LCDC timings need reworking
Here are the problems as follows:
- LY counting is incorrect/off
- We do not fire STAT interrupts at the correct times (needs to be on the edge of a transition to a new mode, not at the start of the new mode)
#12 (closed) #15 (closed) #21 (closed) and the graphical glitches related to #19 (closed) are likely duplicates of this. If there's no objections I'm marking them dupes.