Game capture vs window capture vs display capture

Game capture hooks into the game process to read its frames; window capture and display capture ask Windows to hand over a window or a whole monitor without touching the game. A hook can be the cheapest path but it puts code inside the game, which anti-cheat software may dislike; the Windows methods inject nothing but can change how the game is presented and cost frames. No method is free, and the best one depends on the game and the machine.

Last updated

The three methods

Capture methods for a PC game, as of 2026-11-10
MethodHow it worksStrengthWatch out for
Game captureA hook injected into the game process reads each frame as the game presents itCan follow a game across menus and resolution changes; in OBS it is the usual choice for gamesInjects code into the game; anti-cheat systems may block or flag it; needs the hook to match the game's graphics API
Window captureWindows hands the content of one window to the recorder (Windows.Graphics.Capture)No injection; works with protected games; follows a window you chooseIn BriseStream's measurements the game left direct presentation (independent flip) and was composed by the desktop while its window was captured
Display captureWindows hands over a whole monitor, through Windows.Graphics.Capture or Desktop DuplicationSimple; captures everything on the screen, including overlays you addCaptures notifications and other windows too; one of the two APIs measured much worse than the other in a synthetic test

What BriseStream measured

On one machine (GTX 1050, 1080p), with a synthetic game at about 500 FPS and 95 % GPU load, capturing the game's window with Windows.Graphics.Capture alone lost 10.1 % of FPS, and Desktop Duplication alone lost 18.5 to 19.2 %. BriseStream recording the window lost 10.3 % and recording the monitor 9.6 %. At about 167 FPS the same run gave 15.1 % recording the window and 6.8 % recording the monitor with unstable baselines. The numbers move a lot with the game and the load, so treat them as a reason to test, not as a ranking.

Not a verdict on OBS: OBS's game capture was measured in a different test (see the OBS comparison). These capture-only numbers are from one machine and one synthetic game.

How to choose

  • Start with the method your software recommends for games, then test the other if you see FPS drops or a black screen.
  • If a game with anti-cheat gives a black screen or a warning with a hooking method, switch to window or display capture.
  • Match the capture size to the game's size so the GPU does not rescale every frame; BriseStream keeps a capture smaller than the canvas at its own size.
  • Test with the exact game and settings you stream: results differ by title.

BriseStream does not inject into games on purpose: its "game capture" source follows the foreground game through the operating system's capture, so it avoids the anti-cheat question, at the cost of what the Windows capture itself costs.

Questions and answers

Is game capture safer or riskier than window capture?
Game capture injects code into the game, which anti-cheat software can object to, so window or display capture is the lower-risk choice for games with anti-cheat. Always check the game's rules.
Which capture method costs the fewest FPS?
It depends on the game and machine. In BriseStream's synthetic test on one machine, Desktop Duplication cost about twice what window capture did, but results at other frame rates were unstable. Test your own game.
Why do I get a black screen with game capture?
Usually the hook cannot attach (an anti-cheat block, a graphics API mismatch, or the game running with other privileges). Window or display capture avoids the hook.

Sources

  1. ADR 0012: what recording costs a game, measured against OBS
  2. ADR 0003: Windows capture and hardware encoders
  3. OBS Studio parity audit (docs/obs-parity.md)
  4. OBS forum: "OBS Studio causing FPS drops"