Jump to content


Photo
- - - - -

in-camera histograms


  • Please log in to reply
6 replies to this topic

#1 craig

craig

    Full Moon Rising

  • Super Mod
  • 2826 posts
  • Gender:Male
  • Location:Austin, TX

Posted 03 September 2008 - 05:02 PM

I had a minor victory today. I've been trying to make the histograms on my camera and raw converters match and to make the in-camera histogram actually indicate full exposure.

I've been using the Near UniWB jpeg for the D300 to bypass the RGB gains in the camera. I've also selected Neutral and AdobeRGB, but my in-camera histograms were still showing clipping about 1.5 stops too soon. I discovered that the in-camera picture controls are editable and that you have to modify the Neutral setting to minimize brightness and contrast through a somewhat obscure user interface. Once that is done, the two raw converters I'm using needed a default exposure setting of -1/3 eV before the histograms looked good. I now have an in-camera histogram that actually works for the first time!

I don't know why the camera manufacturers don't offer a raw RGB histogram option. It's obvious, it's easy, and it's useful.
I love it when a plan comes together.
- Col. John "Hannibal" Smith

------
Nikon, Seatool, Nexus, Inon
My Galleries

#2 PapaFly

PapaFly

    Clownfish

  • Member
  • Pip
  • 26 posts

Posted 04 September 2008 - 12:35 AM

Great post, it really is a pain in the butt to come home and find out that the in-camera histogram was way off compared to the raw.
Can this be done for the D80? I'm exclusively using acr, however i have the nikon software installed.

Phil

#3 craig

craig

    Full Moon Rising

  • Super Mod
  • 2826 posts
  • Gender:Male
  • Location:Austin, TX

Posted 04 September 2008 - 04:58 AM

I don't know. Search for UniWB for the D80. Without it, you can't get the histogram to stop lying to you. Even with UniWB, you will need to disable all the other effects to make it work. It helps to have a RAW data viewer to see when the camera actually clips. Rawnalyze is what I used.
I love it when a plan comes together.
- Col. John "Hannibal" Smith

------
Nikon, Seatool, Nexus, Inon
My Galleries

#4 PapaFly

PapaFly

    Clownfish

  • Member
  • Pip
  • 26 posts

Posted 05 September 2008 - 02:47 AM

What helped me was the calibration of acr, by using a macbeth color checker test shot and an online script. Now, when i load my nefs into acr, they look almost identical to the in-camera jpeg. This seems to be a general problem of acr and nikon cameras. Without calibration, contrast, brightness and hue - specially in the red channel - are way off.
Still, the basic problem of camera histogram indicating clipping, when the nef is not, persists. I'll give uniwb a try.

#5 craig

craig

    Full Moon Rising

  • Super Mod
  • 2826 posts
  • Gender:Male
  • Location:Austin, TX

Posted 05 September 2008 - 03:34 AM

I agree, ACR needs some help to get its colors correct especially in the reds. I also run calibration scripts against a ColorChecker. Adobe has new profiles meant to address that. They look hugely improved.

I've convinced myself that ACR does, in fact, botch up shadow detail under certain circumstances and it's noise processing is both damaging and always on to some extent. I re-auditioned Bibble and Aperture 2 but my opinion of them hasn't changed. My best results have come from RPP and Raw Developer. RPP has the worst imaginable user interface but produces really good output. Raw Developer is my favorite overall. I'm now trying to determine how to fit it into an overall workflow plan. I'm undecided whether I will continue with Lightroom or simply use RD with Photoshop. Part of the problem may be integer computation in combination with gamma 1.0 processing. That is unavoidable with Lightroom. :)

Both RPP and Raw Developer are mac-only. They both support custom ICC profiles and I've successfully generated those. I've had very good success generating neutral, accurate ColorChecker results with both of them.
I love it when a plan comes together.
- Col. John "Hannibal" Smith

------
Nikon, Seatool, Nexus, Inon
My Galleries

#6 loftus

loftus

    Blue Whale

  • Member
  • PipPipPipPipPipPip
  • 4571 posts
  • Gender:Male
  • Location:Winter Park, Fl

Posted 06 September 2008 - 03:17 PM

With my D700 I am using Capture NX2 (no RAW convertor yet available for Aperture). The histograms in NX and on the screen seem to be pretty well matched. When I convert to a TIFF and import to Aperture there does not seem to be a significant lateral shift of the histogram, only changes in the amplitude.
Nikon D800, Nikon D7000, Nauticam, Inons, Subtronic Novas. Lens collection - 10-17, 15, 16, 16-35, 14-24, 24-70, 85, 18-200, 28-300, 70-200, 60 and 105, TC's. Macs with Aperture and Photoshop.

#7 craig

craig

    Full Moon Rising

  • Super Mod
  • 2826 posts
  • Gender:Male
  • Location:Austin, TX

Posted 06 September 2008 - 05:40 PM

In ACR, there is a hidden +0.5 stop exposure compensation on the D300. I hear it is similar on the D3 but haven't tested it yet. NX supposedly does the same.

The only way to know for sure is to test your camera's output using software that you know doesn't modify the data. dcraw can be made to do that. RPP promises not to do that, RD has a raw-only mode (and lets you modify the default compensation even!) and Rawnalyze will let you see raw histograms. Using one of those, you can determine the true onset of clipping and then determine how that relates to the in-camera histogram. Based on the default settings on my D300, even with UniWB chosen, I was nearly 2 stops underexposed when the histogram said I was good. That's not at all helpful!

Nothing frustrates me more than working to get a good exposure only to see that I'm a stop off in the raw conversion.

There are hacks that enable the D700 to work with Aperture. :)
I love it when a plan comes together.
- Col. John "Hannibal" Smith

------
Nikon, Seatool, Nexus, Inon
My Galleries