Difference between calculating psnr in uint8 and double.

I have implemented a data hiding algorithm(multilevel histogram modification and sequential recovery).I am getting psnr about 30 if calculated with respect to images in double and 10 psnr if calculated with uint8. Should I proceed with psnr in double or not?

Answer 1

[KSSV]

uint8 is used unsigned 8 bit integer. It is easy to convert double to uint8 or otherway.

let say I have matrix of uin8 A.

then to convert it to double use:

A = double(A);

and to convert it back to uint8 use:

A = uint8(A);

Answer 2

[Walter Roberson]

Psnr involves subtraction. But if you subtract a larger uint8 from a smaller you get 0 rather than a negative number. That makes it look like the locations were the same when they are not.