sRGB support for distance field text.

sRGB support for distance field text.

Add a second distance field adjust table that only applies contrast,
not fake-gamma correction. Store a flag in the batch at creation time,
using the same logic we apply elsewhere (render target format, plus
paint flags).

That gets us close, but not as good as bitmap text. The final step is
to use a linear step function (rather than smoothstep) to map distance
to coverage, when we have sRGB output. Smoothstep's nonlinear response
is actually doing some fake-gamma, so it ends up over-correcting when
the output is already gamma-correct.

Results are now very close between L32 (old table, smoothstep) and S32
(contrast-only table, linstep).

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1885613002

Committed: https://skia.googlesource.com/skia/+/0586f5cc9713268238394411a5daa2c7758b092b

[Brian Osman, 2016-04-12 18:50:44]

Description was changed from

==========
sRGB support for distance field text.

Add a second distance field adjust table that only applies contrast,
not fake-gamma correction. Store a flag in the batch at creation time,
using the same logic we apply elsewhere (render target format, plus
paint flags).

Results aren't as close (L32 vs. S32) as bitmap text, but it's pretty
good. DF text in S32 is still slightly softer - my best guess is that
because the distance affects coverage nonlinearly (due to smoothstep),
we're getting accelerated falloff.

BUG=skia:
==========

to

==========
sRGB support for distance field text.

Add a second distance field adjust table that only applies contrast,
not fake-gamma correction. Store a flag in the batch at creation time,
using the same logic we apply elsewhere (render target format, plus
paint flags).

Results aren't as close (L32 vs. S32) as bitmap text, but it's pretty
good. DF text in S32 is still slightly softer - my best guess is that
because the distance affects coverage nonlinearly (due to smoothstep),
we're getting accelerated falloff.

BUG=skia:
GOLD_TRYBOT_URL=
https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&is...
==========

[Brian Osman, 2016-04-12 18:54:22]

Description was changed from

==========
sRGB support for distance field text.

Add a second distance field adjust table that only applies contrast,
not fake-gamma correction. Store a flag in the batch at creation time,
using the same logic we apply elsewhere (render target format, plus
paint flags).

Results aren't as close (L32 vs. S32) as bitmap text, but it's pretty
good. DF text in S32 is still slightly softer - my best guess is that
because the distance affects coverage nonlinearly (due to smoothstep),
we're getting accelerated falloff.

BUG=skia:
GOLD_TRYBOT_URL=
https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&is...
==========

to

==========
sRGB support for distance field text.

Add a second distance field adjust table that only applies contrast,
not fake-gamma correction. Store a flag in the batch at creation time,
using the same logic we apply elsewhere (render target format, plus
paint flags).

That gets us close, but not as good as bitmap text. The final step is
to use a linear step function (rather than smoothstep) to map distance
to coverage, when we have sRGB output. Smoothstep's nonlinear response
is actually doing some fake-gamma, so it ends up over-correcting when
the output is already gamma-correct.

Results are now very close between L32 (old table, smoothstep) and S32
(contrast-only table, linstep).

BUG=skia:
GOLD_TRYBOT_URL=
https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&is...
==========

[Brian Osman, 2016-04-12 18:56:12]

brianosman@google.com changed reviewers:
+ bsalomon@google.com, bungeman@google.com, jvanverth@google.com

[Brian Osman, 2016-04-12 19:15:24]

The CQ bit was checked by brianosman@google.com to run a CQ dry run

[commit-bot: I haz the power, 2016-04-12 19:15:34]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1885613002/40001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1885613002/40001

[bsalomon, 2016-04-12 19:21:21]

On 2016/04/12 19:15:34, commit-bot: I haz the power wrote:
> Dry run: CQ is trying da patch. Follow status at
>  https://chromium-cq-status.appspot.com/patch-status/1885613002/40001
> View timeline at
>  https://chromium-cq-status.appspot.com/patch-timeline/1885613002/40001

seems ok to me (though Jim will know better). How does it look?

[commit-bot: I haz the power, 2016-04-12 19:29:45]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2016-04-12 19:29:45]

Dry run: This issue passed the CQ dry run.

[jvanverth1, 2016-04-12 19:34:46]

lgtm + comment

https://codereview.chromium.org/1885613002/diff/40001/src/gpu/effects/GrDista...
File src/gpu/effects/GrDistanceFieldGeoProc.cpp (right):

https://codereview.chromium.org/1885613002/diff/40001/src/gpu/effects/GrDista...
src/gpu/effects/GrDistanceFieldGeoProc.cpp:404: fragBuilder->codeAppend("float
val = smoothstep(-afwidth, afwidth, distance);");
The SDF path renderer doesn't do any fake gamma, so it's probably worth seeing
how it looks when you use a linear ramp instead of smoothstep.

[Brian Osman, 2016-04-12 19:46:00]

https://codereview.chromium.org/1885613002/diff/40001/src/gpu/effects/GrDista...
File src/gpu/effects/GrDistanceFieldGeoProc.cpp (right):

https://codereview.chromium.org/1885613002/diff/40001/src/gpu/effects/GrDista...
src/gpu/effects/GrDistanceFieldGeoProc.cpp:404: fragBuilder->codeAppend("float
val = smoothstep(-afwidth, afwidth, distance);");
On 2016/04/12 19:34:46, jvanverth1 wrote:
> The SDF path renderer doesn't do any fake gamma, so it's probably worth seeing
> how it looks when you use a linear ramp instead of smoothstep.

Yeah, I was planning to make that change after testing - seemed like the logic
would still apply.

[Brian Osman, 2016-04-12 19:47:21]

The CQ bit was checked by brianosman@google.com

[commit-bot: I haz the power, 2016-04-12 19:47:26]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1885613002/40001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1885613002/40001

[commit-bot: I haz the power, 2016-04-12 19:48:23]

Description was changed from

==========
sRGB support for distance field text.

Add a second distance field adjust table that only applies contrast,
not fake-gamma correction. Store a flag in the batch at creation time,
using the same logic we apply elsewhere (render target format, plus
paint flags).

That gets us close, but not as good as bitmap text. The final step is
to use a linear step function (rather than smoothstep) to map distance
to coverage, when we have sRGB output. Smoothstep's nonlinear response
is actually doing some fake-gamma, so it ends up over-correcting when
the output is already gamma-correct.

Results are now very close between L32 (old table, smoothstep) and S32
(contrast-only table, linstep).

BUG=skia:
GOLD_TRYBOT_URL=
https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&is...
==========

to

==========
sRGB support for distance field text.

Add a second distance field adjust table that only applies contrast,
not fake-gamma correction. Store a flag in the batch at creation time,
using the same logic we apply elsewhere (render target format, plus
paint flags).

That gets us close, but not as good as bitmap text. The final step is
to use a linear step function (rather than smoothstep) to map distance
to coverage, when we have sRGB output. Smoothstep's nonlinear response
is actually doing some fake-gamma, so it ends up over-correcting when
the output is already gamma-correct.

Results are now very close between L32 (old table, smoothstep) and S32
(contrast-only table, linstep).

BUG=skia:
GOLD_TRYBOT_URL=
https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&is...

Committed:
https://skia.googlesource.com/skia/+/0586f5cc9713268238394411a5daa2c7758b092b
==========

[commit-bot: I haz the power, 2016-04-12 19:48:24]

Committed patchset #3 (id:40001) as
https://skia.googlesource.com/skia/+/0586f5cc9713268238394411a5daa2c7758b092b