1
0
mirror of https://github.com/FFmpeg/FFmpeg.git synced 2024-12-18 03:19:31 +02:00
FFmpeg/libavformat/vividas.c
Limin Wang 386dbf5726 avformat/vividas: simplify, use av_rescale_q() instead
note it'll cause a small difference in accuracy for the pts, please see the testing result below:
 $ wget http://samples.ffmpeg.org/archive/all/unknown+unknown+unknown+unknown+5029_paramount_en_1250.viv
 $ ./ffmpeg -t 0.04 -i ./unknown+unknown+unknown+unknown+5029_paramount_en_1250.viv -f null -
 old:
 pts: 522
 pts: 1044
 pts: 1567
 pts: 3918
 pts: 8097
 pts: 12277
 pts: 16457
 ...

 new:
 pts: 522
 pts: 1045
 pts: 1567
 pts: 3918
 pts: 8098
 pts: 12278
 pts: 16457
 ...

Signed-off-by: Limin Wang <lance.lmwang@gmail.com>
2020-05-10 22:32:33 +08:00

776 lines
20 KiB
C

/*
* Vividas VIV format Demuxer
* Copyright (c) 2012 Krzysztof Klinikowski
* Copyright (c) 2010 Andrzej Szombierski
* based on vivparse Copyright (c) 2007 Måns Rullgård
*
* 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
*/
/**
* @file
* @brief Vividas VIV (.viv) file demuxer
* @author Andrzej Szombierski [qq at kuku eu org] (2010-07)
* @sa http://wiki.multimedia.cx/index.php?title=Vividas_VIV
*/
#include "libavutil/intreadwrite.h"
#include "avio_internal.h"
#include "avformat.h"
#include "internal.h"
#define MAX_AUDIO_SUBPACKETS 100
typedef struct VIV_SB_block {
int size, n_packets;
int64_t byte_offset;
int64_t packet_offset;
} VIV_SB_block;
typedef struct VIV_SB_entry {
int size, flag;
} VIV_SB_entry;
typedef struct VIV_AudioSubpacket {
int start, pcm_bytes;
} VIV_AudioSubpacket;
typedef struct VividasDemuxContext {
int n_sb_blocks;
VIV_SB_block *sb_blocks;
int num_audio;
uint32_t sb_key;
int64_t sb_offset;
int current_sb, current_sb_entry;
uint8_t *sb_buf;
AVIOContext *sb_pb;
int n_sb_entries;
VIV_SB_entry *sb_entries;
int n_audio_subpackets;
int current_audio_subpacket;
int64_t audio_sample;
VIV_AudioSubpacket audio_subpackets[MAX_AUDIO_SUBPACKETS];
} VividasDemuxContext;
static int viv_probe(const AVProbeData *p)
{
if (memcmp(p->buf, "vividas03", 9))
return 0;
return AVPROBE_SCORE_MAX;
}
static const uint8_t keybits[32] = {
20, 52, 111, 10, 27, 71, 142, 53,
82, 138, 1, 78, 86, 121, 183, 85,
105, 152, 39, 140, 172, 11, 64, 144,
155, 6, 71, 163, 186, 49, 126, 43,
};
static uint32_t decode_key(uint8_t *buf)
{
uint32_t key = 0;
for (int i = 0; i < 32; i++) {
unsigned p = keybits[i];
key |= ((buf[p] >> ((i*5+3)&7)) & 1u) << i;
}
return key;
}
static void put_v(uint8_t *p, unsigned v)
{
if (v>>28)
*p++ = ((v>>28)&0x7f)|0x80;
if (v>>21)
*p++ = ((v>>21)&0x7f)|0x80;
if (v>>14)
*p++ = ((v>>14)&0x7f)|0x80;
if (v>>7)
*p++ = ((v>>7)&0x7f)|0x80;
}
static unsigned recover_key(unsigned char sample[4], unsigned expected_size)
{
unsigned char plaintext[8] = { 'S', 'B' };
put_v(plaintext+2, expected_size);
return AV_RL32(sample) ^ AV_RL32(plaintext);
}
static void xor_block(void *p1, void *p2, unsigned size, int key, unsigned *key_ptr)
{
unsigned *d1 = p1;
unsigned *d2 = p2;
unsigned k = *key_ptr;
size >>= 2;
while (size > 0) {
*d2 = *d1 ^ (HAVE_BIGENDIAN ? av_bswap32(k) : k);
k += key;
d1++;
d2++;
size--;
}
*key_ptr = k;
}
static void decode_block(uint8_t *src, uint8_t *dest, unsigned size,
uint32_t key, uint32_t *key_ptr,
int align)
{
unsigned s = size;
char tmp[4];
int a2;
if (!size)
return;
align &= 3;
a2 = (4 - align) & 3;
if (align) {
uint32_t tmpkey = *key_ptr - key;
if (a2 > s) {
a2 = s;
avpriv_request_sample(NULL, "tiny aligned block\n");
}
memcpy(tmp + align, src, a2);
xor_block(tmp, tmp, 4, key, &tmpkey);
memcpy(dest, tmp + align, a2);
s -= a2;
}
if (s >= 4) {
xor_block(src + a2, dest + a2, s & ~3,
key, key_ptr);
s &= 3;
}
if (s) {
size -= s;
memcpy(tmp, src + size, s);
xor_block(&tmp, &tmp, 4, key, key_ptr);
memcpy(dest + size, tmp, s);
}
}
static uint32_t get_v(uint8_t *p, int len)
{
uint32_t v = 0;
const uint8_t *end = p + len;
do {
if (p >= end || v >= UINT_MAX / 128 - *p)
return v;
v <<= 7;
v += *p & 0x7f;
} while (*p++ & 0x80);
return v;
}
static uint8_t *read_vblock(AVIOContext *src, uint32_t *size,
uint32_t key, uint32_t *k2, int align)
{
uint8_t tmp[4];
uint8_t *buf;
unsigned n;
if (avio_read(src, tmp, 4) != 4)
return NULL;
decode_block(tmp, tmp, 4, key, k2, align);
n = get_v(tmp, 4);
if (n < 4)
return NULL;
buf = av_malloc(n);
if (!buf)
return NULL;
*size = n;
n -= 4;
memcpy(buf, tmp, 4);
if (avio_read(src, buf + 4, n) == n) {
decode_block(buf + 4, buf + 4, n, key, k2, align);
} else {
av_free(buf);
buf = NULL;
}
return buf;
}
static uint8_t *read_sb_block(AVIOContext *src, unsigned *size,
uint32_t *key, unsigned expected_size)
{
uint8_t *buf;
uint8_t ibuf[8], sbuf[8];
uint32_t k2;
unsigned n;
if (avio_read(src, ibuf, 8) < 8)
return NULL;
k2 = *key;
decode_block(ibuf, sbuf, 8, *key, &k2, 0);
n = get_v(sbuf+2, 6);
if (sbuf[0] != 'S' || sbuf[1] != 'B' || (expected_size>0 && n != expected_size)) {
uint32_t tmpkey = recover_key(ibuf, expected_size);
k2 = tmpkey;
decode_block(ibuf, sbuf, 8, tmpkey, &k2, 0);
n = get_v(sbuf+2, 6);
if (sbuf[0] != 'S' || sbuf[1] != 'B' || expected_size != n)
return NULL;
*key = tmpkey;
}
if (n < 8)
return NULL;
buf = av_malloc(n);
if (!buf)
return NULL;
memcpy(buf, sbuf, 8);
*size = n;
n -= 8;
if (avio_read(src, buf+8, n) < n) {
av_free(buf);
return NULL;
}
decode_block(buf + 8, buf + 8, n, *key, &k2, 0);
return buf;
}
static int track_header(VividasDemuxContext *viv, AVFormatContext *s, uint8_t *buf, int size)
{
int i, j, ret;
int64_t off;
int val_1;
int num_video;
AVIOContext pb0, *pb = &pb0;
ffio_init_context(pb, buf, size, 0, NULL, NULL, NULL, NULL);
ffio_read_varlen(pb); // track_header_len
avio_r8(pb); // '1'
val_1 = ffio_read_varlen(pb);
for (i=0;i<val_1;i++) {
int c = avio_r8(pb);
for (j=0;j<c;j++) {
if (avio_feof(pb))
return AVERROR_EOF;
avio_r8(pb); // val_3
avio_r8(pb); // val_4
}
}
avio_r8(pb); // num_streams
off = avio_tell(pb);
off += ffio_read_varlen(pb); // val_5
avio_r8(pb); // '2'
num_video = avio_r8(pb);
avio_seek(pb, off, SEEK_SET);
if (num_video != 1) {
av_log(s, AV_LOG_ERROR, "number of video tracks %d is not 1\n", num_video);
return AVERROR_PATCHWELCOME;
}
for (i = 0; i < num_video; i++) {
AVStream *st = avformat_new_stream(s, NULL);
if (!st)
return AVERROR(ENOMEM);
st->id = i;
st->codecpar->codec_type = AVMEDIA_TYPE_VIDEO;
st->codecpar->codec_id = AV_CODEC_ID_VP6;
off = avio_tell(pb);
off += ffio_read_varlen(pb);
avio_r8(pb); // '3'
avio_r8(pb); // val_7
st->time_base.num = avio_rl32(pb); // frame_time
st->time_base.den = avio_rl32(pb); // time_base
st->nb_frames = avio_rl32(pb); // n frames
st->codecpar->width = avio_rl16(pb); // width
st->codecpar->height = avio_rl16(pb); // height
avio_r8(pb); // val_8
avio_rl32(pb); // val_9
avio_seek(pb, off, SEEK_SET);
}
off = avio_tell(pb);
off += ffio_read_varlen(pb); // val_10
avio_r8(pb); // '4'
viv->num_audio = avio_r8(pb);
avio_seek(pb, off, SEEK_SET);
if (viv->num_audio != 1)
av_log(s, AV_LOG_WARNING, "number of audio tracks %d is not 1\n", viv->num_audio);
for(i=0;i<viv->num_audio;i++) {
int q;
AVStream *st = avformat_new_stream(s, NULL);
if (!st)
return AVERROR(ENOMEM);
st->id = num_video + i;
st->codecpar->codec_type = AVMEDIA_TYPE_AUDIO;
st->codecpar->codec_id = AV_CODEC_ID_VORBIS;
off = avio_tell(pb);
off += ffio_read_varlen(pb); // length
avio_r8(pb); // '5'
avio_r8(pb); //codec_id
avio_rl16(pb); //codec_subid
st->codecpar->channels = avio_rl16(pb); // channels
st->codecpar->sample_rate = avio_rl32(pb); // sample_rate
avio_seek(pb, 10, SEEK_CUR); // data_1
q = avio_r8(pb);
avio_seek(pb, q, SEEK_CUR); // data_2
avio_r8(pb); // zeropad
if (avio_tell(pb) < off) {
int num_data;
int xd_size = 0;
int data_len[256];
int offset = 1;
uint8_t *p;
ffio_read_varlen(pb); // val_13
avio_r8(pb); // '19'
ffio_read_varlen(pb); // len_3
num_data = avio_r8(pb);
for (j = 0; j < num_data; j++) {
uint64_t len = ffio_read_varlen(pb);
if (len > INT_MAX/2 - xd_size) {
return AVERROR_INVALIDDATA;
}
data_len[j] = len;
xd_size += len;
}
ret = ff_alloc_extradata(st->codecpar, 64 + xd_size + xd_size / 255);
if (ret < 0)
return ret;
p = st->codecpar->extradata;
p[0] = 2;
for (j = 0; j < num_data - 1; j++) {
unsigned delta = av_xiphlacing(&p[offset], data_len[j]);
if (delta > data_len[j]) {
return AVERROR_INVALIDDATA;
}
offset += delta;
}
for (j = 0; j < num_data; j++) {
int ret = avio_read(pb, &p[offset], data_len[j]);
if (ret < data_len[j]) {
st->codecpar->extradata_size = 0;
av_freep(&st->codecpar->extradata);
break;
}
offset += data_len[j];
}
if (offset < st->codecpar->extradata_size)
st->codecpar->extradata_size = offset;
}
}
return 0;
}
static int track_index(VividasDemuxContext *viv, AVFormatContext *s, uint8_t *buf, unsigned size)
{
int64_t off;
int64_t poff;
int maxnp=0;
AVIOContext pb0, *pb = &pb0;
int i;
int64_t filesize = avio_size(s->pb);
ffio_init_context(pb, buf, size, 0, NULL, NULL, NULL, NULL);
ffio_read_varlen(pb); // track_index_len
avio_r8(pb); // 'c'
viv->n_sb_blocks = ffio_read_varlen(pb);
if (viv->n_sb_blocks < 0 || viv->n_sb_blocks > size / 2)
goto error;
viv->sb_blocks = av_calloc(viv->n_sb_blocks, sizeof(VIV_SB_block));
if (!viv->sb_blocks) {
viv->n_sb_blocks = 0;
return AVERROR(ENOMEM);
}
off = 0;
poff = 0;
for (i = 0; i < viv->n_sb_blocks; i++) {
uint64_t size_tmp = ffio_read_varlen(pb);
uint64_t n_packets_tmp = ffio_read_varlen(pb);
if (size_tmp > INT_MAX || n_packets_tmp > INT_MAX)
goto error;
viv->sb_blocks[i].byte_offset = off;
viv->sb_blocks[i].packet_offset = poff;
viv->sb_blocks[i].size = size_tmp;
viv->sb_blocks[i].n_packets = n_packets_tmp;
off += viv->sb_blocks[i].size;
poff += viv->sb_blocks[i].n_packets;
if (maxnp < viv->sb_blocks[i].n_packets)
maxnp = viv->sb_blocks[i].n_packets;
}
if (filesize > 0 && poff > filesize)
goto error;
viv->sb_entries = av_calloc(maxnp, sizeof(VIV_SB_entry));
return 0;
error:
viv->n_sb_blocks = 0;
av_freep(&viv->sb_blocks);
return AVERROR_INVALIDDATA;
}
static void load_sb_block(AVFormatContext *s, VividasDemuxContext *viv, unsigned expected_size)
{
uint32_t size = 0;
int i;
AVIOContext *pb = 0;
if (viv->sb_pb) {
av_free(viv->sb_pb);
viv->sb_pb = NULL;
}
if (viv->sb_buf)
av_free(viv->sb_buf);
viv->sb_buf = read_sb_block(s->pb, &size, &viv->sb_key, expected_size);
if (!viv->sb_buf) {
return;
}
pb = avio_alloc_context(viv->sb_buf, size, 0, NULL, NULL, NULL, NULL);
if (!pb)
return;
viv->sb_pb = pb;
avio_r8(pb); // 'S'
avio_r8(pb); // 'B'
ffio_read_varlen(pb); // size
avio_r8(pb); // junk
ffio_read_varlen(pb); // first packet
viv->n_sb_entries = viv->sb_blocks[viv->current_sb].n_packets;
for (i = 0; i < viv->n_sb_entries; i++) {
viv->sb_entries[i].size = ffio_read_varlen(pb);
viv->sb_entries[i].flag = avio_r8(pb);
}
ffio_read_varlen(pb);
avio_r8(pb);
viv->current_sb_entry = 0;
}
static int viv_read_header(AVFormatContext *s)
{
VividasDemuxContext *viv = s->priv_data;
AVIOContext *pb = s->pb;
int64_t header_end;
int num_tracks;
uint32_t key, k2;
uint32_t v;
uint8_t keybuffer[187];
uint32_t b22_size = 0;
uint32_t b22_key = 0;
uint8_t *buf = 0;
int ret;
avio_skip(pb, 9);
header_end = avio_tell(pb);
header_end += ffio_read_varlen(pb);
num_tracks = avio_r8(pb);
if (num_tracks != 1) {
av_log(s, AV_LOG_ERROR, "number of tracks %d is not 1\n", num_tracks);
return AVERROR(EINVAL);
}
v = avio_r8(pb);
avio_seek(pb, v, SEEK_CUR);
avio_read(pb, keybuffer, 187);
key = decode_key(keybuffer);
viv->sb_key = key;
avio_rl32(pb);
for (;;) {
int64_t here = avio_tell(pb);
int block_len, block_type;
if (here >= header_end)
break;
block_len = ffio_read_varlen(pb);
if (avio_feof(pb) || block_len <= 0)
return AVERROR_INVALIDDATA;
block_type = avio_r8(pb);
if (block_type == 22) {
avio_read(pb, keybuffer, 187);
b22_key = decode_key(keybuffer);
b22_size = avio_rl32(pb);
}
avio_seek(pb, here + block_len, SEEK_SET);
}
if (b22_size) {
k2 = b22_key;
buf = read_vblock(pb, &v, b22_key, &k2, 0);
if (!buf)
return AVERROR(EIO);
av_free(buf);
}
k2 = key;
buf = read_vblock(pb, &v, key, &k2, 0);
if (!buf)
return AVERROR(EIO);
ret = track_header(viv, s, buf, v);
av_free(buf);
if (ret < 0)
return ret;
buf = read_vblock(pb, &v, key, &k2, v);
if (!buf)
return AVERROR(EIO);
ret = track_index(viv, s, buf, v);
av_free(buf);
if (ret < 0)
return ret;
viv->sb_offset = avio_tell(pb);
if (viv->n_sb_blocks > 0) {
viv->current_sb = 0;
load_sb_block(s, viv, viv->sb_blocks[0].size);
} else {
viv->current_sb = -1;
}
return 0;
}
static int viv_read_packet(AVFormatContext *s,
AVPacket *pkt)
{
VividasDemuxContext *viv = s->priv_data;
AVIOContext *pb;
int64_t off;
int ret;
if (!viv->sb_pb)
return AVERROR(EIO);
if (avio_feof(viv->sb_pb))
return AVERROR_EOF;
if (viv->current_audio_subpacket < viv->n_audio_subpackets) {
AVStream *astream;
int size = viv->audio_subpackets[viv->current_audio_subpacket+1].start - viv->audio_subpackets[viv->current_audio_subpacket].start;
pb = viv->sb_pb;
ret = av_get_packet(pb, pkt, size);
if (ret < 0)
return ret;
pkt->pos += viv->sb_offset + viv->sb_blocks[viv->current_sb].byte_offset;
pkt->stream_index = 1;
astream = s->streams[pkt->stream_index];
pkt->pts = av_rescale_q(viv->audio_sample, av_make_q(1, astream->codecpar->sample_rate), astream->time_base);
viv->audio_sample += viv->audio_subpackets[viv->current_audio_subpacket].pcm_bytes / 2 / astream->codecpar->channels;
pkt->flags |= AV_PKT_FLAG_KEY;
viv->current_audio_subpacket++;
return 0;
}
if (viv->current_sb_entry >= viv->n_sb_entries) {
if (viv->current_sb+1 >= viv->n_sb_blocks)
return AVERROR(EIO);
viv->current_sb++;
load_sb_block(s, viv, 0);
viv->current_sb_entry = 0;
}
pb = viv->sb_pb;
if (!pb)
return AVERROR(EIO);
off = avio_tell(pb);
off += viv->sb_entries[viv->current_sb_entry].size;
if (viv->sb_entries[viv->current_sb_entry].flag == 0) {
uint64_t v_size = ffio_read_varlen(pb);
if (!viv->num_audio)
return AVERROR_INVALIDDATA;
ffio_read_varlen(pb);
if (v_size > INT_MAX)
return AVERROR_INVALIDDATA;
ret = av_get_packet(pb, pkt, v_size);
if (ret < 0)
return ret;
pkt->pos += viv->sb_offset + viv->sb_blocks[viv->current_sb].byte_offset;
pkt->pts = viv->sb_blocks[viv->current_sb].packet_offset + viv->current_sb_entry;
pkt->flags |= (pkt->data[0]&0x80)?0:AV_PKT_FLAG_KEY;
pkt->stream_index = 0;
for (int i = 0; i < MAX_AUDIO_SUBPACKETS - 1; i++) {
int start, pcm_bytes;
start = ffio_read_varlen(pb);
pcm_bytes = ffio_read_varlen(pb);
if (i > 0 && start == 0)
break;
viv->n_audio_subpackets = i + 1;
viv->audio_subpackets[i].start = start;
viv->audio_subpackets[i].pcm_bytes = pcm_bytes;
}
viv->audio_subpackets[viv->n_audio_subpackets].start = (int)(off - avio_tell(pb));
viv->current_audio_subpacket = 0;
} else {
uint64_t v_size = ffio_read_varlen(pb);
if (v_size > INT_MAX)
return AVERROR_INVALIDDATA;
ret = av_get_packet(pb, pkt, v_size);
if (ret < 0)
return ret;
pkt->pos += viv->sb_offset + viv->sb_blocks[viv->current_sb].byte_offset;
pkt->pts = viv->sb_blocks[viv->current_sb].packet_offset + viv->current_sb_entry;
pkt->flags |= (pkt->data[0] & 0x80) ? 0 : AV_PKT_FLAG_KEY;
pkt->stream_index = 0;
}
viv->current_sb_entry++;
return 0;
}
static int viv_read_close(AVFormatContext *s)
{
VividasDemuxContext *viv = s->priv_data;
av_freep(&viv->sb_pb);
av_freep(&viv->sb_buf);
av_freep(&viv->sb_blocks);
av_freep(&viv->sb_entries);
return 0;
}
static int viv_read_seek(AVFormatContext *s, int stream_index, int64_t timestamp, int flags)
{
VividasDemuxContext *viv = s->priv_data;
int64_t frame;
if (stream_index == 0)
frame = timestamp;
else
frame = av_rescale_q(timestamp, s->streams[0]->time_base, s->streams[stream_index]->time_base);
for (int i = 0; i < viv->n_sb_blocks; i++) {
if (frame >= viv->sb_blocks[i].packet_offset && frame < viv->sb_blocks[i].packet_offset + viv->sb_blocks[i].n_packets) {
// flush audio packet queue
viv->current_audio_subpacket = 0;
viv->n_audio_subpackets = 0;
viv->current_sb = i;
// seek to ith sb block
avio_seek(s->pb, viv->sb_offset + viv->sb_blocks[i].byte_offset, SEEK_SET);
// load the block
load_sb_block(s, viv, 0);
// most problematic part: guess audio offset
viv->audio_sample = av_rescale_q(viv->sb_blocks[i].packet_offset, av_make_q(s->streams[1]->codecpar->sample_rate, 1), av_inv_q(s->streams[0]->time_base));
// hand-tuned 1.s a/v offset
viv->audio_sample += s->streams[1]->codecpar->sample_rate;
viv->current_sb_entry = 0;
return 1;
}
}
return 0;
}
AVInputFormat ff_vividas_demuxer = {
.name = "vividas",
.long_name = NULL_IF_CONFIG_SMALL("Vividas VIV"),
.priv_data_size = sizeof(VividasDemuxContext),
.read_probe = viv_probe,
.read_header = viv_read_header,
.read_packet = viv_read_packet,
.read_close = viv_read_close,
.read_seek = viv_read_seek,
};