@type alignment_t() :: :au | :nalu

Describes whether and how buffers are aligned.

:au means each buffer contains one Access Unit - all the NAL units required to decode a single frame of video

:nalu aligned stream ensures that no NAL unit is split between buffers, but it is possible that NALUs required for one frame are in different buffers

@type framerate_t() :: {frames :: pos_integer(), seconds :: pos_integer()}

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}

@type height_t() :: pos_integer()

Height of single frame in pixels.

Allowed values may be restricted by used encoding parameters, for example, when using 4:2:0 chroma subsampling dimensions must be divisible by 2.

@type nalu_in_metadata_t() :: boolean()

When alignment is set to :au, determines whether buffers have NALu info attached in metadata.

If true, each buffer contains the NAL units list under metadata.h264.nalus. The list consists of maps with the following entries:

• prefixed_poslen: {pos, len} - position and length of the NALu within the payload
• unprefixed_poslen: {pos, len} - as above, but omits Annex B prefix
• metadata: metadata - metadata that would be merged into the buffer metadata if alignment was :nal.

@type profile_t() ::
  :constrained_baseline
  | :baseline
  | :main
  | :high
  | :high_10
  | :high_422
  | :high_444
  | :high_10_intra
  | :high_422_intra
  | :high_444_intra

Profiles defining constraints for encoders and requirements from decoders decoding such stream

@type t() :: %Membrane.H264{
  alignment: alignment_t(),
  framerate: framerate_t(),
  height: height_t(),
  nalu_in_metadata?: nalu_in_metadata_t(),
  profile: profile_t(),
  width: width_t()
}

Format definition for H264 video stream.

Regardless of the alignment value, NAL units are always in the Annex B format.

In Annex B (defined in ITU-T H.264 Recommendation](http://www.itu.int/rec/T-REC-H.264-201704-I/en)) each NAL unit is preceded by three or four-byte start code (0x(00)000001) that helps to identify boundaries. Annex B is suitable for writing to a file or streaming with MPEG-TS.

@type width_t() :: pos_integer()

Width of single frame in pixels.

Allowed values may be restricted by used encoding parameters, for example, when using 4:2:0 chroma subsampling dimensions must be divisible by 2.