I'm using LibRaw (through FreeImage, though that could change if necessary) to get the pixel data from raw images which then have effects applied to them, e.g. exposure, white balance.

This works well, except if I understand correctly the pixel values are scaled using black and white levels to the range 0-65535 with values outside the range effectively lost. Is it possible with LibRaw to keep the values between the black and white levels?

What I want to do is end up with the values as floating point numbers, where the black point is 0 and the white point is 1, but allow values outside this range (for recovering highlights).