[Jellby]

Originally Posted by hvas

You might not like unhinted Literata but it's objectively closer to the glyphs of the actual font.

There I agree. The unhinted version seems to be rendered closer to the "real" version. But that's something half-expected. Hints are intended to make all vertical stems the same with, for instance, while with no hints the system will simply try to approach the "true" shape.

Originally Posted by GERGE

I thought so too, at first. But screenshots taken by Kobo devices are almost completely have the same letterforms as what I see on the device itself. I don't know what kind of system they are using but letter forms are almost the same. (strangely, i letter is distorted on the unhinted screenshot.)

Then the problem, if there is any, is with the hinting algorithm itself, and not with the nature of the screen. Maybe they changed the renderer to better fit the eink screen, or maybe that's just the way it is. How do the two versions render in a computer, at similar pixel size?

So, screenshots are not useless.

They are useless to show the difference between screens. Initially it was said (or that's what I interpreted) that the hinting methods used for LCDs are not valid for eink, because they are physically different. A screenshot cannot show that, a photo can.

Of course, subpixel rendering techniques that rely on the display structure are useles in eink, unless there's some method especifically tuned for that.