Assignment 4: Reshoot - Form

My tutor got back to me on assignment 4 two weeks ago. Because I have to work on the Life in the UK exam, I can’t get around to reshoot some of the images.

This is the first tutor report I got from Alan. The comments are generally quite positive. To be honest, it is somewhat too positive for someone at my level. Somehow I wonder if this is how the British education system works. When I was young, I don’t remember getting a nice word unless your work is so stunning that it shocks the teacher. I have a look at the last assessment result in July. Only one person get above 80 for this course, which is consistent with what it was like in my Chinese composition class. My teacher used to say if you get above 90, that is Literature with a big “L”.     

Never mind, let’s focus on the negative comments.

Reshoot

There are three images I am looking at: Texture, shape and form. I will put them into three posts. So let’s start with the shape. Here is the original image.

  

This is Alan’s comment.

“This is a good angle and lighting plan to demonstrate the principle shape of your camera.  Unfortunately, although the manufacturer’s logo on the front panel is sharp, I am afraid that the rewind knob on the right and the lettering on the ring on the front of the lens have both drifted out of focus.   I suspect that the aperture was too wide open to create sufficient depth of focus.”

The re-shoot in theory is simple, but it takes a while to setup the subject, camera and light roughly the same to the original shoot. After some trial and error, here is at f/32, flash at full power (compare to the original image at around f/13).

We can see that the “Nikkor” letter is sharp, while there are more detail on the rewind knob. Still if we look at this at 100% corp, the knob is sharp but not exactly pin-sharp. I think there is limited depth of field (DOF). I will go into this in more discussion later. However, since the DOF is limited, I prefer to have the front “Nikkor” to be as sharp as possible instead of keeping the knob sharp. I think it is too easy to spot if the letter is not sharp.

DOF

I have to admit that depth of field is one of the thing I rarely pay attention of. Why? Because most of the photographs I took are sport related. Most of the time I need to shutter speed to be very high. Trying to keep the ISO noise to a manageable level, there is rarely any room to maneuver the aperture. To be honest, I am not sure how small I should set the aperture in assignment 3 when I do urban landscape piece. It is more of a guess work if this f-stop is good enough.

As a result I look at the DOF calculation on Wikipedia (http://en.wikipedia.org/wiki/Depth_of_field). The calculation is done by using thin lens equation in classical optics, combining with some geometry (similar triangles). I replicate the formulas and put it on a spreadsheet. Here are couple questions I am interested in:

1) What was the DOF of the setting of this shot?
2) Apart of from stopping down, what else can I improve on the DOF?

The first one is simple. At f=70mm, the subject is about 40cm away (near the minimal focus distance). Let’s say the distance is 50cm since the subject has depth. With the circle of confusion set at 0.02mm, I get 3.98cm with f/22 and 5.63cm at f/32. The subject depth (from the “Nikkor” to the knob) is about 10cm. Thus, it explains why the knob is not sharp enough even at f/22. f/32 barely makes it, but as I said earlier, the knob is not exactly pin-sharp.

The second question is more interesting. In the solution to the DOF calculation, there are 4 variables: focal length, subject distance, f-stop (which is the same as focal length/aperture diameter), and circle of confusion. I can adjust each of them to get more DOF. So what is the post and con?

Circle of confusion (CoC): The most foreign term is the fastest way to improve DOF. The CoC is defined as the diameter of a spot to spread to be considered to be unsharp. It has to do with the grain or sensor size. The larger the sensor grid (not the number of pixel, but sensor size divide by the number of pixel), the bigger CoC. Some says viewing condition as well, but it is less controllable. This is a short extract from Wikipedia, with the calculation of the DOF at f/22 and f/32.

Image Format	Frame size	CoC (mm)	DOF at f/22 (mm)	DOF at f/32 (mm)
APS-C Nikon	15.7 mm × 23.6 mm	0.019	40	56
35mm Full Frame	24 mm × 36 mm	0.029	58	82
6×6 Medium Format	56 mm × 56 mm	0.053	106	152
8×10 Large Format	203 mm × 254 mm	0.22	540	999

One can see that, while I only have 5.6cm to maneuver in an APS-C frame, I will have three times more in a 6x6 medium format. Simple! However, as Henry Louis Mencken said, “For every complex problem, there is a simple answer – and it’s wrong.”, here is the price list:

Image Format	Example	Price (GBP)
APS-C Nikon	Nikon D7000	559
35mm Full Frame	Nikon D600	1,429
Medium Format	Hasselblad H4D-60	22,962
Large Format	Deardorff View Camera V8 (film)	3,795

The Hasselblad H4D is not a 6x6, but something like a 4x5. It costs 41-times more than my D7000. There is no digital large format camera around, so I find a film one, it still cost 6 times more than what I have. Then, I also have to process my own film because there is no shop that does large format film processing.

Focal Length and Subject Distance: These two go together. Unless I corp, I cannot move the subject further away and maintain the composition. Assume that I don’t corp, then I have to increase the focal length as I move the subject further away. Re-run this calculation, this is what I get.

Focal Length (mm)	Subject Distance (mm)	DOF at f/22 (mm)
70	500	38.67
135	1,000	41.79
200	1,500	42.91

I am not sure if this works at all, given that I am not getting even 1cm worth of improvement.

Aperture: I went all the way down to f/32. I am glad that my lens is old enough to have f/32. Most modern lens only got as low as f/22. If somebody is making an f/64 lens, I can put it in and get a DOF of 109mm. Here is a problem: someone told me that with small sensor camera (like mine), I have to worry about diffraction when the f-stop goes below f/16. I am not sure how to check diffraction pattern in an image. I can’t see anything very exciting at f/32, but will it be the same at f/64?

The Airy disk function is just a 2D Fourier transform of a circular aperture. Adding two together side by side will give me an idea on how close they have to be for the sensor to fail to detect two peaks. To be honest, I haven’t done math for a while. Therefore, I used the online calculator from Cambridge in Colour (http://www.cambridgeincolour.com/tutorials/diffraction-photography.htm). At f/32, the diameter of the Airy dist is only 41 micron, larger than the CoC of 21 micron. A quick calculation with the thin lens equation assuming that my maximum resolution at the sensor is 41 mircon, with 70mm focal length, at f/32, subject at 50cm, it implies that I will not be able to resolve two points that is 0.252mm apart (only at the 50cm plane).

For the sharp shot, I have to say that I don’t care about a resolution of 0.252mm. It is probably more important if I am doing the texture shot, when the small details make all the difference in the texture.

Anyway, that’s all for the sharp. This analysis is simply too tiring. I am sure I will never finish this course if I keep doing this. I will write less for the next two re-shoots.