In case you want more:
- Since all pins M1,M2,M3,M4 are at PortD of the ATmega328P it is possible to toggle those pins at the same time (using the pin-toggle option by writing to the PINB register while the pin is set as output).
- If you don't need to do other things in your sketch, it is possible to turn off the interrupts and use the cpu clock cycles to delay. The delay will be in steps of 62.5 ns. I have used that some time ago, but I can not find it yet. Since toggling the output pins also takes 62.5ns, it is possible to calculate the delay to have 100% accurate timing.