cdcafbc4c5
add changelog notes
...
and linewrap for 80 column limit
2022-05-07 12:51:58 +01:00
894ed1cb13
Resize improvements; add `ceil` and `gap` options ( #1769 )
...
* Add a round-up option to shrink
* Only leave the final 200% to reduce
* Add gap option to reduce and resize
* Add unit tests
* Incorporate feedback
2022-05-07 12:45:38 +01:00
0337c09700
Cleanup gettext handling ( #2695 )
...
Use GLib's i18n support instead of copying and pasting that
logic into its own header. This deprecates the vips/intl.h
header in favour of glib/gi18n.h.
2022-02-27 15:27:28 +00:00
75632a5641
Merge pull request #1592 from kleisauke/issue-703
...
Fix the pixel shift within reduce (#703 )
2020-06-12 12:36:22 +01:00
4469afaeb0
better handling of "nearest" in resize
...
we now use vips_subsample() for this case
2020-06-12 12:03:15 +01:00
ac358ff4b8
Fix the pixel shift within reduce ( #703 )
2020-06-06 14:01:34 +02:00
ba0dea001d
sort out premultiply rules for upsizing
...
vips_resize() uses vips_affine() for upsizing and vips_reduce() for
downsizing. Affine automaticaly does a vips_premultiply() for images
with an alpha channel, but reduce does not. This meant that we could
sometimes premultiply twice.
This patch adds a "premultiplied" flag for affine which turns automatic
premultiuplication off, vips_resize() uses this to block affine's auto
premul feature, and the resize docs are clarified to stress that the
operation does not do premultiplication for you.
See https://github.com/libvips/libvips/issues/1629
2020-05-19 14:31:34 +01:00
e1baf66f19
"squash" to tiffsave now does lab as well
...
The "squash" option to tiffsave now also squashes 32-bit 3-band float
CIELAB images down to 8 bits.
See https://github.com/libvips/libvips/issues/1499
2019-12-18 17:29:34 +00:00
f45f0ecef4
limit resize if shrink would drop axis <1px
...
vips_resize() will break the aspect ratio and limit resize on an axis if
it would result in an image of less than 1px on that axis
see https://github.com/lovell/sharp/issues/1782#issuecomment-508921306
2019-07-07 10:19:33 +01:00
f5d76b42fc
fix centre sampling for non-int nearest upscale
...
we were not disabling the input offset for NEAREST, whcih is always
centre
thanks edwjusti
see https://github.com/lovell/sharp/issues/1479
2018-12-03 17:13:57 +00:00
3b1e3e1841
add Mitchell kernel
...
see https://github.com/jcupitt/libvips/issues/966
2018-06-12 12:55:10 +01:00
407291c3ff
try a simpler edge detect in canny
...
simple +/- differencing seems to work better
2018-02-08 17:14:26 +00:00
e9b7231ac0
add "extend" option to affine
...
so for upsizing with displacement (like resize) we can prevent dark
edges on bright images
see
https://github.com/jcupitt/php-vips/issues/61
2017-12-07 14:51:07 +00:00
406e228c9b
vips_resize() is always centre convention
...
it was corner for downsize, but should be centre to match imagemagick
2017-11-26 15:24:40 +00:00
ee4186ae8f
revise resize docs
2017-11-25 15:13:20 +00:00
7db13412e5
deprecate centre option to resize
...
it's now centre on upsize, corner on downsize
see https://github.com/jcupitt/libvips/issues/705
2017-11-25 14:22:20 +00:00
cb128eb258
apply --centre to upsize as well as down
...
we were just using the --centre flag in resize to influence downsize
behaviour -- use it to set upsize as well
see https://github.com/jcupitt/libvips/issues/705
2017-11-25 12:09:01 +00:00
68ed42e2fa
make linear and cubic adaptive
2017-10-15 22:09:44 +01:00
cf1e6e3ef8
oh dear, missing breaks in switch
...
too much ruby sigh
2017-10-13 12:16:03 +01:00
de82c9e919
allow much larger reduce factors
...
sometimes you want to use lanczos etc. for everything
2017-10-09 11:48:11 +01:00
ea6d496484
Upsample integral nearest neighour with zoom for accuracy.
...
This mirrors logic from sharp, allowing it to migrate to resize.
See https://github.com/lovell/sharp/issues/752
2017-10-03 19:44:34 +01:00
c24d36fc79
Update GObject introspection annotations for libvips/resample directory
2017-09-30 17:26:55 +01:00
0a5375b6cd
Merge branch 'master' into remove-seq-stalling
2017-03-15 12:26:44 +00:00
d7da3c7de7
fix vips_resize() bug
...
if hscale and vscale were very different, resize could fail with "reduce
factor too large" ... just a silly typo
2017-03-10 10:22:40 +00:00
362c7efdfc
argh reapply seq removal
2017-03-06 11:55:38 +00:00
55840cf131
try setting a seq meta
...
and only caching in shrinkv if we see the tag
2017-03-05 21:59:01 +00:00
20f99b62cf
more ideas
2017-03-05 21:59:01 +00:00
2be0b97dce
switch to g_warning()_/g_info()
...
we had vips_warn() and vips_info(), but they are a bit crappy ... switch
to g_warning() and g_info() instead
see https://github.com/jcupitt/libvips/issues/544
2017-01-03 15:52:27 +00:00
d3cafb3a9f
seems to work
...
add @centre option for downsampling, see
https://github.com/jcupitt/libvips/issues/504
2016-09-10 10:31:33 +01:00
1f403a4add
first attempt
...
seems to all work, test it a bit more
see https://github.com/jcupitt/libvips/issues/491
2016-08-15 15:45:26 +01:00
65105a9442
upsize with something other than nearest
...
vips_resize() uses to just use nearest when upsizing, following standard
practice in repro. This is often unfortunate for image processing, where
small nearest upsizes will produce obvious aliasing.
It now picks a VipsInterpolate which corresponds (roughly) to the
selected VipsKernel and uses that with affine for any upsizing.
2016-06-22 10:33:08 +01:00
f3326c8126
better vips_resize() for cubic/linear/nearest
...
vips_resize() used to do most of a downsize with vips_shrink() and the
final 200 - 300% with vips_reduce(). This was correct for lanczos2/3,
but not right for linear/cubic, which need more shrink and less
reduce to avoid aliasing.
This patch makes vips_resize() leave the final 100 - 200% to
vips_reduce() for linear/cubic, and leave everything to reduce for
nearest.
2016-06-16 09:46:02 +01:00
948d1103d4
note more webp tests needed
2016-05-07 19:53:00 +01:00
1ae92bb15f
make optional args into bullets
...
make docs easier to read
2016-05-02 10:12:37 +01:00
02821ae9e3
expose @kernel setting for vips_resize()
...
a new optional arg lets you pick the downsizing kernel
2016-05-01 12:33:49 +01:00
c2d1de9a26
allow mixed zoom and shrink in resize
...
vips_resize() now allows >1 on one axis, <1 on the other ... previously,
although you could have non-square ratios, you couldn't mix zoom and
shrink
2016-05-01 12:03:13 +01:00
36e0b60ffa
put the resize cache back again
...
stops overcomputation on vips_shrink()
2016-03-22 14:35:44 +00:00
46a18118c6
works!
2016-03-15 16:46:18 +00:00
ff88087a28
try making the reduce mask larger with residual
2016-03-15 14:59:59 +00:00
45c3fea6ac
more resize tuning
2016-03-15 09:49:30 +00:00
ac4e5ce059
return of shrink-less-affine-more
...
though a bit less extreme
2016-03-14 14:58:02 +00:00
fba2ac2f85
fix sizing issues with new reduce code
...
a couple of bugs and some changed conventions
2016-03-13 17:44:24 +00:00
f12fef4aa9
fix compile
2016-03-13 11:35:35 +00:00
70db9aa77b
better rounding in resize
2016-03-12 13:14:34 +00:00
493482c8a7
hey ho, line cache goes back too
2016-03-11 14:56:06 +00:00
08e8c27e38
moore vips_resize() tuning
...
put back the mild post-shrink sharpen, a little more anti-alias
see https://github.com/jcupitt/libvips/issues/404#issuecomment-195297607
2016-03-11 10:17:41 +00:00
5d31d8e824
turn up anti-alias for small shrinks
...
a little more anti-alias for small reductions
see https://github.com/jcupitt/libvips/issues/404#issuecomment-195276154
2016-03-11 09:25:33 +00:00
2ca7b29214
better gauss mat generation
2016-03-10 22:01:55 +00:00
4974a1ed9c
better rounding for vips_resize()
...
we were getting off by one size errors
2016-03-10 19:53:05 +00:00
d351e07a6b
finish up
2016-03-10 17:33:34 +00:00
John Cupitt
Kleis Auke Wolthuizen
Lovell Fuller
Alistair Thomas