What does 'clipping' mean?

[Diurnal Bowler]

Hi, My name is Graham and I live in Bridge of Weir, Scotland. I'm relatively new to photography, but I'm just getting in to HDR. I'm reading a book about it and it keeps referring to clipping of the histogram. Can someone explain this to me because I'm just not getting it.
Regards,

Graham Evans

[Donald]

I've copied Graham's message across from the New Member thread, because I'm sure a lot of people who are new to photography and are looking in will be keen to understand this one as well.

And because he lives just about 30 minutes from me (well, a bit more), he gets to have his first message posted twice .... as a bonus!!

[Jim B.]

Hi Graham,
Clipping
When an image is over or under exposed, its histogram will show a concentration of pixels to the left if under exposed and to the right when over exposed. When the concentration is bunched up against the edge of the histogram, we know that significant detail has been lost in the image. This loss of detail is referred to as clipping.

Sean also has an excellent tutorial on understanding histograms.

[ktuli]

Graham,

Not sure if this will help or not... but if you're used to any kind of audio equipment, clipping is pretty much the same concept in both places. Basically it is a case where there is more signal than the hardware can handle. In audio, the hardware is usually the speakers and when a signal is clipping, you will hear those annoying pops and clicks as the speakers just basically "freak out" trying to figure out how to play a sound that just doesn't make sense. A very generic way to think of it is if the volume is too loud and the speakers just can't keep up.

The same thing happens in a camera. If your sensor is exposed for too long, it collects too much light. Basically, when the light sensitive area on the sensor reaches full capacity, it displays that as completely white. In binary speak, this means that the interpretation is listed as all ones (on an 8 bit sensor, it would be 8 ones). On the histogram, this will look like a peak pushed all the way up against the right edge of the graph. If there is any kind of gap on the right edge of the graph, you are not clipping.

The same happens on the low end of the scale as well. If the sensor has not been able to capture any light, it will display are pure black, and there will be a peak push all the way up against the left edge of the graph. Again, in binary speak, this would be listed as all zeros.

The problem with clipping is that the data there is gone. You can't recover it in post production at all. Some people claim it can be recovered, but in reality, what they are doing is generating a close approximation of what they think *should* be there. They're not adjusting the data that is there, they have to completely create new data out of thin air to "fix" it.

In a regular single exposure, you want to try and keep from clipping so that you don't lose any of your highlights (blow out) or your shadows (under expose). In reality, many scenes just can't be captured that way... which is obviously why you're looking at HDR. When you are collecting a set of images for HDR, you will almost certainly produce intermediary images that clip (basically because if you don't have to, then there is no need to capture the scene as an HDR).

Hope this helps.

- Bill

[ktuli]

Graham,

You might also want to refer to https://www.cambridgeincolour.com/tu...istograms1.htm (which interestingly enough shows up as the top hit when google searching for "histogram clipping").

About 2/3rds of the way into the article is an example of a clipped histogram (though the underexposed clipped histogram is a little hard to see, the overexposed one is a good example).

But read the whole article as it will explain things much much better than I did above.

- Bill

[ktuli]

Graham,

Sorry for the multiple posts, but can you tell us what book you're reading? I find it pretty discouraging that an author would venture to write a book about HDR and not clearly and succinctly explain clipping before getting into any other details. Like I said, clipping is exactly the kind of situation where HDR is useful, so it seems silly to explain how to do it without explaining when and why you would want to!

- Bill

[Diurnal Bowler]

Hello Bill,

The book is called Practical HDR by David Nightingale fist published in the UK in 2009 by ILEX It was bought off Amazon.
I must say that from the responces I've had it is much clearer to me now.

[Steaphany]

Graham,

Here is another way to look at it. Most technology implementations have limits of dynamic range, the span from 0 to an upper limit determined by the number of bits used to represent a numerical value. As an example, the JPG file format has only 8 bits per color channel, so it's dynamic range is 00000000 to 11111111, or more easily dealt and conceptualized 0 to 255.

Clipping is the phenomenon where the values from a scene exceed this range. What if a given exposure yields a midtone gray at 255 and the scene contains areas that correspond to 255 plus 10, 100, or more. If all you have is 8 bits, adding 10, 100, or 500 to 255 will still leave you with 255, i.e. the tonal values which exceed 255 become clipped and all you get is 255.

Likewise, clipping can also occur at the low end, where scene darks end up corresponding to values that would be below 0, but here too, your level stops at 0 and again more clipping.

Since you're new to photography, I recommend that you get a good foundation in the basics and learning how to work within the RAW capabilities of your camera. Then HDR will make more sense and you'll better understand the reasons and techniques employed in creating a HDR image.

[Wirefox]

Excellent explanation Steaphany. I will have to clear some space in my dusty attic of a brain for that one.

[Steve S]

Here is a very simple explanation................

In this imaginary histogram




The area in grey is the tonal range your camera is capable of recording. 0 is pure black and 255 is pure white.

The area in white is the tonal range of the image . As you can see, the tonal range of the image is greater than the camera can record. The area in the red has been clipped and will be rendered (by the camera) pure black. The area in the blue has also been clipped and will be rendered (by the camera ) pure white. The problem is the areas in red and blue are not really pure black and pure white.

So this is where your HDR comes in. If you lower your exposure, the area in blue will shift to the left and will be in the zone the camera can record.(giving you detail in that area) If you increase your exposure , the area in red will shift to the right and be in the zone the camera can record.(giving you detail in that area) And if you shoot one image just as it is pictured, the center portion of the histogram is in the zone the camera can record. Now you take all three photos and combine them to get the full range of tones in the image.

Hope that was a little easier to understand.

[inkista]

One footnote to add. The in-camera histogram and blown highlight warnings you see on the LCD are derived from the JPEG preview, not from the RAW data. This means that you could see clipping/blinkies when you're actually still within the dynamic range of the camera. It's not 100% accurate if you're shooting RAW.