AVSAMPLEFMTFLTP

AVSAMPLEFMTFLTP, in FFmpeg terminology more commonly written as AV_SAMPLE_FMT_FLTP, denotes the planar 32-bit floating-point audio sample format. It is part of FFmpeg’s libavutil and is used to describe how audio samples are stored and accessed in memory.

Planar versus packed data

The FLTP designation indicates planar storage: one separate buffer (plane) per channel. For a multi-channel audio frame, there will be as many data planes as there are channels, with each plane containing the samples for a single channel. This is in contrast to packed (interleaved) formats, such as AV_SAMPLE_FMT_FLT, where samples from all channels are interleaved in a single buffer. Planar formats can simplify per-channel processing and mixing, while packed formats can be more cache-friendly for certain pipelines.

Data representation and usage

Each sample in AV_SAMPLE_FMT_FLTP is a 32-bit floating-point value. In practice, samples are typically normalized around a -1.0 to 1.0 range, though intermediate processing can generate values outside this interval. This format provides high dynamic range and precision, which is advantageous for high-fidelity audio work and certain decoding or transcoding workflows.

Practical considerations

In FFmpeg APIs, a frame using AV_SAMPLE_FMT_FLTP will expose one data plane per channel through the frame’s data pointers. Helpers such as av_sample_fmt_is_planar and av_get_sample_fmt_name help query whether a format is planar and obtain a human-readable name. AV_SAMPLE_FMT_FLTP is commonly used in floating-point pipelines, including many modern decoders and audio processing tools.

Related formats

Related FFmpeg sample formats include AV_SAMPLE_FMT_FLT (packed 32-bit float), AV_SAMPLE_FMT_S16, AV_SAMPLE_FMT_S32, AV_SAMPLE_FMT_DBL and their corresponding planar variants such as AV_SAMPLE_FMT_S16P, AV_SAMPLE_FMT_S32P, and AV_SAMPLE_FMT_DBLP.