Cant read this Histogram :-(

[drsatyendrapaul]

uploading an image i imported into LR4. The histogram's color components i understand, but why there are so few grey pixels in the histogram? Why is the grey mound so small? Any help??




Q. No. 2: If i look at the histograms of same image in color and then in B&W, why the difference in appearance and size in both, as seen below?

[revi]

The gray part in the histograms is just the zone where the areas for all three colours overlap. The same as the yellow zone is overlap between red and green, and the cyan zone is the overlap between blue and green. The gray has no particular significance wrt the image quality.

In the first image, there's very little overlap between the 3 colours, so the gray area is very small, in the second one, there's a lot of overlap, so a large gray area, 3rd image is B/W, so exact overlap of the three curves, so the whole area is gray (but there's no cyan or yellow zone).

What also causes the gray area in the first image to be particularly small is that the image is almost monochrome with a very saturated colour, so the three curves for red, green and blue are at quite different intensities.

[drsatyendrapaul]

Merci beaucoup, Remco. I did not know this fact that the grey part is actually the overlap of all three primaries. This sure does explain a lot to me.
Thanks again.

[Glenn NK]

There are only two things that I "read" on a histogram that mean anything to me and that I can do anything about:

1) if it's clipped on the left I've lost detail in the darker areas,

2) if it's clipped on the right, I've lost detail in the brighter areas.

Glenn

[Inkanyezi]

There are only two things that I "read" on a histogram that mean anything to me and that I can do anything about:

1) if it's clipped on the left I've lost detail in the darker areas,

2) if it's clipped on the right, I've lost detail in the brighter areas.

Glenn

I'd like to add something there.

Clipping, is what we know happens when the light in the brightest part of an image are too bright to be captured, it "runs over" in the sensel, and nothing more can be registered. We reach the maximum digital levels possible and that's it.

The left hand side is different. There is no clipping going on to the left, and i would prefer that the term "clipping" is not used. Instead, we reach a point where any light impression will be drowned in noise, when it falls below the noise threshold. Of course there are light levels that cannot be registered at all, and in a heavily underexposed shot, maybe "clipping" might be considered an adequate term for light that is not registered at all, but before reaching that point, light impressions are unnoticeable in the final image because they fall below the noise threshold.

The left hand side of the histogram mostly is of no interest. The information we can gather from it is either clearly seen in the picture, as much of the data on the left side will only tell us that the image is dark, or if there is also data present all the way to the right side, it will tell us that the scene contrast was higher than the dynamic range of the camera. Whether it is underexposed or not depends on what you wanted to capture, and the important information in the histogram is on the right side, and a crucial point is the very border to the right.

If there is a large part to the right that is flat at the bottom, it indicates that you could expose more to capture more data and to get better tonal rendition and less noise in the darker parts of the image. Underexposure is indicated in the histogram at the right side, just as overexposure is. When you have absolutely no data in a large part to the right, all the way to the end, the image is underexposed. Ideally, the histogram should end just in the lower right corner. Only if there are large areas in the image that you want blown out, should you let the histogram data climb up the right wall a little.

But a low key image?
Now digital is different from film in this respect, and if your final image shall have no bright parts, you can always make it darker in pp, and if you kept the histogram to the right, without climbing the wall, your image will have less noise in its dark parts. The histogram before you adjust brightness, might have an empty part to the left and appear as if it doesn't have any black, but once you set the black point, there will be pitch black areas, with hardly any noise.

The zone system is still alive and kicking in the digital age, but the application of it is very different. We should not care to get dark parts of the scene dark at capture, but concentrate on the bright parts and not let them run over what can be registered on the chip. As long as our camera can cope with the brightness range of the scene, we need not care much about its darker parts. However, when we know just how the camera will register different light levels, and how its dynamic range shrinks when we increase ISO, we may use it to gain shorter shutter time or smaller F-stop without losing neither dark nor bright values, as long as we set it to an ISO that retains dynamic range that can register the total brightness range in the scene. However, doing so, implies that we will have more noise in the darkest levels of the final image than if we use a lower ISO and adjust the black point in pp.