiph_extradata2config(c);
else
av_log(c, AV_LOG_ERROR, "Vorbis configuration info missing\n");
if (!config)
return NULL;
av_strlcatf(buff, size, "a=rtpmap:%d vorbis/%d/%d\r\n"
"a=fmtp:%d configuration=%s\r\n",
payload_type, c->sample_rate, c->channels,
payload_type, config);
break;
case CODEC_ID_THEORA: {
const char *pix_fmt;
if (c->extradata_size)
config = xiph_extradata2config(c);
else
av_log(c, AV_LOG_ERROR, "Theora configuation info missing\n");
if (!config)
return NULL;
switch (c->pix_fmt) {
case PIX_FMT_YUV420P:
pix_fmt = "YCbCr-4:2:0";
break;
case PIX_FMT_YUV422P:
pix_fmt = "YCbCr-4:2:2";
break;
case PIX_FMT_YUV444P:
pix_fmt = "YCbCr-4:4:4";
break;
default:
av_log(c, AV_LOG_ERROR, "Unsupported pixel format.\n");
return NULL;
}
av_strlcatf(buff, size, "a=rtpmap:%d theora/90000\r\n"
"a=fmtp:%d delivery-method=inline; "
"width=%d; height=%d; sampling=%s; "
"configuration=%s\r\n",
payload_type, payload_type,
c->width, c->height, pix_fmt, config);
break;
}
default: default:
/* Nothing special to do here... */ /* Nothing special to do here... */
break; break;

Loading…
Cancel
Save