This article presents new methods for performing two video post-production practices directly on subband compressed images, i.e., without decoding them first. The first post-production operation is composition of two images, i.e., realizing picture-in-picture. If one simply composes the coded subbands of two images, ringing and blurring occur because of the spatially rapid switch from one image to the other. To solve this problem, adaptive box borders are introduced that greatly reduce the ringing. The second post-production operation is overlaying text directly into a subband coded image. Here, only a small section of the image is decoded where the text is to be located, the text is overlaid into this small ima ge fragment, subbands of the fragment are obtained and seamed into the original subbands, and the modified fragment is quantized