for determining the RPM why do you think you need to sample a signal from the ignition coil primary? The signal from one spark plug is sufficient. For determining the spark advance, the spark triggers a timer, when the timer expires, it fires the flash light(normally at top dead centre). The value of this timer is the spark advance value. Measuring the timing for each spark plug seems of little value - most engine analysers allow you to view the signal from each spark plug as this is used to determine the combustion characteristics in each cylinder. The cylinder to cylinder timing is normally fixed mechanically or electrically.

I would suggest you get your hands on a commercial timing light to see how they do this. These can vary from the very cheap to very expensive - the circuitry tends to be similar. Most of the ones I've seen and used have a split ferrite core that clamps around the spark lead and a number of windings for the pickup coil - a simple transformer. For the flash, it is usually a 'ringing choke' style of converter charging a capacitor with about 200-300 VDC and a SCR dumping a small capacitor's chanrge into a trigger transformer for a xenon flash tube.

It is normally easier to measure the period between the sparks rather than the frequency. Derivatives with a PCA timer may help here although you have a capture on timer2. Also remember that the angle is a fraction of the period time value.

I wouldn't waste time using optocouplers - I rarely see use of these in automotive electronics. Best to be careful in pcb layout and protection circuitry. Since the electronics are normally powered from the vehicle electrics, optos add little value.