When considering the problem of audio-to-audio matching, determining musical similarity using low-level features such as Fourier transforms and MFCCs is an extremely difficult task, as there is little semantic information available. Full semantic transcription of audio is an unreliable and imperfect task in the best case, an unsolved problem in the worst. To this end we propose a robust mid-level representation that incorporates both harmonic and rhythmic information, without attempting full transcription. We describe a process for creating this representation automatically, directly from multi-timbral and polyphonic music signals, with an emphasis on popular music. We also offer various evaluations of our techniques. Moreso than most approaches working from raw audio, we incorporate musical knowledge into our assumptions, our models, and our processes. Our hope is that by utilizing this notion of a musically-motivated mid-level representation we may help bridge the gap between symbolic and audio research.