Result of the test: that is now perfect
SomeJoe,
As you said, eac3to should ideally produce the correct values. This discussion might give the idea to madshi to add an option to apply a coef 1001/1000 to chapter timestamps ?
If not, that means a little tool should be written to apply this correction, taking one OGG chapter file as input and producing a corrected OGG chapter file as output. If this tool does not exist, I will develop one myself, that should not be very hard to do.