Filtering is crucial in the field of signal processing, widely applied in communication, electronics, and image processing. Its main function is to filter out interference noise and enhance the energy ratio of useful signals in sample data, thereby achieving a signal-to-noise ratio gain. The mathematical principle can be abstracted as a convolution operation, which is a common formula. In this issue, we still adhere to our original intention, minimizing the listing of formulas and striving to visualize the technical principles, making them easy to understand. We directly created dynamic demonstrations of the principles of time domain, spatial domain, and image domain filtering through Matlab simulations. Here is the video:The essence of time domain filtering is one-dimensional convolution. The video demonstrates the finite impulse response (FIR) filtering process. Of course, there are also infinite impulse response (IIR) filtering and adaptive filtering.Spatial domain filtering, which is beamforming, is demonstrated in the video based on the weighted summation of a linear array, showing the spatial filtering process and effects. It can also suppress interference noise at a specified direction beyond 75°.The last one is image domain filtering, which is essentially a two-dimensional convolution operation. Whether in traditional image processing or current artificial intelligence applications, image domain filtering is indispensable. Its computational load is relatively large and time-consuming; the video only demonstrates a portion of it.