Social contexts are essential for human decision-making, providing the basis for inferences and activities planning. Computationally, this social information is incorporated to the concept of social computing, and applied in very different research fields, such as Medicine, Cyber Security or Economy. The specific aim of this work is to apply social computing in a musical context. We want humans and machines to cooperate for the purpose of generating music from the colors of a digital image; the colors of an image are used as inspiration for the generation of melodies. The social machine is implemented as a Virtual Organization (VO), which contains a case-based reasoning process to select the best combination of sounds extracted from the colors according to the social context. In order to evaluate the present approach, a small society of musical experts rate the sound produced following consonance criteria.