Changing default font (Libra 2)?

[Phssthpok]

I have discovered that the default font on my Libra 2 renders "ǐ" (i with caron, U+01D0) as "ω" (U+03C9), and "ǎ" (a with caron, U+01CE) as "χ" (U+03C7). They show up correctly in the annotations window if I highlight the text containing them. The selected font is "Publisher default", but the book does not specify any font-family in the CSS other than "serif". I assume I'm ending up with some system default font (no idea what, though).

Rather than changing the font individually for each book, is there any way to replace this system default font with something that actually understands Unicode?

[rashkae]

I don't know of any easy way of changing the default font used by "Publisher" on the kobo, but you also don't have to change the font individually for each book. When you select a new font, that will the be default for all new books you open.

However, choosing a font with wide Unicode support for those characters will be limited. The problem you are experiencing can be fixed by sideloading your books to the Kobo in Kepub format. (There are lots of threads here, so I'll skip the howto unless it's needed.) People have difference preferences between epub and kepub on the Kobo, but one of the advantages of Kepub is automatic font substitution (which Epub does not.)

PS, the system default font is Georgia

[Phssthpok]

Originally Posted by rashkae

However, choosing a font with wide Unicode support for those characters will be limited.

I'm surprised that code points in the U+01xx range are unsupported (and that they are mapped to characters which have no seeming relation to the originals).

I'm even more surprised that the font used for the annotations page supports them while the default does not.

PS, the system default font is Georgia

Thanks. I'm astonished that Georgia does not support U+01xx code points.

[Phssthpok]

Originally Posted by rashkae

I don't know of any easy way of changing the default font used by "Publisher" on the kobo, but you also don't have to change the font individually for each book. When you select a new font, that will the be default for all new books you open.

Followup: I've just installed Bitter, Bookerly, ChareInk, Charis SIL, Linux Libertine, Literata and Noto Serif (since all the built-in fonts look frankly disgusting to my eyes) to see how they look, and have discovered a similar problem with Bitter -- i caron displays as ∫, a caron as Δ --but the others seem OK.

[DrChiper]

Originally Posted by Phssthpok

[...]have discovered a similar problem with Bitter -- i caron displays as ∫, a caron as Δ --but the others seem OK.

You could try Bitter Pro, which is more recent than Bitter.

[JSWolf]

Originally Posted by Phssthpok

I have discovered that the default font on my Libra 2 renders "ǐ" (i with caron, U+01D0) as "ω" (U+03C9), and "ǎ" (a with caron, U+01CE) as "χ" (U+03C7). They show up correctly in the annotations window if I highlight the text containing them. The selected font is "Publisher default", but the book does not specify any font-family in the CSS other than "serif". I assume I'm ending up with some system default font (no idea what, though).

Rather than changing the font individually for each book, is there any way to replace this system default font with something that actually understands Unicode?

Don't use Publisher Font for eBooks that do not have embedded fonts. Pick the font that you want.

If you want to find which font is being used when you do have Publisher Font selected when there are no embedded fonts, try ever font in the font list and see which one it is.

[Quoth]

Hmm not so easy for ǐ and ǎ (Caron i & a)

Caron ◌̌ Ǎ ǎB̌ b̌Č čČ̣ č̣Ď ďĚ ěÊ̌ ê̌F̌ f̌Ǧ ǧȞ ȟǏ ǐJ̌ ǰǨ ǩĽ  ľM̌ m̌Ň ňǑ ǒP̌ p̌Q̌ q̌Ř řŘ̩ ř̩Š šṦ ṧŤ ťǓ ǔǙ ǚV̌ v̌ W̌ w̌X̌ x̌Y̌ y̌Ž žǮ ǯ Cyrillic: В̌ в̌Г̌ г̌Ғ̌ ғ̌Д̌ д̌З̌ з̌Р̌ р̌Т̌ т̌Х̌ х̌

and

For legacy reasons, most letters that carry carons are precomposed characters in Unicode, but a caron can also be added to any letter by using the combining character U+030C ◌̌ COMBINING CARON, for example: b̌ q̌ J̌.

The characters Č, č, Ě, ě, Š, š, Ž, ž are a part of the Unicode Latin Extended-A set because they occur in Czech and other official languages in Europe, while the rest are in Latin Extended-B, which often causes an inconsistent appearance.

Unicode also encodes U+032C ◌̬ COMBINING CARON BELOW, for example: p̬.

—Wikipedia

So Ǎ ǎ and Ǐ ǐ are not in the Extended Latin A, which may be why you see the wrong thing.

(All working here on Linux, but I had to look up how to type a caron, it's Compose c <letter>
(Carat is Compose ^ <letter> )

So anyway, now I can test fonts.

You need a font that has extended Latin B as well as regular and extended Latin A.

And I still can't display old Irish dotless i, except with a few old broken fonts (typed letter is regular i). A Turkish dotless i is the wrong character.

[Phssthpok]

Originally Posted by Quoth

You need a font that has extended Latin B as well as regular and extended Latin A.

Right. I still wonder why they get mapped to such disparate code points, though (or I suppose the right code point being mapped to the wrong glyph); can't see any relationship between them.

You'd think that after more than a quarter of a century, Unicode would be Unicode would be Unicode. It's like the old days of using EBCDIC(s).

[Phssthpok]

Originally Posted by DrChiper

You could try Bitter Pro, which is more recent than Bitter.

Thanks.

[Phssthpok]

Originally Posted by JSWolf

Don't use Publisher Font for eBooks that do not have embedded fonts. Pick the font that you want.

Got it. I've moved to a Kobo from a Kindle, and was expecting that font selection would be on a per-book basis.

[Phssthpok]

Originally Posted by Quoth

For legacy reasons, most letters that carry carons are precomposed characters in Unicode, but a caron can also be added to any letter by using the combining character U+030C ◌̌ COMBINING CARON, for example: b̌ q̌ J̌.

After you pointed this out, I decided to edit the book in question to use a combining caron to avoid this sort of problem if I change font again.

[Guego]

Every week or so, there's a post on this forum about problems with unsupported unicode characters. I wonder why kobo sticks to Georgia as the default system font instead of using a font with a better unicode support.

[rantanplan]

Originally Posted by Guego

Every week or so, there's a post on this forum about problems with unsupported unicode characters. I wonder why kobo sticks to Georgia as the default system font instead of using a font with a better unicode support.

Because their main marketplace is north America where special character are rarely needed? Investment costs vs profit increase.

[Quoth]

Originally Posted by Guego

Every week or so, there's a post on this forum about problems with unsupported unicode characters. I wonder why kobo sticks to Georgia as the default system font instead of using a font with a better unicode support.

The first bunch of Kindles (and still true for 4 models) had essentially just basic Roman & Latin A, even though running Linux and it had been a solved problem 10+ years earlier. Even twenty years earlier DOS had better multiligual support than Amazon Kindle till models with azw3 released. Fortunately some early models (K3 oldest, but not DXG) got backported FW updates for azw3.

At least it's solvable now on Kobo and azw3. Still, Kobo is owned by Rakuten (also do Viber Chat) who is Japanese!

[Quoth]

Originally Posted by Phssthpok

After you pointed this out, I decided to edit the book in question to use a combining caron to avoid this sort of problem if I change font again.

On linux to type Combining caron
<letter><Shift Ctrl U>030c<enter>
so for i ǐ and a ǎ
Mostly Compose c <letter> is direct caron ǐ and ǎ

The combining caron looks same as direct caron in text editor. I have Caps Lock as Compose Key, both Shift = Caps Lock (either cancels) and holding Win key lets left mouse drag a window. Often Win is the default Compose Key