Showing posts with label Conclusions. Show all posts
Showing posts with label Conclusions. Show all posts

Tuesday, 10 May 2011

Project 3: a variety of lighting conditions

A similar exercise was required at the beginning of DPP and I was surprised when images I thought were going to produce a low-key image actually appeared to be flat, meaning the shadows were not as dark as I had anticipated. Also some I considered to be low-contrast resulted in full scaled images so it will be interesting to see if my judging by eye has improved. A decision was taken to try and be more adventurous with my shots, previously (mainly due to the horrible weather) objects/areas in my back garden were used just so the exercise could be completed without much thought into the artistic composition. Project 3 has been designed to familiarise you with histograms. The aim was to shoot 5 different images under different lighting conditions which would produce the expected histograms to meet the criteria of full range, high contrast, low contrast, high key and low key. The resulting histograms were then annotated to show why they looked as they did.

Being the beginning of Spring I thought I would try and theme the images to a walk in my local bluebell woods (did cheat with the aquilega - that was in my garden) rather than stage manage setting I knew would meet the instructions. Find below the resulting images.

High Contrast

The scene below was chosen due to the bright sky, which I hoped would provide the high contrast of white tones and the shadows/trees for the darker tones. Again I was anticipating the darker end of the scale to be slightly closer to the end but am guessing this is due to the image not being under-exposed at all. It is still a fairly good representation of a high contrast image with pixels peaking at either end of the scale with the other tonal ranges there but fairly low.




Low Contrast/Flat

This was a little trickier to shoot due to the lighting conditions in the wood providing very dark shade for close-up shot and getting too much "black". Persistence paid off and I found a clump of bluebells in dappled shade, the peak could have been more towards the middle of the scale but with no areas of black or white and no clipping this image is of low contrast.


Full Range

I gauged by eye that this scene should deliver a full range histogram, the shadows were not too deep and the direction of the sun meant the bright sky was not blown. The mid tones in the photograph provide the gradual slope which would possibly be more ideal towards the middle of the chart. However the tones do run from end to end and this is a fair example of a full range image.



High Key

I was not sure if this shot would tonally represent a high key image due to the green in the background. However there is a large number of bright tones indicated by the pixels "climbing" the right hand side of the graph and the majority of the other tones are on the right hand side. Dark tones are there but they are minimal.



Low Key

Whilst out in the local woods I was lucky enough with the weather conditions and the timing to obtain this sunset shot complete with young girl climbing the old mulberry tree. I thought that the histogram values would be more towards the left and the darker tones higher than they were but the black tones are squashed to the left and the rest of the tones are towards the left hand side.


Conclusion

This exercise proves that although at first we assume or are sometimes taught, that the full range histogram is what we expect from a "perfect" image depending on the colours or rather tones in the photograph this isn't always going to happen.The histogram will alter its shape it due to the nature of the scene. Problems only occur when you have anything more than the thinnest of lines line butting up to the extreme left or right sides; indicating excessive black or white. If there is a peak near either side, it just means there’s lots of dark or light tones – which is not a problem.

Although this exercise was using the "luminosity" histogram, my camera is set to use the RGB histogram. This is due to the luminosity histogram not always telling the whole story, it shows an average for the three channels and if a particular colour is clipping (for example the sky) this maybe missed and the exposure not altered to compensate. I found this webpage which helps simplify and gives examples of the RGB histogram and breaks it down.

Sunday, 8 May 2011

Project 1: dynamic range

Camera manufacturers don't like to divulge the dynamic range of their cameras therefore I must endevour to discover it for myself. Dynamic range of the sensor is the difference between the darkest shadows and the brightest highlights that it can capture in one scene without losing detail. If detail is lost this is known as clipping. Within RGB solid black measures 0,0,0 and solid white 255,255,255.

Handy table to remember the f-stop increments...am going to print this out and take it with me!

Shutter
speed
1/8
1/15
1/30
1/60
1/125
1/250
1/500
Aperture
F2.8
F4
F5.6
f.8
F11
F16
F22


Sometimes I think the written instructions are actually more confusing than the task itself...as an LSA and having to write or make information very basic I do double take and think why didn't it just say X? Calculating the DR in DPP was the task I put off for ages, now realising it wasn't that hard to do I wasn't so fussed about this one.
To start- I had a plain white piece of card in daylight. ISO set at its lowest,image size smallest.Used Custom WB and Manual exposure.Filled the frame with the card, adjusted shutter speed for average exposure.Noted the setting. Next I calculated a setting that was 6 f-stops darker by altering the aperture, however due to the lens I was using I found I couldn't change enough stops so increased/decreased exposure by using shutter speed instead. I also had a minor problem with the lighting, being a cloudy day I had to wait for the sun to come back out from behind the clouds to continue enabling me to have consistent lighting conditions. Then it was windy and my card kept falling over!

Apart from that this was fairly straight forward to complete; the mid point metered to F8 @ 250 however, even adjusting my shutter speed I couldn't change down more than 4 stops;  progression was f8 @ 1/500, 1/1000/ 1/2000 and finally 1/4000. In the opposite direction I had no difficulty in the shutter speed f8 @ 1/125, 1/60, 1/30, 1/15,1/8,1/4.

The images were then opened in Photoshop and out of curiosity I took RGB values before converting to greyscale. The black point read (8.8.8) the mid range (110,114,125) which suggests the metering wasn't spot on as mid grey should be as near to (128,128,128) as possible, and the white point did hit (255,255,255)

The ascending scale was the created and here is the resulting chart.

Dynamic Range Canon 400D

Histogram showing tonal peaks


The image was cropped to stop at the black and white points. Once converted the (8,8,8) read as (3,3,3) which I though was quite strange? The mid point (115,115,115) which seems to have averaged out the original readings, and the white point remained the same (255,255,255)

Counting the steps including the cut off points (9,9,9) and (255,255.255) - I didn't count the (3,3,3) due to the fact anything below (5,5,5) would be black with no viewable detail, it would appear that my camera has an 8 stop range which is quite normal. The histogram also appears to reflect this. When testing the dynamic range of my camera for DPP the scene results gave a dynamic range of 7. This test is a little more complex and gives a range of 8. However due to the fact that the lowest point shifted when converted, it may possibly fall between 8/9 and the following website found that the 400D produced a consistent dynamic range of 8.4 with ISO 100-800 and 7.8 at ISO1600.

http://www.dpreview.com/reviews/canoneos400d/page19.asp

This reassures me that I carried out this task correctly (even if my metering was a bit off) and my camera is acting in the way it should. It was also good to note that my mid-point is midway on the chart so should avoid burnt-out highlights and blocked-up shadows. Armed with this information I should be able to judge how much information I can caputure in a given scene, especially one with high contrast.