So, since i try to use the embedded FFC in LR, it does not do the job properly!
Adobe-Support could not- or want not help me, so i'm looking for more experienced users.
1. not every image will process, no message is given about this
2. Some of the ones which will, cause damage like some ghosting (like in the attachment)
3. The other, seemly correct processed images will have a strong gamma-shading, useless to reproduce images in a professional way. (example in attachment)
-Highlights getting dropped down, shaddows pushing up, midtones are "ok". For example, the white field on my Target is about L = 93 if it is set at the center.. at the corner of my image, the same field is pushed down to L = 86. Midtones from around 50 to 51 and shaddows from 18 (for example) to 21.
This is clearly not what i want and even not what it is supposed to be.
Struggling around with this, i noticed that the lens-profiles does the same thing when try to eleminate vignetting, but not as hard as the FFC.
Anybody here with the same problems??
I also asked Adobe if it could be better to use a gray-"card" instead of nearly white to set the reference-image, but received no answer. I'm realy unsatisfied about the support they gave me overall.
When you want to include images for consideration, please use the website and use the appropriate tools provided for including images.
It's me again.. Thanks Jim.. here you can see how it looks like.
I have never done this but if I read the instructions here: https://helpx.adobe.com/lightroom-classic/help/flat-field-correction.html it seems to imply you should use a plastic diffuser card like an expo disc for the reference image. It sounds like you're using gray or white cards. Apparantly a diffusor in front of the lens is what is supposed to work well: https://www.lightroomqueen.com/community/threads/flat-field-correction.37750/
Yes, you're right.. in the case you want to eliminate vignette and colorcast of your Lens+Cam-Combination that's the way to go..
But in my case it is used to correct small uneven illumination too. Which is the usal way to reproduce images in Studio because you simply can't illuminate the surface perfect homogeneus.
Ah! Not knowing anything about how the flat field correction is implemented, my guess would be that the correction is based on a typical spherical model of lens and sensor with at best a quadratic radial curvature of the vignetting and color shift. It probably doesn't work for uneven illumination that doesn't fit this sort of spherical/radial behaviour. I would probably look at Photoshop using the reference images as overlay to color/intensity correct in a layer or using the math functions. That will take much more effort to get right
Sounds logic... BUT 😉 :
First, in this Studio the impact from lens-vignette is higher than the light falloff by illumination.. which i can simply see when i set up the white with warning on.. i can see the spherical overexposure. When i use the lens-profile, it goes the other direction, so you can see that in the middle (highest distance to lightsources) is the darkest point.
Next is, the FFC is used thousands of times for exact the same reason.. and even more.. in further plugin-version you could remove the dust too.
I've used PS for some experiments.. and yes, can work with some tricks, but nothing usefull for a worfklow like i need, it is more of a workaround..
And did you see my examples.. strange things happened..
I see. I would guess that some of the pixels in the nearly white shot are overexposed in the raw data and you would be better off exposing a bit shorter instead so that the brightest pixels are not overexposed in any channel. Hard to tell whether this is the case from the screenshot but possible. Would be good to see the eactula raw file for the reference shot. The ghosting probably is caused by the plugin using the wrong shot for the reference.
ok.. the reference shot is not overexposed (white up to 82 here), all channels are in. And you can see what i was talking about, vignette is over fallout.. second pic is with lens-profile on, exact the same raw-file and set white up to 82 too.
And the ghosting.. i tried many times.. reference first and also as the last frame.. In my first example you can see the reference and the two shots to calibrate.
Sometimes it worked.. in this case it worked when i selected only one shot to the reference, but that was not the plan. With the old plugin-version i never had this problem.
You can't actually tell in camera raw or Lightroom whether any of the raw channels are overexposed. The over indicator shows where the data in the rendered color space is at the highest bit value, not where the raw channels are over. You can sometimes tell if you dial in negative exposure if the histogram stays spikey instead of looks like a Gaussian peak (which it should look like) but otherwise you have to use a raw analysis tool and look at the actual values. I would definitely attempt this with a not as high exposure image as reference. Just dial the shutter faster a stop or two.
regarding the ghosting, that appears to be a bug where it thinks the more gray image is the reference. I would report as a bug at https://feedback.photoshop.com
Yes, you're right again.. for sure.. can't see this in LR or CR application, after virtual colorinterpolation/rendering in colorspace (ecirgb-v2 here) .. but otherwise i have to push up the exposure of my real images much more than i want to. And i cannot imagine that this shot is to bright under this conditions. The raw is just about 50mb, so i can't upload this here, even if i want to..!-But i've the Raw-Histo now :
Otherwise (again) it seems to me like the lenscorrection is doing the same thing, like i wrote in my first post. So, even when the target is placed at the edge, the highlights being dropped and shaddows pushed. And this is strange.. especially when taking a look to the comparisson from the reference shots. I would guess the target would have a higher value like when it is placed in the middle.. but nope, it does not.
Maybe the implementation to correct the falloff is the same or works close to the one on FFC.
I told adobe about the several problems, send them the pics&infos i uploaded here.. but nothing.. 10 peoples in a row.. everytime they starts with questions like what Sys i've running, which version, Camera, Lens, etc.. everyone wanted to have the examples again, ... obviously this wasn't polite and helpfull.
Got a number but no help..
and now my post is gone away..!? Did this happened here before?
Those histograms look good to me. No clipping in any channel indeed so all good. I think you need to note this as a bug in the plugin at https://feedback.photoshop.com. This forum is just fellow users.
Yeah.. that is what i mean.. thanks for your opinion.
So.. i reported this again at Adobe-Support.. with all the infos plus some raw-files.. but they're refused to pass me on to the experts without a video-sequenz from my monitors.. starts to get implausible.
The guy even ask if i can meassure the Grayskale and everything else in the video.. the experts have to see my complete workflow to help me.. what a joke...
Have you actually tried the feedback forum I mention above. They have actual programmers and engineers there roaming the forum and they tend to actually search for bugs there to address. If you already did, my apologies.