The Kowalski project aims to provide a data driven, lightweight cross platform audio solution. The target audience is developers of games and similar applications where real time audio plays an important role. The Kowalski engine, which is the runtime component, relies only on host-specific external libraries to pass the final output buffers to the audio hardware. All other processing is done in the engine code. Features include positional audio (distance attenuation, cone attenuation, Doppler shift, binaural panning), real time Ogg Vorbis decoding, a powerful mix bus system, and tools to build, validate, and view Kowalski data.
FFSampledSP is a free implementation of the javax.sound.sampled.spi interfaces. It is part of the SampledSP collection of javax.sound.sampled libraries. Its main purpose is to decode audio from various formats at high speed using FFmpeg. To do so, it uses the tagtraum FFmpeg package, which offers FFmpeg binaries configured to decode audio while avoiding CODECs with patent problems. In other words, it supports formats like Ogg and FLAC well, while leaving MPEG-4 and MP3 unsupported.