New profile making software: Lumariver Profile Designer

[DustinFrancis]

you can contact the author directly - the software is not bug free

Thank you, I sent an email on 2/27 and one earlier today and am awaiting a response.

[torger]

I have a question I am wondering if anyone can answer.  I have a film scan of a Color Checker SG that I measure the patches of in 3DLUT Creator and then use that data with a Babel Color Patch Tool reference file to create a DNG Profile of in Lumariver Profile Designer.  When I render the profile, the patched definitely change to close to the reference patches...especially the hue, but almost all of them are quite a bit more saturated and either lighter or darker than the reference patches.  I have everything in the Profile Designer set to default and have tried utilizing almost all of the tools and options available, but that only makes things less accurate/pleasing.  My main question is, how do I get the rendered profile patches to match the reference file patches?

Hello. Currently I don't have much time for the forums so support email is the best.

I'm not sure what you are trying to do here, but it seems like you are not trying to make a film simulation profile? That is you use a film scan to make a reference not of the real colorimetric colors, but the film response, and then use that file with a target shot of a regular camera, in order to make a profile that makes that camera look like the film? That's probably not going to work. Lumariver Profile Designer is not designed to be a generic LUT Creator software, but is designed to make camera profiles anchored in realism with only smaller subjective elements. It makes assumptions of how camera/scanner hardware works. This means that it always starts off with a linear matrix and make LUT corrections on top, which make robust smooth profiles, but requires that the system profiled is at large linear. If the behavior of the camera (or the reference file) would be grossly non-linear, that is the camera/reference does not work as real cameras do, the matrix matching will go haywire, and then the LUT correction, which is supposed to make only mild smooth refinements on top, will get into trouble. I haven't really tested these workflows but I wouldn't be surprised if you get crazy results or even if the software fails in optimization and get stuck.

For creating film simulation LUTs there are better tools out there, Lumariver Profile Designer is not intended for that purpose. It would require a different design which does not use a matching matrix at all but instead work with really dense targets (lots and lots of patches to cover the full gamut) and make a complete non-linear match.

If however I have mis-understood your workflow and you are really making a profile in the normal way, the best way is to use the archive project function and send a wetransfer/dropbox/whatever link to the support mail and I can have a look and see what goes wrong. The usual problem in that case is that there is some issue with the target photo (excessive glare, clipping etc), or that the grid is not correctly placed, or that there is some problem with the reference file not matching the target.

[DP]

Currently I don't have much time for the forums

Hi Anders,

if possible - any approximate guesstimate about maintenance releases for LrPD and dcamprof with bugfixes ? thank you.

[torger]

Hi Anders,

if possible - any approximate guesstimate about maintenance releases for LrPD and dcamprof with bugfixes ? thank you.

It's soon time for a new patch release for LrPD, I'd guess within a month or two. DCamProf is harder to estimate, as I prioritize supporting the commercial software. I have a new DCamProf release about 70% ready, it's more documentation than actual code left to do, so there will be a release eventually I just don't know when. In the autumn perhaps.

[DP]

It's soon time for a new patch release for LrPD, I'd guess within a month or two. DCamProf is harder to estimate, as I prioritize supporting the commercial software. I have a new DCamProf release about 70% ready, it's more documentation than actual code left to do, so there will be a release eventually I just don't know when. In the autumn perhaps.

Thank you !

[DP]

as I prioritize supporting the commercial software.

PS: as a side note - you need to change the licensing for LrPD again - it was cracked

[torger]

PS: as a side note - you need to change the licensing for LrPD again - it was cracked

For those that wants to crack software it's relatively easy, it's just about removing checks. It's a little bit harder now than the first release, random crashes here and there will occur if not all counter-measures were compromised, but it's not rocket science, that would just be a waste of my time. All popular software gets at least semi-cracked sooner or later, that's just the way it is, so I guess I can see it as a compliment. A key generator is different from a crack though (it means that you can make unauthorized keys and thus don't need to re-crack every new patch release that comes out). It was for that the software's key scheme changed, and that does not need to be changed for a crack, and even if it would have to I'm done with that now. Sooner or later a key is stolen and you need to block it for the next release etc, make new checks etc. The only comforting in this is that a cracker must spend much more time than I have -- for the sake of my paying customers I'm spending minimal effort on piracy counter-measures, instead I try to provide the best support I can and fix bugs (the actual ones, not the ones caused by a bad crack...). I prefer supporting through email though as I'm not much on the forums currently.

[therightkey]

Hi,

I purchased Lumariver last week, it's something I've looked for a while so thank you to the creator!!

I'm having a lot of trouble creating a usable profile even when I follow the directions. I know i'm missing something crucial in the exporting process but just don't know what.

My results yield a profile where the midtones/highlights are crushed/destroyed severely, where the ICC profiles are not usable at all. I've emailed "info@lumariver.com" a few days ago but haven't heard back.

I've attached pictures of my results. It's an exaggeration of curve to show how the different skin tones are treated. The Generic C1 ICC profile keeps all the colors very linear where as the new Lumariver profile handles them very differently.

My Gear:
Capture One - 11.01
Lumariver -  1.04  (Order # 69150422 )
Camera: Canon 5D Mk IV
Lens: Canon 24-70 f/2.8L II
Color Checker Passport

[Jack Hogan]

My results yield a profile where the midtones/highlights are crushed/destroyed severely, where the ICC profiles are not usable at all. ...

I've attached pictures of my results. It's an exaggeration of curve to show how the different skin tones are treated. The Generic C1 ICC profile keeps all the colors very linear where as the new Lumariver profile handles them very differently.

Hi,

I am sure Torger will chime in in due time.  In the meantime, what do you mean by crushed/destroyed?  Once you start playing with curves 'accuracy' is thrown to the wind: without the raw file nobody can tell you which of the two is more 'accurate' (including you:-).  If you are after 'accuracy' you should not use levels/curves/brightness/contrast at all - all of that should be set in the profile for the given capture/viewing conditions.  On the other hand you are allowed to use the 'Exposure Compensation' slider, assuming it works linearly (ACR/LR current process doesn't, I don't know about C1).

Jack

[sebbe]

...
My results yield a profile where the midtones/highlights are crushed/destroyed severely, where the ICC profiles are not usable at all.
...
Color Checker Passport
...

Did you follow the detailed steps in the manual? http://www.lumariver.com/lrpd-manual/#c1_workflows
Most time the target shots are mediocre. May you share your CCPP-RAWs?

[therightkey]

Hey Everyone,

Yeah Togers got back to me and has been super helpful I'm still having weird issues with colors. Maybe someone here has experienced this. I was able to come down the clipping and generally everything looks great.

I still get super weird uneven color hue changes in the sky that I can't seem to fix. Where blue turns purple.

[longpvo]

I just want to share that for anyone looking to buy a X-Rite ColorChecker Digital SG, you can get one with calibrated/measured references from Image Science Associates. It's a great deal if you don't have a spectrophotometer.

Also CMP now offers a new robust target (early 2018) as they claimed (without optical brightening agents), namely Digital Target Studio Edition (DTSE) and it's shipped with individually calibrated reference data.

While basICColor has a new target called dcam mini (late 2017)

Here's my 2018 targets comparison for anyone considering buying a new target.

Usability (glare, handling, multi-purpose): CCPP (ColorChecker Classic,Photo Enhancement,White Balance Chart*) -> dcam mini -> DTSE -> CCSG/IT 8.7 C1

*At first glance this patch appears to be a large version of the Neutral 8 patch from the original ColorChecker, but this is not the case.
The ColorChecker Classic’s Neutral 8 patch has a spectral reflectance which, although relatively flat across the spectrum, has a slope to it. This slope was due to its colo- rants and gave it a slight color cast, making it unsuitable for neutral balancing a digital camera.
The ColorChecker Passport’s White Balance Chart is very flat across most of the spec- trum. This shows that X-Rite has made a special formulation to remove any color cast

General Accuracy (optimum colour constancy in various lights) CC/CCPP/CCSG (made with a number of pigmented paints applied to paper with a variety of spectral characteristics) -> dcam mini (made with 3 colorants) -> DTSE (made with pigmented inks, on a substrate without OB) -> IT 8.7 C1 (printed)

Built (physical built quality/storage): CCPP/dcam mini -> CCSG/DTSE/IT 8.7 C1

Gamut (ultra saturated colors stuff): DTSE -> IT 8.7 C1 -> dcam mini -> CCSG -> CCPP*
* If you have your own reference data for the spectrum patches, you can profile the whole CCPP and cover some more saturated colors, though not as much as the CCSG.

Reference Matching (some are offered with calibrated reference data at additional cost/free vs measurements from batch): DTSE -> CCSG -> IT 8.7 C1/dcam mini/CCPP

There are also monstrous targets like ISA's Near Neutral/ColorBuild 300 and Library of Congress's Next Generation Camera Calibration Target.

[jrp]

I seen that the new version of MacOS requires that software distributed outside the Mac App Store must be notarized by Apple in order to run on macOS Catalina.  I hope that the notarisation process is under way ...

[Dinarius]

Bringing this to the surface again...

I’m making Repro Profiles out of Capture One.

After applying the profile to a file back in C1, the readings I’m getting from the X-Rite CC (after boosting exposure to account for Linear Mode darkening) are good. But, even tho’ the blues read ok off the CC, they look a bit dead. Adding back in some contrast doesn’t help much.

Monitor is calibrated to D65/2.1/120.

On another note, last time I tried a Lumariver profile file in this https://deltae.picturae.com/index/home (two months ago) they were getting five stars. Now, when I load a file, I’m getting File Not Found. Odd!

Thanks.

D.