DFT filtering (text based image)
A video showing how the removing components from a DFT affects a text based image.
Top row: keeping only the X% most intense components of the DFT
Bottom row: keeping only the X% most intense components of the 8x8 blocked DFT
As less components are used to reconstruct the original image visual fidelity is reduced.
At around 25% of the orignal data (4x reduction in image size) the difference between the two methods becomes aparent. The block based approach sustains image quality better than for the full DFT.
At around 5% of the original data the blocking artefacts become very clear, perfectly highlighting the difference between the two methods.
Pro Tip: Use the left right angled brackets (on the keyboard) to advance the image frame by frame
See: • Visualizing the discrete fourier tran... for a tutorial on how these animations were created!
