Home | News
Lesson Materials
Course Information
Resources

Assignment, Week 2/3: Techniques to Reduce Camera Shake

For a bonus assignment, see Week 3 Bonus Assignment

As you may have noticed from the photos from Week 1 assignment (.zip, 24.3MB), shooting in low-light leads to a few problems. One of the biggest is blur. For a primer on blur, see the course materials.

For a quick summary here, blur is essentially caused when the sensor does not consistently record the same light. Imagine a light source shining directly on an imaging sensor (similar to what I tried to demonstrate in class). Ordinarily, the light is not moving and the sensor is not moving, so we have perfect consistency and end up with sharp images. However, if the sensor were to move, light that was previously falling on one location of the sensor is now falling on another location. Thus, the light isn't recorded at only one point, but is instead recorded over any new area the sensor moves to. This is what occurs when the camera isn't held steadily, and is the source of image blur from camera shake. Conversely, the light source itself could shift position - a person in the photo could be moving for example, thus shifting the light that falls onto the sensor and creating the same sort of blur.

For this first week, we will concentrate on image blur from camera shake. There are numerous technologies and techniques that can help compensate for blur, and this week we'll focus on techniques.

Holding technique: With the advent of digital cameras with rear LCD screens, it's become customary to hold most digital cameras at arms-length, composing and viewing through the LCD. However, this provides the least amount of support for the camera (especially if used one-handed), compared to more traditional grips such as holding up the camera to one's face and shooting through a viewfinder.
Second image vs. first image: When taking a photo, pressing down on the shutter release requires a deliberate force to push down on the button, oftentimes pushing the camera downwards at the same time. This shakes the camera, and oftentimes the first image is affected by this. If a second image is taken in a row after that, however, this second image should not experience any of the shake effects of pressing the shutter button.
Delayed shutter vs. immediate shutter: Along the same vein of pressing the shutter release causing the camera to shake, another way to get around this is to have a delayed shutter. Many cameras have a 10-second or 2-second time, where you can press the button, but the camera won't actually take the picture until several seconds later, avoiding the shake from pressing the shutter button.
Multiple shots: One of the overriding themes of this course (which we actually haven't touched on yet) is statistical probability. Just because your first shot was blurry doesn't mean it's impossible to get the shot. Try taking a series of shots. Chances are, your next shot may easily be even blurrier than your first one. That's okay, take another, and this time it may be less blurry (or not). Like many things, the amount of blur for any given image will have some statistical fluctuation, and with enough samples (taking enough pictures), statistically you should end up with a usable image with acceptable blur, even if it takes 100 tries to get there.
Wider focal length vs. longer focal length: Focal length (also angle of view) will need a little explaining later on, but simply think of this as a small "zoom" vs. a very big "zoom" (for those of you who actually know the terminology, we are comparing short to long focal lengths). Bigger zooms tend to amplify the amount of shake - imagine holding out a really big meterstick at arms length, and then imagine holding a short pencil - with each movement of your arm, you might shift the position of the pencil, but the really long meterstick shifts a whole lot more.

More expansive explanations, and other techniques are detailed under Week 3 lesson plan.

For this assignment, no photo submissions are necessary. The goal is to test the relative effectiveness of techniques, and we do this by taking a series of pictures, and checking how many have acceptable blur (your subjective decision). For example, if you were to test second image vs. first image, you might take 10 series of two shots each (ten 1st shots, ten 2nd shots), then email me (nathanyan@berkeley.edu) telling me that 4 out of 10 first shots were acceptable, while 6 out of 10 second shots were acceptable, for example. The five choices for tests are listed above, and you can also test other technqiues you can think of (email me first though).

No photo submissions are necessary, although if you send me photos I can use them as examples in class. Feel free to email if you have any questions about the testing procedure.

Assignment Summary: