Despite my previous musings that I could probably do without a dedicated lag tester and use the various equipment that I already have to test my displays, I realized that Bodnar's device would come in handy greatly when it comes to fine-tuning a display's options to see what does and does not make a significant difference in lag, so I picked one up in the end. Before I give it a spin on an HDTV that I'm planning to pick up (the first I'm ever going to buy, actually), I thought I'd give it a spin on my two existing gaming displays.
The monitor that I've used for all of my HD gaming for the past 3 years is an Asus VH236H, chosen because it's the standard in fighting game tournaments, which I'm pretty active in (be it livestreaming them, assisting in setup and operation, or occasionally still competing in them). The picture on this thing isn't anything to write home about, and it's a rather obsolete model at this point, but in terms of affordability, convenience, and responsiveness, it's still a well-performing display.
A few things I noticed when using the lag tester:
After calibrating each individual preset mode to the best of my ability for brightness, contrast, sharpness, saturation, etc., the mode that actually produced the best result was the "Night View" mode, which beats out "Game" mode by a miniscule but still perceptible and consistent difference of 0.1 to 0.2 ms. Contrast seems to be better across the board in Night View mode, and I'd wager that that has something to do with how the tester performs.
Cranking up the Overdrive setting to maximum on the VH236H improves the results by another slight 0.1 to 0.2 ms. I do recall reading some Asus marketing materials that mentioned that this feature slightly improves response time, so maybe it wasn't just marketing fluff after all, as miniscule of a difference that it might be.
I did a second test on this monitor, this time using an HDFury 2 to convert from HDMI to VGA:
This returns a result 0.1 ms slower, but given some of the fluctuations, I can probably just write that off as the margin of error. The HDFury 2 is effectively lagless, which is relevant for a third test that I wanted to run:
This time I'm testing my Dell E773c CRT monitor, which is the main display I've been using for SD gaming ever since I picked up my XRGB-3. It's a pretty powerful combo, not only having the best quality picture on any display I've ever owned, but also one that I've noted previously for
actually beating out some CRT TVs when it comes to display lag. In the above pictue, I'm using the HDFury 2 to convert the lag tester's HDMI signal to VGA, and an XCAPTURE-1 to bridge that signal to the monitor (since the monitor's built-in VGA cable and the VGA dongle on the HDFury both have male end connectors).
From what I understand, the theoretical minimums for the lag tester are 0 ms for the very top of the screen, 8.3 ms for the middle, and 16.6 ms for the very bottom, if the the 3 markers are lined up perfectly in those exact positions. Both CRT and LCD progressive scan monitors read in a new frame from top to bottom, and at 60 fps, it takes about 16.7 ms to display a full frame, hence the difference in the readings for those 3 positions. (I believe plasma screens display a picture in a different manner, and that the entire picture is updated at once on them, but I'm not totally sure.)
The E773c then would appear to hit those minimums, at least to the best of my ability to measure it. My best explanation for the bottom reading falling under 16ms and the top reading going over 0 ms is that the viewable area of the screen is only part of the total physical screen, and that there's some lines above and below the borders that are still scanned as part of a single 16.7 ms frame.
Gamers who place an importance on input lag, myself included, often judge the worthiness of a flatscreen display on "how much it lags behind a (processing-free) CRT," so this test is pretty useful for providing a baseline when I get around to testing HDTVs. My personal threshold has always been "no more than a 1 frame difference from a CRT," so all I'd have to do is make sure that the display I'm considering isn't more than 16 ms above the E773c's numbers.