- This topic has 1 reply, 2 voices, and was last updated 6 years, 4 months ago by
.
-
Posts
-
my full worflow is actually this (I need correction matrix, ICM and 5x 3D LUTs):
0) self-cal i1pro
1) create matrix for i1d3
2) do preCal report (verification)
3) interactively adjust whitepoint, keep screenshot
4) do long cal+profile run (e.g. slow+11140 patches)
5) keep screenshot of coverage result (not directly recallable later!) and profile information (can be viewed again later)
6) do postCal report (verification)
7) create Rec709, DCI-P3, BT2020, NTSC, and PAL 3D LUTs (actually 5 steps)
8) manually install all the 3D LUTs in madVR (install button does not work for me, it either changes nothing or assigns it to the wrong monitor/profile in madVR)
9) enjoy
10) “I really have to try this setting and have to improve even more” syndrom kicks in, repeat from 0Ideally 1-9 would be fully automated, but that’s a bit tricky to code, so more simple ideas:
0) there is an option “allow skipping of self cal” – I often use it when just quickly messing around, but IMHO it would be better if this does just that: allow skipping, not always skip it (i.e. a skip button when it prompts you), that button could just always be there and then we could remove that option
1) allow to integrate in total workflow and read patches in parallel from spectro+colorimeter – I don’t know if argyll or displaycal draws the patches etc., maybe it’s relatively easy to just start 2 threads and really let them measure as concurrently as possible (draw, tell both instruments to start at same time) – as what we are doing is using the spectro as transfer standard and correcting the colorimeter, we need as little short time drift (monitor, instruments) as possible, as the corrections are sometimes so small, that this short term drift/noise maybe even large than the correction values calulated (just try to do matrix correction as fast as possible and even then multiple runs will differ sometimes by a factor of 10 (though absolutely small) – this is the combination of display drift, spectro drift, colorimeter drift and random noise in all of them. In high end electrical metrology all the above helps a lot! Maybe even averaging multiple readings should be considered (lowers noise) – i.e. increasing integration time
2+6) checkbox (and – if necessary – filename selection, could be auto generated) that does report before/after cal run, maybe have a choice if the pre run should be before/after whitepoint adjustment (which requires user interaction and also depending on what you want to acheive you may want the pre cal report to be before/after whitepoint adjusment)
3) keep screenshot when pressing “continue to measurement”
4) log file currently only has output per patch for cal, not profiling (i.e. for profiling it just logs first and last patch) – intentional?
progress and log could show more information about last patch read, like expected/actual etc. (basically like verification report) and a total running dE etc. could be displayed, basically the data which is shown at the end. during grey cal it could show basically what’s in the log (dE, repeat/OK etc.) – currently there is very little feedback – sometimes for 10minutes on crappy monitors when adjusting blacks – graphs (gamme/CIE) that update after each measurements – the possibilities are endless 😉5) this just helps with documentation, simple feature would be auto-screenshot, advanced would be generating tailored text/images (e.g. no GUI etc.) and I cannot seem to find the coverage info ever again except by recalcing the profile from data…
6) see 2+6 above
7) if you select “create 3D LUT after cal” it will then not create/install the ICM – I guess there is juse one “finished” dialog. It would be a bit of “boring” GUI work to do this, but ideally one could create infinite 3D LUTs in one step and not manually change settings, wait a few minutes, repeat etc.
8) as stated above “install 3DLUT” is broken for my madVR setup
9) do not make this step shorter! 😉
10) there is no cure.
What do you think? I haven’t done much Python before and only Linux CLI not Windows GUI, but if you could point out some easy targets and hints what parts to look at I could try and also would be happy if you are interested in implementing some of them.
Thanks again for this great piece of software!
Calibrite Display Pro HL on Amazon
Disclosure: As an Amazon Associate I earn from qualifying purchases.there is an option “allow skipping of self cal” – I often use it when just quickly messing around, but IMHO it would be better if this does just that: allow skipping, not always skip it
That option already works in the way described, but there is a timeout of 20 minutes or so (and it is highly recommended to calibrate the instrument each time you get asked, as instrument drift makes the readings inaccurate).
allow to integrate in total workflow and read patches in parallel from spectro+colorimeter
Frankly, I see no general use for such a specialized option.
do preCal report (verification)
What most people fail to understand is that this is of little use – only the whitepoint will change (through the 1D calibration, if any, or the display adjustment). The profile does not change the display response in any way. In favor of a simplified workflow, it is better to leave out as much information as possible that is unnecessary or doesn’t contribute anything.
log file currently only has output per patch for cal, not profiling (i.e. for profiling it just logs first and last patch) – intentional?
All that the logfile functionality does in that case is it logs the output of Argyll’s dispread command line utility. There is no further interesting information to be had (other than progress info we already have).
like expected/actual
That’s a misconception. Characterization (profiling) just records the display response. There are no “expected” values.
6) do postCal report (verification)
Automatic profile verification is planned.
if you select “create 3D LUT after cal” it will then not create/install the ICM – I guess there is juse one “finished” dialog. It would be a bit of “boring” GUI work to do this, but ideally one could create infinite 3D LUTs in one step and not manually change settings
Something like that might be implemented at a later time, but currently it has no priority.
