platf::nv12_zero_device Class Reference

macOS zero-copy encode device that forwards NV12 frames from AVFoundation to FFmpeg. More...

#include <src/platform/macos/nv12_zero_device.h>

Inheritance diagram for platf::nv12_zero_device:

Collaboration diagram for platf::nv12_zero_device:

Public Types
using	pixel_format_fn_t = std::function<void(void *display, int pixelFormat)>
	Callback signature used to update the opaque AVFoundation pixel format.
using	resolution_fn_t = std::function<void(void *display, int width, int height)>
	Callback signature used to update the opaque AVFoundation display resolution.

Public Member Functions
int	convert (img_t &img) override
	Forward a captured AVFoundation frame to FFmpeg without copying.
int	init (void *display, pix_fmt_e pix_fmt, resolution_fn_t resolution_fn, const pixel_format_fn_t &pixel_format_fn)
	Initialize zero-copy NV12 encoding for an AVFoundation display.
int	set_frame (AVFrame *frame, AVBufferRef *hw_frames_ctx) override
	Attach frame resources used by the next conversion or encode operation.
Public Member Functions inherited from platf::avcodec_encode_device_t
virtual void	apply_colorspace ()
	Apply the configured colorspace metadata to the active frame.
virtual void	init_codec_options (AVCodecContext *ctx, AVDictionary **options)
	Provides a hook for allow platform-specific code to adjust codec options.
virtual void	init_hwframes (AVHWFramesContext *frames)
	Initialize the hwframes context.
virtual int	prepare_to_derive_context (int hw_device_type)
	Prepare to derive a context.
Public Member Functions inherited from platf::encode_device_t

Public Attributes
resolution_fn_t	resolution_fn
	Callback stored for later AVFoundation resolution updates.
Public Attributes inherited from platf::avcodec_encode_device_t
void *	data {}
	Backend-specific conversion state.
AVFrame *	frame {}
	FFmpeg frame currently owned by the encode device.
Public Attributes inherited from platf::encode_device_t
video::sunshine_colorspace_t	colorspace
	Colorspace metadata expected by the encoder.

Detailed Description

macOS zero-copy encode device that forwards NV12 frames from AVFoundation to FFmpeg.

Member Function Documentation

convert()

int platf::nv12_zero_device::convert ( platf::img_t & img )

overridevirtual

Forward a captured AVFoundation frame to FFmpeg without copying.

Parameters

img	Image or frame object to read from or populate.

Returns

Conversion status.

Reimplemented from platf::avcodec_encode_device_t.

init()

int platf::nv12_zero_device::init	(	void *	display,
		pix_fmt_e	pix_fmt,
		resolution_fn_t	resolution_fn,
		const pixel_format_fn_t &	pixel_format_fn )

Initialize zero-copy NV12 encoding for an AVFoundation display.

Parameters

display	Display object or identifier associated with the operation.
pix_fmt	Sunshine pixel format to convert or allocate for.
resolution_fn	Callback used to resize the AVFoundation capture output.
pixel_format_fn	Pixel format.

Returns

0 on success; nonzero or negative platform status on failure.

set_frame()

int platf::nv12_zero_device::set_frame ( AVFrame * frame, AVBufferRef * hw_frames_ctx )

overridevirtual

Attach frame resources used by the next conversion or encode operation.

Parameters

frame	Video or graphics frame being processed.
hw_frames_ctx	FFmpeg hardware frames context associated with the frame.

Returns

Status from updating frame.

Reimplemented from platf::avcodec_encode_device_t.

---

The documentation for this class was generated from the following files:

• src/platform/macos/nv12_zero_device.h
• src/platform/macos/nv12_zero_device.cpp