RepoMirrors
/
mpv
mirror of https://github.com/mpv-player/mpv
1
0
Fork 0
Code Issues Releases Wiki Activity

av_common: remove old/unused timestamp passthrough code 

This attempted to pass through double float timestamps in a bit exact
way by reinterpret casting them to int64_t. This usually worked, because
libavcodec (in decoding mode) is mostly not allowed to interpret
timestamps. libavcodec doesn't even know the unit of the timestamps,
unless the API user sets a timebase.

We've stopped doing this, and always set a timebase. Only ad_spdif.c
still used this (indirectly through mp_set_av_packet()), but doesn't
actually need timestamps on the packet. In fact, it's already explicitly
setting the packet timestamp fields to 0 before passing it to FFmpeg
API.

This code is unused, and the passthrough method wasn't terribly elegant
to begin with. Drop this code. Arbitrarily use AV_TIME_BASE_Q as
fallback in situations the passthrough was used.
This commit is contained in:
wm4 2017-01-24 08:36:40 +01:00
parent d16ba1f7f6
commit f762fc4230
1 changed files with 8 additions and 21 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

29
common/av_common.c
View File

@ -85,37 +85,24 @@ AVRational mp_get_codec_timebase(struct mp_codec_params *c)
return tb;
}
// We merely pass-through our PTS/DTS as an int64_t; libavcodec won't use it.
union pts { int64_t i; double d; };
static AVRational get_def_tb(AVRational *tb)
{
return tb && tb->num > 0 && tb->den > 0 ? *tb : AV_TIME_BASE_Q;
}
// Convert the mpv style timestamp (seconds as double) to a libavcodec style
// timestamp (integer units in a given timebase).
//
// If the given timebase is NULL or invalid, pass through the mpv timestamp by
// reinterpret casting them to int64_t. In this case, the timestamps will be
// non-sense for libavcodec, but we expect that it doesn't interpret them,
// and treats them as opaque.
int64_t mp_pts_to_av(double mp_pts, AVRational *tb)
{
assert(sizeof(int64_t) >= sizeof(double));
if (tb && tb->num > 0 && tb->den > 0) {
return mp_pts == MP_NOPTS_VALUE ?
AV_NOPTS_VALUE : llrint(mp_pts / av_q2d(*tb));
}
// The + 0.0 is to squash possible negative zero mp_pts, which would
// happen to end up as AV_NOPTS_VALUE.
return (union pts){.d = mp_pts + 0.0}.i;
AVRational b = get_def_tb(tb);
return mp_pts == MP_NOPTS_VALUE ? AV_NOPTS_VALUE : llrint(mp_pts / av_q2d(b));
}
// Inverse of mp_pts_to_av(). (The timebases must be exactly the same.)
double mp_pts_from_av(int64_t av_pts, AVRational *tb)
{
assert(sizeof(int64_t) >= sizeof(double));
if (tb && tb->num > 0 && tb->den > 0)
return av_pts == AV_NOPTS_VALUE ? MP_NOPTS_VALUE : av_pts * av_q2d(*tb);
// Should libavcodec set the PTS to AV_NOPTS_VALUE, it would end up as
// non-sense (usually negative zero) when unwrapped to double.
return av_pts == AV_NOPTS_VALUE ? MP_NOPTS_VALUE : (union pts){.i = av_pts}.d;
AVRational b = get_def_tb(tb);
return av_pts == AV_NOPTS_VALUE ? MP_NOPTS_VALUE : av_pts * av_q2d(b);
}
// Set dst from mpkt. Note that dst is not refcountable.