105 lines
3.3 KiB
Elixir
105 lines
3.3 KiB
Elixir
defmodule MPEGAudioFrameParser do
|
|
@moduledoc """
|
|
This is the public API for MPEGAudioFrameParser application.
|
|
|
|
MPEGAudioFrameParser is implemented as a GenServer that, when fed consecutive
|
|
packets of binary data (for example, from a file or network source), will
|
|
parse individual MPEG audio frames from the incoming data.
|
|
|
|
No decoding is performed on the audio data. Instead, the resultant frames
|
|
are ready to be fed into a separate decoder, or retransmitted over the
|
|
network.
|
|
"""
|
|
|
|
@server MPEGAudioFrameParser.Server
|
|
|
|
@doc """
|
|
Start the MPEG audio parser server. This must be done before calling the other
|
|
API functions.
|
|
|
|
iex> {:ok, pid} = MPEGAudioFrameParser.start_link()
|
|
...> is_pid(pid)
|
|
true
|
|
"""
|
|
def start_link(name \\ @server) do
|
|
GenServer.start_link(@server, nil, name: name)
|
|
end
|
|
|
|
@doc """
|
|
Add raw binary data to the current stream.
|
|
|
|
Returns: A list of zero or more structs, each representing a complete MPEG
|
|
audio frame. Note that because frames may be split across multiple packets,
|
|
this list may be empty, or contain more than one frame on each call. Any
|
|
leftover bytes will be stored by the server, and prepended to subsequent
|
|
packets.
|
|
|
|
## Example
|
|
|
|
Using a faked 128kbps 44.1k stereo MP3 frame. Note at least two headers must
|
|
be processed to return a single MP3 frame - otherwise frame completeness
|
|
couldn't be assured:
|
|
|
|
iex> packet = <<0b11111111111_11_01_0_1001_00_0_0_00_00_0_0_00::size(32), 1::size(3304)>>
|
|
...> {:ok, _pid} = MPEGAudioFrameParser.start_link()
|
|
...> MPEGAudioFrameParser.add_packet(packet)
|
|
...> MPEGAudioFrameParser.add_packet(packet)
|
|
...> |> length
|
|
1
|
|
|
|
"""
|
|
def add_packet(packet, name \\ @server) do
|
|
GenServer.call(name, {:add_packet, packet})
|
|
end
|
|
|
|
@doc """
|
|
Add raw binary data to the current stream.
|
|
|
|
Similar to `MPEGAudioFrameParser.add_packet/1`, but does not return the
|
|
frames. Instead, they can be retrieved at a later point, or by another
|
|
process.
|
|
|
|
See `MPEGAudioFrameParser.pop_frame/0`.
|
|
"""
|
|
def cast_packet(packet, name \\ @server) do
|
|
GenServer.cast(name, {:add_packet, packet})
|
|
end
|
|
|
|
@doc """
|
|
Pop a single completed frame.
|
|
Useful in combination with `MPEGAudioFrameParser.cast_packet/2`.
|
|
|
|
Returns a struct representing an individual MPEG audio frame, or `nil` if no
|
|
frame is available.
|
|
|
|
## Example
|
|
|
|
Using a faked 128kbps 44.1k stereo MP3 frame. Note at least two headers must
|
|
be processed to return a single MP3 frame - otherwise frame completeness
|
|
couldn't be assured:
|
|
|
|
iex> packet = <<0b11111111111_11_01_0_1001_00_0_0_00_00_0_0_00::size(32), 1::size(3304)>>
|
|
...> {:ok, _pid} = MPEGAudioFrameParser.start_link()
|
|
...> MPEGAudioFrameParser.cast_packet(packet)
|
|
:ok
|
|
...> MPEGAudioFrameParser.cast_packet(packet)
|
|
:ok
|
|
...> frame = MPEGAudioFrameParser.pop_frame()
|
|
...> frame.__struct__
|
|
MPEGAudioFrameParser.Frame
|
|
"""
|
|
def pop_frame(name \\ @server) do
|
|
GenServer.call(name, :pop_frame)
|
|
end
|
|
|
|
@doc """
|
|
Reset the server's state, returning any available complete frames. Any
|
|
additional bytes that are not part of a completed frame are discarded.
|
|
|
|
Returns a list containing any available complete audio frames.
|
|
"""
|
|
def flush(name \\ @server) do
|
|
GenServer.call(name, :flush)
|
|
end
|
|
end
|