Struct CodecCapabilities 

#[non_exhaustive]
pub struct CodecCapabilities {

    pub decode: bool,
    pub encode: bool,
    pub media_type: MediaType,
    pub intra_only: bool,
    pub lossy: bool,
    pub lossless: bool,
    pub hardware_accelerated: bool,
    pub implementation: String,
    pub max_width: Option<u32>,
    pub max_height: Option<u32>,
    pub max_bitrate: Option<u64>,
    pub max_sample_rate: Option<u32>,
    pub max_channels: Option<u16>,
    pub priority: i32,
    pub accepted_pixel_formats: Vec<PixelFormat>,
    pub accepted_sample_rates: Vec<u32>,
    pub accepted_sample_formats: Vec<SampleFormat>,
    pub embedded_blob_kinds: Vec<BlobKind>,
}

What an implementation can do plus how it ranks vs alternatives.

#[non_exhaustive]: build it with audio, video or subtitle and the with_* builders, then read the public fields. Fields are added over time (0.1.38 added the accepted sample rates / formats) and a struct literal in a downstream crate would stop compiling each time.

Fields (Non-exhaustive)

Non-exhaustive structs could have additional fields added in future. Therefore, non-exhaustive structs cannot be constructed in external crates using the traditional Struct { .. } syntax; cannot be matched against without a wildcard ..; and struct update syntax will not work.

decode: bool

Decoding supported by this implementation.

encode: bool

Encoding supported by this implementation.

media_type: MediaType

Media type this implementation handles (audio, video, …).

intra_only: bool

Every coded unit is independently decodable (no inter-frame prediction).

lossy: bool

Supports lossy compression.

lossless: bool

Supports lossless compression. lossy and lossless may both be set for codecs that offer both modes.

hardware_accelerated: bool

Hardware-accelerated implementation (VAAPI/NVENC/QSV/VideoToolbox/…).

implementation: String

Short identifier for this implementation, e.g. “flac_sw”, “h264_qsv”.

max_width: Option<u32>

Restrictions — None means “no constraint”.

max_height: Option<u32>

Maximum supported frame height in pixels; None = unconstrained.

max_bitrate: Option<u64>

Maximum supported bit rate in bits per second; None = unconstrained.

max_sample_rate: Option<u32>

Maximum supported audio sample rate in Hz; None = unconstrained.

max_channels: Option<u16>

Maximum supported audio channel count; None = unconstrained.

priority: i32

Lower numbers are preferred. HW impls should be ~10, SW impls ~100.

accepted_pixel_formats: Vec<PixelFormat>

Pixel formats this implementation accepts (video only). An empty Vec means “any format” — resolution won’t filter on it. When populated, the registry can skip impls whose accepted set does not include the format requested by the caller.

accepted_sample_rates: Vec<u32>

Audio sample rates (Hz) this implementation accepts as encoder input (audio only). An empty Vec means “any rate”. When populated, a pipeline feeding the encoder is expected to resample a stream whose rate is not listed to one of these rates (the listed order is not a preference — consumers pick the closest rate that does not discard bandwidth).

accepted_sample_formats: Vec<SampleFormat>

Audio sample formats this implementation accepts as encoder input (audio only). An empty Vec means “any format”. When populated, the first entry is the preferred conversion target for a stream whose format is not listed.

embedded_blob_kinds: Vec<BlobKind>

Metadata blob kinds this implementation’s encoder embeds in its output: the BlobKinds it reads from the input stream’s CodecParameters::blobs and from frame records and writes into the coded output (an ICC profile into the codestream’s colour box, an Exif block into its application segment, a format’s own custom kind into its native chunk). Empty — the default — means “nothing declared”: a caller cannot tell whether a blob it hands over survives, and a gateway reports that as pass-through. When populated it is a promise in both directions: every listed kind is embedded when present on the input, every unlisted kind is dropped. Lets a caller answer “will the ICC profile survive this save?” before encoding a pixel. Attached with with_embedded_blob_kind / with_embedded_blob_kinds, queried with embeds_blob. Decoder-only registrations leave it empty.

Implementations

impl CodecCapabilities

pub fn audio(implementation: impl Into<String>) -> Self

Construct a software audio decoder/encoder capability set with sensible defaults — adjust fields after creation.

pub fn video(implementation: impl Into<String>) -> Self

Construct a software video decoder/encoder capability set with sensible defaults — adjust fields after creation.

pub fn subtitle(implementation: impl Into<String>) -> Self

Construct a software subtitle decoder/encoder capability set with sensible defaults (intra-only, lossless) — adjust with the with_* builders.

pub fn with_media_type(self, media_type: MediaType) -> Self

Override the media type (for capability sets built from the closest constructor, e.g. data or attachment streams).

pub fn flag_string(&self) -> String

6-character capability flag string (see the module docs for the column layout). Useful for oxideav list-style output.

pub fn with_decode(self) -> Self

Mark this implementation as supporting decode.

pub fn with_encode(self) -> Self

Mark this implementation as supporting encode.

pub fn with_intra_only(self, v: bool) -> Self

Set the intra-frame-only flag.

pub fn with_lossy(self, v: bool) -> Self

Set the lossy-compression flag.

pub fn with_lossless(self, v: bool) -> Self

Set the lossless-compression flag.

pub fn with_hardware(self, v: bool) -> Self

Set the hardware-accelerated flag.

pub fn with_priority(self, p: i32) -> Self

Set the registry ranking priority (lower is preferred; HW ~10, SW ~100).

pub fn with_max_size(self, w: u32, h: u32) -> Self

Constrain the maximum frame size to w × h pixels.

pub fn with_max_bitrate(self, br: u64) -> Self

Constrain the maximum bit rate (bits per second).

pub fn with_max_sample_rate(self, sr: u32) -> Self

Constrain the maximum audio sample rate (Hz).

pub fn with_max_channels(self, ch: u16) -> Self

Constrain the maximum audio channel count.

pub fn with_pixel_format(self, fmt: PixelFormat) -> Self

Add one accepted pixel format. Appends — call multiple times to list several.

pub fn with_pixel_formats(self, fmts: Vec<PixelFormat>) -> Self

Replace the accepted pixel-format set wholesale.

pub fn with_sample_rates(self, rates: Vec<u32>) -> Self

Replace the accepted encoder-input sample-rate set wholesale (see Self::accepted_sample_rates).

pub fn with_sample_formats(self, fmts: Vec<SampleFormat>) -> Self

Replace the accepted encoder-input sample-format set wholesale (see Self::accepted_sample_formats; the first entry is the preferred conversion target).

pub fn with_embedded_blob_kind(self, kind: impl Into<BlobKind>) -> Self

Add one metadata blob kind the encoder embeds (see Self::embedded_blob_kinds). Appends — call several times to list several; a kind already listed is not added twice.

pub fn with_embedded_blob_kinds(self, kinds: Vec<BlobKind>) -> Self

Replace the embedded blob-kind set wholesale (see Self::embedded_blob_kinds). Duplicates are removed, first occurrence kept.

pub fn embeds_blob(&self, kind: &BlobKind) -> bool

true when the encoder declares that it embeds blobs of kind (see Self::embedded_blob_kinds). false both for an unlisted kind and for an encoder that declares nothing — use declares_embedded_blobs to tell the two apart.

pub fn declares_embedded_blobs(&self) -> bool

true when the encoder declares its embedded blob kinds at all, i.e. Self::embedded_blob_kinds is non-empty; false means every answer from embeds_blob is “unknown”, not “no”.

pub fn pick_sample_rate(&self, rate: u32) -> u32

Pick the encoder-input sample rate for a stream at rate Hz: rate itself when it is accepted (or every rate is), otherwise the smallest accepted rate above it (no bandwidth is lost), and failing that the largest accepted rate.

impl CodecCapabilities

pub fn fits_params(&self, p: &CodecParameters, for_encode: bool) -> bool

Whether this implementation’s max-* restrictions are compatible with the requested codec parameters. for_encode is reserved for restrictions that apply asymmetrically. Used by the registry’s make_decoder / make_encoder walker and by out-of-tree selection layers (e.g. oxideav-pipeline’s CodecPreferences filter).

Trait Implementations

impl Clone for CodecCapabilities

fn clone(&self) -> Self

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for CodecCapabilities

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl Display for CodecCapabilities

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.