It's fairly easy to compute the degree of blur of the secondary IR image with a lens that has an IR focusing Index. The blur is the converse of the index. If you use an F2 lens wide-open that has the IR focus index at F5.6, the secondary IR image will be spread across that range of distance. Most Leitz lenses have an IR focus index at about the F2 DOF mark. Others, such as Nikkor, Zeiss, and Canon LTM lenses, are between F4 and F8. It is large enough to show up without pixel peeping. It is blur caused by chromatic dispersion at its worst. Too bad my Calcium Fluorite Pentax 85/4.5 is in M42 mount.
Last edited:
UPHOTO
Member
Yeah that statement is pretty bold and probably not worded the right way.
In comparison to another 10-12 MP camera then I agree. The lack of a AA filter makes a world of difference.
I still don't get why other manufacturers don't get this.
Can you imagine a Canon 5D MKII with no AA filter? Man.........
ampguy
Veteran
Hi Doug,
It's possible that a filter could have helped in your shots, but I also think your WB may be off in your indoor shots, probably needs a custom white balance setting, so you may want to carry a sheet of paper, coffee filter, or expodisc etc.
Check out Maggieo's photo above, everything is fine without filter, as her WB is set correctly.
It's possible that a filter could have helped in your shots, but I also think your WB may be off in your indoor shots, probably needs a custom white balance setting, so you may want to carry a sheet of paper, coffee filter, or expodisc etc.
Check out Maggieo's photo above, everything is fine without filter, as her WB is set correctly.
Hi -- Thanks, and this gives me a good opportunity to clarify any confusion in the first pair of indoor shots a few posts up
... The first is straight from the camera; the second is with a custom white balance set from a white sheet of paper on the first table. There was so much IR in there, including the buffet area at right, I could not find a way to make it acceptable, and just converted to B&W... The lunch shot just above wasn't quite as bad, as there was some natural light coming in too, but my best adjustments still had it looking odd... Skin tones and fabrics are still wrong.
Maggie's magic profile resulted in great color, but we haven't seen it pre-profile.
Generally, my outdoor shots on that trip were fine, even when there was some minor magenta shifts in fabric, but sometimes (as above at the stone ruins) it was just too obvious.
Attachments
Last edited:
ampguy
Veteran
Hi Doug, thanks for the clarification. That last image looks much better, whatever you did to it.
I'm still a bit confused on your first 2 images up there at the buffet tables. It looks like an exposure difference and/or WB change, but the same exposure. It doesn't look like you took one exposure, then changed WB in-camera, and then another photo later. The peoples faces and expressions haven't changed anywhere. I agree, both of those could have used a filter, or something.
I'm still a bit confused on your first 2 images up there at the buffet tables. It looks like an exposure difference and/or WB change, but the same exposure. It doesn't look like you took one exposure, then changed WB in-camera, and then another photo later. The peoples faces and expressions haven't changed anywhere. I agree, both of those could have used a filter, or something.
jaapv
RFF Sponsoring Member.
I shoot my 35/3.5 Summaron sans IR filter (no threads on the lens) and have found Jamie Roberts' color profiles for Capture One to work quite well in taming the IR beast, while still delivering gorgeous colors.
Here's a shot from this weekend, of my pal Nick, inspecting a restored Gibson Firebird that belongs to one of his co-workers:
Is your friend really THAT orange?
They are different post-processing of the same shot, with the second one having a white piece of paper on the first table set to be plain white. That changed the color balance and lightened it at the same time. The IR effects are still way too obvious to be acceptable... due to the IR content of the incandescent lighting and even more intense IR of the heat lamps over at the buffet. I'm convinced an IR cut filter would have done the trick, but I wasn't an IR filter believer then.
alecgold
Established
I don't have the magenta shift, but I do sometimes have the IR problem. Ordered my filters through the owners-section.
But it should be possible to cure the shifts with the ICC-profiles?!
I need to get the profiles into C1, but haven't found out how to do it yet.
time, oh precious time. And with the weather being so nice after months and months of cold and slippery weather, I want to be outdoors, not stuck inside.
But it should be possible to cure the shifts with the ICC-profiles?!
I need to get the profiles into C1, but haven't found out how to do it yet.
time, oh precious time. And with the weather being so nice after months and months of cold and slippery weather, I want to be outdoors, not stuck inside.
Lord Fluff
Established
I don't see how any profile can remove IR contamination. Sure it can remove magenta, but what about the stuff in the scene that did contain magenta?
I'm sure it will help sometimes, but it generally amazes me that people will put up with (and post) pics where the colour really isn't correct - porky-pink faces being the most obvious kind of image.
I'm sure it will help sometimes, but it generally amazes me that people will put up with (and post) pics where the colour really isn't correct - porky-pink faces being the most obvious kind of image.
ampguy
Veteran
looks normal to me
looks normal to me
The kid probably spends a lot of time in front of a computer or practicing the guitar. He could be slightly orange from being embarrassed that his photo is being taken.
looks normal to me
The kid probably spends a lot of time in front of a computer or practicing the guitar. He could be slightly orange from being embarrassed that his photo is being taken.
Is your friend really THAT orange?
jaapv
RFF Sponsoring Member.
Well, not with a profile - but if you are into CS and have begun to understand the first principles of LAB color there are quite a few possiblities to differentiate between IR-induced color shifts, colors that can be defined but cannot exist and real color, and correct for that. But be prepared to invest a lot of time and headache in the learning curve.:bang: And a lot of time in the corrections themselves as well. Even then the result may not meet your expectation.
Sparrow
Veteran
Probably a silly question but why not just remove it with Selective Colour in Photoshop? it takes about 15 seconds
jaapv
RFF Sponsoring Member.
What are you removing then? All colors are affected. The magenta cast is the simplest part of the problem. But for instance green paint will be correct, but the green on the tree in front of it will have shifted to yellow.
Last edited:
J. Borger
Well-known
Is your friend really THAT orange?
I am not anti IR filter.
But in all honesty: I have seen A LOT worse from M8 with IR filter and M9.
jaapv
RFF Sponsoring Member.
So have I. But I have seen a lot worse from a 5D as well. Or any digicam for that matter. Let's not get started on pilot error...
Sparrow
Veteran
Attachments
jaapv
RFF Sponsoring Member.
I don't want to offend, but I can only call those colors horrible.
alecgold
Established
the IR is registered in the dng file and not as black or purple. so if you isolate the ir in the dng file and "tell it to get black" that could work pretty well. At least that is the short version of what I understood about it.
But the magenta isn't a IR problem, its the AWB being off, that is easily corrected.
jaapv
RFF Sponsoring Member.
Excuse me How would that be? The whole point is that the Bayer Filter cannot decode IR light, being RGB, so it cannot register IR light The reason it records as purple is that green is the dominant color in the filter, so the impossible light gets registered as the opposite: magenta. Other colors that hold an IR reflectance, like for instance chlorophyl that turns yellow or heamoglobin that turns red, shift differently. Other colors that look the same as their IR reflecting counterparts are rendered correctly. How on earth would you separate that mess? The DNG file has nothing to do with this at all. Nor has color balance.
How would that be? The whole point is that the Bayer Filter cannot decode IR light, being RGB, so it cannot register IR light The reason it records as purple is that green is the dominant color in the filter, so the impossible light gets registered as the opposite: magenta. Other colors that hold an IR reflectance, like for instance chlorophyl that turns yellow or heamoglobin that turns red, shift differently. Other colors that look the same as their IR reflecting counterparts are rendered correctly. How on earth would you separate that mess? The DNG file has nothing to do with this at all. Nor has color balance.
Last edited: