diff options
| author | Mans Rullgard <mans@mansr.com> | 2012-05-10 12:19:10 +0100 |
|---|---|---|
| committer | Mans Rullgard <mans@mansr.com> | 2012-05-10 22:54:27 +0100 |
| commit | 4f1500689d662cfe08bff6d949f7fb49311024da (patch) | |
| tree | e2056d37f05397ad61d23414869aa58e3ba87bef | |
| parent | 110d0cdc9d1ec414a658f841a3fbefbf6f796d61 (diff) | |
| download | ffmpeg-4f1500689d662cfe08bff6d949f7fb49311024da.tar.gz | |
avconv: use lrint() for rounding double timestamps
Converting the double to float for lrintf() loses precision when
the value is not exactly representable as a single-precision float.
Apart from being inaccurate, this causes discrepancies in some
configurations due to differences in rounding.
Note that the changed timestamp in the vc1-ism test is a bogus,
made-up value.
Signed-off-by: Mans Rullgard <mans@mansr.com>
| -rw-r--r-- | avconv.c | 4 | ||||
| -rw-r--r-- | tests/ref/fate/vc1-ism | 2 |
2 files changed, 3 insertions, 3 deletions
@@ -1697,10 +1697,10 @@ static void do_video_out(AVFormatContext *s, if (delta <= -0.6) nb_frames = 0; else if (delta > 0.6) - ost->sync_opts = lrintf(sync_ipts); + ost->sync_opts = lrint(sync_ipts); break; case VSYNC_PASSTHROUGH: - ost->sync_opts = lrintf(sync_ipts); + ost->sync_opts = lrint(sync_ipts); break; default: av_assert0(0); diff --git a/tests/ref/fate/vc1-ism b/tests/ref/fate/vc1-ism index a9cfb2c5d5..a4d29cc5ca 100644 --- a/tests/ref/fate/vc1-ism +++ b/tests/ref/fate/vc1-ism @@ -118,4 +118,4 @@ 0, 48750000, 48750000, 0, 37440, 0xf0fe8c1c 0, 49170000, 49170000, 0, 37440, 0xc0036222 0, 49590000, 49590000, 0, 37440, 0x3058385c -0, 49798332, 49798332, 0, 37440, 0x68141016 +0, 49798330, 49798330, 0, 37440, 0x68141016 |