Originally Posted by Spooony;13613579
vsync and tripple buffering is two different things! Why do the people always talk about vsync when I said tripple buffering?
Vsync just let the gpu know of your monitors limitation. If the two buffers are full and its waiting for the LCD what's going to happen?
Do not blame a monitor for input lag. Yes it does have a bit of input lag but its not noticeable. Its part of a graphics stage which starts at your hdd and ends at the monitor. You are blaming the monitor for the end result. The code doesn't have to sync with the framerate, a physics' thread could be running at 360Hz, the control thread at 60Hz and rendering at 30Hz. A game of 30fps got more lag than a one at 60fps. So tell me if a LCD ain't fast enough why does it have more lag at a low frame rate?
Input lag comes from a ton of places, but it stacks. Vsync on top of everything else means a crap load of input lag.
And yes, the same applies for triple-buffering: everything that buffers stuff ads lag.
Also, with vsync obviously a game has more lag @ 30fps than @ 60, for the simple reason that you got the same number of frames buffered...but @ 30fps they are twice as long. No wonder it feels more laggy...because you do have more lag.
Originally Posted by De-Zant;13618039
and can be measured by having the same image displayed on a LCD and a CRT side by side, running a millisecond precise counter, and then taking a picture of the thing on a camera with a fast shutter speed. This way, we can see that the LCD is falling behind the CRT by usually 10-30ms.
That ain't correct. If you would be able to shoot to a monitor fast enough, you would see that any counter you watch would work in 16ms jumps (more or less), because it updates at such frequency.
So, if could happen that you would be comparing side by side a CRT and an LCD and got mixed up results: unless both screens were prepared to update at the very same moment, you could shoot at the moment in which CRT had just updated 1ms ago but TFT was about to start: result, 20ms of fake input lag.
Thats why It ain't easy to measure everything. and why most result should be disregarded.