Moiré, Z6 + 70-200mm/f2.8 S

Stopping down to f/8 may not be sufficient.

 

Definitely I could be wrong about this. It's largely a guess.

 

Here's how I came up with f/8. I found the pixel pitch for this specific camera, then estimated the "blur circle" diameter needed to prevent aliasing. ("Blur circle" being the diameter of the smallest detail point.) Since Shun rarely (?) has encountered moire I presume that the camera setup, with blur filter (aka AA filter) is very near to "good enough" with respect to preventing moire. So if the "blur circle" can be very slightly enlarged it should be possible to eliminate very nearly ALL moire. I estimate that stopping down from f/2.8 to f/8 will increase the blur circle diameter by about 20%, and am comletely GUESSING that this will be enough. Alternatively, going all the way to f/11 would increase the blur circle diameter by about 33%, which makes things even more favorable. (Note that these are not general rules - they're specific to a 6 micron pixel pitch and a nearly "good enough" blur filter.)

 

In the real world I'd probably bet a beer at the local tavern that f/8 is good enough. Meaning that I think it's better than a 50:50 likelihood. Going down to f/11 is a significantly better likelihood. And so on. But Shun would probably be the one that would have to determine this, using the same test subject.

That is not a practical solution. I certainly didn't pay for an f2.8 70-200 and carry that weight in order to use it at f8 and f11, which is way too slow indoors.

 

In the studio f/11 is pretty normal. I typically set my lights for f/11 for head and shoulders but for full-body images often f/6.3 to f/8. I use the 24-70/2.8 VR for the full-body images and it's just not that sharp to cause problems. ;-)

 

An older lens such as the 70-200/2.8G VR (1st version) would probably not produce moire at f/2.8. Other lenses that could be tried include the 58/1.4 AF-S Nikkor, the 85mm f/1.4D and the DC Nikkors, or adding a soft-focus filter to the Z 70-200. Use of film would be another way, how about that Contax. :)

 

People often want perfect per pixel sharpness and no artifacts but that's not within the realm of realistic possibility. As one pushes for high contrast near Nyquist frequency invariably some detail has higher frequency than Nyquist and even in the presence of an optical low-pass filter, it can leak through if the contrast is very high at those frequencies.

Edited December 20, 2021 by ilkka_nissila

That is not a practical solution. I certainly didn't pay for an f2.8 70-200 and carry that weight in order to use it at f8 and f11, which is way too slow indoors.

 

Well, your other main option is find out exactly what magnification of a specific fabric causes the problem, and then avoid that. I'm not saying that you have to sit down and calculate things out with different lenses. Rather, just go from your experience with how much something fills the frame. With your problem shirt, it happens when the guy's body is filling roughly 1/2 to 2/3 of the horizontal frame, so just avoid shooting that subject in that range. Head and shoulder shots, or 3/4 body (and longer) shots are probably OK. If you need in-between, crop in on the wider shot.

 

Of course, you are still subject to problems when another fabric shows up. It just kinda depends on what you're willing to deal with. As I said earlier I once spent time screening potential cameras for a studio chain outfit. Virtually every shot had some sort of fabric in it. So I started out screening for moire. If any camera showed moire like this it was immediately out of the running; we just didn't wanna deal with it. (There were cameras available that were completely immune to the Oxford pin-point fabric moire.) If you do a different sort of work, well, you probably weigh your decisions differently.

Here's a simple solution: Exclude that guy or give him minimum coverage unless he puts a jacket over his moire-prone shirt. ;) He would look more professional anyway. Explain nicely and I believe any reasonable person would agree.

 

:p:D:eek:

Edited December 20, 2021 by Mary Doo

An older lens such as the 70-200/2.8G VR (1st version) would probably not produce moire at f/2.8.

 

I'm not a Nikon guy, and can't say for sure. But I'd certainly expect it WOULD have moire problems. Earlier you were thinking that that the original (very high-quality?) lens might still have moire issues at f/8 (even with its diffraction). So it looks like you are suggesting that a (presumably) lesser quality lens at f/2.8 cannot produce a smaller spot of focused light than the f/8 diffraction limit.

 

I wouldn't argue against the older lens having less contrast and lower resolution out towards the edges of the frame, but... failure to deliver smaller spot sizes near the center of the frame than an f/8 diffraction-limited lens? I have a real hard time believing that.

I'm not a Nikon guy, and can't say for sure. But I'd certainly expect it WOULD have moire problems. Earlier you were thinking that that the original (very high-quality?) lens might still have moire issues at f/8 (even with its diffraction). So it looks like you are suggesting that a (presumably) lesser quality lens at f/2.8 cannot produce a smaller spot of focused light than the f/8 diffraction limit.

 

The 70-200/2.8 first-generation F-mount lens is quite soft at f/2.8 and this is what I base my assessment on. I don't have that lens any more. It was markedly soft at 200 mm at f/2.8 on 10 MP DX which is similar to the resolution of the 24 MP sensor in the Z6 (in units of pixels per mm). At shorter focal lengths the softness may be less obvious though, but it always looked more like a beautiful portrait lens than what is available today, the kind of ultra high-contrast high-resolution lens like the FL versions.

Remember I pointed out that my blue and gold macaw image from 2007 was captured with the first version of 70-200mm/f2.8 AF-S VR? Using that lens doesn't always prevent moiré either. Actually I still own that lens. Its main issue is softness around the edges in the FX frame at 200mm. The center is fine, and back in 2007 I was using a D2X, which is a DX body.

 

My current moiré issue is from our company Christmas party; it was not a studio session. I wasn't even thinking about moiré at all, since the last time I had such issues was some 14 years ago in 2007. Unless I become a regular studio, party, or wedding photographer, I am not going to worry about this issue.

Thanks for the update Shun.

I think I'd try a burst whilst gently moving away or towards the target in AF-C.

 

Once aligned and stacked, you'd end up being able to erase through the moire areas to a unmoired bit beneath....:p

It's impossible to avoid moiré completely as long as rasterizing is somewhere part of the proces. Reducing resolution will change the spatial frequency at which it will happen. But given a world full op patterns and nothing that prevents thise patterns appearing in thd critical soatial frequency, moiré is just waiting to happen.

When we notice that it does, we can take steps to avoid or minimize it, in that situation. And if we care about it not showing in that situation, we should. No guarantee that what we do then will prevent moiré in a different situation.

 

All moiré is, after all, is a coïncidence of spatial frequencies of two rasters/grids. One of them is in our camera (or printer - elliminate, or do not create, on-sensor moiré, and it may still appear in print). The other appears in front if your lens.

@ say, 1/250th 'each', I guess the 4 shots, via mirrorless is what, 1/60th?

Not with Sony's implementation. That takes a couple of seconds, regardless of shutter speed. And needs a rock-steady tripod to keep any shake well within 1 pixel distance. That's <3.8 microns of camera movement!

 

FWIW, I recently read about some clever software that interpolates sub-pixel data from a RAW file to get simulated co-sited colour and heightened resolution. Might be worth investigating. I'm intrigued enough to want to find out more.

Edited December 20, 2021 by rodeo_joe|1

Why not just take the moire out in post processing? In Lightroom (at least in Lightroom Version 6) in the Develop module select the adjustment brush which has a moire slider. Martin Evening recommends using a hard brush and auto-mask. According to Mr. Evening, it works best on RAW files.

 

I have never needed to use it, but I have found that Martin Evenings advice is usually valid.

 

Granted, it is best not to have moire, if it can be eliminated without adversely affecting the other elements of the image.

Moire and aliasing in general cannot be removed in post in such a way that you get the original subject's image like it would have been photographed with a correctly designed process (which includes a sensor Nyquist (for 2x2 diagonal) > lens resolution so that aliasing doesn't occur), the information to recover the correct subject detail is simply lost and not recorded in the image. Edited December 21, 2021 by ilkka_nissila