- Messages
- 7,908
- Name
- Terry
- Edit My Images
- Yes
Histograms are as difficult to explain in simple terms, as they are for a beginner to understand.
It is generally thought that they are useful for judging exposure, and in the crudest sense they are. You can instantly see if the exposure consists of too many dark pixels or too many highlight ones, if they are, the graphic will show the graph as bunched up to the left (dark tones) or to the right (highlight tones.)
However these configurations can also represent a perfectly good exposure. As some of the following shots and histograms will show.
In many cameras you can only see the histogram representing the pixels in black and white. In others you can see the pixels representing the red blue green components, which can give you a better Idea of what is going on. I have shown both versions.
In the first diagram I show a typical histogram ranging from some black pixels to some white ones. With the bulk of them in between. The pink column in the centre represents all the pixels of a particular grey tone, stacked one on another as if counting beans.
And this is exactly what the histograms show. That is the total number of pixels of each tone of grey in the image, stacked in columns from left to right. What they do not show is exposure.
histogram-diagram by Terry Andrews, on Flickr
The first image is of an Artillery Sargent taken in about 1954, it is well exposed, but the histogram shows the greatest number of pixels towards the right, this is because those are the tones of the sky and ground which take up the majority of the picture. The Sargent is somewhat darker and is mostly represented by the flattened portion of the graph. Neither the highlights nor the shadows are clipped. (Off the scale). So all the tones are reproduced normally.
hist-A by Terry Andrews, on Flickr
The second shot is also black and white taken in Madrid in 1957. both the photograph and the histogram look very different. Again the peak to the right represents the bright sky and the scorched earth. But this time far more extreme. There are a some pixels right across the scale but they few in number with a small peak in the deep greys. Again neither the highlights nor shadows are clipped. But the highlights very nearly so, in this very high key shot. So again the exposure is correct.
Hist-b by Terry Andrews, on Flickr
The next shot is again different not only because it is in colour but also because it is taken in candlelight, that had no chance at all to fill the vastness of the church. But the exposure was able to show some detail in what it could illuminate. This time the peak represents the vast black area, and the people and candles are represented by the thin black line along the bottom. Again it is the correct exposure to show the scene as it was. Had I taken a long time exposure I could have shown far more detail, but burnt out the candle light, and totally lost the atmosphere. There is no need to be afraid of low key images.
hist-1a by Terry Andrews, on Flickr
The next is a shot taken directly into the light but not a silhouette. It contains a full range of tones from black to near white. It again is correctly exposed. This time the histograms show two peaks one of dark tones and the other of bright tones but with rather fewer middle tones. This sort of shot is perhaps more difficult to expose correctly with out burning out the highlights.
histC by Terry Andrews, on Flickr
The butterfly image is very evenly lit, as can be seen by the background, and is typical of shooting artworks. The histogram in this case is not so much recording differences in light but more the local tones and colour differences in the subject itself. Though it was taken in situ at a university show, I was able to introduce some soft sidelight to emphasise the textures.
histD by Terry Andrews, on Flickr
The histogram of this Autumn show table, is hard to explain in terms of tones, as it is the bright colours that are emphasised in all their glory in the tangled colour histogram. When processing it, I had trouble with the colours clipping, and it was with some difficulty that I managed to retain detail in them all. Never the less the exposure is correct, the problem lay with the extreme colour gamut and saturation.
HistE by Terry Andrews, on Flickr
The last shot of the bright red Jeep also had colour gamut problems in the warm evening light. However apart from the strange colouration in the faces, the exposure just about captures the entire range of tones and highly saturated colours.
hist-F by Terry Andrews, on Flickr
None of these shots display the typically “Correct” histogram nor would I expect them to. However they all span in some fashion the range of tones from dark to light with little or no clipping.
It is more important to try to discover what a histogram is trying to tell you, than try to force it to fit some notion of correct exposure. Most examples of histograms shown to beginners are more simplistic than the ones that I have given, however they are also less representative of real images.
Histogram Addendum
Now that this thread has reached maturity, there are a couple of points that were raised that need clarification.
( I will also add this post back into the OP….. )
@HoppyUK said ” you claim "What they do not show is exposure" when they absolutely do - that is their main purpose and usefulness.”
Answer:
You can derive no Exposure information or settings from examining a histogram,
Two entirely different exposures and illumination, can show the exact same histogram.
A histogram does not take into account your intent, it only shows the tonal values that you have or are about to capture.
What it does do… is to give you a guide as to where those tone fall, and the direction that you would need to adjust exposure or exposure compensation to centralise those tones or avoid clipping.
It gives you no information about the magnitude or value of those adjustments.
Adjustments can only be made by trial and error and observation of the changed histogram.
The information a Histogram gives in this way, is indeed useful to correcting exposures
@DG phototraining said
“Nicely proving what a waste of time histograms are in general use”
and
“They may tell you what is clipped, but only you know if it matters or not, which is why "blinkies" are so much more useful if you follow the ETTR method”
Answer:
Centralising a histogram is especially useful to a beginner shooting Jpegs, who needs to get the image tones pegged, so as that the mid tone is correctly positioned. This may indeed clip some highlights and perhaps shadows, but will give the best representation of the subject matter with in the limits of an in camera jpeg.
More advanced users “shooting to the right” and using raw, will of course have more options and latitude, and can adjust the position of the tones to avoid clipping and maximise quality. However they will need to readjust the image during raw processing, to gain the maximum advantage. (I did not cover this in the original OP as it is not a “beginner” option.)
“Blinkies” are no more accurate at showing the degree of clipping, than the histogram itself, as both rely on the same data from the attached jpeg or screen capture, not the raw data. The main advantage of blinkies is their attention grabbing value. I would suggest that you get the most advantage if they are used in conjunction with the histogram.
I grant that some people seem to be Histogram phobic, and derive nothing from their use.
Thankfully that is not the case for the majority of photographers.
It is generally thought that they are useful for judging exposure, and in the crudest sense they are. You can instantly see if the exposure consists of too many dark pixels or too many highlight ones, if they are, the graphic will show the graph as bunched up to the left (dark tones) or to the right (highlight tones.)
However these configurations can also represent a perfectly good exposure. As some of the following shots and histograms will show.
In many cameras you can only see the histogram representing the pixels in black and white. In others you can see the pixels representing the red blue green components, which can give you a better Idea of what is going on. I have shown both versions.
In the first diagram I show a typical histogram ranging from some black pixels to some white ones. With the bulk of them in between. The pink column in the centre represents all the pixels of a particular grey tone, stacked one on another as if counting beans.
And this is exactly what the histograms show. That is the total number of pixels of each tone of grey in the image, stacked in columns from left to right. What they do not show is exposure.
histogram-diagram by Terry Andrews, on Flickr
The first image is of an Artillery Sargent taken in about 1954, it is well exposed, but the histogram shows the greatest number of pixels towards the right, this is because those are the tones of the sky and ground which take up the majority of the picture. The Sargent is somewhat darker and is mostly represented by the flattened portion of the graph. Neither the highlights nor the shadows are clipped. (Off the scale). So all the tones are reproduced normally.
hist-A by Terry Andrews, on Flickr
The second shot is also black and white taken in Madrid in 1957. both the photograph and the histogram look very different. Again the peak to the right represents the bright sky and the scorched earth. But this time far more extreme. There are a some pixels right across the scale but they few in number with a small peak in the deep greys. Again neither the highlights nor shadows are clipped. But the highlights very nearly so, in this very high key shot. So again the exposure is correct.
Hist-b by Terry Andrews, on Flickr
The next shot is again different not only because it is in colour but also because it is taken in candlelight, that had no chance at all to fill the vastness of the church. But the exposure was able to show some detail in what it could illuminate. This time the peak represents the vast black area, and the people and candles are represented by the thin black line along the bottom. Again it is the correct exposure to show the scene as it was. Had I taken a long time exposure I could have shown far more detail, but burnt out the candle light, and totally lost the atmosphere. There is no need to be afraid of low key images.
hist-1a by Terry Andrews, on Flickr
The next is a shot taken directly into the light but not a silhouette. It contains a full range of tones from black to near white. It again is correctly exposed. This time the histograms show two peaks one of dark tones and the other of bright tones but with rather fewer middle tones. This sort of shot is perhaps more difficult to expose correctly with out burning out the highlights.
histC by Terry Andrews, on Flickr
The butterfly image is very evenly lit, as can be seen by the background, and is typical of shooting artworks. The histogram in this case is not so much recording differences in light but more the local tones and colour differences in the subject itself. Though it was taken in situ at a university show, I was able to introduce some soft sidelight to emphasise the textures.
histD by Terry Andrews, on Flickr
The histogram of this Autumn show table, is hard to explain in terms of tones, as it is the bright colours that are emphasised in all their glory in the tangled colour histogram. When processing it, I had trouble with the colours clipping, and it was with some difficulty that I managed to retain detail in them all. Never the less the exposure is correct, the problem lay with the extreme colour gamut and saturation.
HistE by Terry Andrews, on Flickr
The last shot of the bright red Jeep also had colour gamut problems in the warm evening light. However apart from the strange colouration in the faces, the exposure just about captures the entire range of tones and highly saturated colours.
hist-F by Terry Andrews, on Flickr
None of these shots display the typically “Correct” histogram nor would I expect them to. However they all span in some fashion the range of tones from dark to light with little or no clipping.
It is more important to try to discover what a histogram is trying to tell you, than try to force it to fit some notion of correct exposure. Most examples of histograms shown to beginners are more simplistic than the ones that I have given, however they are also less representative of real images.
Histogram Addendum
Now that this thread has reached maturity, there are a couple of points that were raised that need clarification.
( I will also add this post back into the OP….. )
@HoppyUK said ” you claim "What they do not show is exposure" when they absolutely do - that is their main purpose and usefulness.”
Answer:
You can derive no Exposure information or settings from examining a histogram,
Two entirely different exposures and illumination, can show the exact same histogram.
A histogram does not take into account your intent, it only shows the tonal values that you have or are about to capture.
What it does do… is to give you a guide as to where those tone fall, and the direction that you would need to adjust exposure or exposure compensation to centralise those tones or avoid clipping.
It gives you no information about the magnitude or value of those adjustments.
Adjustments can only be made by trial and error and observation of the changed histogram.
The information a Histogram gives in this way, is indeed useful to correcting exposures
@DG phototraining said
“Nicely proving what a waste of time histograms are in general use”
and
“They may tell you what is clipped, but only you know if it matters or not, which is why "blinkies" are so much more useful if you follow the ETTR method”
Answer:
Centralising a histogram is especially useful to a beginner shooting Jpegs, who needs to get the image tones pegged, so as that the mid tone is correctly positioned. This may indeed clip some highlights and perhaps shadows, but will give the best representation of the subject matter with in the limits of an in camera jpeg.
More advanced users “shooting to the right” and using raw, will of course have more options and latitude, and can adjust the position of the tones to avoid clipping and maximise quality. However they will need to readjust the image during raw processing, to gain the maximum advantage. (I did not cover this in the original OP as it is not a “beginner” option.)
“Blinkies” are no more accurate at showing the degree of clipping, than the histogram itself, as both rely on the same data from the attached jpeg or screen capture, not the raw data. The main advantage of blinkies is their attention grabbing value. I would suggest that you get the most advantage if they are used in conjunction with the histogram.
I grant that some people seem to be Histogram phobic, and derive nothing from their use.
Thankfully that is not the case for the majority of photographers.
Last edited: