Questions about using LCC plates.

[drevil]

Hi Steve. I don't use the 'library' approach, but I think I'm right when I say that the only things that matter are you use a pre-made LCC image that was shot with the same lens, same aperture, and same movements. This would be a pain for me because I might use 5mm of rise and 2mm of shift for one shot, and then some other combo for another. But if you're shooting the same pattern of images (e.g., your 9 LCC frames at f/11) this could work well for you.

one way to work around this is to simply work in the field with 5mm increments(and crop later to liking, when necessary). if work is focused on stitching i see no problem
So pre-fabricated LCCs would work for me.

2 sets of LCCs of 5mm increments, one with F8 and one with F11 (for me)

 

[mristuccia]

This is how I usually place the LCC in the histogram when taking it:



On the field I usually only notice the big peak in the live histogram, I change the exposure time so that it moves to the middle-right zone.

Regarding using higher ISO, it smells a little bit . I also use the LCC to fix the PDAF banding problem, so I want as much definition and less noise as possible on the recorded bandings. I'd probably better take note of the settings and take the LCC in a second moment during the day.

Regarding the presets archive, I'm not pre-building it intentionally. It grows gradually as I take new shots and new LCCs.

 

[rdeloe]

Now that I've managed to get flat field correction (partially) working with Lightroom again I was able to explore the question of the exposure level for the LCC frame.

I shot the same scene three times, keeping the scene exposure the same, but using different exposures for the LCC frames. For the first LCC frame, I did not adjust exposure compensation. For the second I added a stop, and for the third two stops.



Notice how dark the bottom-right corner is in these LCC images. Here's the corresponding corner for each of the three scene images processed with these LCC frames. I'm showing a crop that is 587 pixels on the long edge, from an 11,648 pixel (long edge) file. This puts me deep into the darkest part of the scene. There's no difference that I can detect in noise.

 

[dchew]

As @jng said, I try to remember the LCC at the beginning if it is sunset, or the end if it is sunrise. The one thing I do sometimes to cheat the darkness is swing the camera around to face what little light there is left in order to shoot the LCC. So when taking the LCC, I will swing the camera around to the direction of where the sun is rising/setting. That usually adds enough light to keep the shutter speed the same, or close to it.

Dave

 

[mristuccia]

Now that I've managed to get flat field correction (partially) working with Lightroom again I was able to explore the question of the exposure level for the LCC frame.

I shot the same scene three times, keeping the scene exposure the same, but using different exposures for the LCC frames. For the first LCC frame, I did not adjust exposure compensation. For the second I added a stop, and for the third two stops.

View attachment 219595

Notice how dark the bottom-right corner is in these LCC images. Here's the corresponding corner for each of the three scene images processed with these LCC frames. I'm showing a crop that is 587 pixels on the long edge, from an 11,648 pixel (long edge) file. This puts me deep into the darkest part of the scene. There's no difference that I can detect in noise.

View attachment 219597

That's reassuring.

When I was working with the non-BSI CFV-50c and SK 35XL, I remember Phocus sometimes complaining with a warning saying that the LCC taken was not good enough. That happened either upon an underexposed LCC or with too extreme shift movements. So, somehow Phocus makes a quality check on the data recorded in the LCC when creating a correction profile.

If you have a look at the histogram I shared in a preceding post, you'll see that the LCC extends over a long range, although the majority of data is located at the very right end. I think that the few data on the left are as important as the rightmost peak. So, like in any digital photo, as long as we don't cut any data on the left or on the right, we should be okay, and you demonstrated here that there will be no difference in terms of noise on whether we ETTR or not.

Quick question: did you change the exposure time or the ISO? I'd be curious about noise figures when changing ISO as opposed to exposure time.

 

[rdeloe]

That's reassuring.

When I was working with the non-BSI CFV-50c and SK 35XL, I remember Phocus sometimes complaining with a warning saying that the LCC taken was not good enough. That happened either upon an underexposed LCC or with too extreme shift movements. So, somehow Phocus makes a quality check on the data recorded in the LCC when creating a correction profile.

If you have a look at the histogram I shared in a preceding post, you'll see that the LCC extends over a long range, although the majority of data is located at the very right end. I think that the few data on the left are as important as the rightmost peak. So, like in any digital photo, as long as we don't cut any data on the left or on the right, we should be okay, and you demonstrated here that there will be no difference in terms of noise on whether we ETTR or not.

Quick question: did you change the exposure time or the ISO? I'd be curious about noise figures when changing ISO as opposed to exposure time.

Lightroom also checks the quality of the image. It will not perform the correction if there's hard mechanical vignetting, which makes sense. Here is an example where I shifted a lens past its circle of illumination. The extreme corners at left and right are almost purely black. LR rejected this one -- as it should.



I kept ISO the same (base -- or 100 for my GFX 100S). I shot in aperture priority mode, and let the camera pick the shutter speed. To add exposure, I used the exposure compensation wheel.

With the proviso that I don't actually know what it is doing when it does a FFC... I have a hunch that it is not using the LCC image as "data", in the sense that it is not performing a blend that combines the data of the scene image with the data of the LCC image. Instead, it is simply using the LCC image to check for light falloff and colour shift. This could be as simple as comparing adjacent pixels using a median sampling approach. If it decides there is light falloff or colour shift in a pixel, I think it's using the LCC to decide what change to apply to the scene image.

If I am correct, then exposure, colour temperature, ISO, noise, etc. won't matter (if the LCC image is 'in spec') because it's comparing regions of the LCC image to itself. An LCC image would not be in spec if it had hard mechanical vignetting.

 

[TimoK]

I can confirm that color temperature does not matter but I'm not sure of exposure and noise. I made a small test.

This picture from Actus rear shifted 15 mm down and 10 mm left, 2.5 ° left swing. Apo-Grandagon 55 mm. Jpg out of camera.


Then lcc assigned in RawTherapee. Lcc was taken at the same place. Very cloudy weather, very blue. RT claims 6900K


Then I take a new Lcc under a 4000K lamp ( RT says 3690K). New lcc assigned to the same picture.


Btw. Fortunately Therapee doesn't reject lccs with pure black and that's a good thing. It happens that I go over the image circle or Actus might vignette one corner to black. This picture can still be very useful in panorama.

 

[rdeloe]

Raw Therapee is my reliable backup when all else fails in LR. However, I don't like having to go there because I'm now working in TIFF, and the results are different enough from LR to be an issue.

 

[SrMphoto]

Thinking more about it,

one way to work around this is to simply work in the field with 5mm increments(and crop later to liking, when necessary). if work is focused on stitching i see no problem
So pre-fabricated LCCs would work for me.

2 sets of LCCs of 5mm increments, one with F8 and one with F11 (for me)

What about distance and tilt? Those are variables that change the LCC frame, don't they?

 

[SrMphoto]

Raw Therapee is my reliable backup when all else fails in LR. However, I don't like having to go there because I'm now working in TIFF, and the results are different enough from LR to be an issue.

Since you are working with GFX, you could use C1, which apparently has working batch-processing.

 

[rdeloe]

Since you are working with GFX, you could use C1, which apparently has working batch-processing.

C1 is always an option. If I did this in large volumes, I'd probably go that direction.

But I only need LCC frames with one lens, and I'm not shooting thousands of frames. I just did a set of six pairs and was done in almost as long as it would have taken to select them all and run it once. I have FFC assigned to a hot key, so it's two clicks to select a pair, press Ctrl-Alt-F and hit enter, two clicks to select the next pair, etc. I want Adobe to fix this properly, but personally I've reached the stage I call, "Perfection is the enemy of Done".

 

[guphotography]

You only need one LCC frame per movement, as long as the movements remain the same, you can point the camera at brighter ligth source or simply shine a light onto the card to shorten exposure time, make sure nothing obscure the lcc card that would cast a shadow.

C1 some times warns about inadequate exposure level, but it still allows correction, and the differences are negligible by eye.

 

[Reginald Fluggelding]

You only need one LCC frame per movement, as long as the movements remain the same, you can point the camera at brighter ligth source or simply shine a light onto the card to shorten exposure time, make sure nothing obscure the lcc card that would cast a shadow.

C1 some times warns about inadequate exposure level, but it still allows correction, and the differences are negligible by eye.

I was about to make this same (or very similar comment). I often shoot very dark interiors and there's no way I can do 10, 15, 30 second LCC exposures (Often I am just doing a 1 second exposure but using FA on IQ4, but it still needs a regular LCC exposure). Held at the right distance my iPhone torch gives a consistent 1 second exposure for a properly exposed LCC. The reason you might want to do this versus some of the other methods described is that if you are using the LCC to get rid of dust, they won't work unless focused at the same distance. Two of my lenses don't have any cast on the IQ4, a little bit of vignetting, but 95% of the time there are dust spots.

 

[rdeloe]

Using the LCC to get rid of dust must be a Capture 1 thing because that's not something that I've ever seen discussed (or just seen!) with Lightroom's implementation. Very handy.

 

[John Black]

The dust correction is handy, but I don't do LCC's in the field. I take notes on the shots and if an LCC is needed, I'll recreate it at home with the camera pointed up towards the sky for good illumination.

 

[chriswebb]

I believe Phocus may also do dust correction with LCC frames, but its been a while since I have used it, and it wasn't functionality that I relied upon so I could be mistaken.

Flat files are commonly used in astrophotography to get rid of dust and other problems in the imaging train. Usually though you are taking multiple flats (at the beginning or end of the imaging session) and integrating them together into a master and using that to calibrate your images [along with darks and (dark flats or bias images to calibrate your flats)] as opposed to the single LCC here. Pre-created dark and bias libraries can be used with caveats, but flats should be taken every session for good calibration.

Obviously the calibration used in astrophotography is overkill here, since it's a lot easier generally to spot and correct dust spots in terrestrial photography, and if you can't see the dust spot it probably doesn't need to be corrected.