Measuring the image qualities/ Project 2: understanding the histogram

I touched upon histograms in DPP and learnt how useful they are when measuring the contrast of an image but this project made me look more closely at  how they can teach me what to expect from a scene. Tones squashed to the left indicate loss of shadow detail, squashed to the right-loss of detail in white/bright areas and if there are empty spaces at left/right the image would appear flat.

It was really intriguing to look at the images of Tibet in the materials folder and note which peak corresponded with certain tones.

Mount Kailash correctly exposed property of Open College of Arts © 2004

matching histogram for Mount Kailash property of Open College of Arts © 2004



Obvious areas identified property of Open College of Arts © 2004



Mount Kailash Underexposed property of Open College of Arts © 2004

corresponding histogram for under exposed image property of Open College of Arts © 2004

Project 2: understanding the histogram

This project introduced me to the CD for the first time, my task was to match six images to six histograms about which I felt fairly confident, having covered histograms previously and using them all the time in my photography.

I went for the most obvious first (well to me they were) the Oil jpeg was glass against a white background so I looked for a histogram with the graph where tones bunched mainly to the right. Histogram 5 matched this profile.



Oil Flask histogram

The white background and body of the glass oil jug is represented by the two peaks on the right on the histogram graph; the white background would be the peak furthest to the right showing that it is pure white with no detail at all. The table and light greys would be the smaller bump in the centre, with the dark part of the flask at the neck the smaller values running to the left hand side indicating darker tones were present but not to a great extent.




Carmen histogram



This image was also fairly straight forward to match with it's histogram, with a lot of black and white I was looking for a histogram with values peaking at both ends of the graph. I opted for histogram 2. Because the values touch at both ends it indicates that there is very slight clipping in the light and dark areas, but not overly so because the values are not that high, when you zoom into the image you can still see detail in the dress which is why the values on the right hand side is not high. Also whilst the background is very dark there is still detail in the man's suit. The mid grey tones are represented in the centre of the graph fairly evenly because the tonal match is very similar.



Pyramid histogram

For the pyramid image I was looking for a histogram with tones across the graph due to the deep blacks and white highlight in the doorway, however because the image has very little true black but a large tonal range of dark grey there should be a large group bunching towards the left. This was number 4. The sky/clouds and mid to light greys are there but to a much smaller extent so the graph is very low towards the right. 



Bluebell histogram





Next I decided to take a closer look at the bluebell shot. Straight away I could see no discernible pure blacks or whites. The background was fairly uniform and there was little contrast between the stems and the background. This was a rather flat image so I was looking for the histogram with information missing at both ends of the graph. Histogram 3 fitted the bill perfectly. The mid range tones of the image are represented by the pixels in the middle of the graph and as there are a large number the graph reaches the very top of the chart. The gentle sloping to either side indicates there are other ones represented but they are minimal.

Magnolia histogram



With only two images left it didn't really matter which one I opted for next and chose the white magnolia. On close inspection I could find no black or dark shadows whilst, to my eyes, the magnolia was pure white in places. The sky had no cloud detail and appeared to have uniform tones. This meant I was looking for a histogram that had tone details missing on the left where black would have been, with a high representation of pixels at the far right indicating the pure white detail. The sky would show as a tall fairly wide group of pixels in the middle of the x axis. Histogram 6 was the match with the other tones being shown by the gentle bumps either side of the mid tone peak.



Japanese girl histogram



More by accident than design I was left with the shot of the Japanese girl last. Looking at the image I was expecting the tones to fall off the edge of the graph but although they fill the entire length of the axis they peak just before the very end. This means although the image has a very dark background it is not pure black and will contain some detail still. There is very little clipping in the bright areas and the mid tones are represented consistently across the middle of the graph. I think I would have chosen this histogram for this image had I opted to match this first but would have had to check the other choices before committing because as I say I was expecting the peak to be at the very edge.

On checking the answers I was pleased to have got them all correct. I thought I had, but you never know. Completing this exercise has emphasised how much information you can glean from the on-camera histogram and how important a tool it can be for gaining correct contrast when shooting.

All images shown in this post for information and project purposes are the sole property of Open College of Arts © 2004 http://www.oca-uk.com/