How can I apply a pixel math transform based on 16bit integer

[gvanhau]

Hello
I have a formula I want to apply to a image, but the formula is based on 16bit integer math.
for instance y=0.0104x + 2032.

Since images are converted to 0-1 scale in PixInsight, I'm a little confused.

Regards
Geert

[Carlos Milovic]

Just use:
y= (0.0104x + 2032)/65535

[mschuster]

Maybe x needs to be scale up to 16-bit?

(0.0104 * 65535 * $target + 2032) / 65535

[gvanhau]

Thank you Juan and Mike.

After posting y thought it could be   y=0.0104x + 2032/65536 , very simmilar to what Mike posted (if simplified).

Is it 65535 or 65536 (2**16)

Regards
Geert

[Juan Conejero]

Hi Geert,

The PM expression that you need is probably:

0.0104*$T + 2032/65535

since 0.0104 looks like a scaling factor, and 2032 is an additive pedestal expressed in pixel units.

It is 65535, since you are rescaling from [0,2¹⁶-1] to [0,1]. For a given number of bits n, 2ⁿ-1 is the largest n-bit unsigned integer.

[gvanhau]

Many thanks for the clarification Juan

Regards
Geert