image normalization

/255 is rather hard work

/256 is a lot easier as it is just a right bit shift by 8. In FPGA there is no need to do any bit shifting like this.

--- Quote Start ---

/255 is rather hard work

/256 is a lot easier as it is just a right bit shift by 8. In FPGA there is no need to do any bit shifting like this.

--- Quote End ---

But at the same time I want data bus of the same width(before normalization and after normalization).

Further, Is it possible to re-gain the same image on de-normalization?

Thank You Tricky

If all you are doing is /256 and want an 8 bit fractional, you actually do nothing at all. You just "pretend" you did a divide. In an FPGA there is no real bitshifting when dividing, you just append zeros depending on which way you shift.

So if you had 152, and then divide by 256, then multiply by 1.5 (answer = 228), you just do this:

input = 10011000 (152)

1.5 = 1.1

just append zeros to 1.5 (multiply by 256)

= 110000000

Do the multiply.

You have 8 bits fractional and 8 bits integer. (because your source, 152 was all fractional bits) so you discard the 8 LSBs and you have your answer

so you have:

                   0 1 0 0 1 1 0 0 0
x                  1 1 0 0 0 0 0 0 0
-------------------------------------
=    0 1 0 0 1 1 0 0 0
+  0 1 0 0 1 1 0 0 0
-------------------------------------
   0 1 1 1 0 0 1 0 0|0 0 0 0 0 0 0 0
   --Integer           Fraction
   
 11100100 = 228