Hi,
-"a synthesized voice would simply say "Pulltab", "Nickel", "Iron", "Dime, clad", "Dime, silver", "Not sure" and so on" thats only for the US market, you would need approximately 160 software versions to cover the world market.
-A lot of countries have changed their currencies and coins more times than you can keep track of. E.g. Great Britain had: celtic (a lot of different), roman ( alot of different), sub-roman (..), medival coins (..) and so on. You cant cover this imho.
- A tone gives you a lot of addtional info. If you wiggle the coil and receive short, sharp snappy ding-ding: Ahhh that could a small, regular sized object like a coin and a button.
-Relative orientation (lying flat in ground or on the rom) can change the identfication.
-You will dig a lot of pulltabs giving good signals, a lot of junk lead indicating silver and a lot of tin/zinc particles sounding off as copper/brass/bronze. It is not excact science like math.
I always like it to a tennis racket: As beginner buy a good beginner racket and learn hot to play tennis. If you like the sport buy a better racket. As sidenote due to aforementioned learning curve you play for sometime worse with the better racket...
Greets Namxat