You could manually replicate the concept. You basically need to develop a threshold map to segment your [three] images into the areas you want to use and the areas you want to reject. Recall that Zero Noise did this by representing each acceptable area as a grayscale area in the map. How you chose to do this is up to you.
The images you ultimately combine need to be linear so that they add with no gamma adjustment baked in. If you recall, Zero Noise used dcraw to extract linear RGB TIFF images from the raws you fed it. You could do this as well, with:
dcraw -T -4 -o [your choice here]...
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). These TIFFs would ultimately be combined in PS to yield a single 16bit TIFF.
I just tried the whole process out and it works. 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 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. and a laptop display. The images were shot at the following exposures:
1/2s, 1/8s, 1/60s and 1/1000s.
at f/4 and ISO 800 on a 5DII.
I did all the combining in 32bit, linear space in PSCS6 (AdobeRGB linear profile). I used draw, with -o 2 (AdobeRGB) to make my linear tiffs. 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. 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. 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. I saved the maps and brought them into PS to mask each layer for the composite. Clearly, this is the most difficult part, in terms of getting a clean composite image.
After you get each layer with its mask set up, you have to scale each image to the darkest (shortest) exposure. When using a linear space, it is as simple as applying an Exposure adjustment layer to each image that is not the shortest exposure, and dialing in negative exposure to get the exposure to match the shortest (darkest) exposure. So, for example:
My longest exposure (1/2s) was 9EV above the reference (shortest - 1/1000s), so for that image I would apply an exposure adjustment layer with -9 in the exposure box - you have to make sure that you apply the Exposure adjustment layer CLIPPED to the image you are adjusting so that it only affects that image and not everything in the layer stack under it. My other exposures were 7EV (1/8s) and 4EV (1/60s) above reference, so I dialed in -7EV and -4EV in their respective clipped exposure adjustment layers.
Now all the layers are nice and dark and blend. Because I am in 32bit mode, I can use the 32bit preview slider to examine the blend at the various intensity levels contained within the image.
Anyway, attached is a small image of the test scene - I (very quickly without aesthetic concern!) tonemapped this using the output from what I just described above - also attached is a whacked out curve exposing the area underneath the desk, with no noise (other than that you would expect for a properly exposed ISO800 image from a 5DII). There is massive JPEG artifact so I could fit it to less than 150kb and the masking needs to be cleaned up a little, but consider this a proof of concept.
Or you could just use PixInsight.
kirk
I will put together a step-by-step to give you an idea of what I did, and start a new thread.
Image hosted by forum (
644715)
© kirkt [SHARE LINK] THIS IS A LOW QUALITY PREVIEW. Please log in to see the good quality stuff. Image hosted by forum (
644716)
© kirkt [SHARE LINK] THIS IS A LOW QUALITY PREVIEW. Please log in to see the good quality stuff. 
