I'm writing to ask how the resolution compression algorithm works. In short, when converting a 1280x1024 video to 640x512 with the scale factor of 0.5, is the resulting video just throwing away 75% of the data, or are the final pixel values some average of the input? I'm using these videos to train ...