libswresample

libswresample is a library within the FFmpeg project that provides software-based audio resampling along with conversion between sample formats and channel layouts. It is designed to let applications convert audio data to a different sample rate, bit depth, or speaker configuration, independent of the surrounding decoding or encoding steps.

The core API centers on the SwrContext, an opaque object that holds the resampling state. Typical usage involves creating and configuring the context with swr_alloc() or swr_alloc_set_opts(), initializing it with swr_init(), and performing the conversion with swr_convert(). Additional helpers such as swr_get_delay() help account for processing delay, and swr_free() releases resources. libswresample supports both planar and interleaved data and accepts a range of AVSampleFormat values (for example AV_SAMPLE_FMT_S16, AV_SAMPLE_FMT_FLTP).

libswresample can handle arbitrary input and output sample rates, formats, and channel layouts, performing sample rate conversion, format remapping, and channel reordering as needed. It uses high-quality resampling algorithms, including polyphase filtering, to minimize artifacts while producing the requested output configuration. The library also accommodates delay and buffering concerns to maintain audio continuity across multiple processing steps.

In the FFmpeg ecosystem, libswresample is used by decoders, encoders, and filters to ensure audio data is in the correct form for subsequent processing or playback. It is distributed as part of the FFmpeg libraries and serves as the primary resampling component, replacing older libavresample usage in many contexts.