mirror of https://github.com/FFmpeg/FFmpeg.git
You can not select more than 25 topics
Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
3289 lines
95 KiB
3289 lines
95 KiB
/* |
|
* Various utilities for ffmpeg system |
|
* Copyright (c) 2000, 2001, 2002 Fabrice Bellard |
|
* |
|
* This library 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 of the License, or (at your option) any later version. |
|
* |
|
* This library 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 this library; if not, write to the Free Software |
|
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA |
|
*/ |
|
#include "avformat.h" |
|
|
|
#undef NDEBUG |
|
#include <assert.h> |
|
|
|
/** |
|
* @file libavformat/utils.c |
|
* Various utility functions for using ffmpeg library. |
|
*/ |
|
|
|
static void av_frac_init(AVFrac *f, int64_t val, int64_t num, int64_t den); |
|
static void av_frac_add(AVFrac *f, int64_t incr); |
|
static void av_frac_set(AVFrac *f, int64_t val); |
|
|
|
/** head of registered input format linked list. */ |
|
AVInputFormat *first_iformat = NULL; |
|
/** head of registered output format linked list. */ |
|
AVOutputFormat *first_oformat = NULL; |
|
/** head of registered image format linked list. */ |
|
AVImageFormat *first_image_format = NULL; |
|
|
|
void av_register_input_format(AVInputFormat *format) |
|
{ |
|
AVInputFormat **p; |
|
p = &first_iformat; |
|
while (*p != NULL) p = &(*p)->next; |
|
*p = format; |
|
format->next = NULL; |
|
} |
|
|
|
void av_register_output_format(AVOutputFormat *format) |
|
{ |
|
AVOutputFormat **p; |
|
p = &first_oformat; |
|
while (*p != NULL) p = &(*p)->next; |
|
*p = format; |
|
format->next = NULL; |
|
} |
|
|
|
int match_ext(const char *filename, const char *extensions) |
|
{ |
|
const char *ext, *p; |
|
char ext1[32], *q; |
|
|
|
if(!filename) |
|
return 0; |
|
|
|
ext = strrchr(filename, '.'); |
|
if (ext) { |
|
ext++; |
|
p = extensions; |
|
for(;;) { |
|
q = ext1; |
|
while (*p != '\0' && *p != ',' && q-ext1<sizeof(ext1)-1) |
|
*q++ = *p++; |
|
*q = '\0'; |
|
if (!strcasecmp(ext1, ext)) |
|
return 1; |
|
if (*p == '\0') |
|
break; |
|
p++; |
|
} |
|
} |
|
return 0; |
|
} |
|
|
|
AVOutputFormat *guess_format(const char *short_name, const char *filename, |
|
const char *mime_type) |
|
{ |
|
AVOutputFormat *fmt, *fmt_found; |
|
int score_max, score; |
|
|
|
/* specific test for image sequences */ |
|
if (!short_name && filename && |
|
filename_number_test(filename) >= 0 && |
|
av_guess_image2_codec(filename) != CODEC_ID_NONE) { |
|
return guess_format("image2", NULL, NULL); |
|
} |
|
if (!short_name && filename && |
|
filename_number_test(filename) >= 0 && |
|
guess_image_format(filename)) { |
|
return guess_format("image", NULL, NULL); |
|
} |
|
|
|
/* find the proper file type */ |
|
fmt_found = NULL; |
|
score_max = 0; |
|
fmt = first_oformat; |
|
while (fmt != NULL) { |
|
score = 0; |
|
if (fmt->name && short_name && !strcmp(fmt->name, short_name)) |
|
score += 100; |
|
if (fmt->mime_type && mime_type && !strcmp(fmt->mime_type, mime_type)) |
|
score += 10; |
|
if (filename && fmt->extensions && |
|
match_ext(filename, fmt->extensions)) { |
|
score += 5; |
|
} |
|
if (score > score_max) { |
|
score_max = score; |
|
fmt_found = fmt; |
|
} |
|
fmt = fmt->next; |
|
} |
|
return fmt_found; |
|
} |
|
|
|
AVOutputFormat *guess_stream_format(const char *short_name, const char *filename, |
|
const char *mime_type) |
|
{ |
|
AVOutputFormat *fmt = guess_format(short_name, filename, mime_type); |
|
|
|
if (fmt) { |
|
AVOutputFormat *stream_fmt; |
|
char stream_format_name[64]; |
|
|
|
snprintf(stream_format_name, sizeof(stream_format_name), "%s_stream", fmt->name); |
|
stream_fmt = guess_format(stream_format_name, NULL, NULL); |
|
|
|
if (stream_fmt) |
|
fmt = stream_fmt; |
|
} |
|
|
|
return fmt; |
|
} |
|
|
|
/** |
|
* Guesses the codec id based upon muxer and filename. |
|
*/ |
|
enum CodecID av_guess_codec(AVOutputFormat *fmt, const char *short_name, |
|
const char *filename, const char *mime_type, enum CodecType type){ |
|
if(type == CODEC_TYPE_VIDEO){ |
|
enum CodecID codec_id= CODEC_ID_NONE; |
|
|
|
if(!strcmp(fmt->name, "image2") || !strcmp(fmt->name, "image2pipe")){ |
|
codec_id= av_guess_image2_codec(filename); |
|
} |
|
if(codec_id == CODEC_ID_NONE) |
|
codec_id= fmt->video_codec; |
|
return codec_id; |
|
}else if(type == CODEC_TYPE_AUDIO) |
|
return fmt->audio_codec; |
|
else |
|
return CODEC_ID_NONE; |
|
} |
|
|
|
/** |
|
* finds AVInputFormat based on input format's short name. |
|
*/ |
|
AVInputFormat *av_find_input_format(const char *short_name) |
|
{ |
|
AVInputFormat *fmt; |
|
for(fmt = first_iformat; fmt != NULL; fmt = fmt->next) { |
|
if (!strcmp(fmt->name, short_name)) |
|
return fmt; |
|
} |
|
return NULL; |
|
} |
|
|
|
/* memory handling */ |
|
|
|
/** |
|
* Default packet destructor. |
|
*/ |
|
void av_destruct_packet(AVPacket *pkt) |
|
{ |
|
av_free(pkt->data); |
|
pkt->data = NULL; pkt->size = 0; |
|
} |
|
|
|
/** |
|
* Allocate the payload of a packet and intialized its fields to default values. |
|
* |
|
* @param pkt packet |
|
* @param size wanted payload size |
|
* @return 0 if OK. AVERROR_xxx otherwise. |
|
*/ |
|
int av_new_packet(AVPacket *pkt, int size) |
|
{ |
|
void *data; |
|
if((unsigned)size > (unsigned)size + FF_INPUT_BUFFER_PADDING_SIZE) |
|
return AVERROR_NOMEM; |
|
data = av_malloc(size + FF_INPUT_BUFFER_PADDING_SIZE); |
|
if (!data) |
|
return AVERROR_NOMEM; |
|
memset(data + size, 0, FF_INPUT_BUFFER_PADDING_SIZE); |
|
|
|
av_init_packet(pkt); |
|
pkt->data = data; |
|
pkt->size = size; |
|
pkt->destruct = av_destruct_packet; |
|
return 0; |
|
} |
|
|
|
/** |
|
* Allocate and read the payload of a packet and intialized its fields to default values. |
|
* |
|
* @param pkt packet |
|
* @param size wanted payload size |
|
* @return >0 (read size) if OK. AVERROR_xxx otherwise. |
|
*/ |
|
int av_get_packet(ByteIOContext *s, AVPacket *pkt, int size) |
|
{ |
|
int ret= av_new_packet(pkt, size); |
|
|
|
if(ret<0) |
|
return ret; |
|
|
|
pkt->pos= url_ftell(s); |
|
|
|
ret= get_buffer(s, pkt->data, size); |
|
if(ret<=0) |
|
av_free_packet(pkt); |
|
else |
|
pkt->size= ret; |
|
|
|
return ret; |
|
} |
|
|
|
/* This is a hack - the packet memory allocation stuff is broken. The |
|
packet is allocated if it was not really allocated */ |
|
int av_dup_packet(AVPacket *pkt) |
|
{ |
|
if (pkt->destruct != av_destruct_packet) { |
|
uint8_t *data; |
|
/* we duplicate the packet and don't forget to put the padding |
|
again */ |
|
if((unsigned)pkt->size > (unsigned)pkt->size + FF_INPUT_BUFFER_PADDING_SIZE) |
|
return AVERROR_NOMEM; |
|
data = av_malloc(pkt->size + FF_INPUT_BUFFER_PADDING_SIZE); |
|
if (!data) { |
|
return AVERROR_NOMEM; |
|
} |
|
memcpy(data, pkt->data, pkt->size); |
|
memset(data + pkt->size, 0, FF_INPUT_BUFFER_PADDING_SIZE); |
|
pkt->data = data; |
|
pkt->destruct = av_destruct_packet; |
|
} |
|
return 0; |
|
} |
|
|
|
/* fifo handling */ |
|
|
|
int fifo_init(FifoBuffer *f, int size) |
|
{ |
|
f->buffer = av_malloc(size); |
|
if (!f->buffer) |
|
return -1; |
|
f->end = f->buffer + size; |
|
f->wptr = f->rptr = f->buffer; |
|
return 0; |
|
} |
|
|
|
void fifo_free(FifoBuffer *f) |
|
{ |
|
av_free(f->buffer); |
|
} |
|
|
|
int fifo_size(FifoBuffer *f, uint8_t *rptr) |
|
{ |
|
int size; |
|
|
|
if(!rptr) |
|
rptr= f->rptr; |
|
|
|
if (f->wptr >= rptr) { |
|
size = f->wptr - rptr; |
|
} else { |
|
size = (f->end - rptr) + (f->wptr - f->buffer); |
|
} |
|
return size; |
|
} |
|
|
|
/** |
|
* Get data from the fifo (returns -1 if not enough data). |
|
*/ |
|
int fifo_read(FifoBuffer *f, uint8_t *buf, int buf_size, uint8_t **rptr_ptr) |
|
{ |
|
uint8_t *rptr; |
|
int size, len; |
|
|
|
if(!rptr_ptr) |
|
rptr_ptr= &f->rptr; |
|
rptr = *rptr_ptr; |
|
|
|
if (f->wptr >= rptr) { |
|
size = f->wptr - rptr; |
|
} else { |
|
size = (f->end - rptr) + (f->wptr - f->buffer); |
|
} |
|
|
|
if (size < buf_size) |
|
return -1; |
|
while (buf_size > 0) { |
|
len = f->end - rptr; |
|
if (len > buf_size) |
|
len = buf_size; |
|
memcpy(buf, rptr, len); |
|
buf += len; |
|
rptr += len; |
|
if (rptr >= f->end) |
|
rptr = f->buffer; |
|
buf_size -= len; |
|
} |
|
*rptr_ptr = rptr; |
|
return 0; |
|
} |
|
|
|
/** |
|
* Resizes a FIFO. |
|
*/ |
|
void fifo_realloc(FifoBuffer *f, unsigned int new_size){ |
|
unsigned int old_size= f->end - f->buffer; |
|
|
|
if(old_size < new_size){ |
|
uint8_t *old= f->buffer; |
|
|
|
f->buffer= av_realloc(f->buffer, new_size); |
|
|
|
f->rptr += f->buffer - old; |
|
f->wptr += f->buffer - old; |
|
|
|
if(f->wptr < f->rptr){ |
|
memmove(f->rptr + new_size - old_size, f->rptr, f->buffer + old_size - f->rptr); |
|
f->rptr += new_size - old_size; |
|
} |
|
f->end= f->buffer + new_size; |
|
} |
|
} |
|
|
|
void fifo_write(FifoBuffer *f, const uint8_t *buf, int size, uint8_t **wptr_ptr) |
|
{ |
|
int len; |
|
uint8_t *wptr; |
|
|
|
if(!wptr_ptr) |
|
wptr_ptr= &f->wptr; |
|
wptr = *wptr_ptr; |
|
|
|
while (size > 0) { |
|
len = f->end - wptr; |
|
if (len > size) |
|
len = size; |
|
memcpy(wptr, buf, len); |
|
wptr += len; |
|
if (wptr >= f->end) |
|
wptr = f->buffer; |
|
buf += len; |
|
size -= len; |
|
} |
|
*wptr_ptr = wptr; |
|
} |
|
|
|
/* get data from the fifo (return -1 if not enough data) */ |
|
int put_fifo(ByteIOContext *pb, FifoBuffer *f, int buf_size, uint8_t **rptr_ptr) |
|
{ |
|
uint8_t *rptr = *rptr_ptr; |
|
int size, len; |
|
|
|
if (f->wptr >= rptr) { |
|
size = f->wptr - rptr; |
|
} else { |
|
size = (f->end - rptr) + (f->wptr - f->buffer); |
|
} |
|
|
|
if (size < buf_size) |
|
return -1; |
|
while (buf_size > 0) { |
|
len = f->end - rptr; |
|
if (len > buf_size) |
|
len = buf_size; |
|
put_buffer(pb, rptr, len); |
|
rptr += len; |
|
if (rptr >= f->end) |
|
rptr = f->buffer; |
|
buf_size -= len; |
|
} |
|
*rptr_ptr = rptr; |
|
return 0; |
|
} |
|
|
|
int filename_number_test(const char *filename) |
|
{ |
|
char buf[1024]; |
|
if(!filename) |
|
return -1; |
|
return get_frame_filename(buf, sizeof(buf), filename, 1); |
|
} |
|
|
|
/** |
|
* Guess file format. |
|
*/ |
|
AVInputFormat *av_probe_input_format(AVProbeData *pd, int is_opened) |
|
{ |
|
AVInputFormat *fmt1, *fmt; |
|
int score, score_max; |
|
|
|
fmt = NULL; |
|
score_max = 0; |
|
for(fmt1 = first_iformat; fmt1 != NULL; fmt1 = fmt1->next) { |
|
if (!is_opened && !(fmt1->flags & AVFMT_NOFILE)) |
|
continue; |
|
score = 0; |
|
if (fmt1->read_probe) { |
|
score = fmt1->read_probe(pd); |
|
} else if (fmt1->extensions) { |
|
if (match_ext(pd->filename, fmt1->extensions)) { |
|
score = 50; |
|
} |
|
} |
|
if (score > score_max) { |
|
score_max = score; |
|
fmt = fmt1; |
|
} |
|
} |
|
return fmt; |
|
} |
|
|
|
/************************************************************/ |
|
/* input media file */ |
|
|
|
/** |
|
* Open a media file from an IO stream. 'fmt' must be specified. |
|
*/ |
|
static const char* format_to_name(void* ptr) |
|
{ |
|
AVFormatContext* fc = (AVFormatContext*) ptr; |
|
if(fc->iformat) return fc->iformat->name; |
|
else if(fc->oformat) return fc->oformat->name; |
|
else return "NULL"; |
|
} |
|
|
|
static const AVClass av_format_context_class = { "AVFormatContext", format_to_name }; |
|
|
|
AVFormatContext *av_alloc_format_context(void) |
|
{ |
|
AVFormatContext *ic; |
|
ic = av_mallocz(sizeof(AVFormatContext)); |
|
if (!ic) return ic; |
|
ic->av_class = &av_format_context_class; |
|
return ic; |
|
} |
|
|
|
/** |
|
* Allocates all the structures needed to read an input stream. |
|
* This does not open the needed codecs for decoding the stream[s]. |
|
*/ |
|
int av_open_input_stream(AVFormatContext **ic_ptr, |
|
ByteIOContext *pb, const char *filename, |
|
AVInputFormat *fmt, AVFormatParameters *ap) |
|
{ |
|
int err; |
|
AVFormatContext *ic; |
|
AVFormatParameters default_ap; |
|
|
|
if(!ap){ |
|
ap=&default_ap; |
|
memset(ap, 0, sizeof(default_ap)); |
|
} |
|
|
|
ic = av_alloc_format_context(); |
|
if (!ic) { |
|
err = AVERROR_NOMEM; |
|
goto fail; |
|
} |
|
ic->iformat = fmt; |
|
if (pb) |
|
ic->pb = *pb; |
|
ic->duration = AV_NOPTS_VALUE; |
|
ic->start_time = AV_NOPTS_VALUE; |
|
pstrcpy(ic->filename, sizeof(ic->filename), filename); |
|
|
|
/* allocate private data */ |
|
if (fmt->priv_data_size > 0) { |
|
ic->priv_data = av_mallocz(fmt->priv_data_size); |
|
if (!ic->priv_data) { |
|
err = AVERROR_NOMEM; |
|
goto fail; |
|
} |
|
} else { |
|
ic->priv_data = NULL; |
|
} |
|
|
|
err = ic->iformat->read_header(ic, ap); |
|
if (err < 0) |
|
goto fail; |
|
|
|
if (pb) |
|
ic->data_offset = url_ftell(&ic->pb); |
|
|
|
*ic_ptr = ic; |
|
return 0; |
|
fail: |
|
if (ic) { |
|
av_freep(&ic->priv_data); |
|
} |
|
av_free(ic); |
|
*ic_ptr = NULL; |
|
return err; |
|
} |
|
|
|
/** Size of probe buffer, for guessing file type from file contents. */ |
|
#define PROBE_BUF_MIN 2048 |
|
#define PROBE_BUF_MAX (1<<20) |
|
|
|
/** |
|
* Open a media file as input. The codec are not opened. Only the file |
|
* header (if present) is read. |
|
* |
|
* @param ic_ptr the opened media file handle is put here |
|
* @param filename filename to open. |
|
* @param fmt if non NULL, force the file format to use |
|
* @param buf_size optional buffer size (zero if default is OK) |
|
* @param ap additionnal parameters needed when opening the file (NULL if default) |
|
* @return 0 if OK. AVERROR_xxx otherwise. |
|
*/ |
|
int av_open_input_file(AVFormatContext **ic_ptr, const char *filename, |
|
AVInputFormat *fmt, |
|
int buf_size, |
|
AVFormatParameters *ap) |
|
{ |
|
int err, must_open_file, file_opened, probe_size; |
|
AVProbeData probe_data, *pd = &probe_data; |
|
ByteIOContext pb1, *pb = &pb1; |
|
|
|
file_opened = 0; |
|
pd->filename = ""; |
|
if (filename) |
|
pd->filename = filename; |
|
pd->buf = NULL; |
|
pd->buf_size = 0; |
|
|
|
if (!fmt) { |
|
/* guess format if no file can be opened */ |
|
fmt = av_probe_input_format(pd, 0); |
|
} |
|
|
|
/* do not open file if the format does not need it. XXX: specific |
|
hack needed to handle RTSP/TCP */ |
|
must_open_file = 1; |
|
if (fmt && (fmt->flags & AVFMT_NOFILE)) { |
|
must_open_file = 0; |
|
pb= NULL; //FIXME this or memset(pb, 0, sizeof(ByteIOContext)); otherwise its uninitalized |
|
} |
|
|
|
if (!fmt || must_open_file) { |
|
/* if no file needed do not try to open one */ |
|
if (url_fopen(pb, filename, URL_RDONLY) < 0) { |
|
err = AVERROR_IO; |
|
goto fail; |
|
} |
|
file_opened = 1; |
|
if (buf_size > 0) { |
|
url_setbufsize(pb, buf_size); |
|
} |
|
|
|
for(probe_size= PROBE_BUF_MIN; probe_size<=PROBE_BUF_MAX && !fmt; probe_size<<=1){ |
|
/* read probe data */ |
|
pd->buf= av_realloc(pd->buf, probe_size); |
|
pd->buf_size = get_buffer(pb, pd->buf, probe_size); |
|
if (url_fseek(pb, 0, SEEK_SET) == (offset_t)-EPIPE) { |
|
url_fclose(pb); |
|
if (url_fopen(pb, filename, URL_RDONLY) < 0) { |
|
file_opened = 0; |
|
err = AVERROR_IO; |
|
goto fail; |
|
} |
|
} |
|
/* guess file format */ |
|
fmt = av_probe_input_format(pd, 1); |
|
} |
|
av_freep(&pd->buf); |
|
} |
|
|
|
/* if still no format found, error */ |
|
if (!fmt) { |
|
err = AVERROR_NOFMT; |
|
goto fail; |
|
} |
|
|
|
/* XXX: suppress this hack for redirectors */ |
|
#ifdef CONFIG_NETWORK |
|
if (fmt == &redir_demux) { |
|
err = redir_open(ic_ptr, pb); |
|
url_fclose(pb); |
|
return err; |
|
} |
|
#endif |
|
|
|
/* check filename in case of an image number is expected */ |
|
if (fmt->flags & AVFMT_NEEDNUMBER) { |
|
if (filename_number_test(filename) < 0) { |
|
err = AVERROR_NUMEXPECTED; |
|
goto fail; |
|
} |
|
} |
|
err = av_open_input_stream(ic_ptr, pb, filename, fmt, ap); |
|
if (err) |
|
goto fail; |
|
return 0; |
|
fail: |
|
av_freep(&pd->buf); |
|
if (file_opened) |
|
url_fclose(pb); |
|
*ic_ptr = NULL; |
|
return err; |
|
|
|
} |
|
|
|
/*******************************************************/ |
|
|
|
/** |
|
* Read a transport packet from a media file. |
|
* |
|
* This function is absolete and should never be used. |
|
* Use av_read_frame() instead. |
|
* |
|
* @param s media file handle |
|
* @param pkt is filled |
|
* @return 0 if OK. AVERROR_xxx if error. |
|
*/ |
|
int av_read_packet(AVFormatContext *s, AVPacket *pkt) |
|
{ |
|
return s->iformat->read_packet(s, pkt); |
|
} |
|
|
|
/**********************************************************/ |
|
|
|
/** |
|
* Get the number of samples of an audio frame. Return (-1) if error. |
|
*/ |
|
static int get_audio_frame_size(AVCodecContext *enc, int size) |
|
{ |
|
int frame_size; |
|
|
|
if (enc->frame_size <= 1) { |
|
/* specific hack for pcm codecs because no frame size is |
|
provided */ |
|
switch(enc->codec_id) { |
|
case CODEC_ID_PCM_S32LE: |
|
case CODEC_ID_PCM_S32BE: |
|
case CODEC_ID_PCM_U32LE: |
|
case CODEC_ID_PCM_U32BE: |
|
if (enc->channels == 0) |
|
return -1; |
|
frame_size = size / (4 * enc->channels); |
|
break; |
|
case CODEC_ID_PCM_S24LE: |
|
case CODEC_ID_PCM_S24BE: |
|
case CODEC_ID_PCM_U24LE: |
|
case CODEC_ID_PCM_U24BE: |
|
case CODEC_ID_PCM_S24DAUD: |
|
if (enc->channels == 0) |
|
return -1; |
|
frame_size = size / (3 * enc->channels); |
|
break; |
|
case CODEC_ID_PCM_S16LE: |
|
case CODEC_ID_PCM_S16BE: |
|
case CODEC_ID_PCM_U16LE: |
|
case CODEC_ID_PCM_U16BE: |
|
if (enc->channels == 0) |
|
return -1; |
|
frame_size = size / (2 * enc->channels); |
|
break; |
|
case CODEC_ID_PCM_S8: |
|
case CODEC_ID_PCM_U8: |
|
case CODEC_ID_PCM_MULAW: |
|
case CODEC_ID_PCM_ALAW: |
|
if (enc->channels == 0) |
|
return -1; |
|
frame_size = size / (enc->channels); |
|
break; |
|
default: |
|
/* used for example by ADPCM codecs */ |
|
if (enc->bit_rate == 0) |
|
return -1; |
|
frame_size = (size * 8 * enc->sample_rate) / enc->bit_rate; |
|
break; |
|
} |
|
} else { |
|
frame_size = enc->frame_size; |
|
} |
|
return frame_size; |
|
} |
|
|
|
|
|
/** |
|
* Return the frame duration in seconds, return 0 if not available. |
|
*/ |
|
static void compute_frame_duration(int *pnum, int *pden, AVStream *st, |
|
AVCodecParserContext *pc, AVPacket *pkt) |
|
{ |
|
int frame_size; |
|
|
|
*pnum = 0; |
|
*pden = 0; |
|
switch(st->codec->codec_type) { |
|
case CODEC_TYPE_VIDEO: |
|
if(st->time_base.num*1000LL > st->time_base.den){ |
|
*pnum = st->time_base.num; |
|
*pden = st->time_base.den; |
|
}else if(st->codec->time_base.num*1000LL > st->codec->time_base.den){ |
|
*pnum = st->codec->time_base.num; |
|
*pden = st->codec->time_base.den; |
|
if (pc && pc->repeat_pict) { |
|
*pden *= 2; |
|
*pnum = (*pnum) * (2 + pc->repeat_pict); |
|
} |
|
} |
|
break; |
|
case CODEC_TYPE_AUDIO: |
|
frame_size = get_audio_frame_size(st->codec, pkt->size); |
|
if (frame_size < 0) |
|
break; |
|
*pnum = frame_size; |
|
*pden = st->codec->sample_rate; |
|
break; |
|
default: |
|
break; |
|
} |
|
} |
|
|
|
static int is_intra_only(AVCodecContext *enc){ |
|
if(enc->codec_type == CODEC_TYPE_AUDIO){ |
|
return 1; |
|
}else if(enc->codec_type == CODEC_TYPE_VIDEO){ |
|
switch(enc->codec_id){ |
|
case CODEC_ID_MJPEG: |
|
case CODEC_ID_MJPEGB: |
|
case CODEC_ID_LJPEG: |
|
case CODEC_ID_RAWVIDEO: |
|
case CODEC_ID_DVVIDEO: |
|
case CODEC_ID_HUFFYUV: |
|
case CODEC_ID_FFVHUFF: |
|
case CODEC_ID_ASV1: |
|
case CODEC_ID_ASV2: |
|
case CODEC_ID_VCR1: |
|
return 1; |
|
default: break; |
|
} |
|
} |
|
return 0; |
|
} |
|
|
|
static int64_t lsb2full(int64_t lsb, int64_t last_ts, int lsb_bits){ |
|
int64_t mask = lsb_bits < 64 ? (1LL<<lsb_bits)-1 : -1LL; |
|
int64_t delta= last_ts - mask/2; |
|
return ((lsb - delta)&mask) + delta; |
|
} |
|
|
|
static void compute_pkt_fields(AVFormatContext *s, AVStream *st, |
|
AVCodecParserContext *pc, AVPacket *pkt) |
|
{ |
|
int num, den, presentation_delayed; |
|
/* handle wrapping */ |
|
if(st->cur_dts != AV_NOPTS_VALUE){ |
|
if(pkt->pts != AV_NOPTS_VALUE) |
|
pkt->pts= lsb2full(pkt->pts, st->cur_dts, st->pts_wrap_bits); |
|
if(pkt->dts != AV_NOPTS_VALUE) |
|
pkt->dts= lsb2full(pkt->dts, st->cur_dts, st->pts_wrap_bits); |
|
} |
|
|
|
if (pkt->duration == 0) { |
|
compute_frame_duration(&num, &den, st, pc, pkt); |
|
if (den && num) { |
|
pkt->duration = av_rescale(1, num * (int64_t)st->time_base.den, den * (int64_t)st->time_base.num); |
|
} |
|
} |
|
|
|
if(is_intra_only(st->codec)) |
|
pkt->flags |= PKT_FLAG_KEY; |
|
|
|
/* do we have a video B frame ? */ |
|
presentation_delayed = 0; |
|
if (st->codec->codec_type == CODEC_TYPE_VIDEO) { |
|
/* XXX: need has_b_frame, but cannot get it if the codec is |
|
not initialized */ |
|
if (( st->codec->codec_id == CODEC_ID_H264 |
|
|| st->codec->has_b_frames) && |
|
pc && pc->pict_type != FF_B_TYPE) |
|
presentation_delayed = 1; |
|
/* this may be redundant, but it shouldnt hurt */ |
|
if(pkt->dts != AV_NOPTS_VALUE && pkt->pts != AV_NOPTS_VALUE && pkt->pts > pkt->dts) |
|
presentation_delayed = 1; |
|
} |
|
|
|
if(st->cur_dts == AV_NOPTS_VALUE){ |
|
if(presentation_delayed) st->cur_dts = -pkt->duration; |
|
else st->cur_dts = 0; |
|
} |
|
|
|
// av_log(NULL, AV_LOG_DEBUG, "IN delayed:%d pts:%lld, dts:%lld cur_dts:%lld st:%d pc:%p\n", presentation_delayed, pkt->pts, pkt->dts, st->cur_dts, pkt->stream_index, pc); |
|
/* interpolate PTS and DTS if they are not present */ |
|
if (presentation_delayed) { |
|
/* DTS = decompression time stamp */ |
|
/* PTS = presentation time stamp */ |
|
if (pkt->dts == AV_NOPTS_VALUE) { |
|
/* if we know the last pts, use it */ |
|
if(st->last_IP_pts != AV_NOPTS_VALUE) |
|
st->cur_dts = pkt->dts = st->last_IP_pts; |
|
else |
|
pkt->dts = st->cur_dts; |
|
} else { |
|
st->cur_dts = pkt->dts; |
|
} |
|
/* this is tricky: the dts must be incremented by the duration |
|
of the frame we are displaying, i.e. the last I or P frame */ |
|
if (st->last_IP_duration == 0) |
|
st->cur_dts += pkt->duration; |
|
else |
|
st->cur_dts += st->last_IP_duration; |
|
st->last_IP_duration = pkt->duration; |
|
st->last_IP_pts= pkt->pts; |
|
/* cannot compute PTS if not present (we can compute it only |
|
by knowing the futur */ |
|
} else if(pkt->pts != AV_NOPTS_VALUE || pkt->dts != AV_NOPTS_VALUE || pkt->duration){ |
|
if(pkt->pts != AV_NOPTS_VALUE && pkt->duration){ |
|
int64_t old_diff= ABS(st->cur_dts - pkt->duration - pkt->pts); |
|
int64_t new_diff= ABS(st->cur_dts - pkt->pts); |
|
if(old_diff < new_diff && old_diff < (pkt->duration>>3)){ |
|
pkt->pts += pkt->duration; |
|
// av_log(NULL, AV_LOG_DEBUG, "id:%d old:%Ld new:%Ld dur:%d cur:%Ld size:%d\n", pkt->stream_index, old_diff, new_diff, pkt->duration, st->cur_dts, pkt->size); |
|
} |
|
} |
|
|
|
/* presentation is not delayed : PTS and DTS are the same */ |
|
if (pkt->pts == AV_NOPTS_VALUE) { |
|
if (pkt->dts == AV_NOPTS_VALUE) { |
|
pkt->pts = st->cur_dts; |
|
pkt->dts = st->cur_dts; |
|
} |
|
else { |
|
st->cur_dts = pkt->dts; |
|
pkt->pts = pkt->dts; |
|
} |
|
} else { |
|
st->cur_dts = pkt->pts; |
|
pkt->dts = pkt->pts; |
|
} |
|
st->cur_dts += pkt->duration; |
|
} |
|
// av_log(NULL, AV_LOG_DEBUG, "OUTdelayed:%d pts:%lld, dts:%lld cur_dts:%lld\n", presentation_delayed, pkt->pts, pkt->dts, st->cur_dts); |
|
|
|
/* update flags */ |
|
if (pc) { |
|
pkt->flags = 0; |
|
/* key frame computation */ |
|
switch(st->codec->codec_type) { |
|
case CODEC_TYPE_VIDEO: |
|
if (pc->pict_type == FF_I_TYPE) |
|
pkt->flags |= PKT_FLAG_KEY; |
|
break; |
|
case CODEC_TYPE_AUDIO: |
|
pkt->flags |= PKT_FLAG_KEY; |
|
break; |
|
default: |
|
break; |
|
} |
|
} |
|
} |
|
|
|
void av_destruct_packet_nofree(AVPacket *pkt) |
|
{ |
|
pkt->data = NULL; pkt->size = 0; |
|
} |
|
|
|
static int av_read_frame_internal(AVFormatContext *s, AVPacket *pkt) |
|
{ |
|
AVStream *st; |
|
int len, ret, i; |
|
|
|
for(;;) { |
|
/* select current input stream component */ |
|
st = s->cur_st; |
|
if (st) { |
|
if (!st->need_parsing || !st->parser) { |
|
/* no parsing needed: we just output the packet as is */ |
|
/* raw data support */ |
|
*pkt = s->cur_pkt; |
|
compute_pkt_fields(s, st, NULL, pkt); |
|
s->cur_st = NULL; |
|
return 0; |
|
} else if (s->cur_len > 0 && st->discard < AVDISCARD_ALL) { |
|
len = av_parser_parse(st->parser, st->codec, &pkt->data, &pkt->size, |
|
s->cur_ptr, s->cur_len, |
|
s->cur_pkt.pts, s->cur_pkt.dts); |
|
s->cur_pkt.pts = AV_NOPTS_VALUE; |
|
s->cur_pkt.dts = AV_NOPTS_VALUE; |
|
/* increment read pointer */ |
|
s->cur_ptr += len; |
|
s->cur_len -= len; |
|
|
|
/* return packet if any */ |
|
if (pkt->size) { |
|
got_packet: |
|
pkt->duration = 0; |
|
pkt->stream_index = st->index; |
|
pkt->pts = st->parser->pts; |
|
pkt->dts = st->parser->dts; |
|
pkt->destruct = av_destruct_packet_nofree; |
|
compute_pkt_fields(s, st, st->parser, pkt); |
|
return 0; |
|
} |
|
} else { |
|
/* free packet */ |
|
av_free_packet(&s->cur_pkt); |
|
s->cur_st = NULL; |
|
} |
|
} else { |
|
/* read next packet */ |
|
ret = av_read_packet(s, &s->cur_pkt); |
|
if (ret < 0) { |
|
if (ret == -EAGAIN) |
|
return ret; |
|
/* return the last frames, if any */ |
|
for(i = 0; i < s->nb_streams; i++) { |
|
st = s->streams[i]; |
|
if (st->parser && st->need_parsing) { |
|
av_parser_parse(st->parser, st->codec, |
|
&pkt->data, &pkt->size, |
|
NULL, 0, |
|
AV_NOPTS_VALUE, AV_NOPTS_VALUE); |
|
if (pkt->size) |
|
goto got_packet; |
|
} |
|
} |
|
/* no more packets: really terminates parsing */ |
|
return ret; |
|
} |
|
|
|
st = s->streams[s->cur_pkt.stream_index]; |
|
|
|
s->cur_st = st; |
|
s->cur_ptr = s->cur_pkt.data; |
|
s->cur_len = s->cur_pkt.size; |
|
if (st->need_parsing && !st->parser) { |
|
st->parser = av_parser_init(st->codec->codec_id); |
|
if (!st->parser) { |
|
/* no parser available : just output the raw packets */ |
|
st->need_parsing = 0; |
|
}else if(st->need_parsing == 2){ |
|
st->parser->flags |= PARSER_FLAG_COMPLETE_FRAMES; |
|
} |
|
} |
|
} |
|
} |
|
} |
|
|
|
/** |
|
* Return the next frame of a stream. |
|
* |
|
* The returned packet is valid |
|
* until the next av_read_frame() or until av_close_input_file() and |
|
* must be freed with av_free_packet. For video, the packet contains |
|
* exactly one frame. For audio, it contains an integer number of |
|
* frames if each frame has a known fixed size (e.g. PCM or ADPCM |
|
* data). If the audio frames have a variable size (e.g. MPEG audio), |
|
* then it contains one frame. |
|
* |
|
* pkt->pts, pkt->dts and pkt->duration are always set to correct |
|
* values in AV_TIME_BASE unit (and guessed if the format cannot |
|
* provided them). pkt->pts can be AV_NOPTS_VALUE if the video format |
|
* has B frames, so it is better to rely on pkt->dts if you do not |
|
* decompress the payload. |
|
* |
|
* @return 0 if OK, < 0 if error or end of file. |
|
*/ |
|
int av_read_frame(AVFormatContext *s, AVPacket *pkt) |
|
{ |
|
AVPacketList *pktl; |
|
int eof=0; |
|
const int genpts= s->flags & AVFMT_FLAG_GENPTS; |
|
|
|
for(;;){ |
|
pktl = s->packet_buffer; |
|
if (pktl) { |
|
AVPacket *next_pkt= &pktl->pkt; |
|
|
|
if(genpts && next_pkt->dts != AV_NOPTS_VALUE){ |
|
while(pktl && next_pkt->pts == AV_NOPTS_VALUE){ |
|
if( pktl->pkt.stream_index == next_pkt->stream_index |
|
&& next_pkt->dts < pktl->pkt.dts |
|
&& pktl->pkt.pts != pktl->pkt.dts //not b frame |
|
/*&& pktl->pkt.dts != AV_NOPTS_VALUE*/){ |
|
next_pkt->pts= pktl->pkt.dts; |
|
} |
|
pktl= pktl->next; |
|
} |
|
pktl = s->packet_buffer; |
|
} |
|
|
|
if( next_pkt->pts != AV_NOPTS_VALUE |
|
|| next_pkt->dts == AV_NOPTS_VALUE |
|
|| !genpts || eof){ |
|
/* read packet from packet buffer, if there is data */ |
|
*pkt = *next_pkt; |
|
s->packet_buffer = pktl->next; |
|
av_free(pktl); |
|
return 0; |
|
} |
|
} |
|
if(genpts){ |
|
AVPacketList **plast_pktl= &s->packet_buffer; |
|
int ret= av_read_frame_internal(s, pkt); |
|
if(ret<0){ |
|
if(pktl && ret != -EAGAIN){ |
|
eof=1; |
|
continue; |
|
}else |
|
return ret; |
|
} |
|
|
|
/* duplicate the packet */ |
|
if (av_dup_packet(pkt) < 0) |
|
return AVERROR_NOMEM; |
|
|
|
while(*plast_pktl) plast_pktl= &(*plast_pktl)->next; //FIXME maybe maintain pointer to the last? |
|
|
|
pktl = av_mallocz(sizeof(AVPacketList)); |
|
if (!pktl) |
|
return AVERROR_NOMEM; |
|
|
|
/* add the packet in the buffered packet list */ |
|
*plast_pktl = pktl; |
|
pktl->pkt= *pkt; |
|
}else{ |
|
assert(!s->packet_buffer); |
|
return av_read_frame_internal(s, pkt); |
|
} |
|
} |
|
} |
|
|
|
/* XXX: suppress the packet queue */ |
|
static void flush_packet_queue(AVFormatContext *s) |
|
{ |
|
AVPacketList *pktl; |
|
|
|
for(;;) { |
|
pktl = s->packet_buffer; |
|
if (!pktl) |
|
break; |
|
s->packet_buffer = pktl->next; |
|
av_free_packet(&pktl->pkt); |
|
av_free(pktl); |
|
} |
|
} |
|
|
|
/*******************************************************/ |
|
/* seek support */ |
|
|
|
int av_find_default_stream_index(AVFormatContext *s) |
|
{ |
|
int i; |
|
AVStream *st; |
|
|
|
if (s->nb_streams <= 0) |
|
return -1; |
|
for(i = 0; i < s->nb_streams; i++) { |
|
st = s->streams[i]; |
|
if (st->codec->codec_type == CODEC_TYPE_VIDEO) { |
|
return i; |
|
} |
|
} |
|
return 0; |
|
} |
|
|
|
/** |
|
* Flush the frame reader. |
|
*/ |
|
static void av_read_frame_flush(AVFormatContext *s) |
|
{ |
|
AVStream *st; |
|
int i; |
|
|
|
flush_packet_queue(s); |
|
|
|
/* free previous packet */ |
|
if (s->cur_st) { |
|
if (s->cur_st->parser) |
|
av_free_packet(&s->cur_pkt); |
|
s->cur_st = NULL; |
|
} |
|
/* fail safe */ |
|
s->cur_ptr = NULL; |
|
s->cur_len = 0; |
|
|
|
/* for each stream, reset read state */ |
|
for(i = 0; i < s->nb_streams; i++) { |
|
st = s->streams[i]; |
|
|
|
if (st->parser) { |
|
av_parser_close(st->parser); |
|
st->parser = NULL; |
|
} |
|
st->last_IP_pts = AV_NOPTS_VALUE; |
|
st->cur_dts = 0; /* we set the current DTS to an unspecified origin */ |
|
} |
|
} |
|
|
|
/** |
|
* Updates cur_dts of all streams based on given timestamp and AVStream. |
|
* |
|
* Stream ref_st unchanged, others set cur_dts in their native timebase |
|
* only needed for timestamp wrapping or if (dts not set and pts!=dts) |
|
* @param timestamp new dts expressed in time_base of param ref_st |
|
* @param ref_st reference stream giving time_base of param timestamp |
|
*/ |
|
static void av_update_cur_dts(AVFormatContext *s, AVStream *ref_st, int64_t timestamp){ |
|
int i; |
|
|
|
for(i = 0; i < s->nb_streams; i++) { |
|
AVStream *st = s->streams[i]; |
|
|
|
st->cur_dts = av_rescale(timestamp, |
|
st->time_base.den * (int64_t)ref_st->time_base.num, |
|
st->time_base.num * (int64_t)ref_st->time_base.den); |
|
} |
|
} |
|
|
|
/** |
|
* Add a index entry into a sorted list updateing if it is already there. |
|
* |
|
* @param timestamp timestamp in the timebase of the given stream |
|
*/ |
|
int av_add_index_entry(AVStream *st, |
|
int64_t pos, int64_t timestamp, int size, int distance, int flags) |
|
{ |
|
AVIndexEntry *entries, *ie; |
|
int index; |
|
|
|
if((unsigned)st->nb_index_entries + 1 >= UINT_MAX / sizeof(AVIndexEntry)) |
|
return -1; |
|
|
|
entries = av_fast_realloc(st->index_entries, |
|
&st->index_entries_allocated_size, |
|
(st->nb_index_entries + 1) * |
|
sizeof(AVIndexEntry)); |
|
if(!entries) |
|
return -1; |
|
|
|
st->index_entries= entries; |
|
|
|
index= av_index_search_timestamp(st, timestamp, AVSEEK_FLAG_ANY); |
|
|
|
if(index<0){ |
|
index= st->nb_index_entries++; |
|
ie= &entries[index]; |
|
assert(index==0 || ie[-1].timestamp < timestamp); |
|
}else{ |
|
ie= &entries[index]; |
|
if(ie->timestamp != timestamp){ |
|
if(ie->timestamp <= timestamp) |
|
return -1; |
|
memmove(entries + index + 1, entries + index, sizeof(AVIndexEntry)*(st->nb_index_entries - index)); |
|
st->nb_index_entries++; |
|
}else if(ie->pos == pos && distance < ie->min_distance) //dont reduce the distance |
|
distance= ie->min_distance; |
|
} |
|
|
|
ie->pos = pos; |
|
ie->timestamp = timestamp; |
|
ie->min_distance= distance; |
|
ie->size= size; |
|
ie->flags = flags; |
|
|
|
return index; |
|
} |
|
|
|
/** |
|
* build an index for raw streams using a parser. |
|
*/ |
|
static void av_build_index_raw(AVFormatContext *s) |
|
{ |
|
AVPacket pkt1, *pkt = &pkt1; |
|
int ret; |
|
AVStream *st; |
|
|
|
st = s->streams[0]; |
|
av_read_frame_flush(s); |
|
url_fseek(&s->pb, s->data_offset, SEEK_SET); |
|
|
|
for(;;) { |
|
ret = av_read_frame(s, pkt); |
|
if (ret < 0) |
|
break; |
|
if (pkt->stream_index == 0 && st->parser && |
|
(pkt->flags & PKT_FLAG_KEY)) { |
|
av_add_index_entry(st, st->parser->frame_offset, pkt->dts, |
|
0, 0, AVINDEX_KEYFRAME); |
|
} |
|
av_free_packet(pkt); |
|
} |
|
} |
|
|
|
/** |
|
* Returns TRUE if we deal with a raw stream. |
|
* |
|
* Raw codec data and parsing needed. |
|
*/ |
|
static int is_raw_stream(AVFormatContext *s) |
|
{ |
|
AVStream *st; |
|
|
|
if (s->nb_streams != 1) |
|
return 0; |
|
st = s->streams[0]; |
|
if (!st->need_parsing) |
|
return 0; |
|
return 1; |
|
} |
|
|
|
/** |
|
* Gets the index for a specific timestamp. |
|
* @param flags if AVSEEK_FLAG_BACKWARD then the returned index will correspond to |
|
* the timestamp which is <= the requested one, if backward is 0 |
|
* then it will be >= |
|
* if AVSEEK_FLAG_ANY seek to any frame, only keyframes otherwise |
|
* @return < 0 if no such timestamp could be found |
|
*/ |
|
int av_index_search_timestamp(AVStream *st, int64_t wanted_timestamp, |
|
int flags) |
|
{ |
|
AVIndexEntry *entries= st->index_entries; |
|
int nb_entries= st->nb_index_entries; |
|
int a, b, m; |
|
int64_t timestamp; |
|
|
|
a = - 1; |
|
b = nb_entries; |
|
|
|
while (b - a > 1) { |
|
m = (a + b) >> 1; |
|
timestamp = entries[m].timestamp; |
|
if(timestamp >= wanted_timestamp) |
|
b = m; |
|
if(timestamp <= wanted_timestamp) |
|
a = m; |
|
} |
|
m= (flags & AVSEEK_FLAG_BACKWARD) ? a : b; |
|
|
|
if(!(flags & AVSEEK_FLAG_ANY)){ |
|
while(m>=0 && m<nb_entries && !(entries[m].flags & AVINDEX_KEYFRAME)){ |
|
m += (flags & AVSEEK_FLAG_BACKWARD) ? -1 : 1; |
|
} |
|
} |
|
|
|
if(m == nb_entries) |
|
return -1; |
|
return m; |
|
} |
|
|
|
#define DEBUG_SEEK |
|
|
|
/** |
|
* Does a binary search using av_index_search_timestamp() and AVCodec.read_timestamp(). |
|
* this isnt supposed to be called directly by a user application, but by demuxers |
|
* @param target_ts target timestamp in the time base of the given stream |
|
* @param stream_index stream number |
|
*/ |
|
int av_seek_frame_binary(AVFormatContext *s, int stream_index, int64_t target_ts, int flags){ |
|
AVInputFormat *avif= s->iformat; |
|
int64_t pos_min, pos_max, pos, pos_limit; |
|
int64_t ts_min, ts_max, ts; |
|
int64_t start_pos, filesize; |
|
int index, no_change; |
|
AVStream *st; |
|
|
|
if (stream_index < 0) |
|
return -1; |
|
|
|
#ifdef DEBUG_SEEK |
|
av_log(s, AV_LOG_DEBUG, "read_seek: %d %"PRId64"\n", stream_index, target_ts); |
|
#endif |
|
|
|
ts_max= |
|
ts_min= AV_NOPTS_VALUE; |
|
pos_limit= -1; //gcc falsely says it may be uninitalized |
|
|
|
st= s->streams[stream_index]; |
|
if(st->index_entries){ |
|
AVIndexEntry *e; |
|
|
|
index= av_index_search_timestamp(st, target_ts, flags | AVSEEK_FLAG_BACKWARD); //FIXME whole func must be checked for non keyframe entries in index case, especially read_timestamp() |
|
index= FFMAX(index, 0); |
|
e= &st->index_entries[index]; |
|
|
|
if(e->timestamp <= target_ts || e->pos == e->min_distance){ |
|
pos_min= e->pos; |
|
ts_min= e->timestamp; |
|
#ifdef DEBUG_SEEK |
|
av_log(s, AV_LOG_DEBUG, "using cached pos_min=0x%"PRIx64" dts_min=%"PRId64"\n", |
|
pos_min,ts_min); |
|
#endif |
|
}else{ |
|
assert(index==0); |
|
} |
|
|
|
index= av_index_search_timestamp(st, target_ts, flags & ~AVSEEK_FLAG_BACKWARD); |
|
assert(index < st->nb_index_entries); |
|
if(index >= 0){ |
|
e= &st->index_entries[index]; |
|
assert(e->timestamp >= target_ts); |
|
pos_max= e->pos; |
|
ts_max= e->timestamp; |
|
pos_limit= pos_max - e->min_distance; |
|
#ifdef DEBUG_SEEK |
|
av_log(s, AV_LOG_DEBUG, "using cached pos_max=0x%"PRIx64" pos_limit=0x%"PRIx64" dts_max=%"PRId64"\n", |
|
pos_max,pos_limit, ts_max); |
|
#endif |
|
} |
|
} |
|
|
|
if(ts_min == AV_NOPTS_VALUE){ |
|
pos_min = s->data_offset; |
|
ts_min = avif->read_timestamp(s, stream_index, &pos_min, INT64_MAX); |
|
if (ts_min == AV_NOPTS_VALUE) |
|
return -1; |
|
} |
|
|
|
if(ts_max == AV_NOPTS_VALUE){ |
|
int step= 1024; |
|
filesize = url_fsize(&s->pb); |
|
pos_max = filesize - 1; |
|
do{ |
|
pos_max -= step; |
|
ts_max = avif->read_timestamp(s, stream_index, &pos_max, pos_max + step); |
|
step += step; |
|
}while(ts_max == AV_NOPTS_VALUE && pos_max >= step); |
|
if (ts_max == AV_NOPTS_VALUE) |
|
return -1; |
|
|
|
for(;;){ |
|
int64_t tmp_pos= pos_max + 1; |
|
int64_t tmp_ts= avif->read_timestamp(s, stream_index, &tmp_pos, INT64_MAX); |
|
if(tmp_ts == AV_NOPTS_VALUE) |
|
break; |
|
ts_max= tmp_ts; |
|
pos_max= tmp_pos; |
|
if(tmp_pos >= filesize) |
|
break; |
|
} |
|
pos_limit= pos_max; |
|
} |
|
|
|
if(ts_min > ts_max){ |
|
return -1; |
|
}else if(ts_min == ts_max){ |
|
pos_limit= pos_min; |
|
} |
|
|
|
no_change=0; |
|
while (pos_min < pos_limit) { |
|
#ifdef DEBUG_SEEK |
|
av_log(s, AV_LOG_DEBUG, "pos_min=0x%"PRIx64" pos_max=0x%"PRIx64" dts_min=%"PRId64" dts_max=%"PRId64"\n", |
|
pos_min, pos_max, |
|
ts_min, ts_max); |
|
#endif |
|
assert(pos_limit <= pos_max); |
|
|
|
if(no_change==0){ |
|
int64_t approximate_keyframe_distance= pos_max - pos_limit; |
|
// interpolate position (better than dichotomy) |
|
pos = av_rescale(target_ts - ts_min, pos_max - pos_min, ts_max - ts_min) |
|
+ pos_min - approximate_keyframe_distance; |
|
}else if(no_change==1){ |
|
// bisection, if interpolation failed to change min or max pos last time |
|
pos = (pos_min + pos_limit)>>1; |
|
}else{ |
|
// linear search if bisection failed, can only happen if there are very few or no keframes between min/max |
|
pos=pos_min; |
|
} |
|
if(pos <= pos_min) |
|
pos= pos_min + 1; |
|
else if(pos > pos_limit) |
|
pos= pos_limit; |
|
start_pos= pos; |
|
|
|
ts = avif->read_timestamp(s, stream_index, &pos, INT64_MAX); //may pass pos_limit instead of -1 |
|
if(pos == pos_max) |
|
no_change++; |
|
else |
|
no_change=0; |
|
#ifdef DEBUG_SEEK |
|
av_log(s, AV_LOG_DEBUG, "%"PRId64" %"PRId64" %"PRId64" / %"PRId64" %"PRId64" %"PRId64" target:%"PRId64" limit:%"PRId64" start:%"PRId64" noc:%d\n", pos_min, pos, pos_max, ts_min, ts, ts_max, target_ts, pos_limit, start_pos, no_change); |
|
#endif |
|
assert(ts != AV_NOPTS_VALUE); |
|
if (target_ts <= ts) { |
|
pos_limit = start_pos - 1; |
|
pos_max = pos; |
|
ts_max = ts; |
|
} |
|
if (target_ts >= ts) { |
|
pos_min = pos; |
|
ts_min = ts; |
|
} |
|
} |
|
|
|
pos = (flags & AVSEEK_FLAG_BACKWARD) ? pos_min : pos_max; |
|
ts = (flags & AVSEEK_FLAG_BACKWARD) ? ts_min : ts_max; |
|
#ifdef DEBUG_SEEK |
|
pos_min = pos; |
|
ts_min = avif->read_timestamp(s, stream_index, &pos_min, INT64_MAX); |
|
pos_min++; |
|
ts_max = avif->read_timestamp(s, stream_index, &pos_min, INT64_MAX); |
|
av_log(s, AV_LOG_DEBUG, "pos=0x%"PRIx64" %"PRId64"<=%"PRId64"<=%"PRId64"\n", |
|
pos, ts_min, target_ts, ts_max); |
|
#endif |
|
/* do the seek */ |
|
url_fseek(&s->pb, pos, SEEK_SET); |
|
|
|
av_update_cur_dts(s, st, ts); |
|
|
|
return 0; |
|
} |
|
|
|
static int av_seek_frame_byte(AVFormatContext *s, int stream_index, int64_t pos, int flags){ |
|
int64_t pos_min, pos_max; |
|
#if 0 |
|
AVStream *st; |
|
|
|
if (stream_index < 0) |
|
return -1; |
|
|
|
st= s->streams[stream_index]; |
|
#endif |
|
|
|
pos_min = s->data_offset; |
|
pos_max = url_fsize(&s->pb) - 1; |
|
|
|
if (pos < pos_min) pos= pos_min; |
|
else if(pos > pos_max) pos= pos_max; |
|
|
|
url_fseek(&s->pb, pos, SEEK_SET); |
|
|
|
#if 0 |
|
av_update_cur_dts(s, st, ts); |
|
#endif |
|
return 0; |
|
} |
|
|
|
static int av_seek_frame_generic(AVFormatContext *s, |
|
int stream_index, int64_t timestamp, int flags) |
|
{ |
|
int index; |
|
AVStream *st; |
|
AVIndexEntry *ie; |
|
|
|
if (!s->index_built) { |
|
if (is_raw_stream(s)) { |
|
av_build_index_raw(s); |
|
} else { |
|
return -1; |
|
} |
|
s->index_built = 1; |
|
} |
|
|
|
st = s->streams[stream_index]; |
|
index = av_index_search_timestamp(st, timestamp, flags); |
|
if (index < 0) |
|
return -1; |
|
|
|
/* now we have found the index, we can seek */ |
|
ie = &st->index_entries[index]; |
|
av_read_frame_flush(s); |
|
url_fseek(&s->pb, ie->pos, SEEK_SET); |
|
|
|
av_update_cur_dts(s, st, ie->timestamp); |
|
|
|
return 0; |
|
} |
|
|
|
/** |
|
* Seek to the key frame at timestamp. |
|
* 'timestamp' in 'stream_index'. |
|
* @param stream_index If stream_index is (-1), a default |
|
* stream is selected, and timestamp is automatically converted |
|
* from AV_TIME_BASE units to the stream specific time_base. |
|
* @param timestamp timestamp in AVStream.time_base units |
|
* or if there is no stream specified then in AV_TIME_BASE units |
|
* @param flags flags which select direction and seeking mode |
|
* @return >= 0 on success |
|
*/ |
|
int av_seek_frame(AVFormatContext *s, int stream_index, int64_t timestamp, int flags) |
|
{ |
|
int ret; |
|
AVStream *st; |
|
|
|
av_read_frame_flush(s); |
|
|
|
if(flags & AVSEEK_FLAG_BYTE) |
|
return av_seek_frame_byte(s, stream_index, timestamp, flags); |
|
|
|
if(stream_index < 0){ |
|
stream_index= av_find_default_stream_index(s); |
|
if(stream_index < 0) |
|
return -1; |
|
|
|
st= s->streams[stream_index]; |
|
/* timestamp for default must be expressed in AV_TIME_BASE units */ |
|
timestamp = av_rescale(timestamp, st->time_base.den, AV_TIME_BASE * (int64_t)st->time_base.num); |
|
} |
|
st= s->streams[stream_index]; |
|
|
|
/* first, we try the format specific seek */ |
|
if (s->iformat->read_seek) |
|
ret = s->iformat->read_seek(s, stream_index, timestamp, flags); |
|
else |
|
ret = -1; |
|
if (ret >= 0) { |
|
return 0; |
|
} |
|
|
|
if(s->iformat->read_timestamp) |
|
return av_seek_frame_binary(s, stream_index, timestamp, flags); |
|
else |
|
return av_seek_frame_generic(s, stream_index, timestamp, flags); |
|
} |
|
|
|
/*******************************************************/ |
|
|
|
/** |
|
* Returns TRUE if the stream has accurate timings in any stream. |
|
* |
|
* @return TRUE if the stream has accurate timings for at least one component. |
|
*/ |
|
static int av_has_timings(AVFormatContext *ic) |
|
{ |
|
int i; |
|
AVStream *st; |
|
|
|
for(i = 0;i < ic->nb_streams; i++) { |
|
st = ic->streams[i]; |
|
if (st->start_time != AV_NOPTS_VALUE && |
|
st->duration != AV_NOPTS_VALUE) |
|
return 1; |
|
} |
|
return 0; |
|
} |
|
|
|
/** |
|
* Estimate the stream timings from the one of each components. |
|
* |
|
* Also computes the global bitrate if possible. |
|
*/ |
|
static void av_update_stream_timings(AVFormatContext *ic) |
|
{ |
|
int64_t start_time, start_time1, end_time, end_time1; |
|
int i; |
|
AVStream *st; |
|
|
|
start_time = MAXINT64; |
|
end_time = MININT64; |
|
for(i = 0;i < ic->nb_streams; i++) { |
|
st = ic->streams[i]; |
|
if (st->start_time != AV_NOPTS_VALUE) { |
|
start_time1= av_rescale_q(st->start_time, st->time_base, AV_TIME_BASE_Q); |
|
if (start_time1 < start_time) |
|
start_time = start_time1; |
|
if (st->duration != AV_NOPTS_VALUE) { |
|
end_time1 = start_time1 |
|
+ av_rescale_q(st->duration, st->time_base, AV_TIME_BASE_Q); |
|
if (end_time1 > end_time) |
|
end_time = end_time1; |
|
} |
|
} |
|
} |
|
if (start_time != MAXINT64) { |
|
ic->start_time = start_time; |
|
if (end_time != MININT64) { |
|
ic->duration = end_time - start_time; |
|
if (ic->file_size > 0) { |
|
/* compute the bit rate */ |
|
ic->bit_rate = (double)ic->file_size * 8.0 * AV_TIME_BASE / |
|
(double)ic->duration; |
|
} |
|
} |
|
} |
|
|
|
} |
|
|
|
static void fill_all_stream_timings(AVFormatContext *ic) |
|
{ |
|
int i; |
|
AVStream *st; |
|
|
|
av_update_stream_timings(ic); |
|
for(i = 0;i < ic->nb_streams; i++) { |
|
st = ic->streams[i]; |
|
if (st->start_time == AV_NOPTS_VALUE) { |
|
if(ic->start_time != AV_NOPTS_VALUE) |
|
st->start_time = av_rescale_q(ic->start_time, AV_TIME_BASE_Q, st->time_base); |
|
if(ic->duration != AV_NOPTS_VALUE) |
|
st->duration = av_rescale_q(ic->duration, AV_TIME_BASE_Q, st->time_base); |
|
} |
|
} |
|
} |
|
|
|
static void av_estimate_timings_from_bit_rate(AVFormatContext *ic) |
|
{ |
|
int64_t filesize, duration; |
|
int bit_rate, i; |
|
AVStream *st; |
|
|
|
/* if bit_rate is already set, we believe it */ |
|
if (ic->bit_rate == 0) { |
|
bit_rate = 0; |
|
for(i=0;i<ic->nb_streams;i++) { |
|
st = ic->streams[i]; |
|
bit_rate += st->codec->bit_rate; |
|
} |
|
ic->bit_rate = bit_rate; |
|
} |
|
|
|
/* if duration is already set, we believe it */ |
|
if (ic->duration == AV_NOPTS_VALUE && |
|
ic->bit_rate != 0 && |
|
ic->file_size != 0) { |
|
filesize = ic->file_size; |
|
if (filesize > 0) { |
|
for(i = 0; i < ic->nb_streams; i++) { |
|
st = ic->streams[i]; |
|
duration= av_rescale(8*filesize, st->time_base.den, ic->bit_rate*(int64_t)st->time_base.num); |
|
if (st->start_time == AV_NOPTS_VALUE || |
|
st->duration == AV_NOPTS_VALUE) { |
|
st->start_time = 0; |
|
st->duration = duration; |
|
} |
|
} |
|
} |
|
} |
|
} |
|
|
|
#define DURATION_MAX_READ_SIZE 250000 |
|
|
|
/* only usable for MPEG-PS streams */ |
|
static void av_estimate_timings_from_pts(AVFormatContext *ic) |
|
{ |
|
AVPacket pkt1, *pkt = &pkt1; |
|
AVStream *st; |
|
int read_size, i, ret; |
|
int64_t end_time; |
|
int64_t filesize, offset, duration; |
|
|
|
/* free previous packet */ |
|
if (ic->cur_st && ic->cur_st->parser) |
|
av_free_packet(&ic->cur_pkt); |
|
ic->cur_st = NULL; |
|
|
|
/* flush packet queue */ |
|
flush_packet_queue(ic); |
|
|
|
for(i=0;i<ic->nb_streams;i++) { |
|
st = ic->streams[i]; |
|
if (st->parser) { |
|
av_parser_close(st->parser); |
|
st->parser= NULL; |
|
} |
|
} |
|
|
|
/* we read the first packets to get the first PTS (not fully |
|
accurate, but it is enough now) */ |
|
url_fseek(&ic->pb, 0, SEEK_SET); |
|
read_size = 0; |
|
for(;;) { |
|
if (read_size >= DURATION_MAX_READ_SIZE) |
|
break; |
|
/* if all info is available, we can stop */ |
|
for(i = 0;i < ic->nb_streams; i++) { |
|
st = ic->streams[i]; |
|
if (st->start_time == AV_NOPTS_VALUE) |
|
break; |
|
} |
|
if (i == ic->nb_streams) |
|
break; |
|
|
|
ret = av_read_packet(ic, pkt); |
|
if (ret != 0) |
|
break; |
|
read_size += pkt->size; |
|
st = ic->streams[pkt->stream_index]; |
|
if (pkt->pts != AV_NOPTS_VALUE) { |
|
if (st->start_time == AV_NOPTS_VALUE) |
|
st->start_time = pkt->pts; |
|
} |
|
av_free_packet(pkt); |
|
} |
|
|
|
/* estimate the end time (duration) */ |
|
/* XXX: may need to support wrapping */ |
|
filesize = ic->file_size; |
|
offset = filesize - DURATION_MAX_READ_SIZE; |
|
if (offset < 0) |
|
offset = 0; |
|
|
|
url_fseek(&ic->pb, offset, SEEK_SET); |
|
read_size = 0; |
|
for(;;) { |
|
if (read_size >= DURATION_MAX_READ_SIZE) |
|
break; |
|
/* if all info is available, we can stop */ |
|
for(i = 0;i < ic->nb_streams; i++) { |
|
st = ic->streams[i]; |
|
if (st->duration == AV_NOPTS_VALUE) |
|
break; |
|
} |
|
if (i == ic->nb_streams) |
|
break; |
|
|
|
ret = av_read_packet(ic, pkt); |
|
if (ret != 0) |
|
break; |
|
read_size += pkt->size; |
|
st = ic->streams[pkt->stream_index]; |
|
if (pkt->pts != AV_NOPTS_VALUE) { |
|
end_time = pkt->pts; |
|
duration = end_time - st->start_time; |
|
if (duration > 0) { |
|
if (st->duration == AV_NOPTS_VALUE || |
|
st->duration < duration) |
|
st->duration = duration; |
|
} |
|
} |
|
av_free_packet(pkt); |
|
} |
|
|
|
fill_all_stream_timings(ic); |
|
|
|
url_fseek(&ic->pb, 0, SEEK_SET); |
|
} |
|
|
|
static void av_estimate_timings(AVFormatContext *ic) |
|
{ |
|
int64_t file_size; |
|
|
|
/* get the file size, if possible */ |
|
if (ic->iformat->flags & AVFMT_NOFILE) { |
|
file_size = 0; |
|
} else { |
|
file_size = url_fsize(&ic->pb); |
|
if (file_size < 0) |
|
file_size = 0; |
|
} |
|
ic->file_size = file_size; |
|
|
|
if ((ic->iformat == &mpegps_demux || ic->iformat == &mpegts_demux) && file_size && !ic->pb.is_streamed) { |
|
/* get accurate estimate from the PTSes */ |
|
av_estimate_timings_from_pts(ic); |
|
} else if (av_has_timings(ic)) { |
|
/* at least one components has timings - we use them for all |
|
the components */ |
|
fill_all_stream_timings(ic); |
|
} else { |
|
/* less precise: use bit rate info */ |
|
av_estimate_timings_from_bit_rate(ic); |
|
} |
|
av_update_stream_timings(ic); |
|
|
|
#if 0 |
|
{ |
|
int i; |
|
AVStream *st; |
|
for(i = 0;i < ic->nb_streams; i++) { |
|
st = ic->streams[i]; |
|
printf("%d: start_time: %0.3f duration: %0.3f\n", |
|
i, (double)st->start_time / AV_TIME_BASE, |
|
(double)st->duration / AV_TIME_BASE); |
|
} |
|
printf("stream: start_time: %0.3f duration: %0.3f bitrate=%d kb/s\n", |
|
(double)ic->start_time / AV_TIME_BASE, |
|
(double)ic->duration / AV_TIME_BASE, |
|
ic->bit_rate / 1000); |
|
} |
|
#endif |
|
} |
|
|
|
static int has_codec_parameters(AVCodecContext *enc) |
|
{ |
|
int val; |
|
switch(enc->codec_type) { |
|
case CODEC_TYPE_AUDIO: |
|
val = enc->sample_rate; |
|
break; |
|
case CODEC_TYPE_VIDEO: |
|
val = enc->width && enc->pix_fmt != PIX_FMT_NONE; |
|
break; |
|
default: |
|
val = 1; |
|
break; |
|
} |
|
return (val != 0); |
|
} |
|
|
|
static int try_decode_frame(AVStream *st, const uint8_t *data, int size) |
|
{ |
|
int16_t *samples; |
|
AVCodec *codec; |
|
int got_picture, ret=0; |
|
AVFrame picture; |
|
|
|
if(!st->codec->codec){ |
|
codec = avcodec_find_decoder(st->codec->codec_id); |
|
if (!codec) |
|
return -1; |
|
ret = avcodec_open(st->codec, codec); |
|
if (ret < 0) |
|
return ret; |
|
} |
|
|
|
if(!has_codec_parameters(st->codec)){ |
|
switch(st->codec->codec_type) { |
|
case CODEC_TYPE_VIDEO: |
|
ret = avcodec_decode_video(st->codec, &picture, |
|
&got_picture, (uint8_t *)data, size); |
|
break; |
|
case CODEC_TYPE_AUDIO: |
|
samples = av_malloc(AVCODEC_MAX_AUDIO_FRAME_SIZE); |
|
if (!samples) |
|
goto fail; |
|
ret = avcodec_decode_audio(st->codec, samples, |
|
&got_picture, (uint8_t *)data, size); |
|
av_free(samples); |
|
break; |
|
default: |
|
break; |
|
} |
|
} |
|
fail: |
|
return ret; |
|
} |
|
|
|
/* absolute maximum size we read until we abort */ |
|
#define MAX_READ_SIZE 5000000 |
|
|
|
/* maximum duration until we stop analysing the stream */ |
|
#define MAX_STREAM_DURATION ((int)(AV_TIME_BASE * 3.0)) |
|
|
|
/** |
|
* Read the beginning of a media file to get stream information. This |
|
* is useful for file formats with no headers such as MPEG. This |
|
* function also compute the real frame rate in case of mpeg2 repeat |
|
* frame mode. |
|
* |
|
* @param ic media file handle |
|
* @return >=0 if OK. AVERROR_xxx if error. |
|
* @todo let user decide somehow what information is needed so we dont waste time geting stuff the user doesnt need |
|
*/ |
|
int av_find_stream_info(AVFormatContext *ic) |
|
{ |
|
int i, count, ret, read_size, j; |
|
AVStream *st; |
|
AVPacket pkt1, *pkt; |
|
AVPacketList *pktl=NULL, **ppktl; |
|
int64_t last_dts[MAX_STREAMS]; |
|
int64_t duration_sum[MAX_STREAMS]; |
|
int duration_count[MAX_STREAMS]={0}; |
|
|
|
for(i=0;i<ic->nb_streams;i++) { |
|
st = ic->streams[i]; |
|
if(st->codec->codec_type == CODEC_TYPE_VIDEO){ |
|
/* if(!st->time_base.num) |
|
st->time_base= */ |
|
if(!st->codec->time_base.num) |
|
st->codec->time_base= st->time_base; |
|
} |
|
//only for the split stuff |
|
if (!st->parser) { |
|
st->parser = av_parser_init(st->codec->codec_id); |
|
if(st->need_parsing == 2 && st->parser){ |
|
st->parser->flags |= PARSER_FLAG_COMPLETE_FRAMES; |
|
} |
|
} |
|
} |
|
|
|
for(i=0;i<MAX_STREAMS;i++){ |
|
last_dts[i]= AV_NOPTS_VALUE; |
|
duration_sum[i]= INT64_MAX; |
|
} |
|
|
|
count = 0; |
|
read_size = 0; |
|
ppktl = &ic->packet_buffer; |
|
for(;;) { |
|
/* check if one codec still needs to be handled */ |
|
for(i=0;i<ic->nb_streams;i++) { |
|
st = ic->streams[i]; |
|
if (!has_codec_parameters(st->codec)) |
|
break; |
|
/* variable fps and no guess at the real fps */ |
|
if( st->codec->time_base.den >= 101LL*st->codec->time_base.num |
|
&& duration_count[i]<20 && st->codec->codec_type == CODEC_TYPE_VIDEO) |
|
break; |
|
if(st->parser && st->parser->parser->split && !st->codec->extradata) |
|
break; |
|
} |
|
if (i == ic->nb_streams) { |
|
/* NOTE: if the format has no header, then we need to read |
|
some packets to get most of the streams, so we cannot |
|
stop here */ |
|
if (!(ic->ctx_flags & AVFMTCTX_NOHEADER)) { |
|
/* if we found the info for all the codecs, we can stop */ |
|
ret = count; |
|
break; |
|
} |
|
} else { |
|
/* we did not get all the codec info, but we read too much data */ |
|
if (read_size >= MAX_READ_SIZE) { |
|
ret = count; |
|
break; |
|
} |
|
} |
|
|
|
/* NOTE: a new stream can be added there if no header in file |
|
(AVFMTCTX_NOHEADER) */ |
|
ret = av_read_frame_internal(ic, &pkt1); |
|
if (ret < 0) { |
|
/* EOF or error */ |
|
ret = -1; /* we could not have all the codec parameters before EOF */ |
|
for(i=0;i<ic->nb_streams;i++) { |
|
st = ic->streams[i]; |
|
if (!has_codec_parameters(st->codec)){ |
|
char buf[256]; |
|
avcodec_string(buf, sizeof(buf), st->codec, 0); |
|
av_log(ic, AV_LOG_INFO, "Could not find codec parameters (%s)\n", buf); |
|
break; |
|
} |
|
} |
|
if (i == ic->nb_streams) |
|
ret = 0; |
|
break; |
|
} |
|
|
|
pktl = av_mallocz(sizeof(AVPacketList)); |
|
if (!pktl) { |
|
ret = AVERROR_NOMEM; |
|
break; |
|
} |
|
|
|
/* add the packet in the buffered packet list */ |
|
*ppktl = pktl; |
|
ppktl = &pktl->next; |
|
|
|
pkt = &pktl->pkt; |
|
*pkt = pkt1; |
|
|
|
/* duplicate the packet */ |
|
if (av_dup_packet(pkt) < 0) { |
|
ret = AVERROR_NOMEM; |
|
break; |
|
} |
|
|
|
read_size += pkt->size; |
|
|
|
st = ic->streams[pkt->stream_index]; |
|
st->codec_info_duration += pkt->duration; |
|
if (pkt->duration != 0) |
|
st->codec_info_nb_frames++; |
|
|
|
{ |
|
int index= pkt->stream_index; |
|
int64_t last= last_dts[index]; |
|
int64_t duration= pkt->dts - last; |
|
|
|
if(pkt->dts != AV_NOPTS_VALUE && last != AV_NOPTS_VALUE && duration>0){ |
|
if(duration*duration_count[index]*10/9 < duration_sum[index]){ |
|
duration_sum[index]= duration; |
|
duration_count[index]=1; |
|
}else{ |
|
int factor= av_rescale(duration, duration_count[index], duration_sum[index]); |
|
duration_sum[index] += duration; |
|
duration_count[index]+= factor; |
|
} |
|
if(st->codec_info_nb_frames == 0 && 0) |
|
st->codec_info_duration += duration; |
|
} |
|
last_dts[pkt->stream_index]= pkt->dts; |
|
} |
|
if(st->parser && st->parser->parser->split && !st->codec->extradata){ |
|
int i= st->parser->parser->split(st->codec, pkt->data, pkt->size); |
|
if(i){ |
|
st->codec->extradata_size= i; |
|
st->codec->extradata= av_malloc(st->codec->extradata_size); |
|
memcpy(st->codec->extradata, pkt->data, st->codec->extradata_size); |
|
} |
|
} |
|
|
|
/* if still no information, we try to open the codec and to |
|
decompress the frame. We try to avoid that in most cases as |
|
it takes longer and uses more memory. For MPEG4, we need to |
|
decompress for Quicktime. */ |
|
if (!has_codec_parameters(st->codec) /*&& |
|
(st->codec->codec_id == CODEC_ID_FLV1 || |
|
st->codec->codec_id == CODEC_ID_H264 || |
|
st->codec->codec_id == CODEC_ID_H263 || |
|
st->codec->codec_id == CODEC_ID_H261 || |
|
st->codec->codec_id == CODEC_ID_VORBIS || |
|
st->codec->codec_id == CODEC_ID_MJPEG || |
|
st->codec->codec_id == CODEC_ID_PNG || |
|
st->codec->codec_id == CODEC_ID_PAM || |
|
st->codec->codec_id == CODEC_ID_PGM || |
|
st->codec->codec_id == CODEC_ID_PGMYUV || |
|
st->codec->codec_id == CODEC_ID_PBM || |
|
st->codec->codec_id == CODEC_ID_PPM || |
|
st->codec->codec_id == CODEC_ID_SHORTEN || |
|
(st->codec->codec_id == CODEC_ID_MPEG4 && !st->need_parsing))*/) |
|
try_decode_frame(st, pkt->data, pkt->size); |
|
|
|
if (av_rescale_q(st->codec_info_duration, st->time_base, AV_TIME_BASE_Q) >= MAX_STREAM_DURATION) { |
|
break; |
|
} |
|
count++; |
|
} |
|
|
|
// close codecs which where opened in try_decode_frame() |
|
for(i=0;i<ic->nb_streams;i++) { |
|
st = ic->streams[i]; |
|
if(st->codec->codec) |
|
avcodec_close(st->codec); |
|
} |
|
for(i=0;i<ic->nb_streams;i++) { |
|
st = ic->streams[i]; |
|
if (st->codec->codec_type == CODEC_TYPE_VIDEO) { |
|
if(st->codec->codec_id == CODEC_ID_RAWVIDEO && !st->codec->codec_tag && !st->codec->bits_per_sample) |
|
st->codec->codec_tag= avcodec_pix_fmt_to_codec_tag(st->codec->pix_fmt); |
|
|
|
if(duration_count[i] && st->codec->time_base.num*101LL <= st->codec->time_base.den && |
|
st->time_base.num*duration_sum[i]/duration_count[i]*101LL > st->time_base.den){ |
|
int64_t num, den, error, best_error; |
|
|
|
num= st->time_base.den*duration_count[i]; |
|
den= st->time_base.num*duration_sum[i]; |
|
|
|
best_error= INT64_MAX; |
|
for(j=1; j<60*12; j++){ |
|
error= ABS(1001*12*num - 1001*j*den); |
|
if(error < best_error){ |
|
best_error= error; |
|
av_reduce(&st->r_frame_rate.num, &st->r_frame_rate.den, j, 12, INT_MAX); |
|
} |
|
} |
|
for(j=24; j<=30; j+=6){ |
|
error= ABS(1001*12*num - 1000*12*j*den); |
|
if(error < best_error){ |
|
best_error= error; |
|
av_reduce(&st->r_frame_rate.num, &st->r_frame_rate.den, j*1000, 1001, INT_MAX); |
|
} |
|
} |
|
} |
|
|
|
/* set real frame rate info */ |
|
/* compute the real frame rate for telecine */ |
|
if ((st->codec->codec_id == CODEC_ID_MPEG1VIDEO || |
|
st->codec->codec_id == CODEC_ID_MPEG2VIDEO) && |
|
st->codec->sub_id == 2) { |
|
if (st->codec_info_nb_frames >= 20) { |
|
float coded_frame_rate, est_frame_rate; |
|
est_frame_rate = ((double)st->codec_info_nb_frames * AV_TIME_BASE) / |
|
(double)st->codec_info_duration ; |
|
coded_frame_rate = 1.0/av_q2d(st->codec->time_base); |
|
#if 0 |
|
printf("telecine: coded_frame_rate=%0.3f est_frame_rate=%0.3f\n", |
|
coded_frame_rate, est_frame_rate); |
|
#endif |
|
/* if we detect that it could be a telecine, we |
|
signal it. It would be better to do it at a |
|
higher level as it can change in a film */ |
|
if (coded_frame_rate >= 24.97 && |
|
(est_frame_rate >= 23.5 && est_frame_rate < 24.5)) { |
|
st->r_frame_rate = (AVRational){24000, 1001}; |
|
} |
|
} |
|
} |
|
/* if no real frame rate, use the codec one */ |
|
if (!st->r_frame_rate.num){ |
|
st->r_frame_rate.num = st->codec->time_base.den; |
|
st->r_frame_rate.den = st->codec->time_base.num; |
|
} |
|
} |
|
} |
|
|
|
av_estimate_timings(ic); |
|
#if 0 |
|
/* correct DTS for b frame streams with no timestamps */ |
|
for(i=0;i<ic->nb_streams;i++) { |
|
st = ic->streams[i]; |
|
if (st->codec->codec_type == CODEC_TYPE_VIDEO) { |
|
if(b-frames){ |
|
ppktl = &ic->packet_buffer; |
|
while(ppkt1){ |
|
if(ppkt1->stream_index != i) |
|
continue; |
|
if(ppkt1->pkt->dts < 0) |
|
break; |
|
if(ppkt1->pkt->pts != AV_NOPTS_VALUE) |
|
break; |
|
ppkt1->pkt->dts -= delta; |
|
ppkt1= ppkt1->next; |
|
} |
|
if(ppkt1) |
|
continue; |
|
st->cur_dts -= delta; |
|
} |
|
} |
|
} |
|
#endif |
|
return ret; |
|
} |
|
|
|
/*******************************************************/ |
|
|
|
/** |
|
* start playing a network based stream (e.g. RTSP stream) at the |
|
* current position |
|
*/ |
|
int av_read_play(AVFormatContext *s) |
|
{ |
|
if (!s->iformat->read_play) |
|
return AVERROR_NOTSUPP; |
|
return s->iformat->read_play(s); |
|
} |
|
|
|
/** |
|
* Pause a network based stream (e.g. RTSP stream). |
|
* |
|
* Use av_read_play() to resume it. |
|
*/ |
|
int av_read_pause(AVFormatContext *s) |
|
{ |
|
if (!s->iformat->read_pause) |
|
return AVERROR_NOTSUPP; |
|
return s->iformat->read_pause(s); |
|
} |
|
|
|
/** |
|
* Close a media file (but not its codecs). |
|
* |
|
* @param s media file handle |
|
*/ |
|
void av_close_input_file(AVFormatContext *s) |
|
{ |
|
int i, must_open_file; |
|
AVStream *st; |
|
|
|
/* free previous packet */ |
|
if (s->cur_st && s->cur_st->parser) |
|
av_free_packet(&s->cur_pkt); |
|
|
|
if (s->iformat->read_close) |
|
s->iformat->read_close(s); |
|
for(i=0;i<s->nb_streams;i++) { |
|
/* free all data in a stream component */ |
|
st = s->streams[i]; |
|
if (st->parser) { |
|
av_parser_close(st->parser); |
|
} |
|
av_free(st->index_entries); |
|
av_free(st->codec->extradata); |
|
av_free(st->codec); |
|
av_free(st); |
|
} |
|
flush_packet_queue(s); |
|
must_open_file = 1; |
|
if (s->iformat->flags & AVFMT_NOFILE) { |
|
must_open_file = 0; |
|
} |
|
if (must_open_file) { |
|
url_fclose(&s->pb); |
|
} |
|
av_freep(&s->priv_data); |
|
av_free(s); |
|
} |
|
|
|
/** |
|
* Add a new stream to a media file. |
|
* |
|
* Can only be called in the read_header() function. If the flag |
|
* AVFMTCTX_NOHEADER is in the format context, then new streams |
|
* can be added in read_packet too. |
|
* |
|
* @param s media file handle |
|
* @param id file format dependent stream id |
|
*/ |
|
AVStream *av_new_stream(AVFormatContext *s, int id) |
|
{ |
|
AVStream *st; |
|
|
|
if (s->nb_streams >= MAX_STREAMS) |
|
return NULL; |
|
|
|
st = av_mallocz(sizeof(AVStream)); |
|
if (!st) |
|
return NULL; |
|
|
|
st->codec= avcodec_alloc_context(); |
|
if (s->iformat) { |
|
/* no default bitrate if decoding */ |
|
st->codec->bit_rate = 0; |
|
} |
|
st->index = s->nb_streams; |
|
st->id = id; |
|
st->start_time = AV_NOPTS_VALUE; |
|
st->duration = AV_NOPTS_VALUE; |
|
st->cur_dts = AV_NOPTS_VALUE; |
|
|
|
/* default pts settings is MPEG like */ |
|
av_set_pts_info(st, 33, 1, 90000); |
|
st->last_IP_pts = AV_NOPTS_VALUE; |
|
|
|
s->streams[s->nb_streams++] = st; |
|
return st; |
|
} |
|
|
|
/************************************************************/ |
|
/* output media file */ |
|
|
|
int av_set_parameters(AVFormatContext *s, AVFormatParameters *ap) |
|
{ |
|
int ret; |
|
|
|
if (s->oformat->priv_data_size > 0) { |
|
s->priv_data = av_mallocz(s->oformat->priv_data_size); |
|
if (!s->priv_data) |
|
return AVERROR_NOMEM; |
|
} else |
|
s->priv_data = NULL; |
|
|
|
if (s->oformat->set_parameters) { |
|
ret = s->oformat->set_parameters(s, ap); |
|
if (ret < 0) |
|
return ret; |
|
} |
|
return 0; |
|
} |
|
|
|
/** |
|
* allocate the stream private data and write the stream header to an |
|
* output media file |
|
* |
|
* @param s media file handle |
|
* @return 0 if OK. AVERROR_xxx if error. |
|
*/ |
|
int av_write_header(AVFormatContext *s) |
|
{ |
|
int ret, i; |
|
AVStream *st; |
|
|
|
// some sanity checks |
|
for(i=0;i<s->nb_streams;i++) { |
|
st = s->streams[i]; |
|
|
|
switch (st->codec->codec_type) { |
|
case CODEC_TYPE_AUDIO: |
|
if(st->codec->sample_rate<=0){ |
|
av_log(s, AV_LOG_ERROR, "sample rate not set\n"); |
|
return -1; |
|
} |
|
break; |
|
case CODEC_TYPE_VIDEO: |
|
if(st->codec->time_base.num<=0 || st->codec->time_base.den<=0){ //FIXME audio too? |
|
av_log(s, AV_LOG_ERROR, "time base not set\n"); |
|
return -1; |
|
} |
|
if(st->codec->width<=0 || st->codec->height<=0){ |
|
av_log(s, AV_LOG_ERROR, "dimensions not set\n"); |
|
return -1; |
|
} |
|
break; |
|
} |
|
} |
|
|
|
if(s->oformat->write_header){ |
|
ret = s->oformat->write_header(s); |
|
if (ret < 0) |
|
return ret; |
|
} |
|
|
|
/* init PTS generation */ |
|
for(i=0;i<s->nb_streams;i++) { |
|
int64_t den = AV_NOPTS_VALUE; |
|
st = s->streams[i]; |
|
|
|
switch (st->codec->codec_type) { |
|
case CODEC_TYPE_AUDIO: |
|
den = (int64_t)st->time_base.num * st->codec->sample_rate; |
|
break; |
|
case CODEC_TYPE_VIDEO: |
|
den = (int64_t)st->time_base.num * st->codec->time_base.den; |
|
break; |
|
default: |
|
break; |
|
} |
|
if (den != AV_NOPTS_VALUE) { |
|
if (den <= 0) |
|
return AVERROR_INVALIDDATA; |
|
av_frac_init(&st->pts, 0, 0, den); |
|
} |
|
} |
|
return 0; |
|
} |
|
|
|
//FIXME merge with compute_pkt_fields |
|
static int compute_pkt_fields2(AVStream *st, AVPacket *pkt){ |
|
int b_frames = FFMAX(st->codec->has_b_frames, st->codec->max_b_frames); |
|
int num, den, frame_size; |
|
|
|
// av_log(NULL, AV_LOG_DEBUG, "av_write_frame: pts:%lld dts:%lld cur_dts:%lld b:%d size:%d st:%d\n", pkt->pts, pkt->dts, st->cur_dts, b_frames, pkt->size, pkt->stream_index); |
|
|
|
/* if(pkt->pts == AV_NOPTS_VALUE && pkt->dts == AV_NOPTS_VALUE) |
|
return -1;*/ |
|
|
|
/* duration field */ |
|
if (pkt->duration == 0) { |
|
compute_frame_duration(&num, &den, st, NULL, pkt); |
|
if (den && num) { |
|
pkt->duration = av_rescale(1, num * (int64_t)st->time_base.den, den * (int64_t)st->time_base.num); |
|
} |
|
} |
|
|
|
//XXX/FIXME this is a temporary hack until all encoders output pts |
|
if((pkt->pts == 0 || pkt->pts == AV_NOPTS_VALUE) && pkt->dts == AV_NOPTS_VALUE && !b_frames){ |
|
pkt->dts= |
|
// pkt->pts= st->cur_dts; |
|
pkt->pts= st->pts.val; |
|
} |
|
|
|
//calculate dts from pts |
|
if(pkt->pts != AV_NOPTS_VALUE && pkt->dts == AV_NOPTS_VALUE){ |
|
if(b_frames){ |
|
if(st->last_IP_pts == AV_NOPTS_VALUE){ |
|
st->last_IP_pts= -pkt->duration; |
|
} |
|
if(st->last_IP_pts < pkt->pts){ |
|
pkt->dts= st->last_IP_pts; |
|
st->last_IP_pts= pkt->pts; |
|
}else |
|
pkt->dts= pkt->pts; |
|
}else |
|
pkt->dts= pkt->pts; |
|
} |
|
|
|
if(st->cur_dts && st->cur_dts != AV_NOPTS_VALUE && st->cur_dts >= pkt->dts){ |
|
av_log(NULL, AV_LOG_ERROR, "error, non monotone timestamps %"PRId64" >= %"PRId64"\n", st->cur_dts, pkt->dts); |
|
return -1; |
|
} |
|
if(pkt->dts != AV_NOPTS_VALUE && pkt->pts != AV_NOPTS_VALUE && pkt->pts < pkt->dts){ |
|
av_log(NULL, AV_LOG_ERROR, "error, pts < dts\n"); |
|
return -1; |
|
} |
|
|
|
// av_log(NULL, AV_LOG_DEBUG, "av_write_frame: pts2:%lld dts2:%lld\n", pkt->pts, pkt->dts); |
|
st->cur_dts= pkt->dts; |
|
st->pts.val= pkt->dts; |
|
|
|
/* update pts */ |
|
switch (st->codec->codec_type) { |
|
case CODEC_TYPE_AUDIO: |
|
frame_size = get_audio_frame_size(st->codec, pkt->size); |
|
|
|
/* HACK/FIXME, we skip the initial 0-size packets as they are most likely equal to the encoder delay, |
|
but it would be better if we had the real timestamps from the encoder */ |
|
if (frame_size >= 0 && (pkt->size || st->pts.num!=st->pts.den>>1 || st->pts.val)) { |
|
av_frac_add(&st->pts, (int64_t)st->time_base.den * frame_size); |
|
} |
|
break; |
|
case CODEC_TYPE_VIDEO: |
|
av_frac_add(&st->pts, (int64_t)st->time_base.den * st->codec->time_base.num); |
|
break; |
|
default: |
|
break; |
|
} |
|
return 0; |
|
} |
|
|
|
static void truncate_ts(AVStream *st, AVPacket *pkt){ |
|
int64_t pts_mask = (2LL << (st->pts_wrap_bits-1)) - 1; |
|
|
|
// if(pkt->dts < 0) |
|
// pkt->dts= 0; //this happens for low_delay=0 and b frames, FIXME, needs further invstigation about what we should do here |
|
|
|
pkt->pts &= pts_mask; |
|
pkt->dts &= pts_mask; |
|
} |
|
|
|
/** |
|
* Write a packet to an output media file. |
|
* |
|
* The packet shall contain one audio or video frame. |
|
* |
|
* @param s media file handle |
|
* @param pkt the packet, which contains the stream_index, buf/buf_size, dts/pts, ... |
|
* @return < 0 if error, = 0 if OK, 1 if end of stream wanted. |
|
*/ |
|
int av_write_frame(AVFormatContext *s, AVPacket *pkt) |
|
{ |
|
int ret; |
|
|
|
ret=compute_pkt_fields2(s->streams[pkt->stream_index], pkt); |
|
if(ret<0) |
|
return ret; |
|
|
|
truncate_ts(s->streams[pkt->stream_index], pkt); |
|
|
|
ret= s->oformat->write_packet(s, pkt); |
|
if(!ret) |
|
ret= url_ferror(&s->pb); |
|
return ret; |
|
} |
|
|
|
/** |
|
* interleave_packet implementation which will interleave per DTS. |
|
* packets with pkt->destruct == av_destruct_packet will be freed inside this function. |
|
* so they cannot be used after it, note calling av_free_packet() on them is still safe |
|
*/ |
|
static int av_interleave_packet_per_dts(AVFormatContext *s, AVPacket *out, AVPacket *pkt, int flush){ |
|
AVPacketList *pktl, **next_point, *this_pktl; |
|
int stream_count=0; |
|
int streams[MAX_STREAMS]; |
|
|
|
if(pkt){ |
|
AVStream *st= s->streams[ pkt->stream_index]; |
|
|
|
// assert(pkt->destruct != av_destruct_packet); //FIXME |
|
|
|
this_pktl = av_mallocz(sizeof(AVPacketList)); |
|
this_pktl->pkt= *pkt; |
|
if(pkt->destruct == av_destruct_packet) |
|
pkt->destruct= NULL; // non shared -> must keep original from being freed |
|
else |
|
av_dup_packet(&this_pktl->pkt); //shared -> must dup |
|
|
|
next_point = &s->packet_buffer; |
|
while(*next_point){ |
|
AVStream *st2= s->streams[ (*next_point)->pkt.stream_index]; |
|
int64_t left= st2->time_base.num * (int64_t)st ->time_base.den; |
|
int64_t right= st ->time_base.num * (int64_t)st2->time_base.den; |
|
if((*next_point)->pkt.dts * left > pkt->dts * right) //FIXME this can overflow |
|
break; |
|
next_point= &(*next_point)->next; |
|
} |
|
this_pktl->next= *next_point; |
|
*next_point= this_pktl; |
|
} |
|
|
|
memset(streams, 0, sizeof(streams)); |
|
pktl= s->packet_buffer; |
|
while(pktl){ |
|
//av_log(s, AV_LOG_DEBUG, "show st:%d dts:%lld\n", pktl->pkt.stream_index, pktl->pkt.dts); |
|
if(streams[ pktl->pkt.stream_index ] == 0) |
|
stream_count++; |
|
streams[ pktl->pkt.stream_index ]++; |
|
pktl= pktl->next; |
|
} |
|
|
|
if(s->nb_streams == stream_count || (flush && stream_count)){ |
|
pktl= s->packet_buffer; |
|
*out= pktl->pkt; |
|
|
|
s->packet_buffer= pktl->next; |
|
av_freep(&pktl); |
|
return 1; |
|
}else{ |
|
av_init_packet(out); |
|
return 0; |
|
} |
|
} |
|
|
|
/** |
|
* Interleaves a AVPacket correctly so it can be muxed. |
|
* @param out the interleaved packet will be output here |
|
* @param in the input packet |
|
* @param flush 1 if no further packets are available as input and all |
|
* remaining packets should be output |
|
* @return 1 if a packet was output, 0 if no packet could be output, |
|
* < 0 if an error occured |
|
*/ |
|
static int av_interleave_packet(AVFormatContext *s, AVPacket *out, AVPacket *in, int flush){ |
|
if(s->oformat->interleave_packet) |
|
return s->oformat->interleave_packet(s, out, in, flush); |
|
else |
|
return av_interleave_packet_per_dts(s, out, in, flush); |
|
} |
|
|
|
/** |
|
* Writes a packet to an output media file ensuring correct interleaving. |
|
* |
|
* The packet must contain one audio or video frame. |
|
* If the packets are already correctly interleaved the application should |
|
* call av_write_frame() instead as its slightly faster, its also important |
|
* to keep in mind that completly non interleaved input will need huge amounts |
|
* of memory to interleave with this, so its prefereable to interleave at the |
|
* demuxer level |
|
* |
|
* @param s media file handle |
|
* @param pkt the packet, which contains the stream_index, buf/buf_size, dts/pts, ... |
|
* @return < 0 if error, = 0 if OK, 1 if end of stream wanted. |
|
*/ |
|
int av_interleaved_write_frame(AVFormatContext *s, AVPacket *pkt){ |
|
AVStream *st= s->streams[ pkt->stream_index]; |
|
|
|
//FIXME/XXX/HACK drop zero sized packets |
|
if(st->codec->codec_type == CODEC_TYPE_AUDIO && pkt->size==0) |
|
return 0; |
|
|
|
//av_log(NULL, AV_LOG_DEBUG, "av_interleaved_write_frame %d %Ld %Ld\n", pkt->size, pkt->dts, pkt->pts); |
|
if(compute_pkt_fields2(st, pkt) < 0) |
|
return -1; |
|
|
|
if(pkt->dts == AV_NOPTS_VALUE) |
|
return -1; |
|
|
|
for(;;){ |
|
AVPacket opkt; |
|
int ret= av_interleave_packet(s, &opkt, pkt, 0); |
|
if(ret<=0) //FIXME cleanup needed for ret<0 ? |
|
return ret; |
|
|
|
truncate_ts(s->streams[opkt.stream_index], &opkt); |
|
ret= s->oformat->write_packet(s, &opkt); |
|
|
|
av_free_packet(&opkt); |
|
pkt= NULL; |
|
|
|
if(ret<0) |
|
return ret; |
|
if(url_ferror(&s->pb)) |
|
return url_ferror(&s->pb); |
|
} |
|
} |
|
|
|
/** |
|
* @brief Write the stream trailer to an output media file and |
|
* free the file private data. |
|
* |
|
* @param s media file handle |
|
* @return 0 if OK. AVERROR_xxx if error. |
|
*/ |
|
int av_write_trailer(AVFormatContext *s) |
|
{ |
|
int ret, i; |
|
|
|
for(;;){ |
|
AVPacket pkt; |
|
ret= av_interleave_packet(s, &pkt, NULL, 1); |
|
if(ret<0) //FIXME cleanup needed for ret<0 ? |
|
goto fail; |
|
if(!ret) |
|
break; |
|
|
|
truncate_ts(s->streams[pkt.stream_index], &pkt); |
|
ret= s->oformat->write_packet(s, &pkt); |
|
|
|
av_free_packet(&pkt); |
|
|
|
if(ret<0) |
|
goto fail; |
|
if(url_ferror(&s->pb)) |
|
goto fail; |
|
} |
|
|
|
if(s->oformat->write_trailer) |
|
ret = s->oformat->write_trailer(s); |
|
fail: |
|
if(ret == 0) |
|
ret=url_ferror(&s->pb); |
|
for(i=0;i<s->nb_streams;i++) |
|
av_freep(&s->streams[i]->priv_data); |
|
av_freep(&s->priv_data); |
|
return ret; |
|
} |
|
|
|
/* "user interface" functions */ |
|
|
|
void dump_format(AVFormatContext *ic, |
|
int index, |
|
const char *url, |
|
int is_output) |
|
{ |
|
int i, flags; |
|
char buf[256]; |
|
|
|
av_log(NULL, AV_LOG_INFO, "%s #%d, %s, %s '%s':\n", |
|
is_output ? "Output" : "Input", |
|
index, |
|
is_output ? ic->oformat->name : ic->iformat->name, |
|
is_output ? "to" : "from", url); |
|
if (!is_output) { |
|
av_log(NULL, AV_LOG_INFO, " Duration: "); |
|
if (ic->duration != AV_NOPTS_VALUE) { |
|
int hours, mins, secs, us; |
|
secs = ic->duration / AV_TIME_BASE; |
|
us = ic->duration % AV_TIME_BASE; |
|
mins = secs / 60; |
|
secs %= 60; |
|
hours = mins / 60; |
|
mins %= 60; |
|
av_log(NULL, AV_LOG_INFO, "%02d:%02d:%02d.%01d", hours, mins, secs, |
|
(10 * us) / AV_TIME_BASE); |
|
} else { |
|
av_log(NULL, AV_LOG_INFO, "N/A"); |
|
} |
|
if (ic->start_time != AV_NOPTS_VALUE) { |
|
int secs, us; |
|
av_log(NULL, AV_LOG_INFO, ", start: "); |
|
secs = ic->start_time / AV_TIME_BASE; |
|
us = ic->start_time % AV_TIME_BASE; |
|
av_log(NULL, AV_LOG_INFO, "%d.%06d", |
|
secs, (int)av_rescale(us, 1000000, AV_TIME_BASE)); |
|
} |
|
av_log(NULL, AV_LOG_INFO, ", bitrate: "); |
|
if (ic->bit_rate) { |
|
av_log(NULL, AV_LOG_INFO,"%d kb/s", ic->bit_rate / 1000); |
|
} else { |
|
av_log(NULL, AV_LOG_INFO, "N/A"); |
|
} |
|
av_log(NULL, AV_LOG_INFO, "\n"); |
|
} |
|
for(i=0;i<ic->nb_streams;i++) { |
|
AVStream *st = ic->streams[i]; |
|
int g= ff_gcd(st->time_base.num, st->time_base.den); |
|
avcodec_string(buf, sizeof(buf), st->codec, is_output); |
|
av_log(NULL, AV_LOG_INFO, " Stream #%d.%d", index, i); |
|
/* the pid is an important information, so we display it */ |
|
/* XXX: add a generic system */ |
|
if (is_output) |
|
flags = ic->oformat->flags; |
|
else |
|
flags = ic->iformat->flags; |
|
if (flags & AVFMT_SHOW_IDS) { |
|
av_log(NULL, AV_LOG_INFO, "[0x%x]", st->id); |
|
} |
|
if (strlen(st->language) > 0) { |
|
av_log(NULL, AV_LOG_INFO, "(%s)", st->language); |
|
} |
|
av_log(NULL, AV_LOG_DEBUG, ", %d/%d", st->time_base.num/g, st->time_base.den/g); |
|
av_log(NULL, AV_LOG_INFO, ": %s", buf); |
|
if(st->codec->codec_type == CODEC_TYPE_VIDEO){ |
|
if(st->r_frame_rate.den && st->r_frame_rate.num) |
|
av_log(NULL, AV_LOG_INFO, ", %5.2f fps(r)", av_q2d(st->r_frame_rate)); |
|
/* else if(st->time_base.den && st->time_base.num) |
|
av_log(NULL, AV_LOG_INFO, ", %5.2f fps(m)", 1/av_q2d(st->time_base));*/ |
|
else |
|
av_log(NULL, AV_LOG_INFO, ", %5.2f fps(c)", 1/av_q2d(st->codec->time_base)); |
|
} |
|
av_log(NULL, AV_LOG_INFO, "\n"); |
|
} |
|
} |
|
|
|
typedef struct { |
|
const char *abv; |
|
int width, height; |
|
int frame_rate, frame_rate_base; |
|
} AbvEntry; |
|
|
|
static AbvEntry frame_abvs[] = { |
|
{ "ntsc", 720, 480, 30000, 1001 }, |
|
{ "pal", 720, 576, 25, 1 }, |
|
{ "qntsc", 352, 240, 30000, 1001 }, /* VCD compliant ntsc */ |
|
{ "qpal", 352, 288, 25, 1 }, /* VCD compliant pal */ |
|
{ "sntsc", 640, 480, 30000, 1001 }, /* square pixel ntsc */ |
|
{ "spal", 768, 576, 25, 1 }, /* square pixel pal */ |
|
{ "film", 352, 240, 24, 1 }, |
|
{ "ntsc-film", 352, 240, 24000, 1001 }, |
|
{ "sqcif", 128, 96, 0, 0 }, |
|
{ "qcif", 176, 144, 0, 0 }, |
|
{ "cif", 352, 288, 0, 0 }, |
|
{ "4cif", 704, 576, 0, 0 }, |
|
}; |
|
|
|
/** |
|
* parses width and height out of string str. |
|
*/ |
|
int parse_image_size(int *width_ptr, int *height_ptr, const char *str) |
|
{ |
|
int i; |
|
int n = sizeof(frame_abvs) / sizeof(AbvEntry); |
|
const char *p; |
|
int frame_width = 0, frame_height = 0; |
|
|
|
for(i=0;i<n;i++) { |
|
if (!strcmp(frame_abvs[i].abv, str)) { |
|
frame_width = frame_abvs[i].width; |
|
frame_height = frame_abvs[i].height; |
|
break; |
|
} |
|
} |
|
if (i == n) { |
|
p = str; |
|
frame_width = strtol(p, (char **)&p, 10); |
|
if (*p) |
|
p++; |
|
frame_height = strtol(p, (char **)&p, 10); |
|
} |
|
if (frame_width <= 0 || frame_height <= 0) |
|
return -1; |
|
*width_ptr = frame_width; |
|
*height_ptr = frame_height; |
|
return 0; |
|
} |
|
|
|
/** |
|
* Converts frame rate from string to a fraction. |
|
* |
|
* First we try to get an exact integer or fractional frame rate. |
|
* If this fails we convert the frame rate to a double and return |
|
* an approximate fraction using the DEFAULT_FRAME_RATE_BASE. |
|
*/ |
|
int parse_frame_rate(int *frame_rate, int *frame_rate_base, const char *arg) |
|
{ |
|
int i; |
|
char* cp; |
|
|
|
/* First, we check our abbreviation table */ |
|
for (i = 0; i < sizeof(frame_abvs)/sizeof(*frame_abvs); ++i) |
|
if (!strcmp(frame_abvs[i].abv, arg)) { |
|
*frame_rate = frame_abvs[i].frame_rate; |
|
*frame_rate_base = frame_abvs[i].frame_rate_base; |
|
return 0; |
|
} |
|
|
|
/* Then, we try to parse it as fraction */ |
|
cp = strchr(arg, '/'); |
|
if (!cp) |
|
cp = strchr(arg, ':'); |
|
if (cp) { |
|
char* cpp; |
|
*frame_rate = strtol(arg, &cpp, 10); |
|
if (cpp != arg || cpp == cp) |
|
*frame_rate_base = strtol(cp+1, &cpp, 10); |
|
else |
|
*frame_rate = 0; |
|
} |
|
else { |
|
/* Finally we give up and parse it as double */ |
|
AVRational time_base = av_d2q(strtod(arg, 0), DEFAULT_FRAME_RATE_BASE); |
|
*frame_rate_base = time_base.den; |
|
*frame_rate = time_base.num; |
|
} |
|
if (!*frame_rate || !*frame_rate_base) |
|
return -1; |
|
else |
|
return 0; |
|
} |
|
|
|
/** |
|
* Converts date string to number of seconds since Jan 1st, 1970. |
|
* |
|
* @code |
|
* Syntax: |
|
* - If not a duration: |
|
* [{YYYY-MM-DD|YYYYMMDD}]{T| }{HH[:MM[:SS[.m...]]][Z]|HH[MM[SS[.m...]]][Z]} |
|
* Time is localtime unless Z is suffixed to the end. In this case GMT |
|
* Return the date in micro seconds since 1970 |
|
* |
|
* - If a duration: |
|
* HH[:MM[:SS[.m...]]] |
|
* S+[.m...] |
|
* @endcode |
|
*/ |
|
#ifndef CONFIG_WINCE |
|
int64_t parse_date(const char *datestr, int duration) |
|
{ |
|
const char *p; |
|
int64_t t; |
|
struct tm dt; |
|
int i; |
|
static const char *date_fmt[] = { |
|
"%Y-%m-%d", |
|
"%Y%m%d", |
|
}; |
|
static const char *time_fmt[] = { |
|
"%H:%M:%S", |
|
"%H%M%S", |
|
}; |
|
const char *q; |
|
int is_utc, len; |
|
char lastch; |
|
int negative = 0; |
|
|
|
#undef time |
|
time_t now = time(0); |
|
|
|
len = strlen(datestr); |
|
if (len > 0) |
|
lastch = datestr[len - 1]; |
|
else |
|
lastch = '\0'; |
|
is_utc = (lastch == 'z' || lastch == 'Z'); |
|
|
|
memset(&dt, 0, sizeof(dt)); |
|
|
|
p = datestr; |
|
q = NULL; |
|
if (!duration) { |
|
for (i = 0; i < sizeof(date_fmt) / sizeof(date_fmt[0]); i++) { |
|
q = small_strptime(p, date_fmt[i], &dt); |
|
if (q) { |
|
break; |
|
} |
|
} |
|
|
|
if (!q) { |
|
if (is_utc) { |
|
dt = *gmtime(&now); |
|
} else { |
|
dt = *localtime(&now); |
|
} |
|
dt.tm_hour = dt.tm_min = dt.tm_sec = 0; |
|
} else { |
|
p = q; |
|
} |
|
|
|
if (*p == 'T' || *p == 't' || *p == ' ') |
|
p++; |
|
|
|
for (i = 0; i < sizeof(time_fmt) / sizeof(time_fmt[0]); i++) { |
|
q = small_strptime(p, time_fmt[i], &dt); |
|
if (q) { |
|
break; |
|
} |
|
} |
|
} else { |
|
if (p[0] == '-') { |
|
negative = 1; |
|
++p; |
|
} |
|
q = small_strptime(p, time_fmt[0], &dt); |
|
if (!q) { |
|
dt.tm_sec = strtol(p, (char **)&q, 10); |
|
dt.tm_min = 0; |
|
dt.tm_hour = 0; |
|
} |
|
} |
|
|
|
/* Now we have all the fields that we can get */ |
|
if (!q) { |
|
if (duration) |
|
return 0; |
|
else |
|
return now * int64_t_C(1000000); |
|
} |
|
|
|
if (duration) { |
|
t = dt.tm_hour * 3600 + dt.tm_min * 60 + dt.tm_sec; |
|
} else { |
|
dt.tm_isdst = -1; /* unknown */ |
|
if (is_utc) { |
|
t = mktimegm(&dt); |
|
} else { |
|
t = mktime(&dt); |
|
} |
|
} |
|
|
|
t *= 1000000; |
|
|
|
if (*q == '.') { |
|
int val, n; |
|
q++; |
|
for (val = 0, n = 100000; n >= 1; n /= 10, q++) { |
|
if (!isdigit(*q)) |
|
break; |
|
val += n * (*q - '0'); |
|
} |
|
t += val; |
|
} |
|
return negative ? -t : t; |
|
} |
|
#endif /* CONFIG_WINCE */ |
|
|
|
/** |
|
* Attempts to find a specific tag in a URL. |
|
* |
|
* syntax: '?tag1=val1&tag2=val2...'. Little URL decoding is done. |
|
* Return 1 if found. |
|
*/ |
|
int find_info_tag(char *arg, int arg_size, const char *tag1, const char *info) |
|
{ |
|
const char *p; |
|
char tag[128], *q; |
|
|
|
p = info; |
|
if (*p == '?') |
|
p++; |
|
for(;;) { |
|
q = tag; |
|
while (*p != '\0' && *p != '=' && *p != '&') { |
|
if ((q - tag) < sizeof(tag) - 1) |
|
*q++ = *p; |
|
p++; |
|
} |
|
*q = '\0'; |
|
q = arg; |
|
if (*p == '=') { |
|
p++; |
|
while (*p != '&' && *p != '\0') { |
|
if ((q - arg) < arg_size - 1) { |
|
if (*p == '+') |
|
*q++ = ' '; |
|
else |
|
*q++ = *p; |
|
} |
|
p++; |
|
} |
|
*q = '\0'; |
|
} |
|
if (!strcmp(tag, tag1)) |
|
return 1; |
|
if (*p != '&') |
|
break; |
|
p++; |
|
} |
|
return 0; |
|
} |
|
|
|
/** |
|
* Returns in 'buf' the path with '%d' replaced by number. |
|
* |
|
* Also handles the '%0nd' format where 'n' is the total number |
|
* of digits and '%%'. Return 0 if OK, and -1 if format error. |
|
*/ |
|
int get_frame_filename(char *buf, int buf_size, |
|
const char *path, int number) |
|
{ |
|
const char *p; |
|
char *q, buf1[20], c; |
|
int nd, len, percentd_found; |
|
|
|
q = buf; |
|
p = path; |
|
percentd_found = 0; |
|
for(;;) { |
|
c = *p++; |
|
if (c == '\0') |
|
break; |
|
if (c == '%') { |
|
do { |
|
nd = 0; |
|
while (isdigit(*p)) { |
|
nd = nd * 10 + *p++ - '0'; |
|
} |
|
c = *p++; |
|
} while (isdigit(c)); |
|
|
|
switch(c) { |
|
case '%': |
|
goto addchar; |
|
case 'd': |
|
if (percentd_found) |
|
goto fail; |
|
percentd_found = 1; |
|
snprintf(buf1, sizeof(buf1), "%0*d", nd, number); |
|
len = strlen(buf1); |
|
if ((q - buf + len) > buf_size - 1) |
|
goto fail; |
|
memcpy(q, buf1, len); |
|
q += len; |
|
break; |
|
default: |
|
goto fail; |
|
} |
|
} else { |
|
addchar: |
|
if ((q - buf) < buf_size - 1) |
|
*q++ = c; |
|
} |
|
} |
|
if (!percentd_found) |
|
goto fail; |
|
*q = '\0'; |
|
return 0; |
|
fail: |
|
*q = '\0'; |
|
return -1; |
|
} |
|
|
|
/** |
|
* Print nice hexa dump of a buffer |
|
* @param f stream for output |
|
* @param buf buffer |
|
* @param size buffer size |
|
*/ |
|
void av_hex_dump(FILE *f, uint8_t *buf, int size) |
|
{ |
|
int len, i, j, c; |
|
|
|
for(i=0;i<size;i+=16) { |
|
len = size - i; |
|
if (len > 16) |
|
len = 16; |
|
fprintf(f, "%08x ", i); |
|
for(j=0;j<16;j++) { |
|
if (j < len) |
|
fprintf(f, " %02x", buf[i+j]); |
|
else |
|
fprintf(f, " "); |
|
} |
|
fprintf(f, " "); |
|
for(j=0;j<len;j++) { |
|
c = buf[i+j]; |
|
if (c < ' ' || c > '~') |
|
c = '.'; |
|
fprintf(f, "%c", c); |
|
} |
|
fprintf(f, "\n"); |
|
} |
|
} |
|
|
|
/** |
|
* Print on 'f' a nice dump of a packet |
|
* @param f stream for output |
|
* @param pkt packet to dump |
|
* @param dump_payload true if the payload must be displayed too |
|
*/ |
|
//FIXME needs to know the time_base |
|
void av_pkt_dump(FILE *f, AVPacket *pkt, int dump_payload) |
|
{ |
|
fprintf(f, "stream #%d:\n", pkt->stream_index); |
|
fprintf(f, " keyframe=%d\n", ((pkt->flags & PKT_FLAG_KEY) != 0)); |
|
fprintf(f, " duration=%0.3f\n", (double)pkt->duration / AV_TIME_BASE); |
|
/* DTS is _always_ valid after av_read_frame() */ |
|
fprintf(f, " dts="); |
|
if (pkt->dts == AV_NOPTS_VALUE) |
|
fprintf(f, "N/A"); |
|
else |
|
fprintf(f, "%0.3f", (double)pkt->dts / AV_TIME_BASE); |
|
/* PTS may be not known if B frames are present */ |
|
fprintf(f, " pts="); |
|
if (pkt->pts == AV_NOPTS_VALUE) |
|
fprintf(f, "N/A"); |
|
else |
|
fprintf(f, "%0.3f", (double)pkt->pts / AV_TIME_BASE); |
|
fprintf(f, "\n"); |
|
fprintf(f, " size=%d\n", pkt->size); |
|
if (dump_payload) |
|
av_hex_dump(f, pkt->data, pkt->size); |
|
} |
|
|
|
void url_split(char *proto, int proto_size, |
|
char *authorization, int authorization_size, |
|
char *hostname, int hostname_size, |
|
int *port_ptr, |
|
char *path, int path_size, |
|
const char *url) |
|
{ |
|
const char *p; |
|
char *q; |
|
int port; |
|
|
|
port = -1; |
|
|
|
p = url; |
|
q = proto; |
|
while (*p != ':' && *p != '\0') { |
|
if ((q - proto) < proto_size - 1) |
|
*q++ = *p; |
|
p++; |
|
} |
|
if (proto_size > 0) |
|
*q = '\0'; |
|
if (authorization_size > 0) |
|
authorization[0] = '\0'; |
|
if (*p == '\0') { |
|
if (proto_size > 0) |
|
proto[0] = '\0'; |
|
if (hostname_size > 0) |
|
hostname[0] = '\0'; |
|
p = url; |
|
} else { |
|
char *at,*slash; // PETR: position of '@' character and '/' character |
|
|
|
p++; |
|
if (*p == '/') |
|
p++; |
|
if (*p == '/') |
|
p++; |
|
at = strchr(p,'@'); // PETR: get the position of '@' |
|
slash = strchr(p,'/'); // PETR: get position of '/' - end of hostname |
|
if (at && slash && at > slash) at = NULL; // PETR: not interested in '@' behind '/' |
|
|
|
q = at ? authorization : hostname; // PETR: if '@' exists starting with auth. |
|
|
|
while ((at || *p != ':') && *p != '/' && *p != '?' && *p != '\0') { // PETR: |
|
if (*p == '@') { // PETR: passed '@' |
|
if (authorization_size > 0) |
|
*q = '\0'; |
|
q = hostname; |
|
at = NULL; |
|
} else if (!at) { // PETR: hostname |
|
if ((q - hostname) < hostname_size - 1) |
|
*q++ = *p; |
|
} else { |
|
if ((q - authorization) < authorization_size - 1) |
|
*q++ = *p; |
|
} |
|
p++; |
|
} |
|
if (hostname_size > 0) |
|
*q = '\0'; |
|
if (*p == ':') { |
|
p++; |
|
port = strtoul(p, (char **)&p, 10); |
|
} |
|
} |
|
if (port_ptr) |
|
*port_ptr = port; |
|
pstrcpy(path, path_size, p); |
|
} |
|
|
|
/** |
|
* Set the pts for a given stream. |
|
* |
|
* @param s stream |
|
* @param pts_wrap_bits number of bits effectively used by the pts |
|
* (used for wrap control, 33 is the value for MPEG) |
|
* @param pts_num numerator to convert to seconds (MPEG: 1) |
|
* @param pts_den denominator to convert to seconds (MPEG: 90000) |
|
*/ |
|
void av_set_pts_info(AVStream *s, int pts_wrap_bits, |
|
int pts_num, int pts_den) |
|
{ |
|
s->pts_wrap_bits = pts_wrap_bits; |
|
s->time_base.num = pts_num; |
|
s->time_base.den = pts_den; |
|
} |
|
|
|
/* fraction handling */ |
|
|
|
/** |
|
* f = val + (num / den) + 0.5. |
|
* |
|
* 'num' is normalized so that it is such as 0 <= num < den. |
|
* |
|
* @param f fractional number |
|
* @param val integer value |
|
* @param num must be >= 0 |
|
* @param den must be >= 1 |
|
*/ |
|
static void av_frac_init(AVFrac *f, int64_t val, int64_t num, int64_t den) |
|
{ |
|
num += (den >> 1); |
|
if (num >= den) { |
|
val += num / den; |
|
num = num % den; |
|
} |
|
f->val = val; |
|
f->num = num; |
|
f->den = den; |
|
} |
|
|
|
/** |
|
* Set f to (val + 0.5). |
|
*/ |
|
static void av_frac_set(AVFrac *f, int64_t val) |
|
{ |
|
f->val = val; |
|
f->num = f->den >> 1; |
|
} |
|
|
|
/** |
|
* Fractionnal addition to f: f = f + (incr / f->den). |
|
* |
|
* @param f fractional number |
|
* @param incr increment, can be positive or negative |
|
*/ |
|
static void av_frac_add(AVFrac *f, int64_t incr) |
|
{ |
|
int64_t num, den; |
|
|
|
num = f->num + incr; |
|
den = f->den; |
|
if (num < 0) { |
|
f->val += num / den; |
|
num = num % den; |
|
if (num < 0) { |
|
num += den; |
|
f->val--; |
|
} |
|
} else if (num >= den) { |
|
f->val += num / den; |
|
num = num % den; |
|
} |
|
f->num = num; |
|
} |
|
|
|
/** |
|
* register a new image format |
|
* @param img_fmt Image format descriptor |
|
*/ |
|
void av_register_image_format(AVImageFormat *img_fmt) |
|
{ |
|
AVImageFormat **p; |
|
|
|
p = &first_image_format; |
|
while (*p != NULL) p = &(*p)->next; |
|
*p = img_fmt; |
|
img_fmt->next = NULL; |
|
} |
|
|
|
/** |
|
* Guesses image format based on data in the image. |
|
*/ |
|
AVImageFormat *av_probe_image_format(AVProbeData *pd) |
|
{ |
|
AVImageFormat *fmt1, *fmt; |
|
int score, score_max; |
|
|
|
fmt = NULL; |
|
score_max = 0; |
|
for(fmt1 = first_image_format; fmt1 != NULL; fmt1 = fmt1->next) { |
|
if (fmt1->img_probe) { |
|
score = fmt1->img_probe(pd); |
|
if (score > score_max) { |
|
score_max = score; |
|
fmt = fmt1; |
|
} |
|
} |
|
} |
|
return fmt; |
|
} |
|
|
|
/** |
|
* Guesses image format based on file name extensions. |
|
*/ |
|
AVImageFormat *guess_image_format(const char *filename) |
|
{ |
|
AVImageFormat *fmt1; |
|
|
|
for(fmt1 = first_image_format; fmt1 != NULL; fmt1 = fmt1->next) { |
|
if (fmt1->extensions && match_ext(filename, fmt1->extensions)) |
|
return fmt1; |
|
} |
|
return NULL; |
|
} |
|
|
|
/** |
|
* Read an image from a stream. |
|
* @param gb byte stream containing the image |
|
* @param fmt image format, NULL if probing is required |
|
*/ |
|
int av_read_image(ByteIOContext *pb, const char *filename, |
|
AVImageFormat *fmt, |
|
int (*alloc_cb)(void *, AVImageInfo *info), void *opaque) |
|
{ |
|
uint8_t buf[PROBE_BUF_MIN]; |
|
AVProbeData probe_data, *pd = &probe_data; |
|
offset_t pos; |
|
int ret; |
|
|
|
if (!fmt) { |
|
pd->filename = filename; |
|
pd->buf = buf; |
|
pos = url_ftell(pb); |
|
pd->buf_size = get_buffer(pb, buf, PROBE_BUF_MIN); |
|
url_fseek(pb, pos, SEEK_SET); |
|
fmt = av_probe_image_format(pd); |
|
} |
|
if (!fmt) |
|
return AVERROR_NOFMT; |
|
ret = fmt->img_read(pb, alloc_cb, opaque); |
|
return ret; |
|
} |
|
|
|
/** |
|
* Write an image to a stream. |
|
* @param pb byte stream for the image output |
|
* @param fmt image format |
|
* @param img image data and informations |
|
*/ |
|
int av_write_image(ByteIOContext *pb, AVImageFormat *fmt, AVImageInfo *img) |
|
{ |
|
return fmt->img_write(pb, img); |
|
} |
|
|
|
|