Windows frame presentation triggers at the end of refresh cycle, before the VBI. So if you want less lag with a “non-VSYNC-OFF” RTSS Scanline Sync, you want the tearline as close to VBI (screen bottomedge) but still BEFORE in clock time, as seen in high speed videos at. Might as well turn off Scanline Sync and use plain jane VSYNC ON instead. If the tearline is below bottom edge during VSYNC OFF, then turning on a sync technology can add one refresh delay as since it already missed the end of refresh cycle (last pixel row or first line of VBI), the enabled sync technology it waits until the end of the next refesh cycle before displaying the presented frame, because if the VSYNC OFF tearline is already below screen bottom edge, “the VSYNC is already missed!”, so +1 refresh lag penalty. This is important for sync technologies that waits-til-end-of-refresh (like VSYNC or triple buffering) before showing new frames. You begin to realize tearlines just above bottom edge will take less time to the end of refresh cycle. Now the reason why if you’re using the other sync technologies instead of VSYNC OFF, is that tearline moves downwards on a time-bases, as it follows the Display Scanout (high speed videos).
First, to understand how RTSS Scanline Sync was born: See Tearline Jedi. This is a very old HOWTO that does need to be rewritten, indeed.īut RTSS Scanline Sync is an airplane cockpit compared to other sync technologies and is very hard to explain without raster interrupts knowledge (beam raced tear lines). I think it really would be worth rewriting it properly, in more approachable way, with more situational explanations. Dont take it personal, but I think the how-to in question written in a really confusing and poorly structured way. So why did the how-to settle on these recommendation? On the contrary, if we start to move scanline sync/tear up, we now mostly seeing last frame after sync, but in this case we take greater input latency hit. In this case, we greatly lower input latency (as a consequence), however we lose on frame data actuality (since it is mostly last frame). If we start to move scanline down, towards the bottom edge then we essentially gonna see mostly last frame.
Why is it recommended to move tearline down, just above the bottom edge? Let's take basic VSYNC OFF setup, where we just move tearline away so we dont see it. , 02:26 VSYNC OFF: Calibrate tearline offscreenįast Sync: Calibrate tearline to stay permanently above bottom edgeĮnhanced Sync: Calibrate tearline to stay permanently above bottom edge
0 kommentar(er)
