Activating a LED slightly illuminates other LED's in the row.

[1-max-1]

Only a minor issue. This can be seen best when activating a single red LED. For example:

matrix.setPixel(3, 3, 31, 0, 0);

Turns on the correct LED. However, looking closely, all other red LED's on row 3 are just barely illuminated, when they should be completely off.

If I remove certain columns from the display scan cycle, the LED's in those columns no longer illuminate, as expected. This means that as we scan through the columns, row 3 is still "active" somehow, very briefly providing a path to ground even through the shift register should have deactivated row 3. Not sure why.

Possibly some electrical parasitic aspects, as the PCB design is pretty poor. Maybe crosstalk or self inductance is keeping the current flowing in row 3 when it shouldn't be. Saving this for later as I want to use an oscilloscope to see what's going on.