Unixko wrote: ↑10 Aug 2020, 15:15
I collected many logs to get the time scale similar on both graphs. At first it looked like a pattern(smoothness) was emerging, but nope, after enough logging there is no difference really with ferrite cores on or off according to MouseTester.
However, in-game (CSGO), without the ferrite cores, my mouse movement constantly overshoots, swinging back and forth pass the target. I am used to having the ferrite cores on, so I guess I am over compensating in aim and the ferrite cores have a dampening effect, lowering the floatiness and making the crosshair more controllable on my setup.
Your science problem is you're probably not testing under the same RFI environment.
MouseTester app alone, creates a PC that consumes only 60-100 watts (idling computer with power management)
CS:GO creates a PC that consumes >300 watts (GPU chugging at max power) on a typical modern powerful GPU.
The fields of RFI at 300 watts is much bigger and stronger.
TL;DR: You ideally want to rev your computer to maximum (CPU 100%, GPU 100%, Robust Network Througput, AND Robust Disk Throughput) during RFI testing
Re-test MouseTester under the same RFI conditions (GPU generating 3D at full power). The problem is multitasking MouseTester and CS:GO may not be possible. Maybe primary-monitor CS:GO with secondary-monitor MouseTester? You may have to launch into some disposable super-frantic CS:GO server with a disposable account, so you don't ruin your CS:GO as a sitting duck when Alt+Tabbing to MouseTester. Even that might not work, if the game internally pauses some processing when Alt-Tabbed. Make sure to use VSYNC OFF though, to full-throttle your GPU RFI / CPU RFI to the maximum interference possible. Keep the network spewing hard (maximum network RFI), so be in an online game while running MouseTester.
To understand better, you literally want as many as possible -- transistors and cables and circuit paths -- hot with signals simultaneously, churning those pixels, bytes, polygons, game data, movements, packets, whatnot, etc. Blast the packets, blast the frames, blast the polls, get the GPU 100% CPU 100%... If one of your chips or cables are idle, de-idle it somehow, when running RFI-rejection performance tests. Create the duress condition.
Don't forget to add a Kill-a-Watt to your lab testing equipment if you're an aspiring hardware tester who wants to scientifically reproduce RFI testing conditions. A computer is not a polar Yes/No in RFI behavior. You have to really chug-chug the maximum number of copper wires & transistors possible, to intentionally maximize RFI problems, which means.... Full hardcore network GAMING while mouse-testing.
For RFI testing, redline your computer to the max! Sometimes RFI only shows up at the figurative metaphorical redline region, on the metaphorical RPM meter of your computer. Get all the chips and wires as hot as possible simultaneously. This is where scientific testing of ferrite cores starts to show actual results.
Sherlock Holming the science, baby!
P.S.
MouseTester is open source. Are you a software developer? One could modify MouseTester to use mousehooks (HOOKPROC LowLevelMouseProc via SetWIndowsHookEx), so that you can run it on a secondary monitor while playing the game on a primary monitor (at least on cheat-allowed servers, since mouse hooks are detected by anticheat software). This would make RFI testing easier with MouseTester
