Turns out I had connected rows and columns wrong (swapped really), so the lights were always on and stuck on the very first row. Doh...
After a night of frantic coding, I've managed to do something really weird.
With an input voltage of 12V, I get; 12V out. Despite duty cycling at 1:8.
I measured the voltage without anything connected, between P-channel output to N-channel input, it's rock solid at 12V when active. If I put a light there and measure the same way, it's around 1V since the bulb is stealing the rest of the voltage. The light turns on and off as I rotate the active light in software, so it's not constant on either. I am using pull-ups for the P-Channel MOSFET's, but it does feels a bit too good to be true. :)
|1) Chipkit driven by 12V wall wart, 12V out to P-channel TIP107's via the Vin-pin. The meter says 12V, but my senses says it couldn't really be. But it's bright enough to shine through a very dark smoked glass candle thingy.|