The Histogram comes in a variety of flavors. The one that is most famous and mostly used is the luminosity histogram. Take a look at the following image and the Luminosity histogram that accompanies it.
If you notice, the luminosity histogram (which plots a graph of all pixels based on their luminosity scale from 0 to 255) looks absolutely fine.
Each pixel in your camera contains information in three channels (Red, Green and Blue) so each pixel has an RGB value. For e.g.
A complete white has RGB values of 255,255,255
A complete black pixel has RGB values of 0,0,0
A darkish grey image might have RGB values of 123,123,123
A bright yellow might have an RGB value of 255,255,0 and so on.
Now with this in mind, let’s see how is the Luminosity Histogram created.
The Luminosity histogram is created by taking the RGB values of each pixel and then creating a single reading out of those values with the following equation
0.59 x Green + 0.3 x Red + 0.11 x Blue
So the green gets a 59% weight, red gets 30% and blue gets 11%.
Now assume that a pixel is a vibrant red.
Its pixel value comes to something like 255,50,50
Note that a 255 value in the red channel means that there is an over-exposure of the Red Channel.
When this is shown in the Luminosity Histogram, the value comes to
(50 x 0.59) + (255 x 0.3) + (50 x 0.11) = 111.5
This value of ‘111.5’ will never show you that the Red Channel is clipped/over-exposed.
The Highlight Enable/Alert options in your cameras also generally work on the Luminosity concept so they will not blink on channel specific over-exposure either.
This is where the Channel Specific Histograms come handy.
Take a look at the Channel Specific histograms for the image below and you will quickly realize that the Reds are getting overexposed. The easiest remedy is to reduce the exposure a little and get the colors right.
If your camera has the functionality to display the channel specific histograms as well, make sure you enable them. Helps when you are shooting subjects which have a variety of colors.