Assignment, Week 3: Shake reduction methods

In our review of low-light photos, one of the chief problems you may have encountered in low-light situations is image blur. Generally, blur is caused by inconsistency in where light falls onto the sensor - this might be caused by the sensor itself moving and changing its position in space, or by the subject itself moving, and thus changing where the light falls onto the sensor.

We'll focus on blur due to camera shake first. There are basically three degrees of freedom - three ways that you can "shake": rotationally (pivoting around the camera's center), laterally (side-to-side, or up-down), and fore-aft movements (shifting the camera forward and backward). You can see a more rigorous definition in the powerpoint presentation, but in most cases the lateral and fore-aft movements are negligible in comparison to the rotational movements, and thus we can focus on the rotational shake, which can be boiled down thus:

As can be seen from this equation, the amount of blur you get (the unit for the above is image lengths, or blur as a % of the entire image) is a function of three things: average rotational speed (how fast you shake), exposure time (shutter speed), and angle of view (a function of focal length and the crop multiplier of your sensor). This jives with conventional logic - if we shake the camera faster, we get more blur. We tend to get more blur with longer shutter speeds. And if you use a larger angle of view (when you "zoom out"), you tend to have less blur. An important point to note is that each of these factors are multiplicative, rather than additive - zeroing one will zero the total amount of blur. So if you had a zero rotational speed, it doesn't matter how long your shutter speed is, or how narrow your angle is - the blur will be zero. Similarly, with an extremely fast shutter speed (say 1/10,000s - close enough to zero), we also zero the blur, no matter how fast we shake the camera.

There are several methods you can try out to reduce the amount of blur in your photos. This week will focus on the rotational speed factor - basically, ways to shake the camera less.

For this week, the assignment will be a test of one of the methods (second shot, breathing, angle of view, or improvised tripod) outlined above. Take a few shots (~10) as you normally would, then take a few shots using one of the shake reduction methods. For each set, count the number of shots that came in "acceptable sharp" (your criteria), and post your results in the forum:

/forums/viewtopic.php?id=47

You may also try the multi-shot method, in which case you'll take as many shots as you need to get an acceptly sharp shot, report that number (good shots/shots taken). For the gallery, upload at least three photos: a typical shot from your series (take the first shot, or pick one at random), your best (least blurry) shot, and your worst (most blurry).

My general idea is for you to push the limits of your photography skills in dealing with blur - "zoom in" to super-long focal lengths, and take long 1s, 2s exposures in the dark, where you'd typically get insane amounts of blur, but see if you can utilize the shake reduction techniques discussed above to come out with a usable shot.

Assignment summary