LA subscript seems to be much right in Kannada

[sridatta1]

The LA subscript appears to be at much right in Tiro Kannada in combination with consonants having narrow width.

Similar to Telugu fonts, in Kannada handwriting, hand-painted signages, manuscripts, many printed texts/fonts LA subscript is commonly centred below. Attaching some samples from random books

From an early print, 1859 From a modern newspaper

So with consonants having narrow/ normal width like ಕ್ಲ ಗ್ಲ ತ್ಲ ಪ್ಲ ಶ್ಲ ಬ್ಲ, the la ottu can be centred below. This would also apply to top- vowel signs ಲ್ಲಿ ಲ್ಲೇ With the wider consonants and right side vowel signs, unlike Telugu la ottu can be placed at right/ existing shaping can be retained, but should not be too much to the right.

Examples from some other fonts Earlier version of Noto Serif Kannada

In Some of the Nudi fonts

Adobe Kannada

[sridatta1]

There is one more test case which forgot to include. For the the aspirated consonants having tear drop though might be rare in actual texts, the current implementation seems fine.

[tiroj]

This is a more complicated issue than might first be assumed. Unlike Telugu, Kannada subscript handling typically uses a fairly simple kerning method, in which the subscripts have a positive advance width and an overhang of the subscript shape on the left. A standard overlap on the left enables general kerning below the preceding base without adjustment for most bases, and can be set to avoid collision with the aspirate teardrop. The advance width and proportional sidebearing on the right—taking into account the shape of the individual subscript—permits kerning refinement of the relationship to a following base or to a second subscript.

Handling the –La differently might require abandoning this simple model, at least for this subscript, and moving to the sort of complex mark handling used in Telugu. Unlike mark anchor attachment, kerning has a knock-on horizontal layout effect, which means any moving of the –La subscript further left—especially if optically centered—under bases of varying width also affects the distance between the adjacent bases.

I will think about this some, and see if I can come up with a solution that retains the general simplicity of the Kannada layout model.

What is the source of that 1859 print example? That looks like a very useful resource.

[sridatta1]

This is book from 1859 https://archive.org/details/dli.CiXIV290-33/mode/2up Don't know if other books from this series are digitised or not, didn't find them. Have seen some other books in this type.

[tiroj]

That book is interesting in illustrating almost all subscript letters centered below bases, so much more like Telugu than one sees in most Kannada type.

[sridatta1]

This is book from 1859 https://archive.org/details/dli.CiXIV290-33/mode/2up Don't know if other books from this series are digitised or not, didn't find them. Have seen some other books in this type.

On Sun, 10 Jul, 2022, 11:33 pm John Hudson, @.***> wrote:

This is a more complicated issue than might first be assumed. Unlike Telugu, Kannada subscript handling typically uses a fairly simple kerning method, in which the subscripts have a positive advance width and an overhang of the subscript shape on the left. A standard overlap on the left enables general kerning below the preceding base without adjustment for most bases, and can be set to avoid collision with the aspirate teardrop. The advance width and proportional sidebearing on the right—taking into account the shape of the individual subscript—permits kerning refinement of the relationship to a following base or to a second subscript.

Handling the –La differently might require abandoning this simple model, at least for this subscript, and moving to the sort of complex mark handling used in Telugu. Unlike mark anchor attachment, kerning has a knock-on horizontal layout effect, which means any moving of the –La subscript further left—especially if optically centered—under bases of varying width also affects the distance between the adjacent bases.

I will think about this some, and see if I can come up with a solution that retains the general simplicity of the Kannada layout model.

What is the source of that 1859 print example? That looks like a very useful resource.

— Reply to this email directly, view it on GitHub https://github.com/TiroTypeworks/Indigo/issues/26#issuecomment-1179772923, or unsubscribe https://github.com/notifications/unsubscribe-auth/AFRPT247DBOHD665OBC6LA3VTMGAVANCNFSM53E3XUWQ . You are receiving this because you authored the thread.Message ID: @.***>