i want to make a game using voice as main way to play it. In game the player will be given a word to say, how do I calculate the percentage similarity of pronunciation in the dataset that I have and the pronunciation of the player ? what's the lightest and best methode ?