Unicode Grantha typeface on OS X (with glyphs from the IITM font)

347 views
Skip to first unread message

Ambarish Sridharanarayanan

unread,
May 25, 2014, 10:57:55 PM5/25/14
to granth...@googlegroups.com
Hello all,

As I mentioned on a different thread a few months ago, the Unicode standardisation effort motivated me to work on a typeface for Grantha. Windows support depends on Microsoft's mercy, but not so OS X support, so I picked OS X as my first target.

Prof. R. Kalyana Krishnan kindly granted me permission to reuse the glyphs from the IITMGrantha font he'd created. So I made use of the glyphs and added the required AAT tables. I'm happy to share the news that this typeface (that I've named Sampradaya) renders most text correctly on recent OS X systems. You can get Sampradaya v0.5 from:


Attached is a UTF-8 file with Grantha characters that should render correctly on OS X once the font-file is installed. Warning: does not work on Windows or Linux!

There are a few bugs, and a few areas of improvement:
* missing glyphs for anunāsika (candrabindu) and many combined conjuncts
* mis-rendering certain conjuncts with more than 2 consonants (trya in tryambakaḥ, tyva in mr̥tyvagnī, tpla in utplavaḥ)
* bad kerning for vowel-markers
* hand-drawn glyphs in some cases that could do with polishing
* no Graphite support :-)

Let me know if you tried it out and it worked!
Ambarish
VishnuNamasahasrakaStotram.txt

Shriramana Sharma

unread,
May 28, 2014, 10:14:48 PM5/28/14
to granth...@googlegroups.com
Dear sir,

Thank you for your work on this. A few remarks:

1) Please correct me if I am missing something but it seems you have
omitted a very important details -- to wit, what LICENCE you are
releasing this font under. Obviously, it depends on what LICENCE Dr
Kalyanakrishnan permitted you to use his glyphs under. IIUC most free
fonts use OFL today. It would be good if you could clarify this point,
especially since IIUC the IITM Grantha font has been available for
many years (decades?) but people are not able to distribute modified
versions because it is (was) not under an Open Source licence. So this
is an important issue.

2) Please note that adding Graphite supports means it is supported on
all platforms that LibreOffice runs on. You have written like "Linux
support via Graphite TBD". Obviously we can't expect Microsoft to
hurry up on Grantha support and even if they release it in Windows 10
or such not everyone can buy. But if Graphite support is done, then it
will work on Firefox/XeTeX/LibreOffice on all platforms, since these
use HB-NG (and are usually compiled with Graphite support enabled). So
please correct that wording also.

3) Finally, from what vibes I'm getting from various sources, Mac will
slowly switch over to OT and give up on AAT (though they may retain
for backward compatibility). So I think it is better to concentrate on
Graphite (which is cross-platform) rather than AAT (which is
Mac-specific).

Shriramana Sharma.


--
Shriramana Sharma ஶ்ரீரமணஶர்மா श्रीरमणशर्मा

Ambarish Sridharanarayanan

unread,
Jun 2, 2014, 3:58:52 AM6/2/14
to granth...@googlegroups.com
namaskarōmi,

Thanks for your feedback. Please see inline.

On Wednesday, 28 May 2014 19:14:48 UTC-7, Shriramana Sharma wrote:

1) Please correct me if I am missing something but it seems you have
omitted a very important details -- to wit, what LICENCE you are
releasing this font under. Obviously, it depends on what LICENCE Dr
Kalyanakrishnan permitted you to use his glyphs under. IIUC most free
fonts use OFL today. It would be good if you could clarify this point,
especially since IIUC the IITM Grantha font has been available for
many years (decades?) but people are not able to distribute modified
versions because it is (was) not under an Open Source licence. So this
is an important issue.

I'm releasing this under the OFL. I've mentioned that in the font-file itself, but do you think it would help to include that in the bitbucket page somewhere? I'm a total newbie at this process of releasing fonts, and so am happy to receive feedback. And yes, Prof. Kalyanakrishnan has gave me (e-mail and over the phone) permission to reuse the glyphs and release a font containing them under an "open-source" licence of my choice.
 
2) Please note that adding Graphite supports means it is supported on
all platforms that LibreOffice runs on. You have written like "Linux
support via Graphite TBD". Obviously we can't expect Microsoft to
hurry up on Grantha support and even if they release it in Windows 10
or such not everyone can buy. But if Graphite support is done, then it
will work on Firefox/XeTeX/LibreOffice on all platforms, since these
use HB-NG (and are usually compiled with Graphite support enabled). So
please correct that wording also.

Made a clarification in the overview. I know very little about Graphite, so I'm happy to get any help in that regard (Pull requests welcome!)
 
3) Finally, from what vibes I'm getting from various sources, Mac will
slowly switch over to OT and give up on AAT (though they may retain
for backward compatibility). So I think it is better to concentrate on
Graphite (which is cross-platform) rather than AAT (which is
Mac-specific).

There were a few different reasons I started out with AAT on OS X. I'm a typography newbie and amateur, and I first tried adding OT support, before I realised I'd be beholden to Microsoft's timelines. AAT seemed to be self-contained (all the information is in the font-file itself) and thus a better choice; at that time I didn't know about Graphite.

I now see adding Graphite tables would be useful, but I don't think AAT support is a waste; my primary platform is OS X, and there's something to be said for the font rendering cleanly in every Carbon application, from TextEdit to Finder to Firefox to Chrome. In any case, AAT support is mostly done (modulo bug-fixing), and I'm happy to focus on Graphite support now.

svasti,
Ambarish

Ambarish Sridharanarayanan

unread,
Jun 2, 2014, 4:02:05 AM6/2/14
to granth...@googlegroups.com
On Monday, 2 June 2014 00:58:52 UTC-7, Ambarish Sridharanarayanan wrote:
 
the font rendering cleanly in every Carbon application, from TextEdit to Finder to Firefox to Chrome.

That should have been "every Cocoa application", of course. 

Vinodh Rajan

unread,
Jun 5, 2014, 1:03:56 PM6/5/14
to granth...@googlegroups.com
Hi Ambarish,

Thanks a lot for this work. 

For some reasons vowel sign /o/ is not working. Within Textedit, the /kombu/ is shifted to beginning of the line. Within Safari, the vowel sign doesn't join with the consonant. I am running OS X 10.9.3.

See the screenshots below:






Also, with Pages 5.3, nothing is being displayed.




But when the textfile is opened sing Safari/Firefox/Chrome. The display is working fine.








Regards,

Vinodh Rajan



--
You received this message because you are subscribed to the Google Groups "Grantha கிரந்தம்" group.
To unsubscribe from this group and stop receiving emails from it, send an email to grantha-lipi...@googlegroups.com.
To post to this group, send email to granth...@googlegroups.com.
Visit this group at http://groups.google.com/group/grantha-lipi.
For more options, visit https://groups.google.com/d/optout.



--
http://www.virtualvinodh.com

Ambarish Sridharanarayanan

unread,
Jun 23, 2014, 2:32:30 AM6/23/14
to granth...@googlegroups.com
Thanks, Vinodh. The kombu for 'ō' and 'au' was one of the earlier things I added, and it's quite possible a latter edit has regressed it. I'll take a look, but it may be a few weeks.

Regarding Pages, I don't know if it supports the SMP; I don't have iWork installed, but I wouldn't be surprised if it were a Pages incompatibility.

Glad to hear Firefox works fine.

Regarding Safari, can I ask you to perform a test for me? When you are at Richard Ishida's page on Safari, and notice the kombu not rendering correctly, can you add a Devanagari or Tamil letter to the same line, and see if that makes a difference? I noticed a bug in Safari, where it seems to trigger correct AAT processing only when it sees a glyph from a script that it knows requires AAT table processing. I'm not sure that that is the bug, but it seems to be. Let me know.

Thanks!
Ambarish

Vinodh Rajan

unread,
Jul 1, 2014, 9:28:46 PM7/1/14
to granth...@googlegroups.com

On 23 June 2014 07:32, Ambarish Sridharanarayanan <amba...@ksharanam.net> wrote:
Regarding Safari, can I ask you to perform a test for me? When you are at Richard Ishida's page on Safari, and notice the kombu not rendering correctly, can you add a Devanagari or Tamil letter to the same line, and see if that makes a difference? I noticed a bug in Safari, where it seems to trigger correct AAT processing only when it sees a glyph from a script that it knows requires AAT table processing. I'm not sure that that is the bug, but it seems to be.

Yeah. It appears so.



Shriramana Sharma

unread,
Oct 28, 2016, 6:10:05 AM10/28/16
to granth...@googlegroups.com
Hello. Is there any update on this Grantha font? I got a query from
somebody who is doing a research publication using Mac, so it may be
useful for them if it is published and working.

Thanks!

Ambarish Sridharanarayanan

unread,
Oct 31, 2016, 2:48:15 AM10/31/16
to granth...@googlegroups.com
Hi there,

There are 2 kinds of remaining issues;

1. The ‘o’ and ‘au’ issue Vinodh brought up: these look like OSX rendering issues. I have 3 reasons for believing so:
* The Tamil fonts shipped with OSX have identical tables, but seem to work.
* These are only triggered in TextEdit and Safari; Firefox and Chrome work fine.
* Adding another non-Grantha Indic character in the same line fixes the rendering; it seems to trigger a different rendering path.
Anyway, I asked in the Apple Developer forums about these a while ago, but I didn’t hear back. As above, there’s a workaround often.

2. A small number of remaining bugs when rendering triple or quadruple conjuncts, where the result is technically correct, but abnormal in terms of how Grantha was actually used. I haven’t had time to fix these, but these should be rare: I know of very few such conjucts (kārtsnya, dāridrya, etc.)

So as far as I can see, with a few limitations, the font is usable and working. If they’re willing to, I’d love for your acquaintance to try it out and send me feedback.

Ambarish
> --
> You received this message because you are subscribed to a topic in the Google Groups "Grantha கிரந்தம்" group.
> To unsubscribe from this topic, visit https://groups.google.com/d/topic/grantha-lipi/Rpa-ftV0GFY/unsubscribe.
> To unsubscribe from this group and all its topics, send an email to grantha-lipi...@googlegroups.com.
> To post to this group, send email to granth...@googlegroups.com.
> Visit this group at https://groups.google.com/group/grantha-lipi.

O தோழன்:-}

unread,
Oct 31, 2016, 4:38:22 AM10/31/16
to granth...@googlegroups.com

Namasthe Ambarish ji

Do you have code for vedic notes(swara) for udaaththa,anudaattam and plutham.
Please inform.

Thanks
Kalyan ji


On 31 Oct 2016 6:48 a.m., "Ambarish Sridharanarayanan" <amba...@ksharanam.net> wrote:
Hi there,

There are 2 kinds of remaining issues;

1. The ‘o’ and ‘au’ issue Vinodh brought up: these look like OSX rendering issues. I have 3 reasons for believing so:
    * The Tamil fonts shipped with OSX have identical tables, but seem to work.
    * These are only triggered in TextEdit and Safari; Firefox and Chrome work fine.
    * Adding another non-Grantha Indic character in the same line fixes the rendering; it seems to trigger a different rendering path.
    Anyway, I asked in the Apple Developer forums about these a while ago, but I didn’t hear back. As above, there’s a workaround often.

2. A small number of remaining bugs when rendering triple or quadruple conjuncts, where the result is technically correct, but abnormal in terms of how Grantha was actually used. I haven’t had time to fix these, but these should be rare: I know of very few such conjucts (kārtsnya, dāridrya, etc.)

So as far as I can see, with a few limitations, the font is usable and working. If they’re willing to, I’d love for your acquaintance to try it out and send me feedback.

Ambarish

> On 28 Oct, 2016, at 03:10, Shriramana Sharma <jama...@gmail.com> wrote:
>
> Hello. Is there any update on this Grantha font? I got a query from
> somebody who is doing a research publication using Mac, so it may be
> useful for them if it is published and working.
>
> Thanks!
>
>
> --
> Shriramana Sharma ஶ்ரீரமணஶர்மா श्रीरमणशर्मा
>
> --
> You received this message because you are subscribed to a topic in the Google Groups "Grantha கிரந்தம்" group.
> To unsubscribe from this topic, visit https://groups.google.com/d/topic/grantha-lipi/Rpa-ftV0GFY/unsubscribe.
> To unsubscribe from this group and all its topics, send an email to grantha-lipi+unsubscribe@googlegroups.com.

> To post to this group, send email to granth...@googlegroups.com.
> Visit this group at https://groups.google.com/group/grantha-lipi.
> For more options, visit https://groups.google.com/d/optout.

--

You received this message because you are subscribed to the Google Groups "Grantha கிரந்தம்" group.
To unsubscribe from this group and stop receiving emails from it, send an email to grantha-lipi+unsubscribe@googlegroups.com.

Ambarish Sridharanarayanan

unread,
Nov 1, 2016, 1:16:36 AM11/1/16
to granth...@googlegroups.com
namastē,

Unfortunately, the original typeface I used (the IITM font) doesn’t have a glyph for the pluta marker. And udātta and anudātta haven’t been standardised in Unicode for Grantha yet.

Thanks.

To unsubscribe from this group and all its topics, send an email to grantha-lipi...@googlegroups.com.

Shriramana Sharma

unread,
Nov 1, 2016, 2:13:25 AM11/1/16
to granth...@googlegroups.com
Dear Shri Ambarish,

There will be no separate udatta/anudatta in Unicode for Grantha alone
since the shapes of these characters are generic and not
Grantha-specific. There is only one set of non-script-specific Vedic
markers: http://www.unicode.org/charts/PDF/U1CD0.pdf. Due to
historical reasons, the anudatta and svarita (under the name of STRESS
SIGN UDATTA) have been encoded in the Devanagari block but are to be
used from there for all Indic scripts:
http://www.unicode.org/charts/PDF/U0900.pdf. Please also check the
Devanagari Extended block http://www.unicode.org/charts/PDF/UA8E0.pdf
to ensure that none of the characters there are used in other scripts.
(To my knowledge there are none.)

Ambarish Sridharanarayanan

unread,
Nov 1, 2016, 10:25:01 AM11/1/16
to granth...@googlegroups.com
namastē,

I see; thanks for clarifying. I guess the situation is similar to that of the danḍa, which is also in the Devanagari block despite being used in other scripts.

Ambarish

O தோழன்:-}

unread,
Nov 1, 2016, 10:33:16 AM11/1/16
to granth...@googlegroups.com

Thank you for your reply

Kind regards
Kalyanji

Reply all
Reply to author
Forward
0 new messages