diff options
Diffstat (limited to 'log.txt')
| -rw-r--r-- | log.txt | 213 |
1 files changed, 213 insertions, 0 deletions
@@ -2,6 +2,213 @@ Argyll CMS change log ===================== +Version 1.7.0 (1st May 2015) +------------- + +* Improved gamut mapping to reduce unnecessary changes to less saturated colors + such as skin tones. + +* Add support for DataColor Spyder 5. + +* Ignore any patches that have zero values for creating Display profiles, + unless they are for device zero. This is to avoid measurements below + and instruments measurement capability causing shadow response to be raised. + +* Add better cross compatibility with non-Argyll ICC profiles: + + Use "wrong Von Kries" media white point adapation for non-Argyll non-display profiles + + Optionally create "wrong Von Kries" media white point adapation profiles + using the "ARGYLL_CREATE_WRONG_VON_KRIES_OUTPUT_CLASS_REL_WP" env. variable, + while maintaining ArgyllCMS compatibility using the 'arts' tag. + + Implement proper absolute colorimetric intent for ICCV2 Display profiles + that use the ICCV4 style of have a media white of D50 and storing the + media chromatic trasnform in the 'chad' tag. + + Optionally create such ICCV4 style V2 Display profiles by using the + "ARGYLL_CREATE_DISPLAY_PROFILE_WITH_CHAD" env. variable. + +* Reverted to handling collink Video encoding overrage input (WTW) with + extrapolation that preserves hues, and added an input encoding option + "Video with clipping" to cope with TV's that don't preserver hue + for WTW.. + +* Add X3D and X3DOM support as an alternative to VRML, and make X3DOM the default. + +* Update to use latest MadVR background setting code (v0.87.11) + +* Fix problem with JETI communication overrun and 1201 timeout. + +* Improve JETI specbos 1201 and 1211 compatibility with older firmware. + +* Improve KLEIN robustness by automatically retrying range errors. + +* Fix bug in recognizing Klein K-10 (non A) instrument. + +* Fixed bug in working with newer firmware Klein K10A. + +* Fix gestaltSystemVersion warning on OS X 10.10. + +* Fix crash in targen with -l 100. + +* Fix crash in colprof when -s/S and -p are used together. + +* Fix very major bug in illumread. It wasn't actually working + at all since V1.4.0. + +* Fix poor handling of corrupted .ccmx and .ccss files. + +* Fix endless loop bug in dispcal verify. + +* Fix bug in collink -iaw when the destination profile + is a matrix + gamma type. + +* Add -z option to ccxxmake to allow setting separate display type + for reference instrument. + +* Fix XDG_CONFIG_HOME and XDG_CONFIG_DIRS environment variable names. + +* Fix occasional timout in specbos on reading spectral. + +* Added preliminary support for ColorHug2 + +* Fixed bug in colverify -L if one of the Normalise options isn't used. + +* Added capability of using the Google ChromeCast as a test + patch generator for Video displays. If you are running on + Linux, you might have to check that your firewall is configured + to allow TCP ports 8081 for this to work, if it + falls back to the Default receiver. Accuracy is OK, but + not as good as main display or MadTPG, due to ChromeCast + using YCbCr video encoding and upsizing of rasters sent to it. + +* Fixed problem with certain .ccmx files note + being recognisd for -y option. + +* Added profcheck -s flag to output delta E orderd patch + list with -v2. + +* Fix bug in chartread for DTP20 - it was timing out when + a off-line chart was being loaded, or an on-line chart was being read. + +* Fix bug in xicc/xicc.c & xicc/xlut.c that sometimes causes bad B2A + tables due to the black point being a NaN. + (This was noticed on code compiled on Ubuntu 14.) + +* Changed Linux udev/hotplug files to be more compatible + with colord, so that colord 69-cd-sensors.rules file + can be upgraded by replacing with 55-Argyll.rules. + +* Fixed bug in tiffgamut when handling .jpg files. Some + pixels were being computed with the wrong values. + +* Added collink -I[gG][:p.p]:x.x options that allows an output offset + gamma curve be targetted instead of BT1886. Like dispcal, this + also allows a hybrid between input black offset (BT.1886-like) and + pure output black offset curves. The -I[bB] option also accepts + an option output proportion value (:p.p), and that + -I[bB]:p.p:x.x has exactly the same effect as -I[gG]:p.p:x.x. + Note using output black offset tends to loose shadow + detail compared to input black offset (BT.1886-like). + + The -I[bBgG] flags now substitute a BT.1886-like EOTF + for the native EOTF of the matrix input profile, rather than + applying an Rec709 to BT.1886 transform. This change has no + impact on source colorspaces that use the Rec709 encoding curve. + +* Added dispcal -b black point hack. This forces the + input zero to map to the output zero, which can be useful + with displays that are well behaved and have a very dark + black point, which an instrument is unable to measure precisely. + +* Added collink -b black point hack. This forces the + input zero to map to the output zero for RGB->RGB links, + which can be useful with displays that are well behaved + and have a very dark black point, which an instrument is + unable to measure precisely. + +* Added nonadaptive mode to i1d3. This allows fast measurement + with much lower accuracy for dark colors. + +* Changes to inst API to accomodate display technology + setting and retrival from instrument. + Also fixed setting custom matrix or ccss so that + other required details (display type & cbid) can + be provided. This now means that the instrument will + automatically configured when a ccmx is loaded. + +* Changed dispwin/dispsup patch delay calculations + (including matching code in targen) to adopt a more + sophisticated display settling model using the display + techology type. + +* Changed dispwin/dispsup patch delay calculations to allow + for a separate instrument reaction time parameter. + +* Added preset list of display techologies to select from + in ccxxmake. The -T option is deprecated, and the required + -t option replaces it. Used "ccxxmake -?? or -t?" to list + display technology types to choose from. + +* Added Klein K10-A instrument support. + +* Fixed bug in hcfr driver which prevented operation. + +* Fixed bug in chartread which caused high res mode to + be used when the -T flag is used. + +* Switched dispcal -R Video LUT entry bit depth algorithm. + +* Added IRIDAS .cube 3DLut format support to collink. + +* Added -O option to spotread that makes it do calibration + or one measurement and then exit. + +* Added better support for installing the Spyder1 firmware. + Simplified the Spyder PLD loading for applications + by modifying setup_spyd2(); + +* Added a -P prune option to profcheck, that creates a + .ti3 file pruned of any patches that have a delta E fit error + greater than a threshold. This may be of use in eliminating + bad reading values from a measurement set. + +* Added histogram plot option -h to both profcheck and verify. + +* Added ls2ti3 tool to convert LightSpace .bcs files to .ti3. + +* Fixed some time measurement bugs in OS X and Linux + USB drivers that mainly affect the DTP92 & DTP94. + +* Simplified the OS X USB drivers. + +* Fixed bug in DTP20 driver where it fails to wait for + a calibration to complete. + +* Fixed collink to more accuratly encode values + for the eeColor 3DLut, as well as change the + BT1886 black point mapping to be more accurate. + +* Added a dispread & fakeread -Z option to set the number of bits + to quantize the test values to. This defaults + to 8 for TV encoded ouput, and using the -Z option may improve + profiling accuracy on systems with limited bit depth + control over their displays. + +* Add support for other OEM versions of the i1d2 and i1d3 colorimeter. + +* Fixed bug in targen - the -V dark emphasis wasn't being + applied to OFPS generated points. + +* Added ColorCheckerSG.ti2 file to ref/, so that a reference file + can be created using chartread. + +* Make sure that if an instrument display reading is interrupted by a + forced calibration, that the user is asked to place it back on the + display before resuming the measurements. + +* Make sure that checking whether an instrument wants calibration + (ie. needs_calibration() or get_n_a_cals()) does not change its + state, requiring calibration due to calibrations being + too old. + Version 1.6.3 (26 January 2014) ------------- @@ -102,6 +309,12 @@ Version 1.6.2 - (19th November 2013) * Fix bug in -H flag in ccxxmake, chartread, dispcal, dispread, illumread & spotread so that it works once again. +* Changed OS X VideoLut access code to issue a warning if the + frame buffer bit depth does not match the gamma table depth, + rather than erroring out due to a bug in OS X. The gamma table + depth is the number actually used, and reports are that this + works. + Version 1.6.1 (30th September 2013) ------------- |