What is saturating arithmetic and what are its advantages and disadvantages in typical multimedia applications?

Question

What will be an ideal response?

Answer 1

In saturating arithmetic, operations that yield a value greater than the maximum are clamped at the maximum.
Consequently, 0xFFFF + 1 in 16?bit arithmetic is 0xFFFF and not the 0x0000 carry 1 you would get with
conventional arithmetic. The same applies to subtraction where the minimum value is 0 (if you are using signed
saturating arithmetic the minimum value is the most negative value). Clamping at the maximum and minimum
values mirrors (models well) real live multimedia situations. You can’t get whiter than white or blacker than
black.
Saturating arithmetic has the advantage that it models real?world physical processes by clamping at a maximum
or minimum value. You don’t need saturating arithmetic – you could implement it in software (which would
slow down multimedia operations). There is no disadvantage of saturating arithmetic, but it would be inappropriate for conventional arithmetic operations that rely on a carry?out to indicate overflow (or use
chained arithmetic).