Nikon Z8 Delivers Best Pixel Shift Out There... and botches the win by not doing the obvious
Related: Adobe, Adobe Camera Raw, botched design, computational photography, digital sensor, golden age of photography, imcompetence crisis, Nikon, Nikon mirrorless, Nikon NX Studio, Nikon Z8, pixel shift, please fix this, please fix this camera, please fix this Nikon, please fix this Sony, Sony, true-color sensor or image
re: pixel shift
re: Nikon Z8 and Nikon Df
re: please fix this camera
re: Frame Averaging: How it Should Work on Every Camera
Resolution matters, but it isn’t everything. While Sony A7R V easily beats the Nikon Z8 on resolution, the implementation is botched*, doubly. When and if Nikon delivers a 60+ MP camera, I’ll be a buyer unless Sony gets their act together.
See “Nikon Z8 Technical Guide (Pixel Shift Edition)”.
https://downloadcenter.nikonimglib.com/en/products/616/Z_8.html
Sony could learn a lot from Nikon*. Sony has the higher-res camera, but Nikon does it way better. Nikon Z8 users should be very pleased and with more than pixel shift—as far as I can tell Nikon is serious about delivering better value for cameras already sold even as Sony lets its users hang out to dry (failing to fix bugs or update firmware in any useful way). Seemingly, Sony’s market leadership has resulted in stagnation and laziness.
It’s not quite what I was thinking in Frame Averaging: How it Should Work on Every Camera, but if you have eight (8) pixel shift frames, eliminating problem ones with motion/lighting artifacts and frame averaging the remaining ones (even if there is some low-level artifacting in any one image) is WAY better than just one 4-shot capture.
Nikon has done a lot right:
- Extremely fast captures. That’s a BIG deal outdoors whenlight/scene is changing/moving—the faster the better.
- 4-shot mode is standard pixel shift. Sony has this, Fujifilm GFX 100S has it too. Artifacting issues arise unless there is no subject motion (or camera stability issue) and no lighting changes in the subject. Any changes, and pixel shift artifacts are poor to unusable. Sony’s motion correction software can address this in some cases, but often doesn’t work out.
- 8-shot mode gives you one frame-averaged result. Or two separate 4-shot results in case one has artifacting issues.
- 16-shot mode offers 2X linear resolution, or 4 separate 4-shot results in case one has artifacting issues. Should one of the 4-shot frames be a problem, you can also frame average 2 or 3 of 4 frames if desired.
- 32-shot mode... wow.
* Sony’s implementation forces a huge waste of space via uncompressed RAW, and has only 4-shot and 16-shot options. Chanbces of unusable results outside are so poor that the feature is useless there.
Continues below...
32-shot mode
The Nikon Z8/Nikon Df 32-shot mode shows how well Nikon thought this out. Clearly Nikon is well aware that pixel shift is subject to all sorts of artifacting issues in real photography.
32-shot mode gives numerous possibilities:
- One 32-shot NEFX consisting of two frame-averaged 16-shot double-res images, for the ultimate in low-noise equivalent of ISO 3.1.
- Two separate 16-shot double-res NEFX (in case one has artifacts), each equivalent ISO of 6.2.
- Four separate 8-shot standard-res NEFX, each equivalent to ISO 12.5.
- Eight 4-shot standard-res NEFX (in case one or more has artifacts). Probblem images can be discarded, and the best of 8 selected.
Botched design at the same time
Why do we have to use software to do all this?
Why can’t the camera directly generate NEFX files, with options for some or all variants and/or one or more originals, like picking the best flavor from a box of chocolates? Why make it so tedious and work intensive? I don’t care if the camera has to take 10 or even 20 seconds to do that! Because it will take far longer on the computer.
Furtermore, why can’t the camera tuto-detect areas of motion given 8 shots to choose from? And then produce a best capture result automagically?
