Paddington and Manusse,
First of all, thank you for developing such a great tool!
I am using SC 2.2.2 beta. When SC reads and writes .idx/.sub file, it introduces a progressive delay into the timestamps in the .sub file. The problem doesn't not affect the .idx file, so it cannot be detected using software players, most of which use the .idx file as a reference.
However, this is a big problem for stand alone players, many of which read the timing info from .sub file directly.
The bug is very easy to see, get any .idx/.sub pair, rename the .idx file to a different name, open the .sub directly using SC (there is another bug in the interface that does not allow you to select a .sub file, you have to input the .sub file name manually). At the same time, open the original .idx file in a text editor. Now compare the timestamps displayed in SC with the timestamps displayed in the text editor, you will see the timestamps in SC have a progressive delay compared to original ones.
The weird thing is that if you save this .sub file, a corresponding .idx file will be generated automatically, and that .idx file have the correct timestamps.
I think this suggests SC not only has a bug in reading the .sub file, but also has a bug in writing the .idx file. The combination of these two bugs gives a correct result accidentally.
The bug has also been reported by another user here: