ISO and Highlight Protection

I think if (IF) pixel shift was a great alternative to HDR it would be a feature of a lot of cameras.
It seems like a way to take a kinda, sorta HDR image with one click of the shutter. I would choose to bracket shots in order to get the full range of the image and then combine them to get a complete picture.
 
I think if (IF) pixel shift was a great alternative to HDR it would be a feature of a lot of cameras.
It seems like a way to take a kinda, sorta HDR image with one click of the shutter. I would choose to bracket shots in order to get the full range of the image and then combine them to get a complete picture.

It's not intended to be an alternative to HDR. It exists for an entirely different purpose. It's designed to solve problems created by this:

BayerData.jpg


Joe
 
It's designed to solve problems created by this:

Yup. A 24 megapixel sensor has 12 million "Green", 6 million red, and 6 million blue. The filter above the sensor restricts the wavelengths reaching each pixel. No other color data beyond this is recorded by the sensor at the time of image capture.

I would choose to bracket shots in order to get the full range of the image and then combine them to get a complete picture.

In a standard HDR all those images you stack (as in a single image), the missing data in each is restored by a mathematical guess based on the data found in neighboring pixels. As Joe mentioned above the technical term for this process is demosaicing or interpolation, and it can be carried out by the camera or in RAW development software. There is unfortunately no optimal demosaicing algorithm, and so Bayer sensors always suffer from a slight loss of detail when reproducing an object, especially near edges. By microscopically shifting pixels around at the time of image capture (up, down, left, and right by one pixel relative to the default sensor position) each pixel gets fully exposed to all three color channels across a series of four images. This effectively quadruples the number of red/blue pixels and doubles the number of green pixels, increasing the overall level of recoverable detail in the final processed image. In other words each pixel of the sensor captures all three of the color channels as opposed to the need for demosaicing. Edges are crisper, there's more detail in the shadows, and (part of my original question), supposedly less susceptible to noise at higher ISO's.
 
Ok, tried some quick experimenting with pixel shift at high ISO. The first image was shot at 1/500, F 5.6 ISO 500.
IMGP0757TPF.jpg


The second image was shot at 1/250, F5.6, ISO 200

IMGP0759TPF.jpg


And the last was 1/60, F5.6, ISO 100

IMGP0760TPF.jpg


These are out of the camera, run through PSP, to crop and resize. To me the 200 has better edge detail, but the 400 has more detail in the highlights. The shadows have more noise at the higher ISO's, which was expected. I tried shooting all the way up to ISO 3200 but the noise started becoming significant over 400 and I didn't see any improvements in highlight details.
 
I agree the ISO200 one is best.

But I don't understand why the ISO 500 Anthers are so fuzzy / unsharp if distance and aperture were the same.

I'd test it again of PixelShift vs non-PS at the same settings.
 
why the ISO 500 Anthers are so fuzzy / unsharp if distance and aperture were the same

Camera was on a tripod never moved, aperature stayed the same. Only ISO and shutter changed. I'm pretty sure the shot was in focus, unless something got bumped while processing. I've found PS to be very sensitive to any kind of movement. Heading home so I have to get packed up but I'lly try non PS and another PSP at the same speed later this weekend.
 

Most reactions

Back
Top