Nikon’s new mirrorless system, coming in ... late September 2018

[davidgp]

By the way... Jim has already starting dissecting the camera: https://blog.kasson.com/nikon-z6-7/nikon-z7-edr-vs-iso/ , I enjoy reading his articles although I confess he goes many times in more depth than I can follow...


http://dgpfotografia.com

[Guillermo Luijk]

Where did you get that image from? This looks like an overprocessed image of the image (JPG) published on DPR.

The original JPG is top right for reference. I applied a contrast curve to enhance the visibility of the stripes, which are already visible without any curve in the original image.

The point is not how much over processing was needed, but if stripes are visible where we could still expect to find valid information on a D850 for instance. If this is the case, the Z7 would have less usable DR than the D850, no matter if their SNR curves are similar (pattern 'noise' has a much greater effect on human vision than on statistical noise measurements).

Regards

[Guillermo Luijk]

Now I understand why I was unable to make any good images for 40 years with reversal film.  I could never underexpose by 5 stops and then print the area as neutral gray. Better start ripping the prints off the wall.

This is heard 100 times and it's 100 times wrong. Lifting the shadows by 5 stops doesn't mean underexposing anything. A high contrast scene must be exposed to preserve the highligts. Done that, you'll lift only the dark areas of the scene to make information visible on them. This is what DR in a digital sensor is about.

Regards

Enviado desde mi PRA-LX1 mediante Tapatalk

[BernardLanguillier]

This is heard 100 times and it's 100 times wrong. Lifting the shadows by 5 stops doesn't mean underexposing anything. A high contrast scene must be exposed to preserve the highligts. Done that, you'll lift only the dark areas of the scene to make information visible on them. This is what DR in a digital sensor is about.

Yes, this is indeed true, but so far the only case of slight banding I have been able to reproduce is an underexposed photo of a black object (a Profoto soft box grid - softbox shut down) at ISO 4000 after a +100% shadow lift with +2 stop exposure increase in LR. And even so, it is pretty subtle.

I am as concerned about DR as anybody else, but I don't feel that this issue has real world implications per my current findings.

Cheers,
Bernard

[32BT]

The original JPG is top right for reference. I applied a contrast curve to enhance the visibility of the stripes, which are already visible without any curve in the original image.

The point is not how much over processing was needed, but if stripes are visible where we could still expect to find valid information on a D850 for instance. If this is the case, the Z7 would have less usable DR than the D850, no matter if their SNR curves are similar (pattern 'noise' has a much greater effect on human vision than on statistical noise measurements).

Regards

I'm not a Nikon fanboy by any means, never owned a Nikon. Was tempted by that 85 nikkor though. But i'm going to repeat this ad nauseum: it's a raw converter problem, not a camera problem. All the FFM offerings have a pattern problem of some kind due the pdaf pixels which apparently need to be compensated for.

In that respect i find "unusable" a bit of an exegaration. It would be perfectly usuable, if the raw conversion did what it's supposed to do.

[gkroeger]

This is heard 100 times and it's 100 times wrong. Lifting the shadows by 5 stops doesn't mean underexposing anything. A high contrast scene must be exposed to preserve the highligts. Done that, you'll lift only the dark areas of the scene to make information visible on them. This is what DR in a digital sensor is about.

Regards

Enviado desde mi PRA-LX1 mediante Tapatalk

I expose for highlights and lift the shadows all the time... but raising the minimally exposed pixels to midtones is alway going to produce noisy midtones.  The only question is whether the noise will have a repeatable pattern or random pattern. I do agree that this makes the DR of the Z7 a bit less than the D850. But I also disagree that there is any meaningful "detail" in the regions where the banding appears... perhaps some subtle tonal variations.

[32BT]

Technically, this gives the Z more dynamic range, not less...

[Guillermo Luijk]

But I also disagree that there is any meaningful "detail" in the regions where the banding appears... perhaps some subtle tonal variations.

I didn't say that, I was just wondering myself. The Dpreview JPG file is unfortunately a bad example to judge because the strips appear in soft out of focus areas, so it's difficult to guess that if on those areas we had had fine detail, a D850 would have been capable of recovering it over (regular, not stripped) noise or not.

Perhaps Bernard can do some D850 vs Z7 comparison on a controlled high contrast scene.

Regards

[BernardLanguillier]

Perhaps Bernard can do some D850 vs Z7 comparison on a controlled high contrast scene.

Yes, I could try to do that.

But before that I need to understand what conditions generate this banding, because I don't see any in some images.

Cheers,
Bernard

[armand]

Yes, I could try to do that.

But before that I need to understand what conditions generate this banding, because I don't see any in some images.

Cheers,
Bernard

Do you see it in something like this: sunset, expose for highlights, lift shadows at 100, lift exposure by 1-2 stops and add some contrast and clarity?

On the comparison on dpreview I see the shadows don't take that color shift towards magenta that all my other cameras do if I push the shadows enough, Oly being the worst but both the X-Ts from Fuji and the D750 do it.

[BernardLanguillier]

Do you see it in something like this: sunset, expose for highlights, lift shadows at 100, lift exposure by 1-2 stops and add some contrast and clarity?

On the comparison on dpreview I see the shadows don't take that color shift towards magenta that all my other cameras do if I push the shadows enough, Oly being the worst but both the X-Ts from Fuji and the D750 do it.

I don't yet have such outdoor samples to check, but I will.

Cheers,
Bernard

[Bo_Dez]

I have seen this sort of banding in other cameras. As I understand these times in the past it has related to not enough shielding of electronics. So it could well be something that can be resolved.

[chez]

This is heard 100 times and it's 100 times wrong. Lifting the shadows by 5 stops doesn't mean underexposing anything. A high contrast scene must be exposed to preserve the highligts. Done that, you'll lift only the dark areas of the scene to make information visible on them. This is what DR in a digital sensor is about.

Regards

Enviado desde mi PRA-LX1 mediante Tapatalk

Exactly. Right now on Kauai shooting sunsets with dark lava rock foregrounds. I expose for the sun and lift the dark foreground in post. Will I lift 5 stops...not always, but I'm glad it's available to me if the scene requires it.

Used to always shoot same scenes with a 3 stop ND filter and still lift some in post...but that had issues with uneven foregrounds which clashed with the straight line of the GND filter.

I think many people that poke fun at lifting exposure just don't shoot under conditions that require this...in other words...speaking from ignorance.

[cgarnerhome]

Wouldn't be the first time someone speaks from ignorance!  So far so good with my Z7 but personally I never open shadows that much so not an issue for me.  If would just shoot a couple of exposures and blend them in PS - albeit it's a little more work.

[armand]

Some updates regarding the performance of the 24-70, the pictures make it easier to understand the magnitude of the problem: https://photographylife.com/nikon-z-24-70mm-f-4-s-performance-overview

The vignetting is a little too much, particularly at 35 but something once can deal with.
The sharpness drop is quite obvious but from memory they are likely similar to my 24-120 F4, just starting later. I was hoping for it to be much better than my 24-120, particularly is it doesn't really improve that much with stopping down but I guess there is no free lunch. The comment that it can be an excellent lens if you use a 1.1x crop is interesting.
I'll play with some raws and see if I get any banding.

[Dan Wells]

Even those "weak corners" are marked "the triangles of what you'd expect from a zoom" in the review. They're tiny pieces of the corner that are weak by the standards of the rest of the field. Those corners are equivalent to something like a Nikkor 24-120 - they stand out because everythingexcept the corners is in a different class...

[armand]

I played with few raws from dpreview. Initially as I would normally process them than I pushed them beyond what I would usually do. Most were at iso 64, one at 6400. I obtained many unnatural things but no banding.

[32BT]

Apparently their oversized mount still wasn't large enough...! 

[BernardLanguillier]

Even those "weak corners" are marked "the triangles of what you'd expect from a zoom" in the review. They're tiny pieces of the corner that are weak by the standards of the rest of the field. Those corners are equivalent to something like a Nikkor 24-120 - they stand out because everythingexcept the corners is in a different class...

That summarizes it pretty well. The corners are not weak in absolute terms, they are merely good.

Cheers,
Bernard

[SrMi]

That summarizes it pretty well. The corners are not weak in absolute terms, they are merely good.

Cheers,
Bernard

Let's wait for more tests. In the article mentioned, the tester may have used a polarizer, which may or may not have affected the corners. Hopefully, the tester did not use the EXIF hack (change Z7 -> D850) otherwise built-in profiles are not applied (crops the image slightly).