The GetDPI Photography Forum

Great to see you here. Join our insightful photographic forum today and start tapping into a huge wealth of photographic knowledge. Completing our simple registration process will allow you to gain access to exclusive content, add your own topics and posts, share your work and connect with other members through your own private inbox! And don’t forget to say hi!

a couple of LCC questions

jotloob

Subscriber Member
Thanks Jae

for sharing your findings . Very interesting indeed .
I think , this is particulary good news for long time exposure shooters.
I very seldom do very long exposures , but will keep this in mind when I might need it .
 

yaya

Active member
To make things simple:

High iso = more noise = lower LCC quality

Jae's examples both seem under exposed and IME the Expodisc is not good enough for LCC

When I run out of exposure I would use a flashgun or do the LCC shot at a later stage (after I've written down the aperture and lens displacement)

Yair
 

jotloob

Subscriber Member
I have written lcc software myself so I have looked at the problem in detail. Noise in the lcc shot can be quite high without problems. The shot is blurred before applied so noise is reduced to practically zero, except for dust details that cannot be blurred in order to make dust spot cancellation possible. So I would not worry about a lightly underexposed lcc shot, ie keeping the same exposure should be ok.
To make things simple:
High iso = more noise = lower LCC quality
Jae's examples both seem under exposed and IME the Expodisc is not good enough for LCC ...........
Yair
So ? ? ?
Torgers contribution actually is the total opposite to Yairs contribution .
I use 40-45% translucent acrylglas (2mm) as well as expodisc neutral filters 67+72 mm . Depending on shooting situation .
The only difference I found so far is that the expodisc needs one stop more for correct exposure . In many cases this does not matter at all .
 

jotloob

Subscriber Member
To make things simple:

High iso = more noise = lower LCC quality

Jae's examples both seem under exposed and IME the Expodisc is not good enough for LCC

When I run out of exposure I would use a flashgun or do the LCC shot at a later stage (after I've written down the aperture and lens displacement)

Yair
The expodisc filter is Not Good enough ? ? ? (i am not taking side of any Brand here)
I wonder , what is Good enough then ? ? ?
 

yaya

Active member
The expodisc filter is Not Good enough ? ? ? (i am not taking side of any Brand here)
I wonder , what is Good enough then ? ? ?
The LCC function in Capture One and Leaf Capture has been developed to work with a certain type of diffusers (the ones that we provide, or similar 3mm White opaque perspex) and there are clear guidelines for how they should be used in terms of exposure etc.

Any other device/ method combination is likely to be a compromise. Might work for some though.

I've been working with hundreds of architecture/ landscape photographers since 2000 and have personally taken about a gazilion images as part of the development process and I prefer to use the supplied diffusers and guidelines as I find them to provide the best results

Yair
 

torger

Active member
To complete what I stated earlier, if your scene have bright highlights you will likely need to increase exposure (or add external light source) or else the LCC shot will be too underexposed, but in a lower contrast scene keeping the same exposure should be okay.

I don't know how Capture One has implemented their flatfield correction, but if they've done it smart it should have some sort of blur involved which reduces the quality requirement of the LCC shot greatly.

One could do some testing, shoot a uniform flat surface and make increasingly underexposed LCC shots for that and pixelpeep and see how bad it can be before there is a visible noise increase in the corrected image. There is a risk though that you will start to get subtle problems with color accuracy in the shifts before noise increase becomes visible, if noise supression is good.
 

Wayne Fox

Workshop Member
To make things simple:

High iso = more noise = lower LCC quality

Jae's examples both seem under exposed and IME the Expodisc is not good enough for LCC

When I run out of exposure I would use a flashgun or do the LCC shot at a later stage (after I've written down the aperture and lens displacement)

Yair
ISO on a Phase back is pretty much just a metadata thing, the recorded data is really the same, the ISO is passed on to the processing pipeline so adjustments can be made. The noise is the result of a lower signal to noise ratio because basically when you increase the ISO, you underexpose the signal.

In a pinch you can just shoot the LCC at the same setting as your work, and then just normalize the exposure up a couple of stops before creating the LCC in C1. End results are pretty much identical to adding two stops via ISO adjustment.

However, adding light to the plexi as suggested by you and others does seem the best suggestion.
 

yaya

Active member
In a pinch you can just shoot the LCC at the same setting as your work, and then just normalize the exposure up a couple of stops before creating the LCC in C1. End results are pretty much identical to adding two stops via ISO adjustment.
Sorry but this isn't how the LCC function in Capture One works...you can adjust the exposure of the LCC image before or after you analyse it but it won't make any difference to neither the RAW data nor to the resulting image

The most important 2 things are to expose correctly i.e. right edge of the histogram sits at +2 and is iso kept to the lowest possible setting.

You can use flash or other light-source to boost the exposure or you can create the LCC image at a different time, providing that you have recorded the aperture and lens displacement (and focus settings if you want perfection)

Obviously you can ignore the above and you will still get an image and in most cases it will be useable or every useable but you are making a compromise

Yair
 

Wayne Fox

Workshop Member
Sorry but this isn't how the LCC function in Capture One works...you can adjust the exposure of the LCC image before or after you analyse it but it won't make any difference to neither the RAW data nor to the resulting image

Obviously you can ignore the above and you will still get an image and in most cases it will be useable or every useable but you are making a compromise

Yair
Yes, the only reason I mentioned the adjustment is then "visually' you would see something very much like you would see if you increased the ISO. Just pointing out that increasing the ISO doesn't change the actual data that is captured, it just sends a metadata setting to the post processor. I assumed the LCC worked strictly on the raw data so any post settings don't apply, and any increase in ISO really doesn't have any affect either.

and your right, a perfect LCC would always be the best, but unless your lens has some serious issues (like a Schneider wide on a IQ180 back), results are normally very acceptable.
 

torger

Active member
Made a quick test with heavily underexposed LCC shot. I don't get problems with noise, but what happens is that the digital back centerfold which normally is no problem becomes so prominent in the LCC shot so when I apply that a centerfold artifact appears in the plain sky. Ie centerfold comes from the underexposed LCC, not from the main image.

So there is a limit of how dark you can make your LCCs, and noise may not be the first problem appearing.
 

Pemihan

Well-known member
Made a quick test with heavily underexposed LCC shot. I don't get problems with noise, but what happens is that the digital back centerfold which normally is no problem becomes so prominent in the LCC shot so when I apply that a centerfold artifact appears in the plain sky. Ie centerfold comes from the underexposed LCC, not from the main image.

So there is a limit of how dark you can make your LCCs, and noise may not be the first problem appearing.
That is very interesting!
 

miska

Member
So imagine you have a perfect LCC: well exposed, exact same settings (aperture, focus, ISO), taken just after the shot. How well are color casts corrected ? Do you ever see "a residual" color cast after the LCC is applied, especially if you "push" the image (i.e. pull shadows or highlights, enhance contrast, change white balance) ?

I understand that noise may increase (because if you have vignetting, you pull the shadows), but how well is color cast dealt with ?

Second question: if you forget to take an LCC, what are your tips and tricks to remove color cast ? I tried to generate an LCC with the same aperture I used in the field (and same focusing distance, but very different lighting, as I shot the scene at sun rise but the LCC mid-day) but for some reason, that didn't work so well - it's possible I screwed up something though :)

Experience from both C1 and Lightroom users are interesting to me...
Thanks !
 

torger

Active member
Residual color cast: at least in theory there could be some residual color cast left, because the actual scene has varying color (blue sky, green grass etc), and the sensor will react slightly different depending on the color. The correction is made with a fixed color, ie ambient light filtered through the LCC, and there can thus be a slight mismatch. How large this residual error is I don't know though, one has to make controlled experiments, I might actually do that at some point. The sense I've got from using LCC over a few years in lots of varied conditions, and also implementing and testing LCC algorithms is that you don't need to worry about this error, it's small enough to be negligible.

Then you have crosstalk, which I've studied deeply during the last couple of months. Crosstalk is more common than people may think, in fact even a long(ish) lens as the 60XL will introduce some crosstalk when shifted. Crosstalk is dependent on color and is not corrected with any normal LCC algorithm. The symptom is generally not color cast, but desaturation and in severe cases demosaicing artifacts. Neither Lightroom or Capture One can correct for this, so it will be there after correction, but as it's generally only a slight desaturation of colors it's not too hard to live with. You could also manually make an approximate correction for it in postprocessing with a local saturation increase.

Additional artifacts related to wide angle response that can appear is micro lens shading (a ripple effect) and tiling, but both of these can be corrected well by LCC, exactly how well in various software I have not studied in detail, but I'd suspect Capture One to be considerably better than Lightroom.

Second question: if you forget to shoot the LCC, doing as you describe would be my tip. I've done that a couple of times, with good results. At those times I had similar light when I remade the LCC shot though, the different light may have caused the problems for you.
 

AreBee

Member
Torger,

Then you have crosstalk...as it's generally only a slight desaturation of colors it's not too hard to live with. You could also manually make an approximate correction for it in postprocessing with a local saturation increase.
Applied radially, increasing with radius?
 

torger

Active member
Applied radially, increasing with radius?
Yes, in principle. One problem though is that most sensors have a different amout of crosstalk vertically vs horizontally, due to how pixels are designed. This means that if you shift you have your sensor horizontally and shift it up you may get more (or less) than if you shift it sideways.
 

miska

Member
Is there a way to apply an "almost good LCC" and tweak it by hand, decoupling vignetting from color cast and changing the magnitude of the correction ?
I am a Photoshop ignoramus, so I don't know how to take the LCC, normalize it, divide the shot by the normalized LCC and then perhaps boost / reduce the LCC intensity to get a scaled down/up version of the color correction.
 

Aviv1887

Member
Within C1 they're sliders that give you control of the amount of correction. You're right in your earlier suggestion that the more vignetting occurs (more with SK lenses for example) the more Ci needs to work to bring that back. Noise starts to happen at those places. The uniform light slider in C1 controls the amount of correction. When you pull that back from the auto 100% it also leaves less noise behind. It still is a pretty amazing process the LCC function. If that wasn't possible we would not be shooting tech cams.
 

weinlamm

Member
In 98% of my pictures I take with Tech.Cam I'll make my LCC later. Or better: I use a set of LCC's from my portfolio. I have one with different shoots - but all taken in steps from 5 degress.
That is something I take care when I shoot: always in 5 degree-steps. If it's possible... But if really not - then I have to make a LCC directly after the shoot.

Where is you problem with your shoot exactly? If you had the same aperture and focus the LCC should work.

What I found out for me (I shoot much 2:1 panoramas with a P20 and with this I have to take always 3 pictures) is, that if your LCC is "good" the residual deviation depends on the white balance.
For me I now have a set of different white-balance-shoots, too (every of my 3 pictures with different Kelvin and different Tint).

Perhaps this helps you? If not perhaps you can show a picutre.
 

miska

Member
Yes, it's true that with some white balances, the LCC taken post-shoot were fine. But isn't the LCC independant of the WB, as it is shot in raw ?
 

weinlamm

Member
I'm not sure with the requirement for white balance in raw. I shoot every picture with the same white balance (sun :)). So this is not a real problem for me. I correct all my files in post process - after apply LCC + white balance.

So for my 2:1 panos with P20 for example:
- 3 shoots (+-20mm; 0) if necessary shift, too

C1:
- different LCC for each of this pictures
- different white balance for each (in real sun round about 5k left, 5k middle, 3.5k right); I had taken these originally from the LCC-card-pictures I used for LCC
- If it's "other weather" I have a white balance set with different Kelvin - but the differences in Kelvin-number is somewhere in this size
- export as tiff

PS/Stiching software:
- stiching the pano
- if necessary I now adjust the white balance again (if you don't shoot pano this could done in C1, too)

If I wanted, I could do the last step in C1, too. But then I had to import my final tif to C1 again - and if I'm working with PS I'll stay there.


You asked about Lightroom, too:
If you work with LR: you know the dng flat field plugin?
 
Top