I want this, too. Computational photography seems like the next stage. Automated stacking, focus stacking, HDR, pixel shift. The potential gains in image quality from these technologies appear to be larger right now than what can easily be achieved with improvements in lenses or sensor resolution. Some of this computational stuff is already available on other cameras--and on phones, of course. It will come to Sony cameras, I would imagine. In the meantime, it would be great be able to get some of the same benefits in post-processing.
hiepphotog wrote:
We can do layer mask still in Photoshop, but automated would be convenient. How slow is the 4-shot merge in Image Edge?
Yes but first we have to find the problem areas manually and it takes a long time to find artifacts on the entire image. In a landscapes shot artifacts are everywhere. (Cloud, water, branches, leaves, lighting change, etc.)
...one of the reasons I never use pixel shift for landscapes.
Now, if the software is smart, that would be more useful for this application. (Sony can call it Pixel Shift AI Plug-in )
I don't use pixel shift for landscapes very often because of the motion artifacts, which are sometimes surprising. Even reflections in distant windows show changes in the sky. When I do use pixel shift, I copy one of the four images to a separate layer underneath the ps layer and use the eraser tool to paint out the artifacts. It works fine, but it means going through the image at 100 percent, screen by screen, and finding problem areas.
One of my oddball uses of pixel shift captures is to just use the four shifted images for regular stacking. This can substantially improve some landscapes, taking out noise, artifacts and wavery lines caused by haze. If I do decide to use the pixel shift function with the 4 files, I can process them that way. If not, I just stack the four images, enlarge, align, change opacities, etc.
vdo1 wrote:
I’ve told you. You won’t find rest until you get that Phase One. 150MP of real pixels. All in one single exposure. No subject motion issues. Can’t beat that .
Yes but first we have to find the problem areas manually and it takes a long time to find artifacts on the entire image. In a landscapes shot artifacts are everywhere. (Cloud, water, branches, leaves, lighting change, etc.)
...one of the reasons I never use pixel shift for landscapes.
Now, if the software is smart, that would be more useful for this application. (Sony can call it Pixel Shift AI Plug-in )
did you try the blend mode "difference" in photoshop to make spotting differences and artefacts easier ? you could also add a curve adjustment layer to brighten the differences with a clipping mask to make them even better visible ?
I only use pixelshift in studio settings so I have no files to do some tests but I think it should be possible to automate most of the process for Sony 4x in photoshop although a dedicated software would probably be better.
here are the steps I would try:
open the pixel shift image
place the unshifted image on top as a layer
set blend mode to difference
make a stamp to a new layer
make the layer monochrom
use the curve or levels tool on this new layer to increase contrast until the difference areas are totally clipped
apply the filter enlarge highlights, ( round ) experiment with the settings
add some gaussian blur - the highlights must stay clipped
go to channels an load a selection from this layer
invert thes election
make the help layer build the mask invisible
add the selection as a mask to the unshifted image layer
I just got an S1 (24 MP) and decided to do a pixel-shift comparison against my A7rII using the same Canon 50mm f/1.4 lens. Both cameras shot with equal exposures (Sony required -EV in post), timer-release, electronic shutter, lens @ f/5.6. Processed in LR with sharpening 45/0.7/35. The 200MP versions were upsampled in PS using Preserve Details 2.0. The S1 sample is the best of 3 manual-focus attempts.
)
))