Displays typically don't produce the exact same image across their entire surface. We've quantified the uniformity of the PG278Q by taking a series of luminance readings in different regions of the panel. We set the brightness level at ~200 cd/m² at the center of the screen before starting.
This monitor's 13% variance from the center of the screen to the edges isn't anything you're likely to notice, even while staring directly at the screen and looking for problems. Both of the other Asus displays we've tested have similar light distribution. The PB287Q has a 15% max variance from the center to the edge of the display.
I've chosen to convey backlight bleed using a picture rather than a series of measurements. Trouble is, I never know exactly how this image will end up looking on the reader's own screen. What I see here is a bit of light bleed around the bottom corners of the monitor, with more bleed extending up the left edge. In regular use, it's not a deal-breaker or even terribly noticeable, but this is a little more light bleed than we saw with the PB287Q.
I've taken these pictures in order to demonstrate how the display's color and contrast shift when the viewing angle changes. As you can see, at the angles we've chosen, the PG278Q comes out looking pretty good. The bottom image, looking up at the panel from below, shows a little too much contrast, especially in the sky, but the colors don't shift or invert at this angle like you might expect from a lower-quality TN display.
TN panels tend to be quick, and this one is no exception. The PG278Q's gray-to-gray transition time is one millisecond, substantially quicker than the five-millisecond rating for the IPS-based PB278.
Input lag comes from many sources, including the scaler chip inside the monitor. Nvidia's G-Sync module is a brand-new and comes from a new player in the market, so we'll want to see how it performs. To find out, we compared the PG278Q against my old Dell 3007WFP-HC. The 3007WFP-HC's IPS panel isn't particularly fast, with an 8-ms gray-to-gray spec, but this monitor has no internal scaler chip, so there's zero input lag from that source.
Well, I didn't expect to see the PG278Q running ahead of our ASIC-free reference unit. I suppose the victory could be the result of the PG278Q's faster pixel-switching times, but I suspect the real culprit is the GPU lag inherent to the screen mirroring mode we used. We've seen this same kind of thing before, with a frame or so of lag for one display or the other, depending on which video card output the display's using.
Anyhow, I'm not terribly concerned about that. What this result demonstrates nicely is that the G-Sync module in the PG278Q offers very low latency. It doesn't appear to introduce even a single frame's worth of additional input lag. That's exactly the outcome one would hope to see.
The G-Sync module driving this display is based on an FPGA rather than a custom chip, so I expected it to draw a little more power than the typical monitor's electronics. I'm not sure that it does, though. The three monitors are within a watt of one another at our typical brightness level of 200 cd/m². The PG278Q draws more power at peak and minimum, but we've already established that it's brighter overall at those levels. I'd chalk up the extra power use to the WLED backlight's additional candlepower.