Just treat it like an 8x8 sprite. That's the fastest way because nib{ is basically a combination of 8-bit storing and masking anyway.

If you are drawing a tilemap, it actually might be better to do something clever by adding values and only storing once to each byte, that way you are using 768 byte store commands instead of twice as many nibble storing commands.  You could do something like:

A*256 + B -> {L6+L}

where A is the nibble of the first sprite, B is the nibble of the second sprite, and L is your offset.  Does that make kinda sense?

If you are drawing a tilemap, it actually might be better to do something clever by adding values and only storing once to each byte, that way you are using 768 byte store commands instead of twice as many nibble storing commands.  You could do something like:

A*256 + B -> {L6+L}

where A is the nibble of the first sprite, B is the nibble of the second sprite, and L is your offset.  Does that make kinda sense?

I thought about that way but I couldn't figure out how because I forgot about multiplying to shift. Don't you multiply by 16 to shift it 4 bits though?