Why? Because PixInsight basically does what Zero Noise does and generates a 32bit composite image - with the ability to save the segmentation map for each image it is combining. With normal blend mode, I applied the segmentation mask to each layer above the bottom one - THIS IS WHERE I CHEATED! I used PixInsight to generate the masks to blend each layer. I brought each linear tiff into PS as a layer, with the longest exposure on the bottom of the stack (nicely exposed shadows under the desk) and the shortest exposure on the top of the stack (halogen lamp shade and top of styrofoam head) - 4 layers in all.

I used draw, with -o 2 (AdobeRGB) to make my linear tiffs. I did all the combining in 32bit, linear space in PSCS6 (AdobeRGB linear profile). The images were shot at the following exposures:

I took 4 exposures of a scene that involves a halogen lamp, a desk, deep shadows under the desk, bright highlights in the lamp bulb and a styrofoam headform., etc. Here's a brief description of what I did - and I preface this by telling you straight up that I cheated on the hardest part, the segmentation maps. I just tried the whole process out and it works. These TIFFs would ultimately be combined in PS to yield a single 16bit TIFF. Where -T is for TIFF output and -4 is for linear 16bit output, -o 0 (with no color management) or 1 (sRGB) or 2 (AdobeRGB).

Sns hdr lite tutorial