Chaps2 wrote:Also your result of 3.7ms input lag does not look logical to me. As a 240Hz panel can not show more than 1 frame every 4.16ms (240/1000=4.16) and that is with a pixel speed of 0 (which does not - and will never exist).
You have to understand how an LCD scans out. It's not displayed all at once.
VSYNC ON:
- Top edge is lagless + GtG/monitor processing
- Center is half a refresh cycle lag
- Bottom edge has full refresh cycle lag
- Average input lag is usually same as center-screen input lag (aka half a refresh cycle)
VSYNC OFF:
- Top edge of frameslice is lagless + GtG/monitor processing.
- Center of frameslice is half a frametime long.
- Bottom edge of frameslice is a full frametime long.
Pixels are transmitted from a GPU over the cable sequentially, one pixel at a time (at the dotclock rate, as seen in Custom Resolution Utility), beginning at the upper-left, scanning left-to-right, top-to-bottom (calender style scan sequence), with padding in between (blanking intervals). On a time-basis, VSYNC OFF essentially looks like this:
During VSYNC OFF< between the tearlines are the frameslices, and the top edge of frameslices at 432fps adds only 0ms to the existing subsystem input lag (e.g. monitor processing, GtG, etc). Center of frameslices at 432fps would add 0.5/432sec of scanout lag to the lag number. Bottom of frameslices at 432fps would add 1/432sec of scanout lag to the lag number.
In a high speed videos, the action of an LCD scanning out, can be seen
http://www.blurbusters.com/lightboost/video
This is a high speed video of testufo.com/flicker running in full screen mode, and it shows the top-to-bottom refreshing action of an LCD panel. In many gaming monitors, this is in sync with the cable output, so the pixels can begin its own transitioning process, pretty much immediately upon the arrival of the pixel (especially if it's avoiding full-framebuffering for monitor processing -- many gaming monitors use linebuffered processing -- essentially realtime/"Instant Mode" processing).
As soon as the GPU is outputting pixels, the top edge of the monitor immediately begins pixel transitions for that particular area of the screen. The whole screen does not have to be fully displayed before your the photons of the first pixels at the top edge of the screen begins hitting your eyeballs. This can occur less than half a refresh cycle later. At ~120Hz with 1ms GtG, the GtG zone is approximately 1/8th the screen height as it scans downwards where the pixel-refreshing action lagsbehind the scanout.
We have helped a few monitor manufacturers on how to create strobe backlights, see
Electronics Hacking: Creating a Strobe Backlight so we have an understanding of how LCDs work. This is the stuff that BlurBusters got started with, we were originally "scanningbacklight.com" before our site renamed more mainstream to Blur Busters with the advent of strobe backlights such as LightBoost.
Anyway:
Lag methodology will output different values for:
-- Lag from GPU-side to monitor pixels
-- Lag from monitor input to monitor pixels (excludes cable transmission overheads, e.g. +1ms)
-- Lag from mouse to monitor pixels
-- Lag from keyboard to monitor pixels
And screen location:
-- Lag from VBI to monitor top (ala VSYNC ON input lag)
-- Lag from VBI to monitor center (ala VSYNC ON input lag)
-- Lag from VBI to monitor bottom (ala VSYNC ON input lag)
-- Lag of pixel transmitted from GPU to corresponding pixel shown on monitor (more representative of VSYNC OFF input lag)
And other variables to keep in mind:
-- Lag of a specific Hz (varies from Hz to Hz)
-- Leo Bodnar Tester is lag of VSYNC ON 60Hz
-- SMTT 2.0 is lag-differential between two screens.
-- Lag of VSYNC OFF is also very different from lag of VSYNC ON.
What this means is that the numbers on one sites are only comparable to other numbers on the same site for the same mode. But numbers are generally not comparable between 2 different websites -- because the input lag measuring methodologies vary quite a bit.
We were also the world's first website to successfully test input lag of G-SYNC, see year 2013 article at
http://www.blurbusters.com/gsync/preview2
We're the inventors of a
peer-reviewed monitor testing technique that has been validated by NOKIA, KELTEK and NIST.gov researchers. Several websites
now use our testing techniques.
(click for more info)
So I think you can now understand we've got plenty of credentials to know what we're talking about.
Input lag measurement is a complicated topic.
It is very important for sites to properly & fully document their input lag measuring methodology.