Sunday, 17 March 2013

Histograms

I found this exercise to be more stimulating than I had imagined and was looking forward to exploring it further in the next part. I already had a fairly good idea of how histograms worked but had never considered where they could come in useful or how changing the exposure of a scene would effect them in relation to each other. I also discovered how to recognise how a high or low contrast scene was different to an average contrast one in terms of highlight or shadow clipping.

   The most insightful aspect for me was the trait that with the low contrast the shots I captured, the histograms resembled more of a 'hump' parabola. This was in contrast to the high contrast scenes, where the parabola was opposite forming a valley shape. Of course, as I bracketed the exposures the histograms changed for each of the three shots of the same scene so I've recorded my findings here.

1. High contrast - dark exposure
   The first high contrast image was the darkest exposure so it therefore had the most shadow clipping (which can be seen with the blue overlay). This correlated with the histogram, where the 'block' of values on the left of the histogram meant there were a lot of dark coloured pixels. There weren't many values on the extreme right (which I expected as it was the darkest of the three high contrast exposures). However it was interesting to see the middle of the histogram wasn't crowded at all, which made sense as most of the image was made up of dark and light-coloured pixels.



2. High contrast - medium exposure


Image 2 was quite similar although the whole histogram had been 'shifted' slightly across from left to right in comparison to the first image. Correspondingly, the shadow clipping grew less because some of those values had moved to the right.






3. High contrast - medium exposure





Image 3 continued with this trend but what it introduced was some highlight clipping because as the values moved from left to right some of them had 'hit' the right 'wall' of the histograms graph. These highlight clippings could be seen (barely) by the red overlay around some of the branches.





4. Low contrast - dark exposure
  


   In contrast to the three high contrast images I produced, the three low contrast images 'reversed' the parabola as mentioned above so the values remained within the sides of the histogram almost completely without much shadow or highlight clipping.
5. Low contrast - medium exposure







This was especially true of the medium exposure (Image 5) with the values distributed comfortably within the edges of the histogram and reaching a peak at the middle.

6. Low contrast - light exposure







Image 6 actually produced a tiny bit of highlight clipping as the values shifted to the right as I adjusted the exposure to a stop brighter. However, the histogram still resembled a 'hump' compared to the valley-like parabola of the high contrast images.






7. Average contrast - dark exposure




  7. quite closely resembled Image 3's histogram, although it was slightly less extreme. This was interesting because it showed the only area the two histograms changed were crucially the edges, where any highlight or shadow clipping would occur.






8. Average contrast - medium exposure

However there were a lot more mid-values in Image 8 as I increased the exposure by 1 stop. This meant the histogram lay somewhere between the high contrast and low contrast examples, which made a lot of sense.





9. Average contrast - light exposure

  



Image 9 basically replicated 8 but moved the whole histogram values to the right, while introducing a small bit of highlight clipping because of this.