Photography Mechanics

There’s a lot to taking a photo.  I don’t have decent grasp on much of it, but I do know optics so I can write intelligently about what I call the mechanics of photography, which simply stated is how to make clean photos.  No, I’m not talking about the opposite of porn, I mean transferring the light from the object to the print without losing information.  This in no way means what you are photographing will be interesting in any way shape or form, especially if you’re taking a photo of me.

Lesson 1.  Pixels Don’t Mean Shit

Bought that 12 MP camera since it must be better than the 10 MP one, did ya?  Well in actuality you made things worse.  But that’s ok, don’t throw it away yet, you can throw it away at the end of the article.  There are many limiting factors to getting resolution in an image.  The lens, the amount of light, and the detector pixel count.  No lens is perfect so most will blur an edge over many pixels.  With low light, the image will appear grainy.  And with too few of pixels the image will look, well, pixelated.  The first two will be handled later, so let’s talk MPs.  How many do you need to not look pixelated?  Printing and Web photos will dictate this.  Click the image right to get the full rundown from West Coast Imaging.  The purple means “pristine prints” so you see you can print 8×12 with a 4 MP camera.  How about web photos?  A good screen is 1920 x 1080 which is about 2 MP.  So I recommend buying a 5 MP camera, which of course is impossible, so just go into your settings and change the saved image to about 5 MP to save disk space.

Lesson 2.  The Camera Equation

Yes an equation.  Did you really think I could talk about this without putting in an equation?  It’s a really simple one though.

Study this for a while.  You don’t have to calculate anything just know what happens if you increase or decrease each of the values.  That symbol, means ‘proportional to’.  Every value needs to be balance to bring in a reasonable pixel brightness.  Consider pixel brightness and the scene constant, then change one of the values.  An increase in exposure time requires a decrease in ISO, or an increase in f/#.   So if you have to balance exposure, ISO and f/#, what do you pick?  You pick the balance based on it’s effect on the picture.

Changing the exposure produces these effects, long exposure left, and short on the right.

Next in the equation is ISO.  This is the gain in the detector, or how much the signal is amplified.  More amplification = more noise.  Noise is bad.  Keep it below 800 for most cameras for decent pictures.

The Scene is essentially how much light is coming from the scene.  Use sunlight, reflected sunlight or flash to increase this number.

The f/#.  First off, F-number, F/#, F-stop, and aperture are all referring to the same thing; the focal length divided by the aperture diameter.  It is the cone of light that would image a single dot on the detector.  A fatter cone collects more light.

It’s worth noting that ONLY the cone angle dictates the amount of light, not just the aperture size.   Speed here is referring directly to the f/#.  ‘Fast glass’ is a low f/# lens.

The great thing about the f/# is that it has a big impact on the amount of light let in (since the value is squared), AND governs the depth of focus.  That’s why I shoot in aperture priority mode almost all the time.   I think photographers call it ‘Depth of Field’.  Basically, the narrower the cone of light, the easier it is for objects at different distances to still be in focus.  If you want a very short depth of focus, open the aperture up by using a lower f/#.  This produces ‘Boca’ which is the en-vogue term for blurring out the background as seen below.

Focal Length.  The distance from the aperture (or lens as shown in the above diagram) to the detector at focus.  The longer it is, the more you’re zoomed in, and the more it compresses the scene (in and out of the picture, not left or right).  That means a short focal length is a wide-angle lens, and gives the perspective that you’re right in the photo.  In the photos below the left was taken at f=10 mm and the right at f=30 mm.  Nothing was changed between photos other than time.  When it got dark I had to switch to my 30 mm lens which is capable of f/1.4 and lets in a lot of light without using flash.  The left is with a zoom 10-20 where the lowest f/# is 3.5.

Lesson 3.  RAW vs. JPG

Shoot in RAW.  Shoot in RAW.  Shoot in RAW.  I can’t say it enough.

I chose the below (left) picture because it had a serious deficit, either a blown out sky (saturated white) with an appropriately lit subject (Jason doing the skate-shoe ascent of a V6), or as I chose to do, light the sky appropriately and make the subject dark so I can edit later.  Note:  once the pixel has saturated white, you can’t get it back – it’s always at max.  As you can see from the right picture, it came out reasonably well.  This is no magazine photo since I didn’t capture Jason’s face, the people in the background confuse the subject and… well… a host of other issues – but it’s decent for a web photo!

Now let’s look at what would happen if I take the original, turn it into a jpg 1st (to emulate what the camera would normally do), then apply the same exact post processing edits I did for the above right version.

The result of the same edits is on the left.  That’s a big ballstron. See how the shirt is still black and the rock is looking blown out?  Also notice that the red crash pad is no longer very red, nor is Jenny’s lower shirt.  Also see that Jenny’s outer shirt is pure black but in the RAW edited photo you can see some white detail.  Just for good measure I re-edited the left jpg to the right just to make it look as good as I could get.  Better but there was a lot of information lost.  Exploring further, note how the bush recovered a significant amount of green in the shadow, but for the jpg edited photos that was lost and all you get is black.  Take a look at the rock detail… in the jpg versions you can hardly tell there’s any texture anymore other than the bigger puzzle looking features.

Why is this?  It comes down to bit level.  Every pixel has a certain quantized number of illumination levels.  For jpg you get 256 levels (8-bit or 2^8=256).  For arguments sake if any number of photons between 0 and 64 hit the pixel, you get a value of 0.  If 65 to 130 hit the pixel you get a value of 1.  Now compare this to RAW which for my camera is 14 bit or 2^14=16,384 levels!  So now you get 0-1 photons is a value of 0, 2 photons is a value of 1 etc.   So can your eye discern between 16,000 levels of light intensity.  Abso-fucking-lutely not.  It can discern about 256 actually.  Your eyeball is 8-bit.  So why use RAW?  As shown above you can edit while losing a lot less information, so now you can take the dark stuff and expand it.  In jpg 0-120 photons spans values of 0-1.  If you brighten and expand this you would simply get values like 50 and 52.  Nothing else.  In raw 0-120 gets 120 values.  Brighten and expand and you would get something like 50, 52, 54 etc up to 240.  Then when you output the FINAL image to jpg it will get compressed to 8-bit.  Make sense?  Hope so but for the less-than-technical I can see how this might be unclear.  Comment below and I’ll try to expand on this.  Perhaps the below picture helps too where the red lines would represent a single illumination level.

Lesson 4.  Lens Aberrations

This is whole field of study but I can narrow it down like this:

1. No lens is perfect.
2. More expensive lenses are closer to perfect.
3. Prime lenses (ones without zoom) are easier to make closer to perfect.
4. The higher the f/# the easier it is to make perfect.
5. The center of the picture will be closer to perfect.
6. Medium focal length lenses are easier to make perfect.