@error:
ANY pixel clock past 330.00 REQUIRES the pixel clock patcher! If you are not using it, then it's using the DEFAULT VT (ignoring your custom resolution completely and using the EDID) or in some cases, if you go too high, the refresh rate simply vanishes.
djriful wrote:Oh, nevermind. It didn't go black screen anymore.
So if I use the Lightboost is basically similar same as I use VT Tweak? If so which one is better? I have Nvidia GTX TITAN here.
Vt tweak tries to force a 'faster' scanout by accelerating the update process of the LC panel. What you're doing is making the vertical synchronization phase *longer* (e.g. from 1125 lines to 1500 lines) at the same refresh rate. I'll try to explain the easy way. I'm sure ToastyX or Chief can jump in and correct me.
The monitor is currently calibrated to have a vertical sync total of 1125 lines
The strobe pulse is based on the # sync lines (someone correct me fiercely if I'm wrong here), so at a strobe phase setting of 000, the first strobe pulse occurs at the top of the synchronization phase, which is at the very top edge of the screen, and the NEXT strobe pulse occurs close to the bottom of the screen, but not all the way at the bottom. So with a strobe pulse of 000, you get a reasonably clean top edge, with some parts of crosstalk showing, and a double image thick mess close to the bottom.
Now, by increasing the vertical total, you are now INCREASING the size of the vertical synchronization area. Display-corner calls it "making the vertical synchronization phase LONGER". By doing this, you make the strobe pulse at the top start at the same spot, but the strobe pulse for the next frame start closer to the bottom, and in case of VT 1500, actually pushing the next phase just barely OFF the screen, because by extending the synchronization phase, the next strobe pulse is now offscreen and you have more of the screen area (1080 lines) visible without both crosstalk border pulses on the screen. You are not extending the SIZE of the screen vertically, pixel wise...just the synchronization. I think you can see that this also increases the pixel clock, since now the screen must "work harder."
I'm sure chief can once again correct me, but I think this is almost the same thing as changing the "Vertical hold" on old analog TV's, except on those, instead of changing the vertical hold 'position", now you're actually changing the SIZE of it.
The drawback of this method is the monitor gets confused by the strange timings and switches to the 60hz backlight pulse widths, and plays havoc with the "display mode' settings, and may show faint horizontal line inversion type artifacts on the screen, not from the dot clock directly, but from being pushed to its limits.
Also, because the vertical synchronization phase is longer, changing the strobe pulse now has more of an effect position wise, on the screen, on the strobe pulse, per "point", than without the VT tweak. That is, the strobe boundary moves FARTHER down per pulse (area) value, which makes complete sense ! Because the sync phase is now LONGER, and you're moving the pulse start between sync phase positions, NOT between "pixels" on the monitor.
A strange drawback to this is that now, changing the strobe DUTY (persistence) to a higher value actually makes the TOP crosstalk boundary move down farther (per point) than it would without the VT tweak, also, so for instance, using a persistence of 2.5ms at 120hz to give a bright screen, would push the top strobe pulse down too far and make the top of the screen look nasty (during motion), while without a VT tweak, you can keep it looking decent (but at the expense of the bottom, again).
But continuing : by using a VT 1500 tweak, when you go past strobe phase 047, you are actually going PAST what strobe phase 100 would do, WITHOUT the VT tweak, aka extending the first strobe pulse past the location where the second (bottom) one would start by default. Masterotaku found out that when this happens, you are now rendering one frame "ahead" (1 frame of input lag reduced) BUT the persistence now starts getting lower, because the strobe phase is pushed past where its supposed to go. So at strobe phase 059, you now have a 0.167ms persistence (this would be identical to strobe phase 000 with strobe duty 001 and a VT tweak, except you are one frame ahead (input lag slightly improved). I'm sure you can figure out exactly what happens if you go from 059 strobe phase to 060.
----------------------
Lightboost does basically the same thing, but instead of increasing the vertical synchronization phase time (making it longer), it updates the LC panel timings directly, through hardware, which force the scanout to complete faster (which basically results in the same thing happening; the top of the strobe pulse being at the very top and the bottom being off screen), and has even _more_ of an effect (since it's tweaking the LC panel directly) than a VT 1502 tweak. This puts even MORE stress on the panel than a VT tweak, so you may find even stronger inversion scanline-type artifacts (NOT talking about the pixel transition/gray/green ones you can test for on lagom) than with the VT tweaks.
Lightboost also has per-line overdrive as well, to improve the inverse ghosting/normal ghosting performance, so that's even more work for the panel.
Which is better?
VT with BBR, or Lightboost?
For 3D games, especially FPS, VT tweak for sure--no loss of color fidelity, no loss of contrast ratio, no pink tints, full ability to tweak the image quality
For 2D games: lightboost can be better, because ghosting is usually more apparent here, and lightboost's per line overdrive--ESPECIALLY with contrast "0", greatly improves things by a large margin on the 27" screens, and removes any advantage Tracefree's overdrive (OD Gain in the service menu, for lightboost mode) had on the 24" VG248QE vs the 24" Benq screens. This depends on the game and the color transitions, of course.