diff options
Diffstat (limited to 'log.txt')
| -rw-r--r-- | log.txt | 306 |
1 files changed, 305 insertions, 1 deletions
@@ -2,6 +2,310 @@ Argyll CMS change log ===================== +Version 1.6.3 (26 January 2014) +------------- + +* Added icc/mkDispProf.c sample code + +* Added ProPhoto.icm and ProPhotoLin.icm to ref profiles. + +* Fix bug in xicclu -py conversion. + +* Added code to minimize ICC rounding error on matrix profile white point + accuracy. Re-generated all reference profiles with this change. + +* Changed i1d3 driver to completely ignore any EEPROM checksum + errors for non "A-01" rev. instruments. + +* Made transplot handle RGB->RGB device link. + +* Removed colprof -y option. Use "profcheck -v2" instead, as it is + more developed. + +* Fixed bug in dispcal - it was not using the final measurement + pass to update the calibration curves. + +* Fixed bug in spotread, dispcal & dispread for CCSS capable instruments + where refresh display types was being ignored if a custom observer + was used, and/or the custom observer as being ignored, and/or + a CCMX was being ignored. + Changed instlib semantics for inst_opt_set_ccss_obs :- + this is now set immediately, and applied also to any subsequent + set_disptype() or col_cal_spec_set(). + +* Renamed verify to colverify to avoid clash with MSWin program + of the same name. Made it print the patch location for -v2 + if it is present in the file. + +* Changed targen to ensure that -V and -p options effects are reflected + in the resulting expected CIE values of the .ti1 file. + +* Changed targen so that -V parameter also affects single channel, + grey wedge, grid & body centered grid point distribution. + +* Changed colprof to deal with variable grid distribution + in a more neuanced way, to reduce overshoot artifacts + when the -V parameter is used. + +* Changed colprof to used a power_like function for the + grid distribution shape from the -V parameter, so as + to avoid issues with a power curve infinte slope near zero. + +* Changed colprof to used a scaled down value of the targen -V + parameter as the default for its -V parameter. Documentation + now recommends more moderate values for -V. + +* Added a special case to collink for RGB video input encoding + to (attempt) to fine tune the black point to compensate + for it (probably) not falling on a cLUT grid point. & out encoding + +* Tweaked dispcal to try and improve accuracy of black + point calibration. + +* Switch dispread to use NoClamp readings, so that + average black point value is not biased. + +* Fixed bug introduced into 1.6.2 oeminst that prevents .ccss files + being installed. + + +Version 1.6.2 - (19th November 2013) +------------- + +* Added "dark region emphasis" -V parameter to targen and colprof, + in an attempt to improve the accuracy of display profiles intended + for use with video. This should improve the subsequent black + point accuracy of the profile. + +* Fixed bug and tweaked dispcal black point optimization to + err on the black side. Added -v3 for even more debugging information. + +* Changed i1d3 driver to be more forgiving of EEProm checksum + calculation, so that it works with the latest release "A-02" rev. + i1 display pro & colormunki display instruments. + +* Improved robustness of i1d3 driver in recovering from errors. + +* Fixed race condition bug in OS X HID driver. This + fixes occassional problem with i1d3, and also solves + problem with the ColorHug on OS X. ColorHug is now fully supported. + +* Fixed problem with TV encoded output and dispread -E -k/-K + (Thanks to Florian Hoch). + +* Fixed minor bug in DE94 in icclib. + +* Fixed major bug in illumread - result was being corrupted. + +* Fixed "edges don't match" bug in printarg when -iCM -h -s/-S used. + +* Fix bug in -H flag in ccxxmake, chartread, dispcal, dispread, + illumread & spotread so that it works once again. + +Version 1.6.1 (30th September 2013) +------------- + +* Fix bug in "average -m" + +* Fix oeminst to work with a wider range of i1d3 install + files. + +* Fix bug introduced into ColorMunki (spectro) reflective + measurement that created discrepancies between spot + and strip measurement mode. Changed short wavelength + handling to be more like the manufacturers driver in + duplicating values, reduced reflective strip reading + scan rate to 55Hz to reduce reflective electrical interference + error. + +* Fix bug introduced into DTP94 driver when running on OS X. + +* Add valid (but unsigned) .cat files for MSWin libusb-win32 driver + installation. This seems to remove the need to disable diver signature + enforcement, and makes it possible to install on Win8.1 + +* Fix MadVR test pattern access to look for 64 bit .dll if it's been + compiled as 64 bit, otherwise API will fail. + +* Change instlib API to honour the inst_mode_highres flag in set_mode(). + Note that this means that High Res. mode state will be set by whatever + the last set_mode() or get_set_opt(inst_opt_highres/inst_opt_stdre) was. + + +Version 1.6.0 (16th August 2013) +------------- + +* Fixed crash in ucmm/ucmm.c when loading certain profiles + using dispwin (thanks to Torok Edwin). + +* Fixed JETI specbos 1201 handling so that all refresh mode + support is disabled, since it is not supported by this instrument. + +* Changed collink video level extended range handling to do a simple + extrapolation rather than linearly interpolated to the sync. values. + +* Added extra options to fakeread & verify to support Video verification. + +* Changed spotread -ew mode to be Bradford chromatic transform + rather than XYZ scaling, to better match Argyll ICC abs. vs. rel. + +* Added support to xicclu to lookup colors though CAL files, both + forward and backwards. Will also plot CAL file contents using -g. + +* Changed frequency calibration test pattern to be 80% white, to + avoid some situations where 100% does not exhibit refresh flicker. + Added warning for spotread when refresh calibration is expected. + +* Added -Y R:rate option to spotread, dispcal, dispread & ccxxmake + to allow setting a chosen display refresh rate. This can be used + with the Colormunki display, as well as situations in which refresh rate + measurement is not reliable. + +* Added 't' interactive command to specbos to toggle laser target + for JETI specbos. + +* Turned off B2A table clip map smoothing, as it seems to introduce + reversals for some data sets, and provides little benefit. + +* Changed dispwin & other display test patch programs on Linux/X11 + to use XRandR V1.3 XRRGetScreenResourcesCurrent() if available, + to avoid flickering and blackouts on some displays. + +* Linux profile installation will use colord if libcolordcompat.so + is present on system. + +* Changed dispwin to use XRandR V1.3 XRRGetScreenResourcesCurrent() + rather than XRRGetScreenResources() if available. + +* Added MadVR test patch support to dispwin, dispcal, dispread + and ccxxmake. + +* For those instruments that support it, made ambient readings + honour refresh mode measurements, to improve repeatability + when measuring regularly flickering light sources. + (specbos, i1d3, i1disp). + +* Made allowance for non standard ICC cLUT resolution of 256 + for special case of MadVR device links. + +* Tweaked colprof display profile creation to give more weight + to samples near the black point, in an attempt to improve + the black point accuracy. + +* Added default 4 black patches in targen for Grey & RGB spaces, -B option + to specify the number of black patches. + +* Fixed gamut mapping intent "rl" to really use relative L*a*b* + +* Add Rec2020 YCbCr encoding support (both types) to collink and xicclu. + Added experimental xvYCC encoding support to collink. + +* Add Rec2020 profile. + +* By default printtarg will create PS and EPS files with a CUPS + job ticket to disable color management. Use the -U flag to disable this. + +* !!! Changed instlib API ambient XYZ and spectral units to Lux. In prevous versions + of instlib they were Lux/pi. !!! + +* Changed spotread so that it won't fall back to emissive spot mode if + an ambient reading is requested. + +* Fixed bug in xicclu for Yxy input. + +* Added display update delay calibration support to i1pro and munki (just like + i1disp3), to improve measurement times. + +* Changed dispcal & dispread so that they wait up to 0.5 seconds when reading dark + patches after light ones to allow for display fall time. Added a patch order + optimisation for display patch sets in targen to minimze the extra time. + +* Added MadVR Video calibration support. + +* Added eeColor Video calibration support. + +* Changed dispwin daemon loader mode from -E to -X. + Added -E option to encode test patch colors in Video 16-235 range. + +* Changed dispcal verify option from -E to -z. + Added -E option to encode test patch colors in Video 16-235 range. + +* Added -E option to to dispread to encode test patch colors in Video 16-235 range. + +* Changed enumerated viewing conditions for new Flare/Glare settings, + changed all to Flareless to improve dark image behaviour, while + retaining Glare modelling. Tweaked brightness and ambient values. + +* Changed CIECAM02 Flare model to distinguish between Flare from + the image itself, and Glare from ambient light. This allows + scaling Glare with ambient automatically. + +* Fix bug in cicam02 in V1.5.0 that causes some mapping + problems in the red for collink -ir or -ila. + +* Changed cctiff so that it does lossless JPEG copy + when there is no color transformation. This makes it + more useful for embedding a profile. + +* Fix xicclu so that it works with device links. + +* Add -Et and -E7 option to xicclu to help with testing + Video calibration device links. + +* Made sure that MSWin test colors are not color managed. + This may affect Vista, Win7 and Win8. + +* Changed spectro/dispwin.c so that null transform + color matching is used for displaying test patches + on OS X >= 10.6. This should fix calibrating/profiling + secondary displays on > OS X 10.6. Note that the 32 bit 10.4 + binary will still have problems on > OS X 10.6. + +* Improved i1pro hi-res mode to improve accuracy. + +* Added JETI specbos 1211/1201 instrument support. + +* Fix some corner case bugs in imdi/cctiff when handling + a calibration file colorspace mismatch, and the zero + color transformation case. + +* Tweak B2A cLUT clipping behaviour slighly, to reduce L* delta E. + +* Added Body Centered Cubic grid option to targen. + +* Added -Yn flag to dispcal and dispread, which skips asking + the user to place the instrument on the measuring spot. + +* Fixed bug in shaper/matrix profile curves that caused + random bumpy black behaviour (shaper curve optimization local minumum problem). + +* Increased default smoothing for XYZ cLUT profiles, to make + their smoothness comparable to L*a*b* cLUT profiles. + +* Improve robustness of i1d3 display update measurement code. + +* Don't add colorant tag to .tiff files in cctiff unless + it is a non-standard space, as Photoshop will barf + on such files. + +* Added support for applying calibration curves in collink. + +* Added a Rec709.icm HDTV, EBU3213_PAL and SMPTE_RP145_NTSC profiles to samples. + +* Fix ColorHug driver so that it is backwards compatible with FW 1.1.8 + +* Fix bug with Spyder not being able to break out of + dispcal adjustment loop. + +* Fix bug in xicc/xicclu, -K flag not being recognised. + +* Fix bug in xicc/xmatrix.c introduced in V1.5 that prevents + matrix only profiles from being created. + +* Using libusb is now deprecated in favour of native USB. + +* Added some prelinary BSD USB driver support. [incomplete] + Version 1.5.1 (8th March 2013) ------------- @@ -1001,7 +1305,7 @@ Version 1.1.1 (21 February 2010) in the input .tif file. * Change spotread so that it takes spectral readings by - default so that FWA comensation readings can be used. + default so that FWA compensation readings can be used. * Changed link/collink to apply Y to L* curve if the input or output space is XYZ. Fixed the Y to L* scaling to make sure it only |