Re: [ft-devel] Hinting paths

[Werner LEMBERG]

> However, the difference between autohint on medium and
> full is still there: http://postimg.org/gallery/27mhq5d00
> (full-noauto, full-auto, medium-auto).

This looks to me like

  full-noauto:   CFF hinting
  full-auto: horizontal and vertical auto-hinting
  medium-auto:   vertical (`light') auto-hinting with warping switched
 on?

But you have to compare the rendered strings with output from one of
the FreeType demo programs to get more clues.


Werner

___
Freetype-devel mailing list
Freetype-devel@nongnu.org
https://lists.nongnu.org/mailman/listinfo/freetype-devel

Re: [ft-devel] Hinting paths

[Markus Trippelsdorf]

On 2015.11.05 at 17:13 +0100, Nikolaus Waxweiler wrote:
> It's actually Liberation Sans -- so ClearType rendering. And switching 
> to light and turning on the warper gives a different look.. mh. I'll 
> simply avoid medium in my proposed fontconfig confs.

There is a nice summary of the recent ATypI conference on LWN today:
http://lwn.net/SubscriberLink/662813/4795feb3a4ac97d2/

To quote: »the panelists generally agreed that hinting "needs to die"«

-- 
Markus

___
Freetype-devel mailing list
Freetype-devel@nongnu.org
https://lists.nongnu.org/mailman/listinfo/freetype-devel

Re: [ft-devel] Hinting paths

[Nikolaus Waxweiler]

It's actually Liberation Sans -- so ClearType rendering. And switching 
to light and turning on the warper gives a different look.. mh. I'll 
simply avoid medium in my proposed fontconfig confs.


___
Freetype-devel mailing list
Freetype-devel@nongnu.org
https://lists.nongnu.org/mailman/listinfo/freetype-devel

[ft-devel] rendering mode of Firefox on Windows 10

[Werner LEMBERG]

Folks,


according to browserstack, the current release 41 of Firefox on
Windows 10 seems still to use the GDI framework, which implies font
rasterizer version 37, compatible widths ClearType, no supixel
positioning, and no vertical smoothing (i.e., B/W rasterization along
the vertical axis) – this is exactly the same rendering mode as used
on Win XP.

Is this really true?  For me, it's hard to believe...

  http://freetype.org/freetype2/docs/rasterinfo/rasterinfo.html


 Werner
___
Freetype-devel mailing list
Freetype-devel@nongnu.org
https://lists.nongnu.org/mailman/listinfo/freetype-devel

Re: [ft-devel] rendering mode of Firefox on Windows 10

[Nikolaus Waxweiler]

Firefox defaults to GDI-style rendering for a handful of TrueType fonts. 
The first thing I do when I set up a new Firefox instance on Windows or 
fiddling with other people's computers when they aren't looking is setting


gfx.font_rendering.cleartype_params.force_gdi_classic_for_families = ""

Meaning, to an empty list. That triggers DirectWrite-style rendering for 
all fonts. Maybe that's what you stumbled over? Or did you test 
rasterinfo.html in Browserstack? Firefox Developer Edition on my Windows 
10 installation shows bytecode interpreter version 40.


___
Freetype-devel mailing list
Freetype-devel@nongnu.org
https://lists.nongnu.org/mailman/listinfo/freetype-devel

Re: [ft-devel] rendering mode of Firefox on Windows 10

[Khaled Hosny]

On Thu, Nov 05, 2015 at 07:43:21PM +0100, Werner LEMBERG wrote:
> 
> >> according to browserstack, the current release 41 of Firefox on
> >> Windows 10 seems still to use the GDI framework, which implies font
> >> rasterizer version 37, [...]
> > 
> > I suspect that browserstack uses remote windows desktop connection
> > or some such which I have noticed to be using Windows XP-like font
> > rendering.
> 
> Hmm.  How do you explain then that both Chrome and Edge use version 40
> on browserstack?

I see, Firefox uses DirectWrite only when 3D acceleration is enabled and
GDI otherwise, so this might be the case here.

Regards,
Khaled

___
Freetype-devel mailing list
Freetype-devel@nongnu.org
https://lists.nongnu.org/mailman/listinfo/freetype-devel

Re: [ft-devel] rendering mode of Firefox on Windows 10

[Werner LEMBERG]

> Firefox defaults to GDI-style rendering for a handful of TrueType
> fonts. The first thing I do when I set up a new Firefox instance on
> Windows or fiddling with other people's computers when they aren't
> looking is setting
>
> gfx.font_rendering.cleartype_params.force_gdi_classic_for_families =
> ""

Interesting.  I wasn't aware that this is controllable.

> Or did you test rasterinfo.html in Browserstack?

Exectly this.  Khaled provided a nice explanation of the possible
cause in another e-mail.


Werner

___
Freetype-devel mailing list
Freetype-devel@nongnu.org
https://lists.nongnu.org/mailman/listinfo/freetype-devel

Re: [ft-devel] rendering mode of Firefox on Windows 10

[Werner LEMBERG]

>> according to browserstack, the current release 41 of Firefox on
>> Windows 10 seems still to use the GDI framework, which implies font
>> rasterizer version 37, [...]
> 
> I suspect that browserstack uses remote windows desktop connection
> or some such which I have noticed to be using Windows XP-like font
> rendering.

Hmm.  How do you explain then that both Chrome and Edge use version 40
on browserstack?


Werner

___
Freetype-devel mailing list
Freetype-devel@nongnu.org
https://lists.nongnu.org/mailman/listinfo/freetype-devel

Re: [ft-devel] rendering mode of Firefox on Windows 10

[Werner LEMBERG]

> Don't know if it's important at all, but this line renders
> differently on Chrome with disabled DirectWrite (that gives engine
> 37), comparing to fresh Firefox installation (with engine 40):
>
> ---
> The current PPEM value used in this paragraph (‘0’):
> ---
>
> Version 37 shows 18, while version 40 shows 00018.

This is intentional: If subpixel hinting is active, the MS rendering
engine ignores modifications of the advance width, always using the
glyph's linear width.  Otherwise, I can adjust the advance width
(using bytecode) to suppress leading zero digits.


 Werner
___
Freetype-devel mailing list
Freetype-devel@nongnu.org
https://lists.nongnu.org/mailman/listinfo/freetype-devel

Re: [ft-devel] rendering mode of Firefox on Windows 10

[Werner LEMBERG]

> Just tried that in a VM, and yes, your test page reports version 37,
> compatible ClearType, no symmetrical smoothing.

Thanks for testing.  How comes that Alexei reports something different
on an *older* Windows version?  Note that on browserstack, both Chrome
and Edge use a version 40 engine...

> By the way, is there a way to enable any of that in freetype?  I
> know I can set engine version to 38, does it mean a set of
> parameters will be applied and there's no way to switch dynamically
> from one to another?

Correct.

> E.g. enable symmetrical smoothing per render call, not per library.

No, this is not possible.  Additionally, FreeType has never supported
non-symmetrical smoothing.


Werner

___
Freetype-devel mailing list
Freetype-devel@nongnu.org
https://lists.nongnu.org/mailman/listinfo/freetype-devel

Re: [ft-devel] Freetype-devel Digest, Vol 130, Issue 8

[Hin-Tak Leung]

...
 From: Markus Trippelsdorf 
...
 Subject: Re: [ft-devel] Hinting paths
 Message-ID: <20151105164556.GC325@x4>
...
 There is a nice summary of the recent ATypI conference on
 LWN today:
 http://lwn.net/SubscriberLink/662813/4795feb3a4ac97d2/
...

Interesting reading.

There was a  couple of inaccuracies in the article - the 1st iso opentype
spec was in 2007, not 1996 (and draft exists earlier than 2007, of course);
and about DSIG, interestingly in the 2015 spec out before the conference,
it is no longer 'recommended' but optional, so one of the 'wishes' has already 
sort-of
happened before the panel discussion.

Also, rather strangely Si Daniels of Microsoft doesn't know that
microsoft's font signing tool package also ships a signature checking tool.
It is rather user-unfriendly, and also tightly coupled to the windows trust 
system
though; and it is not 'open'...

As for a tool for examining signatures in fonts in the 'open' world, well, one
just came into existence a few days ago. People interested in that, please 
reply by privately.

There is apparently at least one non-microsoft 3rd party signing tool in
existence also. If anybody else, other than you-know-who-you-are, know of
a 2nd non-microsoft sigining tool, please speak up or reply privately too.

___
Freetype-devel mailing list
Freetype-devel@nongnu.org
https://lists.nongnu.org/mailman/listinfo/freetype-devel

Re: [ft-devel] rendering mode of Firefox on Windows 10

[Cosimo Lupo]

Have you tried forcing Direct2D by setting "gfx.direct2d.force-enabled" to
"true"?
___
Freetype-devel mailing list
Freetype-devel@nongnu.org
https://lists.nongnu.org/mailman/listinfo/freetype-devel

Re: [ft-devel] rendering mode of Firefox on Windows 10

[Nikolay Sivov]

On 05.11.2015 23:40, Nikolaus Waxweiler wrote:

gfx.font_rendering.cleartype_params.force_gdi_classic_for_families
indeed contained something by default, I tried to reset to empty string,
no change. Also if it really applies to specific family my guess would
be that RasterInfo font won't be affected, as it's never on the list by
default.


Did you restart the browser? :)


Sure I did. On every change.



And yes, RasterInfo isn't affected by default on my installation, so
it's probably Khaled's explanation.

___
Freetype-devel mailing list
Freetype-devel@nongnu.org
https://lists.nongnu.org/mailman/listinfo/freetype-devel



___
Freetype-devel mailing list
Freetype-devel@nongnu.org
https://lists.nongnu.org/mailman/listinfo/freetype-devel

Re: [ft-devel] rendering mode of Firefox on Windows 10

[Nikolaus Waxweiler]

gfx.font_rendering.cleartype_params.force_gdi_classic_for_families
indeed contained something by default, I tried to reset to empty string,
no change. Also if it really applies to specific family my guess would
be that RasterInfo font won't be affected, as it's never on the list by
default.


Did you restart the browser? :)

And yes, RasterInfo isn't affected by default on my installation, so 
it's probably Khaled's explanation.


___
Freetype-devel mailing list
Freetype-devel@nongnu.org
https://lists.nongnu.org/mailman/listinfo/freetype-devel

Re: [ft-devel] rendering mode of Firefox on Windows 10

[Khaled Hosny]

On Thu, Nov 05, 2015 at 11:26:44PM +0300, Nikolay Sivov wrote:
> On 05.11.2015 22:06, Khaled Hosny wrote:
> >On Thu, Nov 05, 2015 at 07:43:21PM +0100, Werner LEMBERG wrote:
> >>
> according to browserstack, the current release 41 of Firefox on
> Windows 10 seems still to use the GDI framework, which implies font
> rasterizer version 37, [...]
> >>>
> >>>I suspect that browserstack uses remote windows desktop connection
> >>>or some such which I have noticed to be using Windows XP-like font
> >>>rendering.
> >>
> >>Hmm.  How do you explain then that both Chrome and Edge use version 40
> >>on browserstack?
> >
> >I see, Firefox uses DirectWrite only when 3D acceleration is enabled and
> >GDI otherwise, so this might be the case here.
> 
> Hm, I wonder if that's because they also make use of Direct2D, because I
> don't think DirectWrite necessary talks to d3d system. Although it's
> possible some sophisticated logic to have system-wide font cache service to
> cache generated bitmaps or defer computations, and d3d is used for that
> directly or not, but I never heard about it.
> 
> gfx.font_rendering.directwrite.enabled was false, toggled to true, no
> difference.
> 
> gfx.font_rendering.cleartype_params.force_gdi_classic_for_families indeed
> contained something by default, I tried to reset to empty string, no change.
> Also if it really applies to specific family my guess would be that
> RasterInfo font won't be affected, as it's never on the list by default.

Check about:support, and see of you have 3D acceleration (you might find
something about DirectWrite there too, I don’t remember for sure).

Regards,
Khaled

___
Freetype-devel mailing list
Freetype-devel@nongnu.org
https://lists.nongnu.org/mailman/listinfo/freetype-devel