Skip to content

access: screen: add DXGI capture support on Windows

Steve Lhomme requested to merge robUx4/vlc:screen-dxgi into master
  • rework the screen code so it can use different capture callbacks depending on which open function worked.
  • allow the vout to use a video context from a decoder device it didn't create, as long as it didn't create one The access has no way to communicate/force its own "decoder" device to the vout and we want to avoid a copy
  • a D3D11-based picture is created from a GPU copy during capture and sent as a block, a dummy decoder transforms this block picture into a picture sent to the display

This will only be used on Win8+. On Windows 7 it will fallback to the GDI capture code.

v13: added support for --screen-follow-mouse, disable DXGI capture when --screen-mouse-image or --screen-fragment-size are used (for now, they could both be supported).

v15: split some platform specific code and better use of the defines (that may go away at some point)

Edited by Steve Lhomme

Merge request reports

VideoLAN code repository instance