Assignment of chapters to cores could be made a little more optimal

[pfh-personal]

When processing an AAX with chapter splitting on a system with a large number of cores, the "leftover" tracks all get assigned to the last core.

For example, if a book with 47 chapters is processed on an 8-core system, the first 40 tracks will be divided evenly among all 8 cores, and then the remaining 7 will be given to the last core (as 7 doesn't divide by 8). As a result, the first 7 cores get 5 tracks, while the last gets 12!

It would be more optimal to split the remaining 7 tracks unevenly among all cores, so that the total processing time would be roughly 6 tracks rather than 12.