Colour Clipping

Messages
1,336
Name
Peter
Edit My Images
Yes
I have just been looking at a video on YouTube where they were talking about colour clipping. What on earth would they be talking about.
 
Probably overexposure of one colour channel. This can happen even when the auto exposure thinks the exposure was correct and unclipped if there was one colour which was unnaturally bright. This can happen quite often with fluorescent red, for example, used by women and flowers to attract notice. Fluorescent colours convert some of the UV falling on them to the fluorescent colour, making the colour brighter than the natural illumination of the scene could make it.
 
Probably overexposure of one colour channel. This can happen even when the auto exposure thinks the exposure was correct and unclipped if there was one colour which was unnaturally bright. This can happen quite often with fluorescent red, for example, used by women and flowers to attract notice. Fluorescent colours convert some of the UV falling on them to the fluorescent colour, making the colour brighter than the natural illumination of the scene could make it.
Well put sir.
I always had problems with red stage lighting in the film days...I thought my problems were over when I went digital...sadly not.
 
When a colour channel goes out of range.

Could be over exposure.

Could be a change in colour space

Either way - it leads to Hue shifts.
 
Probably overexposure of one colour channel. This can happen even when the auto exposure thinks the exposure was correct and unclipped if there was one colour which was unnaturally bright. This can happen quite often with fluorescent red, for example, used by women and flowers to attract notice. Fluorescent colours convert some of the UV falling on them to the fluorescent colour, making the colour brighter than the natural illumination of the scene could make it.
I have had this problem occasionally in the past, usually with reds in fungi.

It is possible that these reds are flouresent colours as I know that fluorescence does occur with some fungi. Red would be the most likely fluorescence in a dim forest as the light is generally low in short wavelengths.

The other possibility is that very pure reds are much more likely than pure greens or blues. Blue is rare in anything approaching a pure form, with the sky being the bluest thing we normally see. Green is even rarer in its spectral colour. Almost all green is from chlorophyll which absorbs some of the red and blue, but not all. Most of the remainder is really yellow plus blue and not spectral green (like parrots). Only red in nature is commonly a pure spectral colour (of red, green, blue) and perhaps this is why it confuses cameras.

There are some colours that have always eluded artists and photographers because of the medium with which we see those colours. These are structural colours, found in some butterfly wings or a peacocks tail. They are not pigments, which are the mediums of paints, printers and video screens. A structural colour can reflect a single (or multiple) wavelengths collected from many directions, into a single direction. Thus it is brighter than bright (or whiter than white as the detergent adds used to say when they started to use artificial whiteners) and can never be accurately portrayed by a pigment colour.

Fascinating stuff and bound to confuse the best of cameras.
 
Last edited:
I have had this problem occasionally in the past, usually with reds in fungi.

It is possible that these reds are flouresent colours as I know that fluorescence does occur with some fungi. Red would be the most likely fluorescence in a dim forest as the light is generally low in short wavelengths.

The other possibility is that very pure reds are much more likely than pure greens or blues. Blue is rare in anything approaching a pure form, with the sky being the bluest thing we normally see. Green is even rarer in its spectral colour. Almost all green is from chlorophyll which absorbs some of the red and blue, but not all. Most of the remainder is really yellow plus blue and not spectral green (like parrots). Only red in nature is commonly a pure spectral colour (of red, green, blue) and perhaps this is why it confuses cameras.

There are some colours that have always eluded artists and photographers because of the medium with which we see those colours. These are structural colours, found in some butterfly wings or a peacocks tail. They are not pigments, which are the mediums of paints, printers and video screens. A structural colour can reflect a single (or multiple) wavelengths collected from many directions, into a single direction. Thus it is brighter than bright (or whiter than white as the detergent adds used to say when they started to use artificial whiteners) and can never be accurately portrayed by a pigment colour.

Fascinating stuff and bound to confuse the best of cameras.

Whoa - thanks.

So how does one avoid this clipping.
 
If you have a RGB histogram you'll be able to tell if one of the colours is blowing out and adjust exposure accordingly.

Whether your camera can recover shadows well without introducing lots of noise is a different matter and camera specific.
 
Whoa - thanks.

So how does one avoid this clipping.
Under expose. If you come across structural colours you just have to admire them because you can never reproduce them with conventional materials.
Fortunately (or unfortunately?) these things aren't common or don't attract photography. How would you go about representing a blue flash from a butterfly's wing? A flash that only occurs at one angle? Perhaps an artist could create something with some nano particles, though positioning them accurrely may be tough.
 
Sometimes a tiny bit clipping is almost not noticeable, but quickly it becomes pretty obvious. sRGB colour space is particularly bad for reds. Your original RAW file may be absolutely fine but if exported not carefully it may look very wrong.
 
If you have a RGB histogram you'll be able to tell if one of the colours is blowing out and adjust exposure accordingly.

Whether your camera can recover shadows well without introducing lots of noise is a different matter and camera specific.
Depends on what the histogram is showing? Is it sRGB, is it AdobeRGB, does it correctly change to Rec709/rec2020/rec2100 when shooting video? Is the histogram showing linear light or signal levels?

I have a feeling it is usually sRGB, so if you're delivering to a wider gamut display, it's not very useful.
 
Back
Top