From ad41f64ed930e177399d7a9106640f3284559db1 Mon Sep 17 00:00:00 2001
From: Matthias Grundmann <grundmann.matthias@gmail.com>
Date: Fri, 14 Apr 2017 11:32:16 -0700
Subject: [PATCH] Fix rounding mode inconsistency

double calculation is using rounding, vs. int using truncation.
---
 src/mp4util.cpp | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/src/mp4util.cpp b/src/mp4util.cpp
index 47bd74e..932b22d 100644
--- a/src/mp4util.cpp
+++ b/src/mp4util.cpp
@@ -259,7 +259,8 @@ uint64_t MP4ConvertTime(uint64_t t,
 
     // check if we can safely use integer operations
     if (ilog2(t) + ilog2(newTimeScale) <= 64) {
-        return (t * newTimeScale) / oldTimeScale;
+      // convert with rounding.
+      return (t * newTimeScale + oldTimeScale / 2) / oldTimeScale;
     }
 
     // final resort is to use floating point