0.2 Why flutter shutter

Why Fluttered Shutter?

The fourier transform of the boxcar filter, as shown in blue lines in the spectrum below, is a sinc function. The sinc function has zeros atcertain frequencies, and thus the information contained in those frequencies are lost entirely when the blurred image is generated. Also,the envelope of the sinc function decays rather quickly at high frequencies. We know that the noise has constant magnitude at allfrequencies. So when we try to deblur the image by deconvolution, the noise at high frequencies will be greatly amplified.

By using a fluttered shutter, which opens and closes during camera exposure according to a random binary sequence (a “1” corresponds to“open”, “0” to “close”), the fluttered shutter changes the boxcar filter to a broad-band filter that preserves high frequency spatialdetails, as shown in the above image in red-colored lines. Also, the flutter shutter filter has no zeros in its frequency response. Thisallows direct deconvolution to be possible without immediate destruction of image quality.