Membrane.RawVideo (Membrane: Raw video format v0.4.3)

This module provides a struct (Membrane.RawVideo.t/0) describing raw video frames.

Summary

Types

aligned()

Determines, whether buffers are aligned i.e. each buffer contains one frame.

framerate()

Number of frames per second. To avoid using floating point numbers, it is described by 2 integers number of frames per timeframe in seconds.

height()

Height of single frame in pixels.

pixel_format()

Format used to encode the color of every pixel in each video frame.

t()

width()

Width of single frame in pixels.

Functions

frame_size(raw_video)

Simple wrapper over frame_size/3. Returns the size of raw video frame in bytes for the given raw video stream format.

frame_size(format, width, height)

Returns the size of raw video frame in bytes (without padding).

image_to_payload(image)

Converts Vix.Vips.Image.t/0 struct to raw video frame payload.

payload_to_image(payload, raw_video)

Converts raw video frame to Vix.Vips.Image.t/0 struct.

Types

aligned()

@type aligned() :: boolean()

Determines, whether buffers are aligned i.e. each buffer contains one frame.

framerate()

@type framerate() :: {frames :: non_neg_integer(), seconds :: pos_integer()} | nil

Number of frames per second. To avoid using floating point numbers, it is described by 2 integers number of frames per timeframe in seconds.

For example, NTSC's framerate of ~29.97 fps is represented by {30_000, 1001}. If the information about the framerate is not present in the stream, nil value should be used and the buffers described by this stream format must have non-nil timestamps.

height()

@type height() :: pos_integer()

Height of single frame in pixels.

pixel_format()

@type pixel_format() ::
  :I420
  | :I422
  | :I444
  | :RGB
  | :BGR
  | :BGRA
  | :RGBA
  | :NV12
  | :NV21
  | :YV12
  | :AYUV
  | :YUY2
  | :I420_10LE
  | :I420_10BE
  | :I422_10LE
  | :I422_10BE
  | :I444_10LE
  | :I444_10BE

Format used to encode the color of every pixel in each video frame.

t()

@type t() :: %Membrane.RawVideo{
  aligned: aligned(),
  framerate: framerate(),
  height: height(),
  pixel_format: pixel_format(),
  width: width()
}

width()

@type width() :: pos_integer()

Width of single frame in pixels.

Functions

frame_size(raw_video)

@spec frame_size(t()) :: {:ok, pos_integer()} | {:error, reason}
when reason: :invalid_dimensions | :invalid_pixel_format

Simple wrapper over frame_size/3. Returns the size of raw video frame in bytes for the given raw video stream format.

frame_size(format, width, height)

@spec frame_size(pixel_format(), width(), height()) ::
  {:ok, pos_integer()} | {:error, reason}
when reason: :invalid_dimensions | :invalid_pixel_format

Returns the size of raw video frame in bytes (without padding).

It may result in error when dimensions don't fulfill requirements for the given format (e.g. I420 requires both dimensions to be divisible by 2).

image_to_payload(image)

@spec image_to_payload(Vix.Vips.Image.t()) :: {:ok, binary(), t()} | {:error, term()}

Converts Vix.Vips.Image.t/0 struct to raw video frame payload.

Returns a tuple {:ok, payload, format} where payload is a binary representing the raw video frame and format is a Membrane.RawVideo.t/0 struct describing the format of the frame. Note, that framerate field in the returned format struct is always nil, however the stream may have a framerate associated with it.

Calls Vix.Vips.Image.write_to_binary/1 internally.

This function is available only if :image dependency is present.

payload_to_image(payload, raw_video)

@spec payload_to_image(binary(), t()) ::
  {:ok, Vix.Vips.Image.t()} | {:error, :invalid_pixel_format | term()}

Converts raw video frame to Vix.Vips.Image.t/0 struct.

It requires the pixel format to be :RGB.

Calls Vix.Vips.Image.new_from_binary/5 internally.

This function is available only if :image dependency is present.