Hello. I would like to estimate my maximum lag input, based on what I am going to tell you. Of course, it remains approximate and a little theorical, I know it. - VSYNC OFF (new for me ) - FAST VSYNC OFF. - Max Pre-rendered frame = 1. ) - MINIMUM FPS= at least 25 (night/surnrise/sunset + rain + 50 AI + mostly cloudy + varied mods/cars) - TV 1K 132 cm : theoretical imput lag with game mod = 8 ms. But I believe that in reality it is 10-11 ms. - Purple bar up to 90% MAX but no visible lag vis (not out of real time). - Classic USB connection from T500 RS to PC. - Windows 10, No antivirus, very few software/Windows features installed - RF2 clean installation. HARDWARE i5 2500K OC + 15%, GTX 780 not OC, 16GB RAM 1372Mhz, SSD 65GB empty, PCIe X16 2.0. If I forgot important information, tell me. The calculation I would do is : T500RS to PC = unknown PC to TV = 1/25 (40 ms) TV game mode = 11 ms. TOTAL = 51 ms or more depending on the T500RS/PC link. EITHER 5% of 1 second (it's much simpler for me). Thank's EDIT Nvidia control panel ALL to nvidia defaults, except : - pre rendered maximum frames = 1 - maximum performance - Texture filtering = HIGH quality Game UI : - AA = level 5 - Full HD Game : ALL TO OFF except Full texture, Full details, Anisotropic X16, Player Full.
I think measuring the input lag is the way to go. Get a cheap camera with a high recording frequency, move your wheel and position the camera so that wheel and monitor are filmed. Then analyze the movie and see how many frames the wheel shown on the monitor lags behind the real wheel. Thats how i measured input lag on my old system (using a cheap Sony Playstation Eye camera which has decent recording frequency)
Exactly. Trying to calculate this from specs and settings is a loss of time IMO. As Vittorio suggested you should be able to make a recording at 120hz. Several mobile devices already offer this. Ask your tech fancy friends if you don't own one. Examine it with a video editing software. Windows Movie Maker is enough. Take a reference position for both virtual and real wheels. Start the recording at that reference position with static wheel. Make the recording with fixed camera/mobile to maintain a static scenario through the recording. Last: write down the time stamp when each of the two wheels passes through that reference position. Stablish a constant criteria for both events and dontd cheat yourself when doing this. Do this minimally ten times (best twenty) and cakcukate the average difference between both wheels in order to average the experimental error. If correctly done you can increase recording accuracy by the nonius effect of repeating the measurement many times. It is quite easy to get extra precision of a 1/100th of a second with a multiple 30 fps recording of what you want to measure.
Ok guys, I will do that. Thank you for the tuto. I think I make a huge mistakes using VSYNC since 12 years (RF1&2). I advise to vsync users to try again without vsync.
