Science, Technology, and pawbs
@TerrorBite @tech Sure, it's a hack, but why can't systems be smart about subpixel rendering?
In theory, an EDID can describe its subpixel layout to a system, and text renderers can use that information to decide whether subpixel rendering makes sense.
I guess there's a lot of information passing that would need to take place that can't/doesn't happen.
That seems pretty plausible to me. If you can have two way communication with a TV/monitor over say HDMI, why not implement that?
Though I suspect the answer may be that it's no longer really worth trying to implement it at this point.
@TerrorBite @tech I'd put it in the same bucket as the composite video comb-filter tricks used on 90's era games consoles to get new colours out of a limited palette.
I.e. a cute use of the odd quirks of a particular display technology that's now largely just of historic interest
@TerrorBite @tech We have a CRT, which I'm not sure subpixel rendering even works well on at all (I think it uses subpixels, but they're not aligned with display signal output pixels). We also have a bog-standard 22" 1080p LCD, we'll need to find a cable for that and see what subpixel antialiasing looks like on that.
I did notice that when we were in the OS installer on our laptop (to back up the OS), the text looked weirdly sharp compared to normal; I bet that was subpixel antialiasing in action. So maybe it is still useful.
But I personally think the disadvantages in dealing with screenshots outweigh the advantages of maybe slightly sharper text. Grayscale antialiasing looks fine.