D500 Overexposure

Joined
Feb 22, 2005
Messages
4,377
Location
CT USA
Not sure if this has been asked yet, how are you determining over-exposure? My D500 lcd looks very light but the histogram still shows 1-1.5 stops under (at 0 comp on matrix metering for an average land/cityscape.) Just keep looking at the histogram and expose to the right!
 
Joined
Jan 15, 2010
Messages
3,889
Location
UK
That being said, I'm not sure I agree with you on the definition of overexposed in this case... As 1/3-2/3 stops difference is pretty small error most of the time it could be saved in post but what I meant is that the picture's could benefit color/contrast wise from little less light and fixing in post wouldn't necessarily give the same results.
I do not know the D500 but you may have missed the fact that o.3 - o.75 factor is only a small part of the exposure correction applied in raw editing (this being the baseline exposure calculation) the other part includes the brightening/contrast and black levels which is in the order of +1 EV. Therefore combined correction can be anywhere from 1.3 - 2 EV.
The definition of overexposed is fairly clear in my mind. Quite simply it is the point at which the sensor reaches saturation the maximum being a value of 16383 (16384 values total) with a 14 bit system - note your system may not reach this maximum.
Anything over these values will not be recorded therefore no texture/detail available and this is where your specular highlights fall
I would agree though that some of the pictures which seemed to be overexposed in the field turned out to be ok, so it might just be that I yet to get use to D500's superior dynamic range, I don't know.. Most pictures - even the correctly exposed ones seemed a whole lot brighter than Im used to somehow.
It is fair to say that if you have not made any changes to default settings ACR or LR and you are shooting raw and your images do not appear too bright on initial display and require brightness lowering then it is highly likely that you are under exposing (less than optimal exposure). This due to the fact that your raw converter applies hidden under the hood correction while leaving any controls at zero.

Your camera histogram is of very little help in determining the point of clipping for raw images using spot for instance it is easier to determine the point which is likely to be around +3EV over the area measured before clipping. This must be established by testing. Matrix although pretty good is not as intelligent as many seem to believe and AFAIK is not completely documented
 
Last edited:
Joined
Nov 11, 2005
Messages
3,185
Location
Westchester County, NY
I didn't notice anything special about exposure on my recently acquired D500 until yesterday. I was fooling around with the body shooting backyard birds in all different lighting situations to get a feel for what the camera will do. I found that I didn't notice anything significant in the exposures compared to my D3, D3S or D810 except in one set of shots. I was shooting a female cardinal against a darker than normal background and the subject bird was waaaay "over exposed" in the jpeg. Probably 2-3 stops. I played with the raw file and found that the details were indeed retained in the file and with some adjustments, was able to nicely correct the image.

I shot lots of similar shots around the backyard from various vantage points, but the series of this cardinal came out weird. Shot at 2500 ISO since I was testing out the abilities of the new body. There has been NO noise reduction applied to either by the way. Thought the camera did pretty well at 2500; noise in the background should be easily fixed with Dfine.

Jpeg SOC:
Subscribe to see EXIF info for this image (if available)

Same image cropped, exposure corrected and sharpened in PS> No noise reduction applied.
Subscribe to see EXIF info for this image (if available)
 
Joined
Jan 15, 2010
Messages
3,889
Location
UK
Mine behaves like every other Nikon I've owned
Which would seem to confirm that Nikon have not done anything different with the metering methods and are following ANSI/ISO standards for camera metering using their K factor of 12-13? with the D500. The only changes they may have made I guess would be on how matrix interprets areas of light and dark to apply exposure estimation?

I didn't notice anything special about exposure on my recently acquired D500 until yesterday. I was fooling around with the body shooting backyard birds in all different lighting situations to get a feel for what the camera will do. I found that I didn't notice anything significant in the exposures compared to my D3, D3S or D810 except in one set of shots. I was shooting a female cardinal against a darker than normal background and the subject bird was waaaay "over exposed" in the jpeg. Probably 2-3 stops. I played with the raw file and found that the details were indeed retained in the file and with some adjustments, was able to nicely correct the image.

I shot lots of similar shots around the backyard from various vantage points, but the series of this cardinal came out weird. Shot at 2500 ISO since I was testing out the abilities of the new body. There has been NO noise reduction applied to either by the way. Thought the camera did pretty well at 2500; noise in the background should be easily fixed with Dfine.
The two images IMHO are a good example of how a raw would appear when correctly exposed and presented in ACR. The JPEG mimicking the too bright view that you may get with a raw image in ACR prior to adjustment and the bottom image after correction.

A rough estimate would be metering the bird for JPEG and opening up by +2.5EV max to make sure no clipping - this would mean the raw image underexposed. For raw you may be able to open by 3+EV prior to clipping whereas using this for JPEG would certainly clip
 
Joined
Dec 21, 2007
Messages
135
Location
Israel
My point exactly(the thing with 0.75 actually being more than 0.75), even if you can fix it in post the overall quality would probably be degraded -
which means that getting the correct exposure out of the camera does matter even if details can be saved in post as there are negative consequences,
therefore in my eyes at least overexposure should also include the range of pictures which are salvageable.

What I understand from what you're saying is the that the matrix metering is biased towards exposing correctly the scene as a whole - which to be expected,
except that ignoring the subject to expose correctly the entire scene wasn't that obvious previously.
 
Joined
Nov 11, 2005
Messages
3,185
Location
Westchester County, NY
The two images IMHO are a good example of how a raw would appear when correctly exposed and presented in ACR. The JPEG mimicking the too bright view that you may get with a raw image in ACR prior to adjustment and the bottom image after correction.

A rough estimate would be metering the bird for JPEG and opening up by +2.5EV max to make sure no clipping - this would mean the raw image underexposed. For raw you may be able to open by 3+EV prior to clipping whereas using this for JPEG would certainly clip
Tony - I don't understand your point. Not disagreeing - just not understanding what you mean. You refer to metering for jpeg. Are you referring to the photographer metering for jpeg or do you mean that the algorithm in the camera meters for jpeg? I have no idea what "metering for jpeg" means. How would "metered for jpeg" and opening up +2.5 stops not clip the highlights? Can't figure out what you are describing and would appreciate your explaining it to me.
 
Joined
May 11, 2006
Messages
42,110
Location
CHARLOTTE
Real Name
Randy
I didn't notice anything special about exposure on my recently acquired D500 until yesterday. I was fooling around with the body shooting backyard birds in all different lighting situations to get a feel for what the camera will do. I found that I didn't notice anything significant in the exposures compared to my D3, D3S or D810 except in one set of shots. I was shooting a female cardinal against a darker than normal background and the subject bird was waaaay "over exposed" in the jpeg. Probably 2-3 stops. I played with the raw file and found that the details were indeed retained in the file and with some adjustments, was able to nicely correct the image.

I shot lots of similar shots around the backyard from various vantage points, but the series of this cardinal came out weird. Shot at 2500 ISO since I was testing out the abilities of the new body. There has been NO noise reduction applied to either by the way. Thought the camera did pretty well at 2500; noise in the background should be easily fixed with Dfine.

Jpeg SOC:
Subscribe to see EXIF info for this image (if available)

Same image cropped, exposure corrected and sharpened in PS> No noise reduction applied.
Subscribe to see EXIF info for this image (if available)
great save
 
Joined
Jan 15, 2010
Messages
3,889
Location
UK
Tony - I don't understand your point. Not disagreeing - just not understanding what you mean. You refer to metering for jpeg. Are you referring to the photographer metering for jpeg or do you mean that the algorithm in the camera meters for jpeg? I have no idea what "metering for jpeg" means. How would "metered for jpeg" and opening up +2.5 stops not clip the highlights? Can't figure out what you are describing and would appreciate your explaining it to me.
Sorry Rick probably a poor choice to quote the look of a bright looking JPEG resembling an optimally exposed raw in ACR with default processing applied.

What I was trying to get across is the fact that an optimally exposed JPEG will be an underexposed raw image and conversely an optimally exposed raw will be an overexposed JPEG probably exhibiting clipping.

As you know the camera meter is calibrated to a standard, let's assume that it is 18% (it will be lower than this - I think Nikon about 12.5). This means that we will have nearly +2.5 EV (actually 2.474) from metering an even toned subject before clipping occurs.

Therefore if we take a reading from the important highlight area that requires detail then open up by a maximum of up to 2.5 EV (in practice this would be 2 1/3 rd EV) then we should have the best exposure retaining detail in the important highlights and have lifted the shadow values as high as possible while still protecting those highlights. This is of course based on what you see on the camera LCD including the histogram and therefore is related to an sRGB 8 bit JPEG.

Your raw image will have more highlight headroom maybe over 3EV.

All this assumes ideally spot metering your ROI and adjusting exposure accordingly. Matrix metering applies its own algorithms of course and when looking at the scene evaluates quite differently although it is quite likely that it still tends to average to the manufacturers K factor
 
Joined
Nov 11, 2005
Messages
3,185
Location
Westchester County, NY
Thanks for lesson. I think I can grasp around 40% of what you have graciously explained to me, ,but considering I almost flunked out of physics 101 in college, I am not surprised. The one big take away on this for me is to never trash a photo based on what it looks like on the screen. Open in PS and make sure that it isn't salvageable in ACR.

Thank Tony.
 
Joined
Feb 18, 2015
Messages
1,107
Location
New York State
Tony's explanation is correct: there is a lot more headroom available if you are shooting RAW 14-bit files than if you are shooting 8-bit sRGB JPEGs.

The in-camera Histogram reflects ONLY the JPG Data which was captured by the camera using its in-camera processing controls (Picture Controls) in the way that you have configured them.

Conversely, the RAW file contains all of the Data which was captured in a non-determined but widest possible Color Space and the NEF is not affected by any in-camera Picture Controls (although the data does record the WB which was in use at the time.)

Nikon appears to use a 12% Grey for metering (instead of the traditional 18% Kodak Grey Card) and Matrix metering averages to 12% Grey and the Spot-metering exposes to record a 12% Grey. This default metering prevents the HLs from blowing-out in the restricted sRGB 8-bit used for JPGs but is much too over-cautious for NEFs.

If you build your own Camera Profiles (and you certainly should do that for every type of lighting under which you shoot), ACR can use the appropriate Profile.

If you edit in ACR, and choose to work in the ProColor RGB 16-bit Color Space, you will find that you have nearly a full EV of additional data in the HLs than the in-camera Histogram indicates.
(I don't use LR but it can be configured to work in a similar way to ACR.)

I find it worthwhile to program my meter-settings in the camera to match the realities of shooting in 14-bit RAW because I very very rarely shoot a JPEG and "exposing to the right" keeps my shadows further away from the left-wall of the Histogram where Noise always lurks.
 
Joined
May 5, 2005
Messages
21,669
Location
SW Virginia
Nikon appears to use a 12% Grey for metering (instead of the traditional 18% Kodak Grey Card) and Matrix metering averages to 12% Grey and the Spot-metering exposes to record a 12% Grey. This default metering prevents the HLs from blowing-out in the restricted sRGB 8-bit used for JPGs but is much too over-cautious for NEFs.
How many people here know what "12% Grey" actually means?
 
Joined
Feb 18, 2015
Messages
1,107
Location
New York State
>>>>>
How many people here know what "12% Grey" actually means?>>>>

In this context, it means that the Nikon's meters, when set to their default settings, will expose the targeted area to equal +4.66 EV brighter than total Black.

Kodak's 18% Grey Cards set the exposure for +5.33 EV brighter than total Black (or minimal recorded density on negative film).
 
Joined
Jan 15, 2010
Messages
3,889
Location
UK
How many people here know what "12% Grey" actually means?
Ok, a longer answer perhaps than you anticipated, but hopefully all still relevant. ;):)

Light meter calibration to 18% grey is a long standing myth i.e. the oft quoted 'exposure meters are calibrated to reproduce the density of an 18% grey card' is plain wrong.

K factor recommended by ANSI/ISO 2720-1974 is 10.6 to 13.4 (Previously ANSI PH3.49-1971) with an allowable error of plus or minus 1-2%. This is the standard that AFAIK all camera and light meter manufacturers comply with.

These standards mean that when we measure a so called 18% grey card or for that matter any even toned subject that the actual grey value will fall approximately 1/2 EV under an 18% value and requires the opening up by this amount to fall correctly on the exposure scale i.e. half way point. Without this correction a histogram is likely to show a spike left of centre (note depends on what your raw converter does with the data prior to showing rendered image)

Kodak published information for 18% card stating that increasing exposure and positioning the card that equates to generally increasing the exposure by 1/2 stop. In other words effectively adjusting the exposure to that of a 12.5% grey card. Unfortunately at some time in the past these instructions were omitted and caused some confusion. I understand that the information was eventually put back in the instructions.

The only published K factor ( K=reflective C= incident) for Nikon I have seen is K=12.5, which happens to be the same as Sekonic

Note that there is another myth that RGB values of 128 equal the mid density point in Photoshop. AFAIK there is no colour space in common use that equates to 128.

18% is a L*a*b* value of approx. 50,0,0. Equating to sRGB values of approx. 117 and for ProPhoto RGB = 100 (note figures rounded)

Nikon @ 12.5% would give a L*a*b* value of 42,0,0. Equal to sRGB values of 99 and ProPhoto = 80

IMO it is better to think of these values in terms of how much room we have between what the meter sees as 'mid grey' and the clipping point in our files for a very light object with texture. For 12.5% meters and for JPEG you will have approx +3EV before clipping and if your meter system really used 18% you would have approx +2.474 before clipping. Please note that it is important to realise that these figures refer to a theoretical JPEG image, for raw you really would need to test for sensor saturation point vs camera metering for mid tone
 
Last edited:
Joined
Nov 11, 2005
Messages
3,185
Location
Westchester County, NY
Regarding my posted photo earlier of the overexposed jpeg of the female cardinal that was saved by some PS work - I was just looking at some of the other images that I shot that same day and noticed that there were some interesting "overexposing" going on. In this photo (which did not look blown out in jpeg), I see that when I played with it in post, the result was a nicely rendered male cardinal, but with overexposed legs and overexposed wooden feeder. Didn't notice it at first, but when I get back to the office I will have a look at the originals and try to figure out what went on. If this were a photo I wanted to keep, I would do some burning on that feeder and legs.

Subscribe to see EXIF info for this image (if available)


Odd that I didn't notice the "overexposing" tendency before it was mentioned by the OP. I still have trouble understanding the physics as meticulously described by Tony, but now that I am aware of a potential issue, I will figure out a work around when shooting with the D500. Just need to run through some more practice sets until I get to know the body as well as I want to. I can expose with my D810 and D3 and D3s without thinking, but the D500 is new to me and I still have to get the feel of what it does behind the scenes.

Very interesting post, that is for sure.
 
Joined
Feb 18, 2015
Messages
1,107
Location
New York State
My terms of reference and experiences concerning the Kodak 18% Grey Cards predate the current K factor (as recommended by ANSI/ISO 2720-1974 at 10.6 to 13.4 )—(Previously ANSI PH3.49-1971) — by more than 20 years!

Prior to the 1970s, Kodak 18% Grey Cards were the standard reference that was used to determine exposure levels for Film Processing; and our exposure meters were calibrated, under the earlier ASA standards, to work in conjunction with those cards.

This was the way it was done when Ansel Adams promoted his Zone System in an attempt to make the subject more intelligible to inexperienced photographers.

Adam's "Zone 5" correlated with the 18% Grey Card reading and offered a safe exposure level for black and white photographs of subjects of normal ranges of contrast which were to be developed under normal conditions.
 
Joined
Sep 17, 2006
Messages
12,141
Location
Hamilton , New Zealand
For example:
1/13, f5.6, ISO 200 @ 11mm, on Aperture mode

1/20, f5.6, ISO 500 @ 11mm, on Aperture mode
3D matrix metering has to decide at some what what the subject is. Much of this is decided by the focus distance and where the active focus point is. In a high contrast scene like yours where the majority of the scene is in the shade it has to ignore blown highlights at some stage. Have a look at some tests I did with my D90 and 3D matrix metering. I don't see over-exposure in your 'subject' at the centre of the frame - if the camera exposed for the bright corners then the majority of the scene would be very dark indeed.
As per my blog in the link I provided the active focus point tells the camera which part of the scene gets priority and if the subject is somewhere around the 3-5m mark and the bright background is beyond that mark it will concentrate on the 3-5m part of the scene.
 

Latest posts

Links on this page may be to our affiliates. Sales through affiliate links may benefit this site.
Nikon Cafe is a fan site and not associated with Nikon Corporation.
Forum post reactions by Twemoji: https://github.com/twitter/twemoji
Copyright © 2005-2019 Amin Forums, LLC
Top Bottom