Running 1500 VT pushes the panel to its limits similar to what lightboost mode does, and forces as fast of a scanout as possible, but still not as fast as lightboost, as LB has a LC panel timing change (on the hardware level) so it does this without needing VT tweaks. But as below, accelerated scanout by either method introduces a faint scanlines effect you may see.
If you run testUFO with 120hz lightboost (you need an Nvidia card or laptop surrogate to unlock lightboost on the Z monitors; strobelight wont work) what you will notice is that the ghosting crosstalk you get at the top of the screen is virtually identical to the Benq at 100hz and 1.0ms persistence with the VT 1500 tweak. You will see the very top of the double image crosstalk band at the very top of the screen, while the rest of the top looks great.
It will look identical to how the Z monitor looks with VT 1502, except the BBR lacks per-line overdrive, so the image overdrive ghosting will be significantly worse than lightboost (this is the highest possible value btw) at 100hz, with between a 0.750-1.0ms persistence (although the Z monitor will be darker).
What's funny is, 100hz lightboost pushes that very top edge of the crosstalk field completely OFF the screen. Using VT tweaks with 80 hz and lower on benq blur reduction *Do the exact same thing*. Now if you were able to unlock lightboost 100hz, and set the contrast to 0, you would notice absolutely NO ghosting at all in testUFO. And none on the desktop either! Just some slight inverse ghosting at the bottom of the screen. Remember lightboost uses per-line overdrive, which Benq BR does not use--that's how lightboost gets rid of almost all the ghosting if you lower the contrast.
But yeah both lightboost and increasing the vertical total force the panel to update the screen and let the pixel transitions complete faster between the strobe phases. Lightboost does this by a LC panel update on the hardware side, while increasing the vertical total pushes the pixel dot clock higher, which forces the panel to update faster and thus helps the panel complete the pixel transitions faster, between the strobe pulses. (the blurry double image mess that you saw is a strobe pulse settling error; without the VT tweaks, you had the one at the top (with strobe phase 000 its at the very top) and the one at the bottom. The VT tweaks do not change the position of the pulse at the top, which is almost offscreen *IF* the persistence is low enough, but the VT tweaks push the bottom of the pulse offscreen. The panel switches to 60 hz backlight pulse widths when it's run out of specification like this as per :
http://display-corner.epfl.ch/index.php/BenQ_XL2411Z ; that's why the OSD is reporting 60 hz.
Because the panel is updating faster, both lightboost mode and VT tweaks in BBR mode may cause a faint scanline pattern to appear, usually around the higher parts of the screen. This is unavoidable and is because the accelerated scanout is pushing the panel harder than it was designed. You will notice that lightboost mode has a larger scanline pattern, because Lightboost's accelerated scanout is more intense than the VT tweak (that's why in lightboost mode, the top and bottom of the crosstalk field are even farther off the screen than the VT1502 BBR tweak.
Note that even if BBR is turned off, you may see the scanline effect even though non strobing operation (AMA) isn't affected in a measurable way.
The main issue still remaining with Benq BR mode is the overdrive artifacts, since there is no per-line overdrive. Benq doesn't have an option to adjust the overdrive gain like the VG248QE does (via the service menu), so you either have to deal with AMA off (normal ghosting) or AMA on (a bit too aggressive ghosting), but considering the lack of per line overdrive, Benq actually set things the best they could. AMA high in benq blur reduction mode (which is actually a hidden setting called "MBR/high" is actually LESS intense than normal AMA high (if you do testUFO and then toggle AMA off and high with BBR ALREADY ON, you will see much more intense inverse ghosting).
Now you may ask "well, why didn't they tune it so the inverse ghosting is completely gone, like lightboost mode?"
Well...look at the bottom of the testufo screen with the VT tweaks.
You see that normal ghosting?
That normal ghosting would turn into 'regular ghosting" even worse than it is now, (because you cant fix BOTH inverse ghosting AND normal ghosting at the same time). EVEN at lightboost 100hz, while the top of the screen is pristine (contrast 0 on Benq 2720Z, contrast 45 on VG248QE), the bottom has an inverse ghosting pattern.
If you go into the VG248QE service menu and adjust the OD gain from 0F (most aggressive) down to 0C, the BOTTOM of the screen goes from faint inverse ghosting, to becoming perfect (!) while the TOP of the screen goes from perfect to showing faint NORMAL ghosting. (yes, OD gain can be adjusted in lightboost mode). And this is with lightboost's per line overdrive.
Due to per line overdrive, some 2D games, where the inverse ghosting can be apparent (on sharp transitions) in MBR mode, will look better in lightboost mode instead (if you can unlock it), so if you play 2D windows side scrollers, if you ever get an Nvidia laptop or cheap surrogate card to unlock lightboost (1138 VT at 100hz and 1149 VT at 120hz activates, but does not "Unlock" lightboost), you may want lightboost for those games. Most 3D games will look better in BBR mode, as overdrive ghosting isn't as obvious (just the crosstalk field ghosting) and having better colors and contrast makes more of a difference.
*edit*
you asked if it's safe.
Yes, it's as safe as using lightboost (accelerated scanout is accelerated scanout), as long as you do not exceed the persistence ratings for the backlight LED's. Remember that using the VT tweaks make the monitor use the 60hz backlight pulse widths and as shown here:
http://display-corner.epfl.ch/index.php/BenQ_XL2411Z
60hz refresh rate itself has a 5.0ms maximum persistence (strobe duty 030), while 120hz has a 2.5ms maximum persistence.
(These are WITHOUT VT Tweaks--this is how the monitor was designed to be run).
Using the VT tweaks forces the 60hz backlight pulse widths, and now you can see a problem--120hz now has a 5.0ms persistence setting possible (the windows blur reduction utility is based on using the VT tweaks btw!), when 120hz is only rated for 2.5ms
100hz has a 5.0ms persistence possible also (using the 60hz pulses), but is rated for 3.0ms. So just don't go past 2.5 to 3.0ms persistence unless you're at 60hz! Using 5.0ms persistence can probably shorten the lifetime of the LED', and besides you would not even WANT to use 5.0ms persistence! If you switched blur reduction OFF, the screen would actually get dimmer than with BBR on (!). There's absolutely no reason to use higher than a 1.5ms persistence, and as I explained in another post (different thread), raising the persistence (strobe duty) LOWERS the position of the top crosstalk field, so you want the persistence as low as possible so that blurry mess is as high up as possible (anything past 1.0ms makes the top of the screen too annoying).
The other drawback to VT tweaks is you can't use the display mode aspect settings anymore (in the OSD), because the screen switching to 60hz pulses causes strange stuff to happen to the OSD's aspect sizes. Even using a VT tweak at 60 hz itself causes things to act weird with the aspect settings. You can use the display mode settings by forcing a custom resolution that doesn't have the VT tweaks; for directX 9 games, you can do that with custom resolution utility easily, then just launch the game with the custom refresh rate matching the same refresh rate as the deskop, and the monitor will pick up the custom refresh rate (a few combinations will give cable errors if you're using the pixel clock patcher, due to the single link DVI limit being patched; there's a version of the toastyX patcher that does not patch the single link limits on a dual link port, to avoid that). For DX 10/11 games, though, you either need a BIN file without the VT tweaks or you need a DETAILED resolution in CRU without the VT tweaks for your game (like 1680x1050@120), for some reason, DX10/11 games refuse to use any established or standard resolution listed in CRU, and will use the first available non standard or non established resolution. (importing a custom bin file you saved, or using restart64.exe is very quick and useful for this stuff).
.
(it's a proper "following camera" photo, by the way).
). Darker games are benefited the most, because very light colors are the most problematic in this mode.