notofonts / bengali Goto Github PK

Title

Half forms are not to be used in Bengali

Font

Noto Sans Bengali & Noto Serif Bengali

Where the font came from, and when

Tested on https://fonts.google.com/noto/fonts site:
June 26, 2022

Issue

Half forms are not the feature of Bengali script, they are not found traditonally. For supporting arbitrary conjunct sequences, some of font makers have tried inventing Devanagari-like half forms for Bengali. The conjuncts with smaller half forms look very incorrect.
So for the other conjuncts where there are not ligatures in the font, either the ligatures can be added wherever it's necessary or just the virama should be used for the fallback.
e.g.
Text: স্স গ্ঘ ক্শ প্ফ ক্খ ম্হ
2. Observed results

Tiro Bangla font

Also the following comment on Bengali half-forms can be looked into
#1 (comment)
#1 (comment)

It is customary to write rṛ as the independent_vocalic_r with a repha instead of ra with vowel sign vocalic r in Devanagari and most other Indic scripts.
In Noto Sans Devangari ( Ver. 1.05) रृ is

I'm a native user of Devanagari and Kannada , I have seen rṛ being written as independent_vocalic_r with a repha in the words like निरृत नैरृत्य निरृति only.
In Noto Serif Devanagari these occur correctly

This is also true for other scripts as well.
Similar forms should occur in Sans and Serif fonts for Kannada, Bengali, Odia, Gujarati scripts which is not the current case.
Modern Telugu and Malayalam texts don't use repha.
For Gujarati-

For other scripts see section 5 of https://unicode.org/L2/L2017/17423-core-spec-indic.pdf

Font

NotoSansBengali-*
All NotoSans Bengali fonts

Where the font came from, and when

https://notofonts.github.io/bengali/
Date: 2023/2/22

Font Version

Version 2.003; ttfautohint (v1.8.4.7-5d5b)

OS name and version

Windows 11 Pro 10.0.22000 build 22000

Application name and version

Google Chrome 110.0.5481.100

Issue

Related to the already closed issue #6.
Noto Sans Bengali does not appear correctly the TTRA character as a ligature.
With recent updates the following strings now appear as:
ত্ত্র
U+09A4, TA
U+09CD, VIRAMA
U+09A4, TA
U+09CD, VIRAMA
U+09B0, RA

Unfortunately this glyph is identical to the TTA + RA glyph.
ট্র
U+099F, TTA
U+09CD, VIRAMA
U+09B0, RA

TTA(U+099F) and TA(U+09A4) represent different sounds. In other words, this update is not desirable.
On the other hand, the Noto Serif fonts display the ligature correctly.

Font

NotoSerifBengali-Regular.ttf

Where the font came from, and when

Site: https://github.com/googlefonts/noto-fonts/blob/9c74a9735acb66df9a41ad86e58020430ede9a9d/phaseIII_only/hinted/ttf/NotoSerifBengali/NotoSerifBengali-Regular.ttf
Date: 2019-11-04

Font version

Version 2.000; ttfautohint (v1.8.2)

Issue

Candrabindu is not properly centered above anji.

Character data

ঀঁ
U+0980 BENGALI ANJI
U+0981 BENGALI SIGN CANDRABINDU

Screenshot

Font

NotoSansBengali-Regular.ttf

Where the font came from, and when

Site: https://github.com/googlefonts/noto-fonts/blob/46eb8fa0c46d16a87eee3e763baa3efb7423a8b4/unhinted/NotoSansBengali/NotoSansBengali-Regular.ttf
Date: 2019-10-17

Font version

Version 2.000

Issue

Ba with a nukta has the same subjoined form as ra in this font.

Character data

ক্ব়
U+0995 BENGALI LETTER KA
U+09CD BENGALI SIGN VIRAMA
U+09AC BENGALI LETTER BA
U+09BC BENGALI SIGN NUKTA

Screenshot

I'm here at the TDi at Reading University with Marek, and he asked me to critique the Noto Serif Bengali. For a little background, I grew up in Bangladesh where I have worked for the last decade as a graphic designer with Latin and Bengali designs. My feedback on the Noto Serif Bengali font:

• The ও strikes the eye as feeling too large
  

• The থ head (curl/ball) feels too small and cramped
  

• The স্থ (sth) feels very stunted and short
  

• The এ feels like it extends too high
  

• The bottom tail of the হ/ই feels too short and cramped into the upper body
  

• The perfectly vertical initial stem of the দ seems too geometric
  

• The chandrabindu is over the consonant instead of the vowel, which is wrong
  

• The 'গ' is too cramped together, it feels like it should extend further down
  

• The 'গ্র' (ga with raphala) compresses this even more, and the contrast between the compressed and non-compressed looks even more awkward
  

• The 'ষ্ট' feels like it's falling forward
  

• The rikaar is too compressed into the 'ত' ('তৃ')
  

• The flourish in the 'i-kaar' seems pulled high
  

• The ফ্র feels too condensed vertically
  

• The reph feels too small and weak
  

• The top left elbow of the জ feels weak and indecisive
  

Honestly, I wouldn't use it the way it stands. I do use the Noto Sans Bengali now and then, but I really don't like their weak "ণ" in the sans.

Hope that helps!

I live in Bangladesh, and sometimes have to create documents/slides with both 
English and Bengali writings. But Noto font family has separate fonts for 
English and Bengali.

How can I merge them together? Or if you could provide a merged font with both 
English and Bengali support, that would be great.

Original issue reported on code.google.com by [email protected] on 1 Jan 2015 at 7:53

Font

NotoSerifBengali-Regular.otf

Where the fonts came from, and when

Site: https://github.com/googlefonts/noto-fonts/blob/4d7396a4fae475975b367e225f2e8f49292d8837/unhinted/otf/NotoSerifBengali/NotoSerifBengali-Regular.otf
Date: 2022-04-16

Font version

Version 2.001

Issue

The pre-base vowel glyph in ⟨প্টে⟩ is not kerned as much as it is in ⟨প্টৌ⟩. The presence or absence of the post-base glyph vowel should not affect kerning on the other side of the cluster.

Character data

প্টে
U+09AA BENGALI LETTER PA
U+09CD BENGALI SIGN VIRAMA
U+099F BENGALI LETTER TTA
U+09C7 BENGALI VOWEL SIGN E
প্টৌ
U+09AA BENGALI LETTER PA
U+09CD BENGALI SIGN VIRAMA
U+099F BENGALI LETTER TTA
U+09CC BENGALI VOWEL SIGN AU

Screenshot

Defect Report

Hi, some characters are needed in Bengali font. I want ্ৱ as ্ব with a line or dot below (as in ৱ or its older form র). Also an Unicode characters is missing in Noto Sans Bengali font: u+0980 (anji). A virama that doesn't create any conjuncts is also needed. It is u+09cf in Kohinoor Bangla font.

Font

NotoSansBengali-Regular.ttf.

Source

Site: https://noto-website-2.storage.googleapis.com/pkgs/NotoSansBengali-hinted.zip
Date: 2017-01-12

OS name and version

Android 6.0.1

Font

NotoSansBengali-*
All NotoSans Bengali fonts

Where the font came from, and when

https://github.com/googlefonts/noto-fonts/tree/main/hinted/ttf/NotoSansBengali　
Date: 2021/3/1

OS name and version

Windows 10 1909

Application name and version

Google Chrome 89.0.4389.90

Issue

Noto Sans Bengali does not display the TTRA character as a ligature.
The character form in NotoSerifBengali and prior version of NotoSansBengali had the character displaying correctly.
According to our records the version 2.000 (2017/9/20) had the character form as the following:
https://github.com/googlefonts/noto-fonts/blob/bc9353da6f625b7a47035de952c2e8fd9e3ae889/hinted/NotoSansBengali-Regular.ttf

However, in Version 1.03 (2015/9/26), the font displays the character as a ligature as the following:
https://github.com/googlefonts/noto-fonts/blob/ad153f5a14c838afb16c83f0a02784c7605f1e21/hinted/NotoSansBengali-Regular.ttf

On the other hand, the Noto Serif fonts display the ligature correctly.

Character data

ত্ত্র
U+09A4, TA
U+09CD, VIRAMA
U+09A4, TA
U+09CD, VIRAMA
U+09B0, RA

Harfbuzz hb-view and hb-shape

[tabeng=0+707|viramabeng=0@-108,0+0|tarabeng=2+715]

Defect Report

Hi, traditionally and also in current handwritten and serif fonts many Bengali characters have knots which are essential in Bengali/Assamese script. But many sans fonts
including Noto Sans Bengali don't have knots in Bengali fonts while they have knots in other scripts (example: Devanagari ढ, Gujarati જ, Tamil அ etc), except for one Bengali character: ঙ. Knots should be added to Bengali font too.
The characters that need knots are অ আ ই ঈ এ ঐ ও ঔ ক খ ঞ ট ঢ ণ ত থ ধ ন ফ ভ ম ল শ হ ঢ় ১ ৩ and ৯ .

Font

NotoSansBengali-Regular.ttf.

Source of the font

Site: https://noto-website-2.storage.googleapis.com/pkgs/NotoSansBengali-hinted.zip
Date: 2017-01-12

It looks like the UI font should have (but no longer has) different feature code to the non-UI font. In particular, deep conjuncts should be squished and lower u-vowel marks should be replaced with UI variants:

$ shape ~/Downloads/android_fonts-master/api_level/30/NotoSansBengaliUI-Regular.otf 'কফ্রুড্রুম্রুদ্বুদ্বৃট্টুক'
[kabeng=0+807|pharasquishbeng=1+877|uvowelsignvattuUIbeng=1@-221,0+0|ddarasquishbeng=5+700|uvowelsignvattuUIbeng=5@-5,0+0|marasquishbeng=9+684|uvowelsignvattuUIbeng=9@-4,0+0|dabasquishbeng=13+633|uvowelsignvattuUIbeng=13@1,0+0|dabasquishbeng=17+633|rvocalicvowelsignvattuUIbeng=17@1,0+0|ttattasquishbeng=21+652|uvowelsignvattuUIbeng=21@-72,0+0|kabeng=25+807]

(But they are not been squished in current UI builds)

$ shape variable_ttf/NotoSansBengaliUI-VF.ttf 'কফ্রুড্রুম্রুদ্বুদ্বৃট্টুক'
[kabeng=0+807|pharabeng=1+899|uvowelsignlowbeng=1@-221,0+0|ddarabeng=5+712|uvowelsignlowerbeng=5@-4,0+0|marabeng=9+658|uvowelsignlowbeng=9@-4,74+0|dababeng=13+617|uvowelsignlowbeng=13+0|dababeng=17+617|rvocalicvowelsignlowbeng=17+0|ttattabeng=21+648|uvowelsignlowbeng=21@-111,0+0|kabeng=25+807]

Additionally, some conjuncts should not be formed in the UI build if they end up too deep:

$ shape ~/Downloads/android_fonts-master/api_level/30/NotoSansBengaliUI-Regular.otf 'কহ্লহ্লু'
[kabeng=0+807|halabeng=1+623|hahalfbeng=4+530|labeng=4+731|uvowelsignbeng=4+0]

(But they are still being formed in current UI builds)

$ shape variable_ttf/NotoSansBengaliUI-VF.ttf 'কহ্লহ্লু'
[kabeng=0+807|halabeng=1+623|halabeng=4+623|uvowelsignlowerbeng=4@-41,0+0]

see #14, in particular the following image:

https://cloud.githubusercontent.com/assets/20583790/17035451/3121b4e2-4f7f-11e6-910c-0a32f586ee89.jpg

The letter doesn't extend far down enough; the curve ends too early. This may have been an attempt to match the curvature of the other similar strokes in the ন and ম, but it needs to be different here. I hope the image makes it clear.

From @lemzwerg via e-mail: GDI hinting [of NotoSansBengali], as shown e.g. on this page, gets clipped at the top. This could be fine-tuned with ttfautohint options, but I think it would be better to adjust NotoSansBengali's value of usWinAscent in the OS/2 table – if you look at this image you can see that even with manual hinting the top gets cut off also at 9px. The same happens for NotoSansDevanagari, BTW.

Font

NotoSerifBengali-Regular.ttf

Where the font came from, and when

Site: https://noto-website-2.storage.googleapis.com/pkgs/NotoSerifBengali-unhinted.zip
Date: 2018-02-28

Font Version

Version 1.01 uh

Issue

The reph is not centered above the khanda ta, as it is in Noto Sans Bengali.

Character data

র্ৎ
U+09B0 BENGALI LETTER RA
U+09CD BENGALI SIGN VIRAMA
U+09CE BENGALI LETTER KHANDA TA

Screenshot

Phase II Bengali TTF fonts need U+02BC 'MODIFIER LETTER APOSTROPHE' (it's already in SOW15).

Alternatively: we need to ship Phase III Bengali

Font

NotoSerifBengali-Regular.ttf.

Where the font came from, and when

Site: https://github.com/googlefonts/noto-fonts/blob/master/phaseIII_only/hinted/ttf/NotoSerifBengali/NotoSerifBengali-Regular.ttf
Date: 2019-12-09

Font Version

2.000

OS name and version

Ubuntu Bionic amd64 is what I am running, but the issue is the names of the glyphs in the TTF, so I don't think which OS will have an affect.

Issue

The source for this font has friendly names such as k_tt_ra-beng. When the TTF is built, this glyph gets the name uni099509CD099709CD09B0, but it should be uni099509CD099F09CD09B0.

1. Steps to reproduce
   View TTF in some font viewer program.
2. Observed results
   uni099509CD099709CD09B0 for k_tt_ra-beng
3. Expected results
   uni099509CD099F09CD09B0 for k_tt_ra-beng
4. Additional information
   The full list of issues I found is

   <key>public.postscriptNames</key>
   <dict>
     <key>k_tt_ra-beng</key>
-    <string>uni099509CD099709CD09B0</string>
+    <string>uni099509CD099F09CD09B0</string>
     <key>k_tta_iMatra-beng</key>
-    <string>uni099509CD099509BF</string>
+    <string>uni099509CD099F09BF</string>
     <key>l_pa-beng</key>
-    <string>uni09B209CD09A9</string>
+    <string>uni09B209CD09AA</string>
     <key>m_pa-beng</key>
-    <string>uni09AE09CD09A9</string>
+    <string>uni09AE09CD09AA</string>
     <key>n_tt_iMatra-beng</key>
-    <string>uni09A809CD099F09BF</string>
+    <string>uni09A809CD099F09CD09BF</string>
     <key>p_pa-beng</key>
-    <string>uni09AA09CD09A9</string>
+    <string>uni09AA09CD09AA</string>
     <key>s_pa-beng</key>
-    <string>uni09B809CD09A9</string>
+    <string>uni09B809CD09AA</string>
     <key>s_tta-beng</key>
-    <string>uni09B809CD0997</string>
+    <string>uni09B809CD099F</string>
     <key>ss_pa-beng</key>
-    <string>uni09B709CD09A9</string>
+    <string>uni09B709CD09AA</string>
     <key>ss_tta-beng</key>
-    <string>uni09B709CD0997</string>
+    <string>uni09B709CD099F</string>
     <key>ss_tta_iMatra-beng</key>
-    <string>uni09B709CD099509BF</string>
+    <string>uni09B709CD099F09BF</string>
   </dict>

The GlyphsApp source file for this font was converted to UFO, so a derivative found could be produced. The above diff is from the diff between the original and fixed files of lib.plist in the UFO.

Note that some of the incorrect AGLFN names reference U+09A9, which is not an assigned character in Unicode.

Hello!

This is an automatically-generated report about possible interpolation problems in NotoSerifBengali[wdth,wght].ttf, as found in the Google Fonts catalog.

The particular version of the font that was tested was NotoSerifBengali-v2.003/NotoSerifBengali-v2.003.zip.

To download a PDF version of this report with helpful visuals of the problems, click here; Or to view it on the GitHub website, click here.

The report follows:

Glyph uni09C1.alt was not compatible:
  Masters: 'wdth=62.5', 'wdth=62.5 wght=100.0':
    Contour 0 start point differs: 0 in 'wdth=62.5', 18 in 'wdth=62.5 wght=100.0'; reversed: True
Glyph uni09C4 was not compatible:
  Masters: 'wdth=62.5', 'wdth=62.5 wght=100.0':
    Contour 1 structures differ: 'wdth=62.5', 'wdth=62.5 wght=100.0'

This report was generated using the fonttools varLib.interpolatable tool. We understand that sometimes the tool generates false-positives. Particularly for more complicated font designs. If you did not find this report useful, please apologize and ignore & close it.

To give feedback about this report, please file an issue or open a discussion at fonttools.

I noticed one problem of NotoSansBengali in rendering "ঁ". e.g. তাঁদের should be rendered like this.

tader

It would be great if we can fix this rendering issue.
.

Defect Report

Noto fonts can be obtained from https://github.com/googlefonts/noto-fonts/tree/master/unhinted/otf/ (unhinted OTF) or https://github.com/googlefonts/noto-fonts/tree/master/hinted/ttf/ (hinted TTF). If you find a problem with either of these then please use this template to file a defect report. For feature requests and other matters, you can use part of the template and delete what you don't need.

Title

Issues with displaying this "হ্যাঁ" character when using Noto sans bengali

Font

Noto sans bengali

Where the font came from, and when

Site: https://fonts.google.com/noto/specimen/Noto+Sans
Date: 2022-05-16 (preferred format)

Font Version

• Win - but we are not sure if it's limited just to this platform

OS name and version

Windows 10

Application name and version

We can see the issues in most tools like INDD, Illustrator, word

Issue

When this character হ্যাঁ is used in text and Noto sans Bengali is picked as font family, it seems that the character is corrupted, I believe that the screenshot below shows the issue quite well.

Character data

Unfortunately not sure about the position of the font within the table

Screenshot

Font

NotoSerifBengali-Regular.otf

Where the fonts came from, and when

Site: https://github.com/googlefonts/noto-fonts/blob/4d7396a4fae475975b367e225f2e8f49292d8837/unhinted/otf/NotoSerifBengali/NotoSerifBengali-Regular.otf
Date: 2022-04-16

Font version

Version 2.001

Issue

Noto Serif Bengali usually places U+0951 DEVANAGARI STRESS SIGN UDATTA at a constant high height above the cluster. In clusters that use the glyphs uni09BF09B009CD and uni09BF0981, however, the U+0951 is positioned too low. This is because an 'abvs' lookup introduces the glyph uni200B, which blocks positioning.

Character data

র্ঢি॑ঢিঁ॑
U+09B0 BENGALI LETTER RA
U+09CD BENGALI SIGN VIRAMA
U+09A2 BENGALI LETTER DDHA
U+09BF BENGALI VOWEL SIGN I
U+0951 DEVANAGARI STRESS SIGN UDATTA
U+09A2 BENGALI LETTER DDHA
U+09BF BENGALI VOWEL SIGN I
U+0981 BENGALI SIGN CANDRABINDU
U+0951 DEVANAGARI STRESS SIGN UDATTA

Screenshot

Defect Report

Lack of knots in the Tirhuta and Bengali Noto Sans fonts is not just wrong but also misleading learners. As a teacher, I am tired of telling learners not to trust the Google noto font for Assamese/Bengali and Tirhuta. I addressed this issue in 2017, but the developers didn't care amd even made another wrong font, Noto Sans Tirhuta. But weirdly they didn't do the same for many other fonts like Noto Sans Modi where knots are correctly added.

Title

Knots needed in Noto Sans Tirhuta and Noto Sans Bengali.

Font

NotoSansTirhuta-Regular.ttf, NotoSansBengali.ttf

Where the font came from, and when

Date: 2018-01-10

Screenshot

Tirhuta consonants. Also look at the ম. The ম in thes noto sans fonts is totally wrong.

Defect Report

In bengali, হ্ + ণ = U+09B9 U+09CD U+09A3 should be

and হ্ + ন = U+09B9 U+09CD U+09A8 should be

but the font now shows the opposite.

Font

Full file name, for example 'NotoSansBengali-Regular.ttf'.

Adobe Bengali, Galada and Atma fonts have got it right but others have got them wrong.