The goal of the Two!Ears project is to develop an intelligent, active computational model of auditory perception and experience in a multi-modal context. The Auditory front-end represents the first stage of the system architecture and concerns bottom-up auditory signal processing, which transforms binaural signals into multi-dimensional auditory representations. The output provided by this consists of several transformed versions of ear signals enriched by perception-based descriptors which form the input to the higher model stages. Specific emphasis is given on the modularity of the software framework, making this more than just a collection of models documented in the literature. Bottom-up signal processing is implemented as a collection of processor modules, which are instantiated and routed by a manager object. A variety of processor modules is provided to compute auditory cues such as rate-maps, interaural time and level differences, interaural coherence, onsets and offsets. An object-oriented approach is used throughout, giving benefits of reusability, encapsulation and extensibility. This affords great flexibility, and allows modification of bottom-up processing in response to feedback from higher levels of the system during run time. Such top-down feedback could, for instance, lead to on-the-fly changes in parameter values of peripheral modules, like the filter bandwidths of the basilar-membrane filters. In addition, the object-oriented framework allows direct switching between alternative peripheral filter modules, while keeping all other components unchanged, allowing for a systematic comparison of alternative processors. Finally, the framework supports online processing of the two-channel ear signals.