#include "avformat.h"
Go to the source code of this file.
Data Structures | |
struct | AVParserStreamState |
structure to store parser state of one AVStream More... | |
struct | AVParserState |
structure to store parser state of AVFormat More... | |
Functions | |
int64_t | ff_gen_syncpoint_search (AVFormatContext *s, int stream_index, int64_t pos, int64_t min_ts, int64_t ts, int64_t max_ts, int flags) |
Search for the sync point of all active streams. | |
AVParserState * | ff_store_parser_state (AVFormatContext *s) |
Store current parser state and file position. | |
void | ff_restore_parser_state (AVFormatContext *s, AVParserState *state) |
Restore previously saved parser state and file position. | |
void | ff_free_parser_state (AVFormatContext *s, AVParserState *state) |
Free previously saved parser state. |
void ff_free_parser_state | ( | AVFormatContext * | s, | |
AVParserState * | state | |||
) |
int64_t ff_gen_syncpoint_search | ( | AVFormatContext * | s, | |
int | stream_index, | |||
int64_t | pos, | |||
int64_t | min_ts, | |||
int64_t | ts, | |||
int64_t | max_ts, | |||
int | flags | |||
) |
Search for the sync point of all active streams.
This routine is not supposed to be called directly by a user application, but by demuxers.
A sync point is defined as a point in stream, such that, when decoding start from this point, the decoded output of all streams synchronizes closest to the given timestamp ts. This routine also takes timestamp limits into account. Thus, the output will synchronize no sooner than ts_min and no later than ts_max.
stream_index | stream index for time base reference of timestamps | |
pos | approximate position where to start searching for key frames | |
min_ts | minimum allowed timestamp (position, if AVSEEK_FLAG_BYTE set) | |
ts | target timestamp (or position, if AVSEEK_FLAG_BYTE set in flags) | |
max_ts | maximum allowed timestamp (position, if AVSEEK_FLAG_BYTE set) | |
flags | if AVSEEK_FLAG_ANY is set, seek to any frame, otherwise only to a keyframe. If AVSEEK_FLAG_BYTE is set, search by position, not by timestamp. |
void ff_restore_parser_state | ( | AVFormatContext * | s, | |
AVParserState * | state | |||
) |
Restore previously saved parser state and file position.
Saved state will be invalidated and freed by this call, since internal structures will be relinked back to the stored state instead of being deeply-copied.
s | context to which to restore state (same as used for storing state) | |
state | state to restore |
AVParserState* ff_store_parser_state | ( | AVFormatContext * | s | ) |
Store current parser state and file position.
This function can be used by demuxers before a destructive seeking algorithm to store the parser state. Depending on the outcome of the seek, either the original state can be restored or the new state kept and the original state freed.
s | context from which to save state |