Survey of Audio Feature Representation using Encoding Techniques

Digital music has become prolific in the web in recent decades. Automated recommendation systems are necessary for users to discover music they love and for artists to reach suitable audience. When manual annotations and user preference data is lacking (e. g. for new artists) these systems must rely on content based methods. Besides powerful machine learning tools for classification and retrieval, a key elements for successful recommendation is the audio content representation. Good representations should catch informative musical patterns in the audio signal of songs. These representations should be to the point, to enable efficient (low storage, easy indexing, fast search) management of huge music repositories, and should also be easy and fast to assess, to enable real-time interaction with a user supplying new songs to the system. Before designing new audio features, the usage of traditional local features are explored, while adding a stage of encoding with a pre-computed codebook and a stage of pooling to get compact vectorial representations.