Commercial software synthesizer programming interfaces are usually complex and tedious. They discourage novice users from exploring timbers outside the confines of “factory presets,” and they take expert users out of their flow state. We present a new interface for programming synthesizers that enables both novices and experts to quickly find their target sound in the large, generative timber spaces of synthesizers. The interface does not utilize knobs and sliders that directly control synthesis parameters. It instead utilizes a query-by-example based approach combined with relevance feedback and active learning to create an interactive, personalized search that allows for exploration. We analyze the effectiveness of this new interface through a user study with four populations of varying types of experience in which we compare this approach to a traditional synthesizer interface. [Work supported by the National Science Foundation Graduate Research Fellowship.]