The Higher Order Ambisonic (HOA) technology is now very popular, thanks to the rise of immersive media. However, from a production perspective, some fundamental operations remain difficult to perform in the HOA framework. In this paper, we focus on the challenge of editing HOA sound scenes, by which we mean amplifying or attenuating sound for specific directions, or extracting the sound incoming from a given direction and moving it to another direction. We describe an algorithm for editing HOA sound scenes, which is based on a plane-wave decomposition. The performance of this algorithm is assessed through numerical simulations. In particular, we examine issues arising when editing HOA sound scenes recorded with a microphone array.