Skip navigation
Currently Being Moderated

Nikon D3s RAW files are underexposed by Lightroom / Camera RAW

Jun 5, 2012 2:02 PM

Tags: #lightroom #raw #camera #jpg #nef #nikon #d3s #d700 #underexposure #underexposed

I'm used to shot RAW files only.

I recently bought a D3s, alongside with my D700 and I noticed D3s files were darker than D700 ones, by about 1/2 or even 2/3 stop. Initially I thought the problem was one of the cameras, but after some tests I found the problem is my RAW converter. Infact this happens only with NEF (RAW) files, and if I shot a RAW+JPG image, the D3s RAW file opened in Lightroom 4 (or Camera RAW) is underexposed if compared with its JPG on-camera version.

I think no camera setting is affecting this behaviour. I reset either the D3s and the D700. No D-Lighting is set on-camera. This happens only with the D3s and only with Camera RAW and Lightroom.

Other RAW converters display the D3s RAW file correctly exposed, the same as its JPG version, and the same as the D700 RAW and JPG. I tested so far Capture One, DxO Pro Optics and Raw Therapee.

Color profile used is custom, but I tried using Adobe Standard and the result is the same.

Default develop settings affect only color profile and sharpening, and I got the same problem on a fresh new installation of Lightroom where no default develop settings are set.

 

If anybody with a D3s can check if his NEFs are underexposed if compared to their JPG versions when opened in LR or Camera RAW, I would appreciate that much.

 

Any idea about the origins of my problem would be great.

 

Thank you.

 
Replies
  • Currently Being Moderated
    Jun 5, 2012 2:19 PM   in reply to alessandroavenali

    Do you have Active D-Lighting turned on in your camera?

     
    |
    Mark as:
  • Currently Being Moderated
    Jun 5, 2012 3:52 PM   in reply to alessandroavenali

    I noticed the D3s supports "in-camera raw processing." Check the in-camera menus to see if there is a setting causing the problem you're seeing in LR. I believe these settings affect the "in-camera" created JPEG, so perhaps there is one set for Exposure compensation +2/3 EV making the raw file look darker compared to the JPEG file.

     

    If so this would indicate you are probably shooting your pictures underexposed.

     
    |
    Mark as:
  • Currently Being Moderated
    Jun 12, 2012 2:53 PM   in reply to alessandroavenali

    I recently bought a D3s and noticed the same thing comparing it to my D3.  Images from the D3s are not only a 1/2 stop darker than the D3, they are also much greener, and since the D3 files look more correct, I'm going to say that they are dark and have a color cast.  The histogram even shifts significantly to the left. 

     

    Both cameras have all settings the same, and both cameras were set to the same custom WB setting (to make sure Nikon didn't change the "cloudy" WB parameters).  All other settings were the same.  I just returned the D3s to BH because I thought something must be wrong with it.  But today I downloaded Nikon ViewNX2, and the images look exactly the same using Nikon's software.  In fact, images from the D3s were ever so slightly brighter, only noticeable in the histogram.  But for all real world purpsoes, they looked the same. 

     

    So the problem lies in Adobe's RAW conversion software.  They ought to be more precise than this, the images look way different.  I understand they are different cameras, but I expect more accuracy or at least more precision from Adobe.  I can post my samples if anyone is interested.  I shot against a gray wall, so it's easy to see the color and exposure change. 

     
    |
    Mark as:
  • Currently Being Moderated
    Jun 12, 2012 4:29 PM   in reply to rhythmdevils

    What do the D3s NEFs look like in LR4.1 final realease version using Process Version 2010 with its default settings?

     
    |
    Mark as:
  • Currently Being Moderated
    Jun 12, 2012 4:40 PM   in reply to trshaner

    I do not have LR4 yet.  Not sure I'll get it, i'm not sure it's worth 200 dollars for better recovery and fill. 

     

    I'd love to hear from someone who has LR4 and both cameras. 

     
    |
    Mark as:
  • Currently Being Moderated
    Jun 12, 2012 5:37 PM   in reply to alessandroavenali

    @Original post.

     

     

    You say " Color profile used is custom".?  cfould you explain

    When you import the raw files without any develop settings do the Basic panel settings look like this (see screenshot)

    Image00106.jpg

    If not are you applying "Auto tone adjustments on import" ?

     

    Could you post a screenshot of your basic panel for us to see what is happening?

    Have you tried the "Camera" profiles in the Calibration Panel, see the screenshot.

     

    Image00107.jpg

     
    |
    Mark as:
  • Currently Being Moderated
    Jun 12, 2012 5:51 PM   in reply to DdeGannes

    I'm not sure if you're talking to me or the original poster, but I have no custom camera calibration set (mine looks identical to yours except that I'm using LR3 and therefore RAW conversion 2010).  The only "auto tone adustments" I've set are to leave "black clipping" at 0 instead of the standard +5 on import (which baffles me why that would be standard, who wants their blacks clipping in all their images?)

     

    But beyond that, I imported the images from both cameras using the exact same settings, at the same time. 

     
    |
    Mark as:
  • Currently Being Moderated
    Jun 12, 2012 5:54 PM   in reply to rhythmdevils

    I am talking to the Original Post.

     
    |
    Mark as:
  • Currently Being Moderated
    Jun 12, 2012 5:57 PM   in reply to DdeGannes

    I'm not sure why you want to ask him specifically.  I have noticed the same exact thing, and do not have any custom camera calibration settings. 

     
    |
    Mark as:
  • Currently Being Moderated
    Jun 12, 2012 6:15 PM   in reply to rhythmdevils

    The original poster is talking about LR V4 and you stated that you have not upgraded. So you are both talking about something quite different. Also because the OP has given limited info it is possible that he may be applying some adjustments other than the system defaults.

    He talks about a custom profile? and is saying that Lightroom is not making a proper rendition of the raw data.

     
    |
    Mark as:
  • Currently Being Moderated
    Jun 12, 2012 6:24 PM   in reply to DdeGannes

    We're not talking about something "quite different" we are describing the exact same thing by comparing the same cameras (D700 has same sensor as D3).  Either we're doing something wrong, or Adobe is making the same mistake with both LR3 and LR4. 

     

    I'll put some sample images on my idisk so we can see if other people notice the same thing with their LR setups. 

     
    |
    Mark as:
  • Currently Being Moderated
    Jun 12, 2012 6:49 PM   in reply to rhythmdevils

    I uploaded 2 sample images onto my website, one from the D3 and one from the D3s.  Let me know if it doesn't work, I can host them elsewhere. 

     

    http://www.whitneydafoe.com/d3vsd3s/

     

    Would love to hear from some  people whether the images from the D3s come out darker and greener than the D3 in LR3 or LR4

     
    |
    Mark as:
  • Currently Being Moderated
    Jun 12, 2012 6:51 PM   in reply to rhythmdevils

    All raw conversion software is different, and will provide different renditions of the raw data. A raw file is not a viewable file and requires processing either on the LCD screen of your camera or on your computer screen. Your camera firmware firmware uses their own propritory process to provide this on your camera lcd screen and to save this process as a jpeg/tiff file to the card. Some cameras permit the capture of the raw data to the card also.

     

    Your camera manufacturer provides you with software that allows you to use the same process to render the raw data and provide you with jpeg/tiff files on your computer. Their intention is to provide you with an appealing rendition of the capture and not neccessarily an accurate or realistic one.

     

    Other third party raw conversion providers; Adobe Photoshop Lightroom/ACR, Capture One, Silkypix, DxO and many others provide you with an alternative means (recipes) to process the raw data. Their objective is to enable you to be creative and produce a superior rendition for your use. Why would you pay extra for a process that would provide you with what you already have for free. Process the raw data with the camera's software and use Adobe Photoshop, Paintshop Pro etc for further processing.

    Have a look at the following renditions from my camera an Olympus E510 of a test file using a number of raw processing software I use.

    http://www.flickr.com/photos/baxter43/sets/72157625678052039/

     
    |
    Mark as:
  • Currently Being Moderated
    Jun 12, 2012 7:13 PM   in reply to alessandroavenali

    OK, It looks as though things are back to reality.

     

    My thinking is there are differences but the tools are there for you to be creative, when you make the choice on your workflow and software tools do not keep going back and forth to different processes. For me I am convinced that LR 4 process version 2012 is the superior process and PV 2010 is history.

     

    That said I am not a professional and time and cost are not important to me, so I can tinker with different software, but if I needed to make a choice then make a clean cut with what works best for you.

     

    When I revisit old files I will make necessary adjustments to my likeling.

     
    |
    Mark as:
  • Currently Being Moderated
    Jun 12, 2012 7:21 PM   in reply to alessandroavenali

    My thinking would be to capture the scene with the best exposure, lighting and composure possible and then the rest can be none in post processing.

     
    |
    Mark as:
  • Currently Being Moderated
    Jun 12, 2012 7:47 PM   in reply to DdeGannes

    You seem to be missing the point DdeGannes.  Why are the results so different in LR between 2 top of the line Nikon cameras?  I would think their RAW conversion would be more precise than that.  I returned my D3s because I thought there was no way they could be off by a whole 1/2 stop plus a green color cast. 

     

    Unless we're doing something wrong.  Would anyome mind opening up the images I hosted in LR to see what results they get?

     
    |
    Mark as:
  • Currently Being Moderated
    Jun 13, 2012 2:10 AM   in reply to rhythmdevils

    I've had a quick look, using LR4. On import, using normal Adobe Standard default settings, there IS a noticeable difference, the D3s is darker and greener. However, if I apply some of the Camera Calibration profiles (such as Camera Standard, Camera Neutral, Camera Vivid, etc), both images then look more or less identical....so you could conclude from that that one of the two default Adobe Standard conversions is not as accurate as the other.

     

    Interestingly, using the camera calibration profiles adjusts the D3 image more towards the D3s initial conversion than vice versa.....implication of that is that it's the D3 Adobe Standard conversion that's off, not the D3s.

     

    But I wasn't there, so I don't know how accurate the camera calibration profiles are....all I can say is they produce the same results.

     
    |
    Mark as:
  • Currently Being Moderated
    Jun 13, 2012 6:02 AM   in reply to alessandroavenali

    alessandroavenali wrote:

     

    I know, DdeGannes, but a different rendition of 2 pro cameras at the same settings is a bit frustrating.

    Would it be better to underexpose the D700 while shooting, or setting a default -0,5EV on every imported picture, if the problem goes on?

    Moreover, seems the Nikon D3s ISOs need to be pushed +2/3 to obtain the same exposure as the D700 in LR workflow. So the advantage of this great sensor could be lost.

    I'd like to know more about D700 and D3s rendition comparison in LR/ACR at same settings.

     

    It's not unusual for two camera bodies of even the same exact make and model to have Exposure differences of up to 1/3 EV, which is very noticeable. Ditto for  White Balance differences. For Example:

     

    Camera 1 is +1/3 EV and Camera 2 is -1/3 EV = 2/3 EV Difference

    Camera 1 White Balance is -500KTemp and Camera 2 is +500K Temp = 1000K Difference

     

    Tack on minor differences between the LR Camera Profiles and it will get much worse if heading in opposite directions, as above.

     

    The only way to compare the raw NEF files from two cameras is to use a RAW converter that applies NO profiles, NO Gamma Correction, or any other modifications to the raw pixel data. See may post here for information on 'Rawnalyze:'

     

    http://forums.adobe.com/message/3952812#3952812

     

    You can download it here:

     

    http://dave-anderson-photo.com/files/software/Rawnalyze/Rawnalyze_2.10 .4.0.zip

     

    Online manual here:

     

    http://www.oitregor.com/numeric/Rawnalyze/Rawnalyse_doc/datas/Rawnalyz eGuide.htm

     

    None of the external links in the manual work, but nothing to worry about. If you have any questions on how to use it let me know.

     

    I believe the Nikon D700, D3, and D3s cameras are supported in Rawnalyze 2.10.4.0. If not you can convert the NEF to DNG.

     
    |
    Mark as:
  • Currently Being Moderated
    Jun 13, 2012 7:03 AM   in reply to rhythmdevils

    I've checked your (rhythmdevils) NEF files, and exposure is different in LR, ACDSee Pro and all free tools based on the dcraw.

    You can download RawDigger and check actual pixel values in the RAW file.

    It looks like RAW data is actually have different exposure, it's not rendering difference.

     
    |
    Mark as:
  • Currently Being Moderated
    Jun 13, 2012 10:13 AM   in reply to rhythmdevils

    rhythmdevils wrote:

     

    I uploaded 2 sample images onto my website, one from the D3 and one from the D3s.  Let me know if it doesn't work, I can host them elsewhere. 

     

    http://www.whitneydafoe.com/d3vsd3s/

     

    I checked both files in Rawnalyze and in LR4.1 final release using default settings and Adobe Standard. The good news is that the files only differ by  ~1/3 EV, but the White Balance setting is significantly different. The below Histogram scale is exponential with the thinner lines at 1/3 EV points.

    NikonD3versusD3s.jpg

    Inside LR the 'As Shot' White Balance is: 

     

    D3: 7300 Temp and +2 Tint

    D3s:7800 Temp and +18 Tint

     

    and inside Rawnalyze:

     

    As Shot Setting RGB

    D3   2.2461 1.0000 1.0859

    D3S 2.2539 1.0000 1.0898

     

    The D3 and D3s in-camera WB settings are NOT being set the same! Using LR settings for D3 WB to 4900/+33 and D3s 5000/+45 with +.33EV the two look near identical.

     

    If you want to get better color matching I would suggest using the ColorChecker Passport. I use it with my Canon 300D, 5D MKII, and 600D. The custom DNG proifles it creates are far superior to the Adobe and Camera Standard profiles. I create standard camera profiles for each body for Daylight, Cloudy, Tungsten, and will sometimes include it in a critical assignment to create a profile for that locaton's lighting. IMHO well worth the $100.

     
    |
    Mark as:
  • Currently Being Moderated
    Jun 13, 2012 10:15 AM   in reply to Jim Wilde

    I don't think the D3 profile is the one that is off, I think the D3s Adobe conversion is at least more to blame. 

     

    Reason is that I took some other test shots using the "shade" WB setting on both cameras, since I shot these in the shade on a sunny day.  The wall I did this test on is flat gray.  The images from the D3 looked much closer to the actual wall color than the D3s, which was very green. 

     

    Maybe this is just a matter of needing to make a custom profile, but I don't know how adding +.5 EV to every shot plays into your latitude for making actual adjustments (not just camera correction).  And I'm not sure why Adobe would be this far off esp for Nikon's TOTL camera.  Still makes me wonder if it's a problem on my end. 

     
    |
    Mark as:
  • Currently Being Moderated
    Jun 13, 2012 10:47 AM   in reply to rhythmdevils

    We must of "cross-posted." I don't think there is anything wrong with either camera. You're dealing with normal "manufacturing tolerances" that occur with camera image sensors & electronics, and a difference between the D3 and D3s in-camera 'Cloudy' WB settings.

     

    FYI - There is no need to set the in-camera white balance when shooting raw, it's only needed for in-camera JPEGs. All my bodies are set on 'Daylight' and the other WB settings have never even seen any light.

     

    The Adobe LR Camera Profiles are created using most likely only one or a small number of camera bodies for each model. For the reasons I stated there are variations within the same model, and the D3s camera Adobe used may be different than your D3s camera. You will also find that different lenses affect WB, as well as UV and Skylight filters, etc. Most of my Canon lenses Daylight WB are right on at 5250 and +28, but my one Sigma 50mm macro requires 5050 and +13. Otherwise it looks too warm sort of like what you see with the D3s.

     

    Tolerances are additive: Lighting + Sensor + In-Camera WB + Lens WB + Filter WB + Camera Profile variations = ???? WB Setting

     

    ERGO: My suggestion to use a ColorChecker PassPort.

     
    |
    Mark as:
  • Currently Being Moderated
    Jun 13, 2012 3:07 PM   in reply to trshaner

    I'm not afraid of something wrong with the camera anymore, since the images look exactly the same in Nikon NX2.  And since they looked the same in Nikon's software, I dont think this is due to variation on Nikon's end.  I can't imagine Nikon having more variation between differnet D3s's than between a D3 purchased 4 years ago and a D3s.  So it seems to me that it's on Adobe's end. 

     

    But to be clear, the images I uploaded were taken using a custom k WB setting, can't remember what I set it to, I wasn't too worried about it being accurate because I was just checking precision between the two.  But I can't imagine that changing between models, 5000k should be the same on any camera right?  I used the same lens on both cameas (switched the same lens back and forth(, so that can't be an issue either. 

     

    I will look into the colorchecker passport, I've been wanting to get one for a long time.  I don't just use daylight because I like them to look purdy on the LCD.  I know...

     
    |
    Mark as:
  • Currently Being Moderated
    Jun 13, 2012 4:09 PM   in reply to rhythmdevils

    The D3 and D3s downloaded NEFs are within less than 1/3 EV, which is very good. Just make sure you are setting the WB in LR (Custom), instead of using 'As Shot.' You need to do this when creating custom Camera Profiles, and even when using the ColorChecker Passport. At least you can do it with accuracy using the eyedropper tool in LR4:

     

    http://feedback.photoshop.com/photoshop_family/topics/white_balance_ey edropper_tool_option

     

    Then create a new Develop Preset for each camera profile containing WB, Process Version, and Calibration.

     

    I also noticed the D3 Adobe Standard LR profile raises EV slightly compared to the Camera Standard v4 profile. I suggest using only the v4 profiles. All of the D3s profiles look fine and it sounds like you can raise your exposure a bit. This will help reduce noise in the shadows, especially at high ISO. I think you will find the ColorChecker created camera profiles much more accurate, and there's nothing stopping you from shooting with ANY lighting condition. The CC PP can even create dual-illuminate (EX- Daylight + Tungsten) camera profiles! Enjoy!

     
    |
    Mark as:
  • Currently Being Moderated
    Jun 25, 2012 3:37 PM   in reply to rhythmdevils

    Rhythdevils,

    You are correct I have also noticed this problem with adobe standard for the d3s is always darker by 0.5+ stops and in both LR3 and LR4 and is also present when trying to do a conversion in DNG Profile editor with colorchecker passport its impossible to get the right exposure when comparing the jpeg file. You are 100% correct its the ADOBE profile at fault - End off

     
    |
    Mark as:
  • Currently Being Moderated
    Jun 25, 2012 3:51 PM   in reply to Nicholas Nicolaou

    Thanks for your input. 

     

    So if we correct the problem by adding .5 EV, is that .5 part of the 2 stop latitude, or is it just making the file look like it should?  In other words, do you still have as much latitude for increasing exposure on an image because of this?  Or are we now limited to 1.5 if 2 is considered the most you can push an image.  (I know there's no rule that applies to all images, but for the sake of communicating I'm simplifying it)

     
    |
    Mark as:
  • Currently Being Moderated
    Jun 25, 2012 3:59 PM   in reply to Nicholas Nicolaou

    There was an issue with the initial D3 profiles, which is apparently still the case with both the D3 and D3s Adobe Standard profiles. As I suggested use only the v4 Camera profiles, which do not have this issue

     

    Nicholas Nicolaou wrote:

     

    when trying to do a conversion in DNG Profile editor with colorchecker passport its impossible to get the right exposure when comparing the jpeg file. You are 100% correct its the ADOBE profile at fault - End off

    When creating ColorChecker Pasport custom camera profiles the LR camera profile is not applied. When using the ColorChecker created camera profile you should not see an exposure difference as with Adobe Standard. It sounds like you are using the CC PP incorrectly.

     
    |
    Mark as:
  • Currently Being Moderated
    Aug 16, 2012 2:37 PM   in reply to trshaner

    So can someone explain what's happening with the Adobe profile a bit more here?  The Adobe profile is making the image darker, therefore britening the image back up in LR isn't going to introduce noise or reduce latitude for further corrections in LR?  For example, if I have to add .5 to every image, am I then less able to correct for underexposed images?

     

    I already returned my previous D3s before I found both cameras images to look identical in NX2 proving Adobe was causing the difference, not Nikon.  So I'm about to buy another D3s. 

     
    |
    Mark as:
  • Currently Being Moderated
    Aug 17, 2012 7:30 AM   in reply to rhythmdevils

    Do NOT use the LR Nikon D3s Adobe profile, which has issues! Use only the LR Nikon D3s 'v4 Camera Profiles,' Better yet buy yourself a ColorChecker Passport and make custom camera profiles for each of your camera bodies.

     

    You will have absolutely no exposure or color issues like you are getting with the current LR Adobe Standard profile, which should eliminate your concerns. Please go back and re-read my posts in this thread for more information on the ColorChecker usage. IMHO-It's the best $100 you'll ever spend on camera equipment.

     
    |
    Mark as:
  • Currently Being Moderated
    Oct 14, 2012 12:55 AM   in reply to alessandroavenali

    I also have a D3s and D700 and find that the NEF RAW files from both of them come out really dark in ACR and Lr3 when importing them. Much darker than they look in camera, and almost all of the photos are underexposed. Worse, clicking AUTO adjust makes them even darker (I know auto doesn't always work, but it seems to never work for me no matter what photo) Would love to know why this is as the exposures always look ok in camera when I'm shooting and the histograms look fine too.

     
    |
    Mark as:
  • Currently Being Moderated
    Oct 14, 2012 1:14 AM   in reply to david_burden

    If you have ADL turned on in camera, then turn it off.  It purposely shoots dimmer to avoid blowing the highlights, then digitally brightens the darker areas.  LR doesn’t know how to do this Nikon-specific function so the imported pictures show the true exposure, not the Nikon-enhanced exposure.

     
    |
    Mark as:

More Like This

  • Retrieving data ...

Bookmarked By (0)

Answers + Points = Status

  • 10 points awarded for Correct Answers
  • 5 points awarded for Helpful Answers
  • 10,000+ points
  • 1,001-10,000 points
  • 501-1,000 points
  • 5-500 points