Resolution

[Blazco]

Did the resolution test on testufo

It says 2561 x 1441 when it should say 2560 x 1440 surely, added picture below.

That a glitch or normal?

[RealNC]

That a glitch or normal?

It's normal when you don't use 100% zoom factor in your browser.

[Chief Blur Buster]

Did the resolution test on testufo

It says 2561 x 1441 when it should say 2560 x 1440 surely, added picture below.

That a glitch or normal?

Your web browser is not at 100% zoom so it's off-by-one rounding behavior at certain zoom ratios.

Web browsers make it extremely difficult for websites to display things at 1:1 native resolution during a zoomed state. However, I was able to create algorithms to un-do the browser zoom + un-do the OS DPI zoom for certain TestUFO tests. It works most of the time, but it is sometimes off-by-one.

No other websites successfully accurately do the 1:1 native resolution during these odd browser zooms, so I'm doing much more accurately.

Set your browser to a more common zoom of an even-numbered multiple or divisor of 100% (100% zoom or 200% zoom) and then www.testufo.com/resolution will be normal. It will often be slightly off for odd slight zooms like 90% or 110% zoom.

[RealNC]

Your web browser is not at 100% zoom so it's off-by-one rounding behavior at certain zoom ratios.

It's off by one for me even at 100%. This appears to be due to Windows 10 DPI scaling (125% here.)

[Chief Blur Buster]

Your web browser is not at 100% zoom so it's off-by-one rounding behavior at certain zoom ratios.

It's off by one for me even at 100%. This appears to be due to Windows 10 DPI scaling (125% here.)

Yes, it's a combination of (browser zoom & Windows DPI), not just browser zoom.

Web browsers make it extremely difficult to get perfect 1:1 native resolution because the Javascript API's give me a 'fake resolution' and I have to apply a custom scaling factor.

Even when off by one, a test pattern will still (usually) be 1:1 native resolution by omitting a pixel row or pixel column.

I could hypothetically round off to the nearest clean even number (e.g. 1441 -> 1440) and compensate that way. I may do that (eventually) but for now, the ability to single-click to a 1:1 native test pattern -- without mandatorily requiring user to reconfigure zoom/DPI -- is still a neat new TestUFO 3.0 feature.