ffmpeg/libavformat/rmdec.c

833 lines
27 KiB
C
Raw Normal View History

/*
* "Real" compatible demuxer.
* Copyright (c) 2000, 2001 Fabrice Bellard
*
* 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 "libavutil/avstring.h"
#include "libavutil/intreadwrite.h"
#include "avformat.h"
#include "rm.h"
struct RMStream {
AVPacket pkt; ///< place to store merged video frame / reordered audio data
int videobufsize; ///< current assembled frame size
int videobufpos; ///< position for the next slice in the video buffer
int curpic_num; ///< picture number of current frame
int cur_slice, slices;
int64_t pktpos; ///< first slice position in file
/// Audio descrambling matrix parameters
int64_t audiotimestamp; ///< Audio packet timestamp
int sub_packet_cnt; // Subpacket counter, used while reading
int sub_packet_size, sub_packet_h, coded_framesize; ///< Descrambling parameters from container
int audio_framesize; /// Audio frame size from container
int sub_packet_lengths[16]; /// Length of each subpacket
};
typedef struct {
int nb_packets;
int old_format;
int current_stream;
int remaining_len;
int audio_stream_num; ///< Stream number for audio packets
int audio_pkt_cnt; ///< Output packet counter
} RMDemuxContext;
static inline void get_strl(ByteIOContext *pb, char *buf, int buf_size, int len)
{
int i;
char *q, r;
q = buf;
for(i=0;i<len;i++) {
r = get_byte(pb);
if (i < buf_size - 1)
*q++ = r;
}
if (buf_size > 0) *q = '\0';
}
static void get_str8(ByteIOContext *pb, char *buf, int buf_size)
{
get_strl(pb, buf, buf_size, get_byte(pb));
}
static void rm_read_metadata(AVFormatContext *s, int wide)
{
char buf[1024];
int i;
for (i=0; i<FF_ARRAY_ELEMS(ff_rm_metadata); i++) {
int len = wide ? get_be16(s->pb) : get_byte(s->pb);
get_strl(s->pb, buf, sizeof(buf), len);
av_metadata_set(&s->metadata, ff_rm_metadata[i], buf);
}
}
RMStream *ff_rm_alloc_rmstream (void)
{
RMStream *rms = av_mallocz(sizeof(RMStream));
rms->curpic_num = -1;
return rms;
}
void ff_rm_free_rmstream (RMStream *rms)
{
av_free_packet(&rms->pkt);
}
static int rm_read_audio_stream_info(AVFormatContext *s, ByteIOContext *pb,
AVStream *st, RMStream *ast, int read_all)
{
char buf[256];
uint32_t version;
/* ra type header */
version = get_be32(pb); /* version */
if (((version >> 16) & 0xff) == 3) {
int64_t startpos = url_ftell(pb);
url_fskip(pb, 14);
rm_read_metadata(s, 0);
if ((startpos + (version & 0xffff)) >= url_ftell(pb) + 2) {
// fourcc (should always be "lpcJ")
get_byte(pb);
get_str8(pb, buf, sizeof(buf));
}
// Skip extra header crap (this should never happen)
if ((startpos + (version & 0xffff)) > url_ftell(pb))
url_fskip(pb, (version & 0xffff) + startpos - url_ftell(pb));
st->codec->sample_rate = 8000;
st->codec->channels = 1;
st->codec->codec_type = CODEC_TYPE_AUDIO;
st->codec->codec_id = CODEC_ID_RA_144;
} else {
int flavor, sub_packet_h, coded_framesize, sub_packet_size;
/* old version (4) */
get_be32(pb); /* .ra4 */
get_be32(pb); /* data size */
get_be16(pb); /* version2 */
get_be32(pb); /* header size */
flavor= get_be16(pb); /* add codec info / flavor */
ast->coded_framesize = coded_framesize = get_be32(pb); /* coded frame size */
get_be32(pb); /* ??? */
get_be32(pb); /* ??? */
get_be32(pb); /* ??? */
ast->sub_packet_h = sub_packet_h = get_be16(pb); /* 1 */
st->codec->block_align= get_be16(pb); /* frame size */
ast->sub_packet_size = sub_packet_size = get_be16(pb); /* sub packet size */
get_be16(pb); /* ??? */
if (((version >> 16) & 0xff) == 5) {
get_be16(pb); get_be16(pb); get_be16(pb);
}
st->codec->sample_rate = get_be16(pb);
get_be32(pb);
st->codec->channels = get_be16(pb);
if (((version >> 16) & 0xff) == 5) {
get_be32(pb);
get_buffer(pb, buf, 4);
buf[4] = 0;
} else {
get_str8(pb, buf, sizeof(buf)); /* desc */
get_str8(pb, buf, sizeof(buf)); /* desc */
}
st->codec->codec_type = CODEC_TYPE_AUDIO;
if (!strcmp(buf, "dnet")) {
st->codec->codec_id = CODEC_ID_AC3;
st->need_parsing = AVSTREAM_PARSE_FULL;
} else if (!strcmp(buf, "28_8")) {
st->codec->codec_id = CODEC_ID_RA_288;
st->codec->extradata_size= 0;
ast->audio_framesize = st->codec->block_align;
st->codec->block_align = coded_framesize;
if(ast->audio_framesize >= UINT_MAX / sub_packet_h){
av_log(s, AV_LOG_ERROR, "ast->audio_framesize * sub_packet_h too large\n");
return -1;
}
av_new_packet(&ast->pkt, ast->audio_framesize * sub_packet_h);
} else if ((!strcmp(buf, "cook")) || (!strcmp(buf, "atrc")) || (!strcmp(buf, "sipr"))) {
int codecdata_length;
get_be16(pb); get_byte(pb);
if (((version >> 16) & 0xff) == 5)
get_byte(pb);
codecdata_length = get_be32(pb);
if(codecdata_length + FF_INPUT_BUFFER_PADDING_SIZE <= (unsigned)codecdata_length){
av_log(s, AV_LOG_ERROR, "codecdata_length too large\n");
return -1;
}
if(sub_packet_size <= 0){
av_log(s, AV_LOG_ERROR, "sub_packet_size is invalid\n");
return -1;
}
if (!strcmp(buf, "cook")) st->codec->codec_id = CODEC_ID_COOK;
else if (!strcmp(buf, "sipr")) st->codec->codec_id = CODEC_ID_SIPR;
else st->codec->codec_id = CODEC_ID_ATRAC3;
st->codec->extradata_size= codecdata_length;
st->codec->extradata= av_mallocz(st->codec->extradata_size + FF_INPUT_BUFFER_PADDING_SIZE);
get_buffer(pb, st->codec->extradata, st->codec->extradata_size);
ast->audio_framesize = st->codec->block_align;
st->codec->block_align = ast->sub_packet_size;
if(ast->audio_framesize >= UINT_MAX / sub_packet_h){
av_log(s, AV_LOG_ERROR, "rm->audio_framesize * sub_packet_h too large\n");
return -1;
}
av_new_packet(&ast->pkt, ast->audio_framesize * sub_packet_h);
} else if (!strcmp(buf, "raac") || !strcmp(buf, "racp")) {
int codecdata_length;
get_be16(pb); get_byte(pb);
if (((version >> 16) & 0xff) == 5)
get_byte(pb);
st->codec->codec_id = CODEC_ID_AAC;
codecdata_length = get_be32(pb);
if(codecdata_length + FF_INPUT_BUFFER_PADDING_SIZE <= (unsigned)codecdata_length){
av_log(s, AV_LOG_ERROR, "codecdata_length too large\n");
return -1;
}
if (codecdata_length >= 1) {
st->codec->extradata_size = codecdata_length - 1;
st->codec->extradata = av_mallocz(st->codec->extradata_size + FF_INPUT_BUFFER_PADDING_SIZE);
get_byte(pb);
get_buffer(pb, st->codec->extradata, st->codec->extradata_size);
}
} else {
st->codec->codec_id = CODEC_ID_NONE;
av_strlcpy(st->codec->codec_name, buf, sizeof(st->codec->codec_name));
}
if (read_all) {
get_byte(pb);
get_byte(pb);
get_byte(pb);
rm_read_metadata(s, 0);
}
}
return 0;
}
int
ff_rm_read_mdpr_codecdata (AVFormatContext *s, ByteIOContext *pb,
AVStream *st, RMStream *rst, int codec_data_size)
{
unsigned int v;
int size;
int64_t codec_pos;
av_set_pts_info(st, 64, 1, 1000);
codec_pos = url_ftell(pb);
v = get_be32(pb);
if (v == MKTAG(0xfd, 'a', 'r', '.')) {
/* ra type header */
if (rm_read_audio_stream_info(s, pb, st, rst, 0))
return -1;
} else {
int fps, fps2;
if (get_le32(pb) != MKTAG('V', 'I', 'D', 'O')) {
fail1:
av_log(st->codec, AV_LOG_ERROR, "Unsupported video codec\n");
goto skip;
}
st->codec->codec_tag = get_le32(pb);
// av_log(s, AV_LOG_DEBUG, "%X %X\n", st->codec->codec_tag, MKTAG('R', 'V', '2', '0'));
if ( st->codec->codec_tag != MKTAG('R', 'V', '1', '0')
&& st->codec->codec_tag != MKTAG('R', 'V', '2', '0')
&& st->codec->codec_tag != MKTAG('R', 'V', '3', '0')
&& st->codec->codec_tag != MKTAG('R', 'V', '4', '0')
&& st->codec->codec_tag != MKTAG('R', 'V', 'T', 'R'))
goto fail1;
st->codec->width = get_be16(pb);
st->codec->height = get_be16(pb);
st->codec->time_base.num= 1;
fps= get_be16(pb);
st->codec->codec_type = CODEC_TYPE_VIDEO;
get_be32(pb);
fps2= get_be16(pb);
get_be16(pb);
st->codec->extradata_size= codec_data_size - (url_ftell(pb) - codec_pos);
if(st->codec->extradata_size + FF_INPUT_BUFFER_PADDING_SIZE <= (unsigned)st->codec->extradata_size){
//check is redundant as get_buffer() will catch this
av_log(s, AV_LOG_ERROR, "st->codec->extradata_size too large\n");
return -1;
}
st->codec->extradata= av_mallocz(st->codec->extradata_size + FF_INPUT_BUFFER_PADDING_SIZE);
if (!st->codec->extradata)
return AVERROR(ENOMEM);
get_buffer(pb, st->codec->extradata, st->codec->extradata_size);
// av_log(s, AV_LOG_DEBUG, "fps= %d fps2= %d\n", fps, fps2);
st->codec->time_base.den = fps * st->codec->time_base.num;
switch(((uint8_t*)st->codec->extradata)[4]>>4){
case 1: st->codec->codec_id = CODEC_ID_RV10; break;
case 2: st->codec->codec_id = CODEC_ID_RV20; break;
case 3: st->codec->codec_id = CODEC_ID_RV30; break;
case 4: st->codec->codec_id = CODEC_ID_RV40; break;
default: goto fail1;
}
}
skip:
/* skip codec info */
size = url_ftell(pb) - codec_pos;
url_fskip(pb, codec_data_size - size);
return 0;
}
static int rm_read_header_old(AVFormatContext *s, AVFormatParameters *ap)
{
RMDemuxContext *rm = s->priv_data;
AVStream *st;
rm->old_format = 1;
st = av_new_stream(s, 0);
if (!st)
return -1;
st->priv_data = ff_rm_alloc_rmstream();
return rm_read_audio_stream_info(s, s->pb, st, st->priv_data, 1);
}
static int rm_read_header(AVFormatContext *s, AVFormatParameters *ap)
{
RMDemuxContext *rm = s->priv_data;
AVStream *st;
ByteIOContext *pb = s->pb;
unsigned int tag;
int tag_size;
unsigned int start_time, duration;
char buf[128];
int flags = 0;
tag = get_le32(pb);
if (tag == MKTAG('.', 'r', 'a', 0xfd)) {
/* very old .ra format */
return rm_read_header_old(s, ap);
} else if (tag != MKTAG('.', 'R', 'M', 'F')) {
return AVERROR(EIO);
}
get_be32(pb); /* header size */
get_be16(pb);
get_be32(pb);
get_be32(pb); /* number of headers */
for(;;) {
if (url_feof(pb))
return -1;
tag = get_le32(pb);
tag_size = get_be32(pb);
get_be16(pb);
#if 0
printf("tag=%c%c%c%c (%08x) size=%d\n",
(tag) & 0xff,
(tag >> 8) & 0xff,
(tag >> 16) & 0xff,
(tag >> 24) & 0xff,
tag,
tag_size);
#endif
if (tag_size < 10 && tag != MKTAG('D', 'A', 'T', 'A'))
return -1;
switch(tag) {
case MKTAG('P', 'R', 'O', 'P'):
/* file header */
get_be32(pb); /* max bit rate */
get_be32(pb); /* avg bit rate */
get_be32(pb); /* max packet size */
get_be32(pb); /* avg packet size */
get_be32(pb); /* nb packets */
get_be32(pb); /* duration */
get_be32(pb); /* preroll */
get_be32(pb); /* index offset */
get_be32(pb); /* data offset */
get_be16(pb); /* nb streams */
flags = get_be16(pb); /* flags */
break;
case MKTAG('C', 'O', 'N', 'T'):
rm_read_metadata(s, 1);
break;
case MKTAG('M', 'D', 'P', 'R'):
st = av_new_stream(s, 0);
if (!st)
return AVERROR(ENOMEM);
st->id = get_be16(pb);
get_be32(pb); /* max bit rate */
st->codec->bit_rate = get_be32(pb); /* bit rate */
get_be32(pb); /* max packet size */
get_be32(pb); /* avg packet size */
start_time = get_be32(pb); /* start time */
get_be32(pb); /* preroll */
duration = get_be32(pb); /* duration */
st->start_time = start_time;
st->duration = duration;
get_str8(pb, buf, sizeof(buf)); /* desc */
get_str8(pb, buf, sizeof(buf)); /* mimetype */
st->codec->codec_type = CODEC_TYPE_DATA;
st->priv_data = ff_rm_alloc_rmstream();
if (ff_rm_read_mdpr_codecdata(s, s->pb, st, st->priv_data,
get_be32(pb)) < 0)
return -1;
break;
case MKTAG('D', 'A', 'T', 'A'):
goto header_end;
default:
/* unknown tag: skip it */
url_fskip(pb, tag_size - 10);
break;
}
}
header_end:
rm->nb_packets = get_be32(pb); /* number of packets */
if (!rm->nb_packets && (flags & 4))
rm->nb_packets = 3600 * 25;
get_be32(pb); /* next data header */
return 0;
}
static int get_num(ByteIOContext *pb, int *len)
{
int n, n1;
n = get_be16(pb);
(*len)-=2;
n &= 0x7FFF;
if (n >= 0x4000) {
return n - 0x4000;
} else {
n1 = get_be16(pb);
(*len)-=2;
return (n << 16) | n1;
}
}
/* multiple of 20 bytes for ra144 (ugly) */
#define RAW_PACKET_SIZE 1000
static int sync(AVFormatContext *s, int64_t *timestamp, int *flags, int *stream_index, int64_t *pos){
RMDemuxContext *rm = s->priv_data;
ByteIOContext *pb = s->pb;
int len, num, res, i;
AVStream *st;
uint32_t state=0xFFFFFFFF;
while(!url_feof(pb)){
*pos= url_ftell(pb) - 3;
if(rm->remaining_len > 0){
num= rm->current_stream;
len= rm->remaining_len;
*timestamp = AV_NOPTS_VALUE;
*flags= 0;
}else{
state= (state<<8) + get_byte(pb);
if(state == MKBETAG('I', 'N', 'D', 'X')){
int n_pkts, expected_len;
len = get_be32(pb);
url_fskip(pb, 2);
n_pkts = get_be32(pb);
expected_len = 20 + n_pkts * 14;
if (len == 20)
/* some files don't add index entries to chunk size... */
len = expected_len;
else if (len != expected_len)
av_log(s, AV_LOG_WARNING,
"Index size %d (%d pkts) is wrong, should be %d.\n",
len, n_pkts, expected_len);
len -= 14; // we already read part of the index header
if(len<0)
continue;
goto skip;
}
if(state > (unsigned)0xFFFF || state < 12)
continue;
len=state;
state= 0xFFFFFFFF;
num = get_be16(pb);
*timestamp = get_be32(pb);
res= get_byte(pb); /* reserved */
*flags = get_byte(pb); /* flags */
len -= 12;
}
for(i=0;i<s->nb_streams;i++) {
st = s->streams[i];
if (num == st->id)
break;
}
if (i == s->nb_streams) {
skip:
/* skip packet if unknown number */
url_fskip(pb, len);
rm->remaining_len = 0;
continue;
}
*stream_index= i;
return len;
}
return -1;
}
static int rm_assemble_video_frame(AVFormatContext *s, ByteIOContext *pb,
RMDemuxContext *rm, RMStream *vst,
Fix index generation in the way that it was supposed to be used. See the discussion in the ML thread "[PATCH] rmdec.c: merge old/new packet reading code". Over time, this code broke somewhat, e.g. seq was never actually written into (and was thus always 1, therefore the seq condition was always true), whereas it was supposed to be set to the sequence number of the video slice in case the video frame is divided over multiple RM packets (slices). The problem of this is that packets other than those containing the beginning of a video frame would be indexed as well. Secondly, flags&2 is supposed to be true for video keyframes and for these audio packets containing the start of a block. For some codecs (e.g. AAC), that is every single packet, whereas for others (e.g. cook), that is the packet containing the first of a series of scrambled packets that are to be descrambled together. Indexing any of the following would lead to incomplete and thus useless frames. Problem here is that flags would be reset to 2 to indicate that the first packet is ready to be returned, and in addition if no data was left to be returned (which is always true for the first packet), then we wouldn't actually write the index entry anyway. All in all, the idea was good and it probably worked at some point, but that is long ago. This patch should at the very least make it likely for this code to be executed again at the right times, i.e. the way it was originally intended to be used. Originally committed as revision 17993 to svn://svn.ffmpeg.org/ffmpeg/trunk
2009-03-15 21:14:25 +01:00
AVPacket *pkt, int len, int *pseq)
{
int hdr, seq, pic_num, len2, pos;
int type;
hdr = get_byte(pb); len--;
type = hdr >> 6;
if(type != 3){ // not frame as a part of packet
seq = get_byte(pb); len--;
}
if(type != 1){ // not whole frame
len2 = get_num(pb, &len);
pos = get_num(pb, &len);
pic_num = get_byte(pb); len--;
}
if(len<0)
return -1;
rm->remaining_len = len;
if(type&1){ // frame, not slice
if(type == 3) // frame as a part of packet
len= len2;
if(rm->remaining_len < len)
return -1;
rm->remaining_len -= len;
if(av_new_packet(pkt, len + 9) < 0)
return AVERROR(EIO);
pkt->data[0] = 0;
AV_WL32(pkt->data + 1, 1);
AV_WL32(pkt->data + 5, 0);
get_buffer(pb, pkt->data + 9, len);
return 0;
}
//now we have to deal with single slice
Fix index generation in the way that it was supposed to be used. See the discussion in the ML thread "[PATCH] rmdec.c: merge old/new packet reading code". Over time, this code broke somewhat, e.g. seq was never actually written into (and was thus always 1, therefore the seq condition was always true), whereas it was supposed to be set to the sequence number of the video slice in case the video frame is divided over multiple RM packets (slices). The problem of this is that packets other than those containing the beginning of a video frame would be indexed as well. Secondly, flags&2 is supposed to be true for video keyframes and for these audio packets containing the start of a block. For some codecs (e.g. AAC), that is every single packet, whereas for others (e.g. cook), that is the packet containing the first of a series of scrambled packets that are to be descrambled together. Indexing any of the following would lead to incomplete and thus useless frames. Problem here is that flags would be reset to 2 to indicate that the first packet is ready to be returned, and in addition if no data was left to be returned (which is always true for the first packet), then we wouldn't actually write the index entry anyway. All in all, the idea was good and it probably worked at some point, but that is long ago. This patch should at the very least make it likely for this code to be executed again at the right times, i.e. the way it was originally intended to be used. Originally committed as revision 17993 to svn://svn.ffmpeg.org/ffmpeg/trunk
2009-03-15 21:14:25 +01:00
*pseq = seq;
if((seq & 0x7F) == 1 || vst->curpic_num != pic_num){
vst->slices = ((hdr & 0x3F) << 1) + 1;
vst->videobufsize = len2 + 8*vst->slices + 1;
av_free_packet(&vst->pkt); //FIXME this should be output.
if(av_new_packet(&vst->pkt, vst->videobufsize) < 0)
return AVERROR(ENOMEM);
vst->videobufpos = 8*vst->slices + 1;
vst->cur_slice = 0;
vst->curpic_num = pic_num;
vst->pktpos = url_ftell(pb);
}
if(type == 2)
len = FFMIN(len, pos);
if(++vst->cur_slice > vst->slices)
return 1;
AV_WL32(vst->pkt.data - 7 + 8*vst->cur_slice, 1);
AV_WL32(vst->pkt.data - 3 + 8*vst->cur_slice, vst->videobufpos - 8*vst->slices - 1);
if(vst->videobufpos + len > vst->videobufsize)
return 1;
if (get_buffer(pb, vst->pkt.data + vst->videobufpos, len) != len)
return AVERROR(EIO);
vst->videobufpos += len;
rm->remaining_len-= len;
if(type == 2 || (vst->videobufpos) == vst->videobufsize){
vst->pkt.data[0] = vst->cur_slice-1;
*pkt= vst->pkt;
vst->pkt.data= NULL;
vst->pkt.size= 0;
if(vst->slices != vst->cur_slice) //FIXME find out how to set slices correct from the begin
memmove(pkt->data + 1 + 8*vst->cur_slice, pkt->data + 1 + 8*vst->slices,
vst->videobufpos - 1 - 8*vst->slices);
pkt->size = vst->videobufpos + 8*(vst->cur_slice - vst->slices);
pkt->pts = AV_NOPTS_VALUE;
pkt->pos = vst->pktpos;
return 0;
}
return 1;
}
static inline void
rm_ac3_swap_bytes (AVStream *st, AVPacket *pkt)
{
uint8_t *ptr;
int j;
if (st->codec->codec_id == CODEC_ID_AC3) {
ptr = pkt->data;
for (j=0;j<pkt->size;j+=2) {
FFSWAP(int, ptr[0], ptr[1]);
ptr += 2;
}
}
}
int
ff_rm_parse_packet (AVFormatContext *s, ByteIOContext *pb,
AVStream *st, RMStream *ast, int len, AVPacket *pkt,
int *seq, int *flags, int64_t *timestamp)
{
RMDemuxContext *rm = s->priv_data;
if (st->codec->codec_type == CODEC_TYPE_VIDEO) {
rm->current_stream= st->id;
Fix index generation in the way that it was supposed to be used. See the discussion in the ML thread "[PATCH] rmdec.c: merge old/new packet reading code". Over time, this code broke somewhat, e.g. seq was never actually written into (and was thus always 1, therefore the seq condition was always true), whereas it was supposed to be set to the sequence number of the video slice in case the video frame is divided over multiple RM packets (slices). The problem of this is that packets other than those containing the beginning of a video frame would be indexed as well. Secondly, flags&2 is supposed to be true for video keyframes and for these audio packets containing the start of a block. For some codecs (e.g. AAC), that is every single packet, whereas for others (e.g. cook), that is the packet containing the first of a series of scrambled packets that are to be descrambled together. Indexing any of the following would lead to incomplete and thus useless frames. Problem here is that flags would be reset to 2 to indicate that the first packet is ready to be returned, and in addition if no data was left to be returned (which is always true for the first packet), then we wouldn't actually write the index entry anyway. All in all, the idea was good and it probably worked at some point, but that is long ago. This patch should at the very least make it likely for this code to be executed again at the right times, i.e. the way it was originally intended to be used. Originally committed as revision 17993 to svn://svn.ffmpeg.org/ffmpeg/trunk
2009-03-15 21:14:25 +01:00
if(rm_assemble_video_frame(s, pb, rm, ast, pkt, len, seq))
return -1; //got partial frame
} else if (st->codec->codec_type == CODEC_TYPE_AUDIO) {
if ((st->codec->codec_id == CODEC_ID_RA_288) ||
(st->codec->codec_id == CODEC_ID_COOK) ||
(st->codec->codec_id == CODEC_ID_ATRAC3) ||
(st->codec->codec_id == CODEC_ID_SIPR)) {
int x;
int sps = ast->sub_packet_size;
int cfs = ast->coded_framesize;
int h = ast->sub_packet_h;
int y = ast->sub_packet_cnt;
int w = ast->audio_framesize;
if (*flags & 2)
y = ast->sub_packet_cnt = 0;
if (!y)
ast->audiotimestamp = *timestamp;
switch(st->codec->codec_id) {
case CODEC_ID_RA_288:
for (x = 0; x < h/2; x++)
get_buffer(pb, ast->pkt.data+x*2*w+y*cfs, cfs);
break;
case CODEC_ID_ATRAC3:
case CODEC_ID_COOK:
for (x = 0; x < w/sps; x++)
get_buffer(pb, ast->pkt.data+sps*(h*x+((h+1)/2)*(y&1)+(y>>1)), sps);
break;
}
if (++(ast->sub_packet_cnt) < h)
return -1;
ast->sub_packet_cnt = 0;
rm->audio_stream_num = st->index;
rm->audio_pkt_cnt = h * w / st->codec->block_align;
} else if (st->codec->codec_id == CODEC_ID_AAC) {
int x;
rm->audio_stream_num = st->index;
ast->sub_packet_cnt = (get_be16(pb) & 0xf0) >> 4;
if (ast->sub_packet_cnt) {
for (x = 0; x < ast->sub_packet_cnt; x++)
ast->sub_packet_lengths[x] = get_be16(pb);
rm->audio_pkt_cnt = ast->sub_packet_cnt;
ast->audiotimestamp = *timestamp;
} else
return -1;
} else {
av_get_packet(pb, pkt, len);
rm_ac3_swap_bytes(st, pkt);
}
} else
av_get_packet(pb, pkt, len);
pkt->stream_index = st->index;
#if 0
if (st->codec->codec_type == CODEC_TYPE_VIDEO) {
if(st->codec->codec_id == CODEC_ID_RV20){
int seq= 128*(pkt->data[2]&0x7F) + (pkt->data[3]>>1);
av_log(s, AV_LOG_DEBUG, "%d %"PRId64" %d\n", *timestamp, *timestamp*512LL/25, seq);
seq |= (*timestamp&~0x3FFF);
if(seq - *timestamp > 0x2000) seq -= 0x4000;
if(seq - *timestamp < -0x2000) seq += 0x4000;
}
}
#endif
pkt->pts= *timestamp;
if (*flags & 2)
pkt->flags |= PKT_FLAG_KEY;
return st->codec->codec_type == CODEC_TYPE_AUDIO ? rm->audio_pkt_cnt : 0;
}
int
ff_rm_retrieve_cache (AVFormatContext *s, ByteIOContext *pb,
AVStream *st, RMStream *ast, AVPacket *pkt)
{
RMDemuxContext *rm = s->priv_data;
assert (rm->audio_pkt_cnt > 0);
if (st->codec->codec_id == CODEC_ID_AAC)
av_get_packet(pb, pkt, ast->sub_packet_lengths[ast->sub_packet_cnt - rm->audio_pkt_cnt]);
else {
av_new_packet(pkt, st->codec->block_align);
memcpy(pkt->data, ast->pkt.data + st->codec->block_align * //FIXME avoid this
(ast->sub_packet_h * ast->audio_framesize / st->codec->block_align - rm->audio_pkt_cnt),
st->codec->block_align);
}
rm->audio_pkt_cnt--;
pkt->flags = 0;
pkt->stream_index = st->index;
return rm->audio_pkt_cnt;
}
static int rm_read_packet(AVFormatContext *s, AVPacket *pkt)
{
RMDemuxContext *rm = s->priv_data;
AVStream *st;
int i, len, res, seq = 1;
int64_t timestamp, pos;
Fix index generation in the way that it was supposed to be used. See the discussion in the ML thread "[PATCH] rmdec.c: merge old/new packet reading code". Over time, this code broke somewhat, e.g. seq was never actually written into (and was thus always 1, therefore the seq condition was always true), whereas it was supposed to be set to the sequence number of the video slice in case the video frame is divided over multiple RM packets (slices). The problem of this is that packets other than those containing the beginning of a video frame would be indexed as well. Secondly, flags&2 is supposed to be true for video keyframes and for these audio packets containing the start of a block. For some codecs (e.g. AAC), that is every single packet, whereas for others (e.g. cook), that is the packet containing the first of a series of scrambled packets that are to be descrambled together. Indexing any of the following would lead to incomplete and thus useless frames. Problem here is that flags would be reset to 2 to indicate that the first packet is ready to be returned, and in addition if no data was left to be returned (which is always true for the first packet), then we wouldn't actually write the index entry anyway. All in all, the idea was good and it probably worked at some point, but that is long ago. This patch should at the very least make it likely for this code to be executed again at the right times, i.e. the way it was originally intended to be used. Originally committed as revision 17993 to svn://svn.ffmpeg.org/ffmpeg/trunk
2009-03-15 21:14:25 +01:00
int old_flags, flags;
for (;;) {
if (rm->audio_pkt_cnt) {
// If there are queued audio packet return them first
st = s->streams[rm->audio_stream_num];
ff_rm_retrieve_cache(s, s->pb, st, st->priv_data, pkt);
} else {
if (rm->old_format) {
RMStream *ast;
st = s->streams[0];
ast = st->priv_data;
timestamp = AV_NOPTS_VALUE;
len = !ast->audio_framesize ? RAW_PACKET_SIZE :
ast->coded_framesize * ast->sub_packet_h / 2;
flags = (seq++ == 1) ? 2 : 0;
} else {
len=sync(s, &timestamp, &flags, &i, &pos);
st = s->streams[i];
}
if(len<0 || url_feof(s->pb))
return AVERROR(EIO);
old_flags = flags;
res = ff_rm_parse_packet (s, s->pb, st, st->priv_data, len, pkt,
&seq, &flags, &timestamp);
if((old_flags&2) && (seq&0x7F) == 1)
av_add_index_entry(st, pos, timestamp, 0, 0, AVINDEX_KEYFRAME);
if (res)
continue;
}
if( (st->discard >= AVDISCARD_NONKEY && !(flags&2))
|| st->discard >= AVDISCARD_ALL){
av_free_packet(pkt);
} else
break;
}
return 0;
}
static int rm_read_close(AVFormatContext *s)
{
int i;
for (i=0;i<s->nb_streams;i++)
ff_rm_free_rmstream(s->streams[i]->priv_data);
return 0;
}
static int rm_probe(AVProbeData *p)
{
/* check file header */
if ((p->buf[0] == '.' && p->buf[1] == 'R' &&
p->buf[2] == 'M' && p->buf[3] == 'F' &&
p->buf[4] == 0 && p->buf[5] == 0) ||
(p->buf[0] == '.' && p->buf[1] == 'r' &&
p->buf[2] == 'a' && p->buf[3] == 0xfd))
return AVPROBE_SCORE_MAX;
else
return 0;
}
static int64_t rm_read_dts(AVFormatContext *s, int stream_index,
int64_t *ppos, int64_t pos_limit)
{
RMDemuxContext *rm = s->priv_data;
int64_t pos, dts;
int stream_index2, flags, len, h;
pos = *ppos;
if(rm->old_format)
return AV_NOPTS_VALUE;
url_fseek(s->pb, pos, SEEK_SET);
rm->remaining_len=0;
for(;;){
int seq=1;
AVStream *st;
len=sync(s, &dts, &flags, &stream_index2, &pos);
if(len<0)
return AV_NOPTS_VALUE;
st = s->streams[stream_index2];
if (st->codec->codec_type == CODEC_TYPE_VIDEO) {
h= get_byte(s->pb); len--;
if(!(h & 0x40)){
seq = get_byte(s->pb); len--;
}
}
if((flags&2) && (seq&0x7F) == 1){
// av_log(s, AV_LOG_DEBUG, "%d %d-%d %"PRId64" %d\n", flags, stream_index2, stream_index, dts, seq);
av_add_index_entry(st, pos, dts, 0, 0, AVINDEX_KEYFRAME);
if(stream_index2 == stream_index)
break;
}
url_fskip(s->pb, len);
}
*ppos = pos;
return dts;
}
AVInputFormat rm_demuxer = {
"rm",
NULL_IF_CONFIG_SMALL("RealMedia format"),
sizeof(RMDemuxContext),
rm_probe,
rm_read_header,
rm_read_packet,
rm_read_close,
NULL,
rm_read_dts,
};
AVInputFormat rdt_demuxer = {
"rdt",
NULL_IF_CONFIG_SMALL("RDT demuxer"),
sizeof(RMDemuxContext),
NULL,
NULL,
NULL,
rm_read_close,
};