Did you use the pixel clock patcher before doing all of this? It must be installed (then you need to run the driver restarter, it is part of ToastyX CRU, so you can get the restarter from there, to avoid having to reboot PC after running the patcher).
And crosstalk is hard to explain.
The reason you have crosstalk is because of refresh.
Refresh rates is how many time the monitor refreshes a second
Monitors always refresh top to bottom. The refresh starts at the top and goes downwards. That is why the input lag (on any monitor review where they test input lag--like this one:
(Go down to the input lag tests and you can see how top has lower input lag than the bottom, since the top refreshes first during a refresh sync).
How does this relate to crosstalk?
Well notice that you have the most crosstalk at the bottom of the screen?
Where is the input lag highest?
And that's where you have high crosstalk.
You can look for crosstalk on this test I recommend:
press F11 to make it full screen.
The crosstalk is NOT the after image 'ghosting' of the UFO or red bar, with an inverted primary and secondary trail--that's just RTA/overdrive ghosting. The crosstalk is where you start seeing TWO frames jumbling together with a really super thick massive double/triple image...(and that's at the bottom).
The reason why this happens is because the same reason the input lag is highest at the bottom. The backlight shuts off and on every refresh (so at 120hz it shuts on and off 120 times a second). The problem is, at the bottom of the screen, since the input lag is highest, the bottom pixels have not finished completing updating the current frame before the strobe is done! So you actually have data from the *NEXT* frame mixed with the CURRENT frame. (next is key here). Actually, if the strobe phase is set to 000, you are actually having 1 frame of higher input lag than a high strobe phase.
As you increase the strobe phase, (you need to do this using the service menu; the windows slider utility is limited to strobe phase 047 to avoid shutting off the backlight), notice what happens.
service menu codes are here:
You start seeing a crosstalk field at the TOP of the screen move DOWNWARDS. Notice what you see above that crosstalk field and below it. Do you notice that the UFOs' and the vertical red bars seem to move AHEAD to the right one frame, ABOVE the crosstalk field, and "back" one frame below the crosstalk field? Since the image is moving from left to right, what you are seeing is the frame shifted 1 frame towards the destination -->the right--thus you are GAINING--improving input lag by 1 frame by raising the strobe phase!
So, to get the absolute lowest input lag, you must RAISE the strobe phase as high as possible before the backlight shuts off!
Now you might wonder "why..why do I have to do all of this?"
its just due to limitations with current monitor panels and response time of refreshing pixels. It's impossible to have NO crosstalk unless you have monitors with instant response time--CRT"s, Quantum LED's or OLED's.
So what exactly does strobe phase do ?
it changes where DURING the refresh that the strobe starts and ends!
As you increase the strobe phase, you make the monitor start the strobe **SOONER** in the refresh. So you actually push the strobe phase into the PREVIOUS refresh -thus LOWERING your input lag--you gain 1 frame of improved input lag. This may not make much sense as to why you are pushing the strobe into the "previous" frame (which is actually the DESIRED CURRENT FRAME YOU WANT), until you realize that the default strobe phase is not 000. It's 100.
Now you might still be asking...."well....wait. Why is the default strobe phase set to 000 if it gives me 1 frame worse (higher) input lag?
Well...the default strobe phase is actually 100!!
When you flash a firmware or do a factory reset (recall all in the factory menu, or reset all, or set "Internet café" to enabled and turn off the monitor, it sets the strobe phase to 100. This gives you the lowest input lag possible, and you have absolutely NO crosstalk at the top of the screen. But because the BOTTOM of the screen has MORE INPUT LAG from panel response time than the refresh rate ITSELF (refresh rate frame lag is equal to 1000 divided by refresh rate in milliseconds)., you get a lot of crosstalk at the bottom of the screen.
Using a vertical total tweak causes the monitor to switch to 60hz backlight pulse widths, while 'extending' the blanking period of the strobe, which reduces strobe crosstalk, but also LIMITS the maximum strobe duty allowed--the higher the refresh rate (up to 128hz), the lower the maximum strobe duty. If you still want more information, and about WHY this happens I have math to explain it.
But here is a fun fact for you to think about in the meantime.
Maximum STROBE DUTY is 030. There are 30 values of persistence. Persistence is equal to the REFRESH RATE milliseconds divided by 100 --for example: 60hz is 16.7 milliseconds (1000 divided by 60), then divide 16.7 by 100 and you get .167, so each point of strobe duty at 60hz increases persistence (strobe width) by 0.167ms. (so without a VT tweak, the higher the refresh rate, the lower the persistence and the dimmer the screen).
Maximum strobe PHASE is 100. 100 values of strobe phase. All work perfectly.
At 120hz same thing. 8.3ms response time from refresh rate, 0.083 ms persistence per point of strobe duty.
But if you use a Vertical total tweak, 120hz uses the 60hz persistence values instead of the 120hz persistence values.
Maximum strobe PHASE then gets cut in HALF. Strobe phase 050 and higher turns off the backlight.
I'll let you think about that for a bit.Edited by Falkentyne - 11/10/15 at 10:11am