the chip uses TDM to output the channels one by one
so we just run the chip 6 times and output at 1/6th the rate, which
means less work for blip_buf and less C++ overhead
so... it appears that pitch slides in Genesis system aren't truly linear
instead, they are "kind of linear". the frequency isn't linear, and when
it overflows the range of an octave, the high bit changes and the
frequency is shifted right
volume slide seems to be kind of quirky:
- sliding all the way down and then up suddenly sets the vol to max
- apparently there is an overflow and the slide stops when it reaches
its boundaries (instead of continuing)
- round to nearest instead of round to zero in SMS period calc
- implement arpeggio
- implement arp macro (kinda)
- polish the SMS platform
- correct the DIV_CMD_PITCH range to 1 semitone
- fix PSG volume in Genesis
- use a better register write strat in Genesis
- fix a bug caused by legacy code
- implement ECxx command
- implement EDxx command
- implement SN noise mode command
- vibrato table is now 64 positions long (instead of 60)
