Support for generic multi-stream key frame finding for new seek API.

Originally committed as revision 19680 to svn://svn.ffmpeg.org/ffmpeg/trunk
release/0.6
Ivan Schreter 16 years ago
parent 50b1785a57
commit 101036adb9
  1. 2
      libavformat/Makefile
  2. 578
      libavformat/seek.c
  3. 97
      libavformat/seek.h
  4. 2
      libavformat/utils.c

@ -5,7 +5,7 @@ FFLIBS = avcodec avutil
HEADERS = avformat.h avio.h
OBJS = allformats.o cutils.o metadata.o metadata_compat.o options.o os_support.o sdp.o utils.o
OBJS = allformats.o cutils.o metadata.o metadata_compat.o options.o os_support.o sdp.o seek.o utils.o
# muxers/demuxers
OBJS-$(CONFIG_AAC_DEMUXER) += raw.o id3v1.o id3v2.o

@ -0,0 +1,578 @@
/*
* Utility functions for seeking for use within FFmpeg format handlers.
*
* Copyright (c) 2009 Ivan Schreter
*
* This file is part of FFmpeg.
*
* FFmpeg is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
* version 2.1 of the License, or (at your option) any later version.
*
* FFmpeg is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with FFmpeg; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
*/
#include "seek.h"
#include "libavutil/mem.h"
// NOTE: implementation should be moved here in another patch, to keep patches
// separated.
extern void av_read_frame_flush(AVFormatContext *s);
/**
* Helper structure to store parser state of AVStream.
*/
typedef struct AVStreamState {
// Saved members of AVStream
AVCodecParserContext *parser;
AVPacket cur_pkt;
int64_t last_IP_pts;
int64_t cur_dts;
int64_t reference_dts;
const uint8_t *cur_ptr;
int cur_len;
int probe_packets;
} AVStreamState;
/**
* Helper structure to store parser state of AVFormat.
*/
struct AVParserState {
int64_t fpos; ///< File position at the time of call.
// Saved members of AVFormatContext
AVStream *cur_st; ///< Current stream.
AVPacketList *packet_buffer; ///< Packet buffer of original state.
AVPacketList *raw_packet_buffer; ///< Raw packet buffer of original state.
int raw_packet_buffer_remaining_size; ///< Remaining size available for raw_packet_buffer.
// Saved info for streams.
int nb_streams; ///< Number of streams with stored state.
AVStreamState *stream_states; ///< States of individual streams (array).
};
/**
* Helper structure describing keyframe search state of one stream.
*/
typedef struct {
int64_t pos_lo; ///< Position of the frame with low timestamp in file or INT64_MAX if not found (yet).
int64_t ts_lo; ///< Frame presentation timestamp or same as pos_lo for byte seeking.
int64_t pos_hi; ///< Position of the frame with high timestamp in file or INT64_MAX if not found (yet).
int64_t ts_hi; ///< Frame presentation timestamp or same as pos_hi for byte seeking.
int64_t last_pos; ///< Last known position of a frame, for multi-frame packets.
int64_t term_ts; ///< Termination timestamp (which TS we already read).
AVRational term_ts_tb; ///< Timebase for term_ts.
int64_t first_ts; ///< First packet timestamp in this iteration (to fill term_ts later).
AVRational first_ts_tb;///< Timebase for first_ts.
int terminated; ///< Termination flag for current iteration.
} AVSyncPoint;
/**
* Compare two timestamps exactly, taking into account their respective time bases.
*
* @param ts_a timestamp A.
* @param tb_a time base for timestamp A.
* @param ts_b timestamp B.
* @param tb_b time base for timestamp A.
* @return -1. 0 or 1 if timestamp A is less than, equal or greater than timestamp B.
*/
static int compare_ts(int64_t ts_a, AVRational tb_a, int64_t ts_b, AVRational tb_b)
{
int64_t a, b, res;
if (ts_a == INT64_MIN)
return ts_a < ts_b ? -1 : 0;
if (ts_a == INT64_MAX)
return ts_a > ts_b ? 1 : 0;
if (ts_b == INT64_MIN)
return ts_a > ts_b ? 1 : 0;
if (ts_b == INT64_MAX)
return ts_a < ts_b ? -1 : 0;
a = ts_a * tb_a.num * tb_b.den;
b = ts_b * tb_b.num * tb_a.den;
res = a - b;
if (res == 0)
return 0;
else
return (res >> 63) | 1;
}
/**
* Compute a distance between timestamps.
*
* Distances are only comparable, if same time bases are used for computing
* distances.
*
* @param ts_hi high timestamp.
* @param tb_hi high timestamp time base.
* @param ts_lo low timestamp.
* @param tb_lo low timestamp time base.
* @return representation of distance between high and low timestamps.
*/
static int64_t ts_distance(int64_t ts_hi, AVRational tb_hi, int64_t ts_lo, AVRational tb_lo)
{
int64_t hi, lo;
hi = ts_hi * tb_hi.num * tb_lo.den;
lo = ts_lo * tb_lo.num * tb_hi.den;
return hi - lo;
}
/**
* Partial search for keyframes in multiple streams.
*
* This routine searches for the next lower and next higher timestamp to
* given target timestamp in each stream, starting at current file position
* and ending at position, where all streams have already been examined
* (or when all higher key frames found in first iteration).
*
* This routine is called iteratively with exponential backoff to find lower
* timestamp.
*
* @param s format context.
* @param timestamp target timestamp (or position, if AVSEEK_FLAG_BYTE).
* @param timebase time base for timestamps.
* @param flags seeking flags.
* @param sync array with information per stream.
* @param keyframes_to_find count of keyframes to find in total.
* @param found_lo pointer to count of already found low timestamp keyframes.
* @param found_hi pointer to count of already found high timestamp keyframes.
* @param first_iter flag for first iteration.
*/
static void search_hi_lo_keyframes(AVFormatContext *s,
int64_t timestamp,
AVRational timebase,
int flags,
AVSyncPoint *sync,
int keyframes_to_find,
int *found_lo,
int *found_hi,
int first_iter)
{
AVPacket pkt;
AVSyncPoint *sp;
AVStream *st;
int idx;
int flg;
int terminated_count = 0;
int64_t pos;
int64_t pts, dts; // PTS/DTS from stream
int64_t ts; // PTS in stream-local time base or position for byte seeking
AVRational ts_tb; // Time base of the stream or 1:1 for byte seeking
for (;;) {
if (av_read_frame(s, &pkt) < 0) {
// EOF or error, make sure high flags are set
for (idx = 0; idx < s->nb_streams; ++idx) {
if (s->streams[idx]->discard < AVDISCARD_ALL) {
sp = &sync[idx];
if (sp->pos_hi == INT64_MAX) {
// No high frame exists for this stream
(*found_hi)++;
sp->ts_hi = INT64_MAX;
sp->pos_hi = INT64_MAX - 1;
}
}
}
break;
}
idx = pkt.stream_index;
st = s->streams[idx];
if (st->discard >= AVDISCARD_ALL) {
// This stream is not active, skip packet.
continue;
}
sp = &sync[idx];
flg = pkt.flags;
pos = pkt.pos;
pts = pkt.pts;
dts = pkt.dts;
if (pts == AV_NOPTS_VALUE) {
// Some formats don't provide PTS, only DTS.
pts = dts;
}
av_free_packet(&pkt);
// Multi-frame packets only return position for the very first frame.
// Other frames are read with position == -1. Therefore, we note down
// last known position of a frame and use it if a frame without
// position arrives. In this way, it's possible to seek to proper
// position. Additionally, for parsers not providing position at all,
// an approximation will be used (starting position of this iteration).
if (pos < 0) {
pos = sp->last_pos;
} else {
sp->last_pos = pos;
}
// Evaluate key frames with known TS (or any frames, if AVSEEK_FLAG_ANY set).
if (pts != AV_NOPTS_VALUE && ((flg & PKT_FLAG_KEY) || (flags & AVSEEK_FLAG_ANY))) {
if (flags & AVSEEK_FLAG_BYTE) {
// For byte seeking, use position as timestamp.
ts = pos;
ts_tb.num = 1;
ts_tb.den = 1;
} else {
// Get stream time_base.
ts = pts;
ts_tb = st->time_base;
}
if (sp->first_ts == AV_NOPTS_VALUE) {
// Note down termination timestamp for the next iteration - when
// we encounter a packet with the same timestamp, we will ignore
// any further packets for this stream in next iteration (as they
// are already evaluated).
sp->first_ts = ts;
sp->first_ts_tb = ts_tb;
}
if (sp->term_ts != AV_NOPTS_VALUE && compare_ts(ts, ts_tb, sp->term_ts, sp->term_ts_tb) > 0) {
// We are past the end position from last iteration, ignore packet.
if (!sp->terminated) {
sp->terminated = 1;
++terminated_count;
if (sp->pos_hi == INT64_MAX) {
// No high frame exists for this stream
(*found_hi)++;
sp->ts_hi = INT64_MAX;
sp->pos_hi = INT64_MAX - 1;
}
if (terminated_count == keyframes_to_find)
break; // all terminated, iteration done
}
continue;
}
if (compare_ts(ts, ts_tb, timestamp, timebase) <= 0) {
// Keyframe found before target timestamp.
if (sp->pos_lo == INT64_MAX) {
// Found first keyframe lower than target timestamp.
(*found_lo)++;
sp->ts_lo = ts;
sp->pos_lo = pos;
} else if (sp->ts_lo < ts) {
// Found a better match (closer to target timestamp).
sp->ts_lo = ts;
sp->pos_lo = pos;
}
}
if (compare_ts(ts, ts_tb, timestamp, timebase) >= 0) {
// Keyframe found after target timestamp.
if (sp->pos_hi == INT64_MAX) {
// Found first keyframe higher than target timestamp.
(*found_hi)++;
sp->ts_hi = ts;
sp->pos_hi = pos;
if (*found_hi >= keyframes_to_find && first_iter) {
// We found high frame for all. They may get updated
// to TS closer to target TS in later iterations (which
// will stop at start position of previous iteration).
break;
}
} else if (sp->ts_hi > ts) {
// Found a better match (actually, shouldn't happen).
sp->ts_hi = ts;
sp->pos_hi = pos;
}
}
}
}
// Clean up the parser.
av_read_frame_flush(s);
}
int64_t ff_gen_syncpoint_search(AVFormatContext *s,
int stream_index,
int64_t pos,
int64_t ts_min,
int64_t ts,
int64_t ts_max,
int flags)
{
AVSyncPoint *sync, *sp;
AVStream *st;
int i;
int keyframes_to_find = 0;
int64_t curpos;
int64_t step;
int found_lo = 0, found_hi = 0;
int64_t min_distance, distance;
int64_t min_pos = 0;
int first_iter = 1;
AVRational time_base;
if (flags & AVSEEK_FLAG_BYTE) {
/* For byte seeking, we have exact 1:1 "timestamps" - positions */
time_base.num = 1;
time_base.den = 1;
} else {
if (stream_index >= 0) {
/* We have a reference stream, which time base we use */
st = s->streams[stream_index];
time_base = st->time_base;
} else {
/* No reference stream, use AV_TIME_BASE as reference time base */
time_base.num = 1;
time_base.den = AV_TIME_BASE;
}
}
// Initialize syncpoint structures for each stream.
sync = (AVSyncPoint*) av_malloc(s->nb_streams * sizeof(AVSyncPoint));
if (!sync) {
// cannot allocate helper structure
return -1;
}
for (i = 0; i < s->nb_streams; ++i) {
st = s->streams[i];
sp = &sync[i];
sp->pos_lo = INT64_MAX;
sp->ts_lo = INT64_MAX;
sp->pos_hi = INT64_MAX;
sp->ts_hi = INT64_MAX;
sp->terminated = 0;
sp->first_ts = AV_NOPTS_VALUE;
sp->term_ts = ts_max;
sp->term_ts_tb = time_base;
sp->last_pos = pos;
st->cur_dts = AV_NOPTS_VALUE;
if (st->discard < AVDISCARD_ALL)
++keyframes_to_find;
}
if (keyframes_to_find == 0) {
// No stream active, error.
av_free(sync);
return -1;
}
// Find keyframes in all active streams with timestamp/position just before
// and just after requested timestamp/position.
step = 1024;
curpos = pos;
for (;;) {
url_fseek(s->pb, curpos, SEEK_SET);
search_hi_lo_keyframes(s,
ts, time_base,
flags,
sync,
keyframes_to_find,
&found_lo, &found_hi,
first_iter);
if (found_lo == keyframes_to_find && found_hi == keyframes_to_find)
break; // have all keyframes we wanted
if (curpos == 0)
break; // cannot go back anymore
curpos = pos - step;
if (curpos < 0)
curpos = 0;
step *= 2;
// switch termination positions
for (i = 0; i < s->nb_streams; ++i) {
st = s->streams[i];
st->cur_dts = AV_NOPTS_VALUE;
sp = &sync[i];
if (sp->first_ts != AV_NOPTS_VALUE) {
sp->term_ts = sp->first_ts;
sp->term_ts_tb = sp->first_ts_tb;
sp->first_ts = AV_NOPTS_VALUE;
}
sp->terminated = 0;
sp->last_pos = curpos;
}
first_iter = 0;
}
// Find actual position to start decoding so that decoder synchronizes
// closest to ts and between ts_min and ts_max.
pos = INT64_MAX;
for (i = 0; i < s->nb_streams; ++i) {
st = s->streams[i];
if (st->discard < AVDISCARD_ALL) {
sp = &sync[i];
min_distance = INT64_MAX;
// Find timestamp closest to requested timestamp within min/max limits.
if (sp->pos_lo != INT64_MAX
&& compare_ts(ts_min, time_base, sp->ts_lo, st->time_base) <= 0
&& compare_ts(sp->ts_lo, st->time_base, ts_max, time_base) <= 0) {
// low timestamp is in range
min_distance = ts_distance(ts, time_base, sp->ts_lo, st->time_base);
min_pos = sp->pos_lo;
}
if (sp->pos_hi != INT64_MAX
&& compare_ts(ts_min, time_base, sp->ts_hi, st->time_base) <= 0
&& compare_ts(sp->ts_hi, st->time_base, ts_max, time_base) <= 0) {
// high timestamp is in range, check distance
distance = ts_distance(sp->ts_hi, st->time_base, ts, time_base);
if (distance < min_distance) {
min_distance = distance;
min_pos = sp->pos_hi;
}
}
if (min_distance == INT64_MAX) {
// no timestamp is in range, cannot seek
av_free(sync);
return -1;
}
if (min_pos < pos)
pos = min_pos;
}
}
url_fseek(s->pb, pos, SEEK_SET);
av_free(sync);
return pos;
}
AVParserState *ff_store_parser_state(AVFormatContext *s)
{
int i;
AVStream *st;
AVStreamState *ss;
AVParserState *state = (AVParserState*) av_malloc(sizeof(AVParserState));
if (!state)
return NULL;
state->stream_states = (AVStreamState*) av_malloc(sizeof(AVStreamState) * s->nb_streams);
if (!state->stream_states) {
av_free(state);
return NULL;
}
state->fpos = url_ftell(s->pb);
// copy context structures
state->cur_st = s->cur_st;
state->packet_buffer = s->packet_buffer;
state->raw_packet_buffer = s->raw_packet_buffer;
state->raw_packet_buffer_remaining_size = s->raw_packet_buffer_remaining_size;
s->cur_st = NULL;
s->packet_buffer = NULL;
s->raw_packet_buffer = NULL;
s->raw_packet_buffer_remaining_size = RAW_PACKET_BUFFER_SIZE;
// copy stream structures
state->nb_streams = s->nb_streams;
for (i = 0; i < s->nb_streams; i++) {
st = s->streams[i];
ss = &state->stream_states[i];
ss->parser = st->parser;
ss->last_IP_pts = st->last_IP_pts;
ss->cur_dts = st->cur_dts;
ss->reference_dts = st->reference_dts;
ss->cur_ptr = st->cur_ptr;
ss->cur_len = st->cur_len;
ss->probe_packets = st->probe_packets;
ss->cur_pkt = st->cur_pkt;
st->parser = NULL;
st->last_IP_pts = AV_NOPTS_VALUE;
st->cur_dts = AV_NOPTS_VALUE;
st->reference_dts = AV_NOPTS_VALUE;
st->cur_ptr = NULL;
st->cur_len = 0;
st->probe_packets = MAX_PROBE_PACKETS;
av_init_packet(&st->cur_pkt);
}
return state;
}
void ff_restore_parser_state(AVFormatContext *s, AVParserState *state)
{
int i;
AVStream *st;
AVStreamState *ss;
av_read_frame_flush(s);
if (!state)
return;
url_fseek(s->pb, state->fpos, SEEK_SET);
// copy context structures
s->cur_st = state->cur_st;
s->packet_buffer = state->packet_buffer;
s->raw_packet_buffer = state->raw_packet_buffer;
s->raw_packet_buffer_remaining_size = state->raw_packet_buffer_remaining_size;
// copy stream structures
for (i = 0; i < state->nb_streams; i++) {
st = s->streams[i];
ss = &state->stream_states[i];
st->parser = ss->parser;
st->last_IP_pts = ss->last_IP_pts;
st->cur_dts = ss->cur_dts;
st->reference_dts = ss->reference_dts;
st->cur_ptr = ss->cur_ptr;
st->cur_len = ss->cur_len;
st->probe_packets = ss->probe_packets;
st->cur_pkt = ss->cur_pkt;
}
av_free(state->stream_states);
av_free(state);
}
static void free_packet_list(AVPacketList *pktl)
{
AVPacketList *cur;
while (pktl) {
cur = pktl;
pktl = cur->next;
av_free_packet(&cur->pkt);
av_free(cur);
}
}
void ff_free_parser_state(AVFormatContext *s, AVParserState *state)
{
int i;
AVStreamState *ss;
if (!state)
return;
for (i = 0; i < state->nb_streams; i++) {
ss = &state->stream_states[i];
if (ss->parser)
av_parser_close(ss->parser);
av_free_packet(&ss->cur_pkt);
}
free_packet_list(state->packet_buffer);
free_packet_list(state->raw_packet_buffer);
av_free(state->stream_states);
av_free(state);
}

@ -0,0 +1,97 @@
/*
* Utility functions for seeking for use within FFmpeg format handlers.
*
* Copyright (c) 2009 Ivan Schreter
*
* This file is part of FFmpeg.
*
* FFmpeg is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
* version 2.1 of the License, or (at your option) any later version.
*
* FFmpeg is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with FFmpeg; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
*/
#ifndef AVFORMAT_SEEK_H
#define AVFORMAT_SEEK_H
#include "avformat.h"
/// Opaque structure for parser state.
typedef struct AVParserState AVParserState;
/**
* Search for sync point of all active streams.
*
* This is not supposed to be called directly by a user application,
* but by demuxers.
*
* A sync point is a point in stream, so that decoding from this point,
* output of decoders of all streams synchronizes closest to given timestamp
* ts (but taking timestamp limits into account, i.e., no sooner than ts_min
* and no later than ts_max).
*
* @param stream_index stream index for time base reference of timestamps.
* @param pos approximate position where to start searching for key frames.
* @param min_ts minimum allowed timestamp (position, if AVSEEK_FLAG_BYTE set).
* @param ts target timestamp (or position, if AVSEEK_FLAG_BYTE set in flags).
* @param max_ts maximum allowed timestamp (position, if AVSEEK_FLAG_BYTE set).
* @param 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.
* @return < 0 if no such sync point could be found, otherwise stream position
* (stream is repositioned to this position).
*/
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);
/**
* Store current parser state and file position.
*
* This function can be used by demuxers before destructive seeking algorithm
* to store parser state. After the seek, depending on outcome, original state
* can be restored or new state kept and original state freed.
*
* @note As a side effect, original parser state is reset, since structures
* are relinked to stored state instead of being deeply-copied (for
* performance reasons and to keep code simple).
*
* @param s context from which to save state.
* @return parser state object or NULL if memory could not be allocated.
*/
AVParserState *ff_store_parser_state(AVFormatContext *s);
/**
* 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 stored state instead of being
* deeply-copied.
*
* @param s context to which to restore state (same as used for storing state).
* @param state state to restore.
*/
void ff_restore_parser_state(AVFormatContext *s, AVParserState *state);
/**
* Free previously saved parser state.
*
* @param s context to which the state belongs (same as used for storing state).
* @param state state to free.
*/
void ff_free_parser_state(AVFormatContext *s, AVParserState *state);
#endif /* AVFORMAT_SEEK_H */

@ -1140,7 +1140,7 @@ int av_find_default_stream_index(AVFormatContext *s)
/**
* Flush the frame reader.
*/
static void av_read_frame_flush(AVFormatContext *s)
void av_read_frame_flush(AVFormatContext *s)
{
AVStream *st;
int i;

Loading…
Cancel
Save