D7100 'streaking'

Started Apr 7, 2013 | Discussions
Jack Hogan Veteran Member • Posts: 6,507
Re: ISO 100 vs 200 equiv exposure

Horshack wrote:

Jack Hogan wrote:

fotolopithecus wrote:

Horshack wrote:

My original D7100 shots above were 14-bit lossy. Here are converted raws from 14-bit lossless, this time at both ISO 100 and 400. I don't see it at ISO 400.

I think I can see it in both shots Adam, but only just in the second shot.

Horshack, you changed exposure (1/320 vs 1/1250), that would make the ISO 400 image noisier, and the streaking could get lost in the dithering.  I can just barely see some horizontal banding in the ISO 400 image against the texture of the wall, but I am not sure it is aligned with the blinds.

ISO 100 pushed 5 stops (left), ISO 200 pushed 4 stops (right)

No perceivable streaking there, with more light (1/200).

krikman Regular Member • Posts: 418
Re: ISO 100 vs 200 equiv exposure

I wasn't enough accurate with my statements.

Now with better monitor i see:

1. Stripes aren't totally disappear at ISO800, so processing is same.

2. stripes go both sides beside lighter area, so there aren't 'lost bits'.

3. ADL was OFF from beginning.

4. 12 or 14 bit gives same.

5. will see another camera, i.e. is it camera specific?

Thanks to all involved.

_sem_ Veteran Member • Posts: 5,022
Re: D7100 'streaking' - doesn't look like blooming or flare

krikman wrote:

My whole point was in-camera ADL work. I can strongly conclude from my D300s behaviour that in-camera ADL works selective. It not just compresses data range but lengthen exposition for darks and somehow restores lights. At levels that impossible for written RAW data. All this required some 'area recognition' for dark and light regions.

Don't let the impression fool you. ADL is a black-box magic packaged thing but is no black magic. Several raw converters and HDR tonemapping programs let you do similar rendering, although not exactly the same, mainly because ADL does not disclose which elementary operations (fill-light, brightness, contrast, local contrast...) it applies and to which degree - there seems to be a sort of a scene-recognition indeed that sets the values, and of course it is totally undocumented. But the main operation is certainly the "fill-light" (shadows protection slider in NX), which was once done via exposure mask (USM) in Photoshop before it got automated.

Notice LR4/ACR allow one to boost shadows much more than VNX2 with ADL.

krikman Regular Member • Posts: 418
Re: D7100 'streaking' - doesn't look like blooming or flare

_sem_ wrote:

krikman wrote:

My whole point was in-camera ADL work. I can strongly conclude from my D300s behaviour that in-camera ADL works selective. It not just compresses data range but lengthen exposition for darks and somehow restores lights. At levels that impossible for written RAW data. All this required some 'area recognition' for dark and light regions.

Don't let the impression fool you. ADL is a black-box magic packaged thing but is no black magic. Several raw converters and HDR tonemapping programs let you do similar rendering, although not exactly the same, mainly because ADL does not disclose which elementary operations (fill-light, brightness, contrast, local contrast...) it applies and to which degree - there seems to be a sort of a scene-recognition indeed that sets the values, and of course it is totally undocumented. But the main operation is certainly the "fill-light" (shadows protection slider in NX), which was once done via exposure mask (USM) in Photoshop before it got automated.

Notice LR4/ACR allow one to boost shadows much more than VNX2 with ADL.

In my tests on D300s ADL max makes 2x longer exposition with better highligths and potent to recover. I feel magic here...

dougster1979 Regular Member • Posts: 106
Re: "I love how it forces me to become a better photographer"

Sure i understand what your saying. But the D7100 has a much improved focussing system.  Far superior to the d40, d70,d90, or d7000 even. My original post wasn`t a criticism , i was questioning weather better equipment makes a better photographer. Otherwise we would always want the "better" camera. Better resolution yes, better sharpness yes. Does it make one a better photographer, most definably not.

_sem_ Veteran Member • Posts: 5,022
Re: D7100 'streaking' - doesn't look like blooming or flare

krikman wrote:

_sem_ wrote:

krikman wrote:

My whole point was in-camera ADL work. I can strongly conclude from my D300s behaviour that in-camera ADL works selective. It not just compresses data range but lengthen exposition for darks and somehow restores lights. At levels that impossible for written RAW data. All this required some 'area recognition' for dark and light regions.

Don't let the impression fool you. ADL is a black-box magic packaged thing but is no black magic. Several raw converters and HDR tonemapping programs let you do similar rendering, although not exactly the same, mainly because ADL does not disclose which elementary operations (fill-light, brightness, contrast, local contrast...) it applies and to which degree - there seems to be a sort of a scene-recognition indeed that sets the values, and of course it is totally undocumented. But the main operation is certainly the "fill-light" (shadows protection slider in NX), which was once done via exposure mask (USM) in Photoshop before it got automated.

Notice LR4/ACR allow one to boost shadows much more than VNX2 with ADL.

In my tests on D300s ADL max makes 2x longer exposition with better highligths and potent to recover. I feel magic here...

I guess you were not careful enough. ADL typically exposes raw data up to about a stop lower than non-ADL (but this is not evident in the in-camera and CNX/VNX previews). This may indeed result in a stop less blown highlights if you'd blow them with the original setting. But this is assuming that you don't watch the histogram and adjust exposure as to not blow the highlights. If you do, you're likely to end up with captures made at about 2 stops less than what is recoverable from raw with highlights maxed to the right (in RawDigger raw histogram), which results in the lifted shadows being noisier than they'd need to be. With ADL, the image indeed looks brighter, despite the raw data being exposed lower, due to the processing applied instantly, and the same processing is also applied upon opening in VNX2/CNX2. Some 3rd-party converters may show the lower exposure - but some may normalize exposure (Picasa) or apply a complex rendering preset (Adobe).

Get the first the free VNX and Raw Therapee, then trials of CNX, LR4, DxO OP (well, you don't need all at once). Play with the sliders and watch the histograms, and eventually you'll get the impression what is going on. There is one interesting exercise: open images with/without ADL in CNX2/VNX2, drag the exposure slider and watch the histogram. Then repeat the exercise with ADL replaced by the fill-light. Mind you can only change the ADL level (the post-processing part of it) in CNX2 and only if it was on while shooting. VNX1.5 used to discard ADL if the Exposure Compensation or maybe some other sliders were moved; I hear this is no longer the case with VNX2.

mosswings Veteran Member • Posts: 8,705
Re: Another set @ 14-bit lossless compressed

Jack Hogan wrote:

You know what I find interesting is that in your new ISO100 NEF, the streaking seems to be more apparent next to the bottom half of the window, where there is less light coming through.  Perhaps the streaking is not a consequence of the bright areas, but of the dark ones.

I had been thinking that as well - in Mattilla's shots, the dark bands were noticeably shifted towards the green. Horshack noted this behavior in his original shadow pushing tests; a consequence of black clipping?

 mosswings's gear list:mosswings's gear list
Olympus XZ-1 Olympus Stylus 1 Nikon D90 Nikon D7100 Nikon AF-S DX Nikkor 18-105mm f/3.5-5.6G ED VR +5 more
rhlpetrus Forum Pro • Posts: 24,923
It's all Nikon/Sony's fault

A few years ago nobody would be pushing shadows 4-5 stops ...

Anyway, keep up th egood owrk, it's good to know everything about the new sensor. Nikon and Toshiba will certainly need to look into banding and this unrelated issue, which may be lens-related, since I doubt one would have this type of sensor "blooming".

-- hide signature --

Renato.
http://www.flickr.com/photos/rhlpedrosa/
OnExposure member
http://www.onexposure.net/
Good shooting and good luck
(after Ed Murrow)

 rhlpetrus's gear list:rhlpetrus's gear list
Leica D-Lux (Typ 109) Nikon D7000 Nikon 1 V1 Nikon AF-S DX Nikkor 12-24mm f/4G ED-IF Nikon AF Nikkor 35mm f/2D +4 more
fotolopithecus Senior Member • Posts: 1,699
Re: Another set @ 14-bit lossless compressed

Jack Hogan wrote:

You know what I find interesting is that in your new ISO100 NEF, the streaking seems to be more apparent next to the bottom half of the window, where there is less light coming through.  Perhaps the streaking is not a consequence of the bright areas, but of the dark ones.

I was starting to wonder the same thing. could this not be more banding?

 fotolopithecus's gear list:fotolopithecus's gear list
Nikon D610 Tamron SP AF 90mm F/2.8 Di Macro Nikon AF-S Nikkor 50mm f/1.8G
mosswings Veteran Member • Posts: 8,705
Re: It's all Nikon/Sony's fault

rhlpetrus wrote:

A few years ago nobody would be pushing shadows 4-5 stops ...

Anyway, keep up th egood owrk, it's good to know everything about the new sensor. Nikon and Toshiba will certainly need to look into banding and this unrelated issue, which may be lens-related, since I doubt one would have this type of sensor "blooming".

Which begs the question wanting to be asked:

Of what importance is this to practical picture taking, and how does the EXMOR sensor technology fare in similar situations?

The general opinion amongst many D7000/5100 owners appears to be that their sensing chain has certain warts in low level color accuracy and resolution, but in general the EXMOR technology appears to limit pretty gracefully. By contrast it seems the Toshiba sensor stumbles more...or we are intent on demonstrating that it does.

Certainly the deeper question underlying these investigations is why Nikon felt it necessary to move away from such a robust technology in the first place.  We'll never know the answer to that. These discoveries make me want to grumble, but also make me realize that I haven't sufficient perspective on their significance to justify that grumbling.

The best thing we can do is quantify these behaviors and relegate them to the deep shadows (so to speak) whence they came and get on with taking pictures...

 mosswings's gear list:mosswings's gear list
Olympus XZ-1 Olympus Stylus 1 Nikon D90 Nikon D7100 Nikon AF-S DX Nikkor 18-105mm f/3.5-5.6G ED VR +5 more
mosswings Veteran Member • Posts: 8,705
Re: Another set @ 14-bit lossless compressed

fotolopithecus wrote:

Jack Hogan wrote:

You know what I find interesting is that in your new ISO100 NEF, the streaking seems to be more apparent next to the bottom half of the window, where there is less light coming through.  Perhaps the streaking is not a consequence of the bright areas, but of the dark ones.

I was starting to wonder the same thing. could this not be more banding?

There may be some mechanisms shared, but this streaking is scene geometry dependent.  Banding isn't, and it has a much finer pattern.

However, it is interesting that it shows up at about the same RAW levels as banding does, and requires the roughly the same level of pulling to observe.

 mosswings's gear list:mosswings's gear list
Olympus XZ-1 Olympus Stylus 1 Nikon D90 Nikon D7100 Nikon AF-S DX Nikkor 18-105mm f/3.5-5.6G ED VR +5 more
krikman Regular Member • Posts: 418
Re: D7100 'streaking' - doesn't look like blooming or flare
I guess you were not careful enough. ADL typically exposes raw data up to about a stop lower than non-ADL (but this is not evident in the in-camera and CNX/VNX previews).
This may indeed result in a stop less blown highlights if you'd blow them with the original setting. But this is assuming that you don't watch the histogram and adjust exposure as to not blow the highlights. If you do, you're likely to end up with captures made at about 2 stops less than what is recoverable from raw with highlights maxed to the right (in RawDigger raw histogram), which results in the lifted shadows being noisier than they'd need to be. With ADL, the image indeed looks brighter, despite the raw data being exposed lower, due to the processing applied instantly, and the same processing is also applied upon opening in VNX2/CNX2. Some 3rd-party converters may show the lower exposure - but some may normalize exposure (Picasa) or apply a complex rendering preset (Adobe).

Yep. I've been wrong. It is truly decreases exposure time and push shadows up.

I've make measurments 3 years ago and forgot details. But since ADL processing takes additional time and slows shot sequense I've got wrong impression since these.

Horshack Veteran Member • Posts: 6,168
Streaking gets worse with lower exposure

Jack Hogan wrote:

You know what I find interesting is that in your new ISO100 NEF, the streaking seems to be more apparent next to the bottom half of the window, where there is less light coming through.  Perhaps the streaking is not a consequence of the bright areas, but of the dark ones.

Here are two exposures, 1EV apart, shot at both ISO 100 and 200. The lower exposure is much worse, indicating it is something with the shadow+channel balancing/black clipping for rows which also have saturated pixels

Base Exposure, ISO 100

-1EV Exposure, ISO 100

Base Exposure, ISO 200

-1EV Exposure, ISO 200

Mako2011
MOD Mako2011 Forum Pro • Posts: 24,585
not an issue

I just checked with the Nikon staff at the NAB Nikon booth in Vegas. They confirm that your conclusions regards what's going on regards the streaking example here are pretty far off base. Unfortunately, they won't give me any D400 hints. Sorry.

-- hide signature --

My opinions are my own and not those of DPR or its administration. They carry no 'special' value (except to me and Lacie of course)

mosswings Veteran Member • Posts: 8,705
Re: not an issue

Mako2011 wrote:

I just checked with the Nikon staff at the NAB Nikon booth in Vegas. They confirm that your conclusions regards what's going on regards the streaking example here are pretty far off base. Unfortunately, they won't give me any D400 hints. Sorry.

-- hide signature --

My opinions are my own and not those of DPR or its administration. They carry no 'special' value (except to me and Lacie of course)

That's cool, thanks.

But perhaps the more salient question would be why is this happening in the first place. Do you think they might hazard a guess?

 mosswings's gear list:mosswings's gear list
Olympus XZ-1 Olympus Stylus 1 Nikon D90 Nikon D7100 Nikon AF-S DX Nikkor 18-105mm f/3.5-5.6G ED VR +5 more
JimPearce
JimPearce Veteran Member • Posts: 9,188
Special pleadings or hot-house flowers...
1

Banding, streaking and left autofocus point alignment are being conjured up by specialized testing far outside of the realm of normal picture taking with a DX camera, even at the professional level.

-- hide signature --

Jim

 JimPearce's gear list:JimPearce's gear list
Nikon D7100 Nikon D500
mosswings Veteran Member • Posts: 8,705
Re: Streaking gets worse with lower exposure

Wow, those color shifts at ISO 100 are pretty visible.

 mosswings's gear list:mosswings's gear list
Olympus XZ-1 Olympus Stylus 1 Nikon D90 Nikon D7100 Nikon AF-S DX Nikkor 18-105mm f/3.5-5.6G ED VR +5 more
_sem_ Veteran Member • Posts: 5,022
Re: Streaking gets worse with lower exposure
1

Horshack wrote:

Here are two exposures, 1EV apart, shot at both ISO 100 and 200. The lower exposure is much worse, indicating it is something with the shadow+channel balancing/black clipping for rows which also have saturated pixels

I think in the higher exposure there's streaking in both the top and the bottom half, with, just a few lines left out below. But in the lower exposure, there is full streaking save a few lines above, and much leass streaking below, therefore the "contrast".

>> I just checked with the Nikon staff at the NAB Nikon booth in Vegas. They confirm that your conclusions regards what's going on regards the streaking example here are pretty far off base.

But perhaps the more salient question would be why is this happening in the first place. Do you think they might hazard a guess?

That'd mean admitting the problem, and I wouldn't thing salespeople would be authorized for this. Unless Nikon already had a fix for it

Banding, streaking and left autofocus point alignment are being conjured up by specialized testing far outside of the realm of normal picture taking with a DX camera, even at the professional level.

I'm not so sure about this. Banding is indeed weak, but the streaking might become disturbing in practice. Wide DR has been one of the things keeping folks from switching a lighter OM-D set, for those who like simple HDR-like processing in LR4 without having to bracket exposures... not everyones cup of tea but one of Nikon DX advantages.

As for the left AF, I still find it almost too bizarre to be true... who needs that extra resolution if the camera doesn't nail focus when the subject is on the left

JimPearce
JimPearce Veteran Member • Posts: 9,188
I'll try to be nice here...

There's a world of difference between using a test as a diagnostic tool where there is a problem, and using the test to "discover" a problem. My D300, D300s and D7100 all nail focus on the left in my normal practice - which is 300mm to 700mm at f3.5 to f7.1. I don't have an issue with using this test on a D800 where shooting wide and fast is part of the rationale for the camera - as long as we know that the lens being used is itself well-behaved. But I have a real problem with testing the D7100 at f1.8 with a 35mm f1.8. And who knows how a D3, D300, D700 etc. - all of the previous cameras with a 51 point autofocus - would fare on this test?

-- hide signature --

Jim

 JimPearce's gear list:JimPearce's gear list
Nikon D7100 Nikon D500
fotolopithecus Senior Member • Posts: 1,699
Re: Special pleadings or hot-house flowers...

JimPearce wrote:

Banding, streaking and left autofocus point alignment are being conjured up by specialized testing far outside of the realm of normal picture taking with a DX camera, even at the professional level.

-- hide signature --

Jim

It's become part of the interest in digital photography to compare, and try to figure out things like, was Nikons move from Sony, to Toshiba wise, unwise, or a wash. I can well imagine that it's no fun for early adopters of the D7100, but don't forget the blizzard of BS the D7000 went through as well. You're right that no one was pushing like this before D7000 though, but that was one of it's strengths for a DX anyway.

 fotolopithecus's gear list:fotolopithecus's gear list
Nikon D610 Tamron SP AF 90mm F/2.8 Di Macro Nikon AF-S Nikkor 50mm f/1.8G
Keyboard shortcuts:
FForum MMy threads