Assume I have two 16 bit fixed point numbers in following format being multiplied.
Q3.13 * Q10.6
What is the correct way to multiply them, i.e do I right shift the first number to align it as Q10.6 or left shift the second number to align it as Q3.13? In both cases I will lose data.
May be the correct way is to put them into 32 bits and then shift them and then multiply them. In this case, no data is lost but result will be 64 bits after multiplication.
What Q format the result have after multiplying Q3.13 and Q10.6?