Bug 100869 - poor type 1c font rendering
Summary: poor type 1c font rendering
Status: RESOLVED MOVED
Alias: None
Product: poppler
Classification: Unclassified
Component: general (show other bugs)
Version: unspecified
Hardware: Other All
: medium normal
Assignee: poppler-bugs
QA Contact:
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2017-04-28 14:03 UTC by Germán Poo-Caamaño
Modified: 2018-08-21 11:01 UTC (History)
0 users

See Also:
i915 platform:
i915 features:


Attachments
PDF Test case (944.37 KB, application/pdf)
2017-04-28 14:03 UTC, Germán Poo-Caamaño
Details
evince 100% (98.03 KB, image/png)
2017-04-28 14:04 UTC, Germán Poo-Caamaño
Details
evince 134% (102.48 KB, image/png)
2017-04-28 14:04 UTC, Germán Poo-Caamaño
Details
evince 150% (101.32 KB, image/png)
2017-04-28 14:05 UTC, Germán Poo-Caamaño
Details
firefox 100% (138.81 KB, image/png)
2017-04-28 14:05 UTC, Germán Poo-Caamaño
Details
firefox 140% (100.17 KB, image/png)
2017-04-28 14:06 UTC, Germán Poo-Caamaño
Details
firefox 150% (106.79 KB, image/png)
2017-04-28 14:06 UTC, Germán Poo-Caamaño
Details
macos 100% (335.07 KB, image/png)
2017-04-28 14:07 UTC, Germán Poo-Caamaño
Details
macos 150% (292.29 KB, image/png)
2017-04-28 14:07 UTC, Germán Poo-Caamaño
Details

Description Germán Poo-Caamaño 2017-04-28 14:03:16 UTC
Created attachment 131121 [details]
PDF Test case

This was reported in https://bugzilla.gnome.org/show_bug.cgi?id=776924

The following pdf document is being rendered rather ugly in evince:

http://www.scottaaronson.com/papers/pnp.pdf

All fonts are embedded. They're all type 1c fonts.

$ pdffonts pnp.pdf 
name                                 type              encoding         emb sub uni object ID
------------------------------------ ----------------- ---------------- --- --- --- ---------
VXEZKP+CMSS17                        Type 1C           Builtin          yes yes no       4  0
YDPTNV+CMR12                         Type 1C           Builtin          yes yes no       5  0
JLCEZQ+CMR17                         Type 1C           Builtin          yes yes no       6  0
NUVJJI+CMSY8                         Type 1C           Builtin          yes yes yes      7  0
VBIHCP+CMBX10                        Type 1C           Builtin          yes yes no       8  0
CXZLTS+CMR10                         Type 1C           Builtin          yes yes yes      9  0
JGIYMZ+CMSS10                        Type 1C           Builtin          yes yes yes     10  0
AGGZPP+CMR7                          Type 1C           Builtin          yes yes no      11  0
RGQKMS+CMSY10                        Type 1C           Builtin          yes yes yes     12  0
BYIFKE+CMBX12                        Type 1C           Builtin          yes yes no      13  0
FNGYIT+CMR8                          Type 1C           Builtin          yes yes yes     15  0
JFBFYB+CMSY6                         Type 1C           Builtin          yes yes yes     32  0
OBXTKK+CMR9                          Type 1C           Builtin          yes yes yes     33  0
GVHKBA+CMTI10                        Type 1C           Builtin          yes yes no      87  0
RVPORJ+CMMI10                        Type 1C           Builtin          yes yes yes     89  0
JKFOWN+CMSS12                        Type 1C           Builtin          yes yes no      93  0
KIISBX+CMR6                          Type 1C           Builtin          yes yes yes     99  0
WWJJES+CMMI9                         Type 1C           Builtin          yes yes yes    100  0
DRXVBH+CMSS9                         Type 1C           Builtin          yes yes no     101  0
IWKPFB+CMMI8                         Type 1C           Builtin          yes yes yes    106  0
IYGFKD+CMMI6                         Type 1C           Builtin          yes yes yes    110  0
SGXEVV+CMTI9                         Type 1C           Builtin          yes yes no     129  0
DNNGNS+CMEX9                         Type 1C           Builtin          yes yes yes    131  0
FPKZIJ+CMSY9                         Type 1C           Builtin          yes yes yes    132  0
DOOAXC+CMSS8                         Type 1C           Builtin          yes yes no     170  0
OLTLXQ+CMCSC10                       Type 1C           Builtin          yes yes no     179  0
JXRWPD+CMEX10                        Type 1C           Builtin          yes yes yes    188  0
KSYIAZ+CMEX7                         Type 1C           Builtin          yes yes no     224  0
UROALT+CMMI5                         Type 1C           Builtin          yes yes no     228  0
SKNWGA+MSBM10                        Type 1C           Builtin          yes yes yes    239  0
PFOPPM+CMSY5                         Type 1C           Builtin          yes yes no     310  0
EFHBMR+CMR5                          Type 1C           Builtin          yes yes no     311  0
PLAHEP+CMEX8                         Type 1C           Builtin          yes yes no     690  0
UEHLMU+MSBM7                         Type 1C           Builtin          yes yes no     852  0

I wonder how the embedded pdf viewer of firefox can do a better job with this document than evince. Most other pdf documents look better in evince.

By ugly the reporter means:

So here are a couple of screenshots. As you can see evince has the lowest font rendering quality, because it's blurry and the stroke width is not concise (especially when it comes to the equal sign in the headline).

The font rendering of the internal pdf viewer of firefox also suffers from being blurry, but at least the stroke width is more concise.

I also added some screenshots taken under macOS. They are not blurry, but the resolution is also much higher. So it's more like a reference.
Comparing evince's font rendering in the pdf with the font rendering in the outline window on the left, you can see that it is very well possible to get a better result with the lower linux resolution.

And it's not like I was searching for a zoom factor where evince has problems. Sure, the quality might improve with other zoom factors, but it also declines with others as well. In the end the zoom factor cannot "fix" the font rendering quality; the font rendering is still clearly better in firefox when "optimizing" the zoom factor in both applications.
Comment 1 Germán Poo-Caamaño 2017-04-28 14:04:39 UTC
Created attachment 131122 [details]
evince 100%
Comment 2 Germán Poo-Caamaño 2017-04-28 14:04:58 UTC
Created attachment 131123 [details]
evince 134%
Comment 3 Germán Poo-Caamaño 2017-04-28 14:05:18 UTC
Created attachment 131124 [details]
evince 150%
Comment 4 Germán Poo-Caamaño 2017-04-28 14:05:44 UTC
Created attachment 131125 [details]
firefox 100%
Comment 5 Germán Poo-Caamaño 2017-04-28 14:06:14 UTC
Created attachment 131126 [details]
firefox 140%
Comment 6 Germán Poo-Caamaño 2017-04-28 14:06:32 UTC
Created attachment 131127 [details]
firefox 150%
Comment 7 Germán Poo-Caamaño 2017-04-28 14:07:00 UTC
Created attachment 131128 [details]
macos 100%
Comment 8 Germán Poo-Caamaño 2017-04-28 14:07:25 UTC
Created attachment 131129 [details]
macos 150%
Comment 9 GitLab Migration User 2018-08-21 11:01:06 UTC
-- GitLab Migration Automatic Message --

This bug has been migrated to freedesktop.org's GitLab instance and has been closed from further activity.

You can subscribe and participate further through the new bug through this link to our GitLab instance: https://gitlab.freedesktop.org/poppler/poppler/issues/468.


Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.