Not a problem, I can see how my previous thread could confuse you; I was still figuring it all out. And, anyway, these comments are useful for gauging how clear or unclear a particular point of mine is. And you just brought another one up; the input latency charts.hammelgammler wrote:You are totally right, I'm sorry about my comment. It was a bit confusing with your other thread where you mentioned the 1ms G-Sync penalty and to cap the FPS even lower to not have tearing. There are also some charts which may be misleading, like the 120 FPS cap on 144Hz. I was a bit confused because of that, which Information is the correct one. :/
The one thing I wanted to mention that I do had the same experience, that you need to cap your FPS much lower with V-Sync OFF. I had noticably tearing with 162 FPS, and I'm pretty sure I felt more Input Lag as well in comparison to 150 FPS. All my testing is with Overwatch and the built in limiter.
But I find it interesting that in your newest charts from this thread, that V-Sync OFF is a tiny bit faster then ON (about 1ms). So isn't it technically a bit faster? I will try it out with V-Sync ON though and write my experience with it.
The answer to your question is a resounding NO; 1ms is well within margin of error. If what you suggest where true, then it would mean G-Sync + v-sync off + 142 fps limit has less input latency than standalone v-sync off + 300 fps limit, which is, of course, not possible. With G-Sync + v-sync on, G-Sync + v-sync off, and v-sync off at the same fps limit, frames reach the middle level of the screen within the same time-frame, thus virtually identical input latency.
I will also try to make the margin of error clearer in my OP. If there were a 2-3 ms difference over several runs, then yes, there may (or may not) be an appreciable difference. Again though, 1ms is entirely within margin of error. At 144 Hz, for instance, if G-Sync + v-sync off + 142 fps actually had less input latency than G-Sync + v-sync on + 142 fps, you'd be seeing up to a full frame of a difference (6.9ms) or more on average.