Histogram in Nikon D7000

[ksasidhar]

Hi all,

I was taking pictures and trying to understand their histograms using my Nikon D7000. After a while I realized that the histogram of white light looks very close to histogram of green. Very little seem to have changed due to red and blue histograms.

Does is it mean that white histogram is taking most of its data from green histogram. Later I read on a website that Nikon used to take all its data from green histogram and neglect red and blue.

If that is the case, how can I understand the data correctly from the histogram by checking the white light? It does not seem to be possible to try and analyze all three histograms at same time, especially if the scenery has slightly higher color of one of the three colors.

Sasidhar

 

[Ysarex]

Read this: Luminance versus RGB Histograms - Photo Tips @ Earthbound Light

In playback the D7000 will display a single histogram, which appears to be an RGB composite, or 4 histograms -- the composite again along with each channel. When a camera displays a single histogram it's usually either an RGB composite or a luminosity histogram. Getting the camera maker to tell you which they're displaying can be a little tricky but I'm pretty sure your D7000 displays an RGB composite. My Fuji displays a luminosity histogram.

Because the RGB composite is calculated from the values of all the three separate channels it's going to look different than any one channel. Depending on how the camera maker does that calculation (many variations exist) the RGB histogram may or may not indicate clipping if only one color channel is clipped. You should be able to run tests to determine this; you can always examine all three color channel histograms.

All of the histograms displayed on your D7000 are calculated for the camera processed JPEG and do not reflect the data saved in a raw file.

Joe

 

[Dave442]

Nice link above, so if green has the most effect on luminance that makes sense that it would be similar to a luminance histogram.

When I do turn on my histogram I have it set to show the three, but as Joe noted these are generated from a file the camera has processed. So in the end the histogram is just another tool like the camera meter to help in establishing an exposure.

If the requirement is for some type of critical work then probably need to camera tethered to a computer and use some software that can give better histograms.

 

[astroNikon]

The d7000 shows a composite histogram
and when the option is selected, a RGB histogram which shows all 4 histograms on the screen - composite, then RGB

I'm not sure if the one is an RGC Composite or a luminosity.
The description in the manual for the single one is
histogram (RGB channels), in all histograms, horizontal axis gives pixel brightness, vertical axis number of pixels)

 

[astroNikon]

couple years ago Sparky wrote a thread about histograms too. Might want to read that ==> Demystifying Histograms | Photography Forum

 

[Ysarex]

The d7000 shows a composite histogram
and when the option is selected, a RGB histogram which shows all 4 histograms on the screen - composite, then RGB

I'm not sure if the one is an RGC Composite or a luminosity.

Every indication is that the single histogram is an RGB composite and not a Luminosity histogram. A Luminosity histogram is what you get if the photo were B&W -- the tone response of the photo without color. Luminosity histograms won't show single or even double color channel clipping whereas RGB composite histograms do tend to show single color channel clipping at least to a degree.

Joe

The description in the manual for the single one is
histogram (RGB channels), in all histograms, horizontal axis gives pixel brightness, vertical axis number of pixels)

 

[ksasidhar]

Educative links Ysarex and atsroNikon. Thanks all for your replies. I now get an idea of RGB histograms.

I still did not understand what ow RGB histograms can be controlled using the camera.

The website Ysarex send mentions "you have to take a more active role in avoiding problems with exposure on a per-channel basis" and "But if you're shooting a more real-world subject that may rely more strongly on one color or another at various points in the image, you'll be better served by an RGB histogram"

We can control the RGB composite by controlling the exposure settings in the camera. But how do we control the R, G, and B channels separately? To my understanding it cannot be controlled in the post processing tool either?

Can you clarify the remarks in the website. Thank you.

 

[Ysarex]

Educative links Ysarex and atsroNikon. Thanks all for your replies. I now get an idea of RGB histograms.

I still did not understand what ow RGB histograms can be controlled using the camera.

The website Ysarex send mentions "you have to take a more active role in avoiding problems with exposure on a per-channel basis" and "But if you're shooting a more real-world subject that may rely more strongly on one color or another at various points in the image, you'll be better served by an RGB histogram"

We can control the RGB composite by controlling the exposure settings in the camera.

Yes -- more or less exposure will cause the camera to process a brighter or darker photo and an RGB composite or luminosity histogram will reflect that change (as will the single channel Red, Green, and Blue histograms).

But how do we control the R, G, and B channels separately?

In camera you don't. The camera does not provide control access that would allow you to adjust for example only the blue channel or only the red channel without effecting the other channels.

To my understanding it cannot be controlled in the post processing tool either?

It's very easy to control in post processing of a raw (NEF) file or an uncompressed RGB file. I do it every day.

Joe

Can you clarify the remarks in the website. Thank you.

 

[Ysarex]

We can control the RGB composite by controlling the exposure settings in the camera. But how do we control the R, G, and B channels separately? To my understanding it cannot be controlled in the post processing tool either?

Can you clarify the remarks in the website. Thank you.

OK, here's an example: I have an attractive nuisance that's overrun my tomatoes -- morning glories. Might as well take some photos while I'm pulling them up. Here's a photo from a couple days ago:

That's the JPEG that my camera created. I did nothing except re-size it for this page. The red channel is blown. The software in my camera did that. This camera has a live view histogram that I can see before I take the photo. If I had looked at it I would have noticed the clipping. The camera displays a live RGB composite histogram. I can't show you that now but I can show you the RGB composite and RGB channels for the photo displayed by Photoshop:

You can see that the red channel is blown out and the blue just slightly clipped while the green channel is OK. The RGB composite does indicate the highlight clipping.

Here's the luminosity histogram for that same photo:

No clipping is indicated but we already know that the red channel is badly blown out.

The camera software is responsible for that photo above -- I didn't do that. The histograms that we're seeing and any histogram displayed on the camera are histograms for that photo above. I have very little control over what my camera's processing software is going to do. I could access the camera's picture controls and start adjusting things like contrast and saturation to try and coax the camera software to do better and of course I could alter the exposure.

BUT IT'S MY PHOTO! And I'm not about to compromise my exposure because the camera software can't process it right. There is no clipping in the sensor exposure for this photo. The raw file exposure is a text book example of nailed. It's important to note that the camera software and I are working from the same raw data. The camera's effort (along with histograms) is above. Here's my effort processing the same raw file (some cropping and additional touch-ups):

I knew when I took the photo that I had the exposure right and that I would be able to process the photo to an excellent end result. How did I determine the exposure? I carefully test my cameras and my cameras' exposure meters and I rely on the results of those tests. Since the camera histograms are all based on the processing output from the camera software I do not rely on those histograms. If I had relied on the camera live view histogram I would have reduced the exposure until the histogram no longer indicated clipping. Less exposure would have resulted in an unnecessary underexposure of the sensor and less recorded data -- why do that if I don't have to?

Joe

 

[ksasidhar]

Thanks Joe for taking effort to explain the process. Just couple of questions that I have.

1. You show the blue histogram and said that you observe clipping. Is it because of the peak that you observe in blue channel and the a large band at the right end of RGB composite?

2. I downloaded your 1st picture and checked the histograms in GIMP. I have almost similar histograms for individual R, G, and B channels. However the RBG composite looks quite different from the one you posted here. Shown below are the histograms for R, G, B, and RGB.

 

[Ysarex]

Thanks Joe for taking effort to explain the process. Just couple of questions that I have.

1. You show the blue histogram and said that you observe clipping. Is it because of the peak that you observe in blue channel and the a large band at the right end of RGB composite?

There is very slight highlight clipping in the blue channel. You can see it in the blue channel histogram -- it's the bright inside section of the flower. Clipping on the shadow end is a different issue with very different implications. Here's a threshold test using the blue channel:

2. I downloaded your 1st picture and checked the histograms in GIMP. I have almost similar histograms for individual R, G, and B channels. However the RBG composite looks quite different from the one you posted here. Shown below are the histograms for R, G, B, and RGB.

Yep, that's why further up in the thread I used imprecise language when I said "...RGB composite histograms do tend to show single color channel clipping at least to a degree." We can't say for sure until we know how the engineers calculated that RGB composite; they don't all do it the same way. The nice thing about GIMP is you can get the code and look it up. Adobe isn't going to be very helpful. This is a problem with RGB composites as we generally don't know how they're derived. You can fairly assume that just as you've noticed this discrepancy between GIMP and Adobe similar discrepancies exist between camera manufacturers. Just what is an RGB composite telling you? How are the three separate color channels combined into one? If you could actually reach someone at Nikon who would understand that question, I doubt they'd be able to get you an answer.

Joe