Why are the corners RED?

[JeffreyTotaro]

Advertisement (gone after registration)

Hello:

I took this shot last night with the WATE and with the UV/IR filter on, and lens detection UV/IR on at 18mm. Why are ceiling corners red? Looks like Lens Cast on a Phase back. I processed this one in ACR, but have similar results in C1 Pro.

[Advertisement]

The Leica Q3 is available for pre-order at B&H Photo Video and Adorama

[marknorton]

You can get red corners is there's over-correction for cyan, such as there would be if you told the camera there was a filter on the lens and there wasn't OR if you were shooting with the lens set to a longer focal length than you had specified in the menu.

[JeffreyTotaro]

Thanks Mark,

I'm very sure everything was set properly. 18mm on the lens, and 18mm in the menu.

[sean_reid]

The WATE corrections, while good, are not perfect under all lighting conditions.

 

Cheers,

 

Sean

[robsteve]

Was it a Leica brand filter? With the WATE, the Leica brand seems to work the best.

[Albert4321]

This is a troubling news to me as I have the WATE on order. I thought I am getting the best super wide lens money can get as some of my super wide shoots are critical.

 

Jeffrey, I will be interested to know your findings.

[robsteve]

Advertisement (gone after registration)

Do you have the latest firmware installed? It has been mentioned that the latest firmware has fine tuned the vignette correction to match the Leica UV-IR filters.

 

Robert

[sean_reid]

This is a troubling news to me as I have the WATE on order. I thought I am getting the best super wide lens money can get as some of my super wide shoots are critical.

 

Jeffrey, I will be interested to know your findings.

 

Albert,

 

This shouldn't be news as I've been writing about it for months now.

 

Its virtually impossible for any firmware correction to work perfectly under all lighting conditions. There are just too many variables. Leica can keep tweaking it but, with UW lenses especially, there will be situations where the in-camera correction doesn't quite nail things. That could be true even if the M8 cost $50,000. Correcting cyan drift is shooting at a target that is almost never the same twice.

 

Jeffrey just happened to find conditions where the firmware over-corrects a bit.

 

If you really want the most precise corrections for those files, use the free Cornerfix program and create profiles for every type of subject/lighting condition you normally encounter.

 

If we did super rigorous filter testing (which is not needed, IMO) I'd bet we'd see some very minor variation from example to example of any given filter. But, of course, the firmware can only be tuned to one particular filter per specified lens.

 

Fortunately, the in-camera corrections work just fine under many conditions. The UW lenses, however, push the envelope because they trigger the most cyan drift. Long story short then:

 

1) Leica can keep tweaking but the firmware will never be able to perform perfectly for every condition.

 

2) If the firmware isn't doing the trick for a given shoot, CornerFix makes sense.

 

3) Very often, the in-camera corrections will be just fine.

 

Cheers,

[ho_co]

Jeffrey--

I assume you're using the Milich filter adapter?

 

If not, this could be a reflection through the Leica adapter's viewfinder cutouts. Though as for that, it's unlikely both because the pattern is so regular and because you would likely be using the cutout mask with the Leica filter adapter.

 

Sean's response is probably the best one. If you re-shoot, you might want to try varying the focal length settings in the camera from those on the lens. In other words, maybe a mix-and-match scenario would wipe out or reduce the effect you've got here.

 

Certainly an interesting happenstance. Glad you made us aware of it!

 

--HC

[sandymc]

A few comments:

 

One issue here is that judging from the image, you were under fluorescent lights - fluorescent lights have a line spectrum, not a continuous spectrum, so are a problem to correct in a "one size fits all" form such as in the firmware. The required correction varies with the type of lamp - different lamps have very different spectra. Combined with the firmware correction which isn't too good for the WATE even in natural light, you can get results like this.

 

CornerFix can help here in a number of ways. Firstly, you can shoot a reference image under the fluorescent lights in question, which will give you close to 100% correction. But even if you were using a profile generated under natural light, Cornerfix would probably give better results: (a) for the WATE, the basic correction will be better especially if you generate the profile specifically for your lens/filter/camera combination, so the corners won't be so far out, but ( you can use the chroma correction slider to manually get rid of any remaining red corners.

 

If you need to correct these images, there are two things you can do with CornerFix:

 

1. Go back to the location, and shoot a profile under the same conditions that generated the red corners - CornerFix can use this to correct. This will get you close to 100% right.

 

2. If you don't have access to the location, you can try the following - shoot a reference image using the WATE without the IR filter under natural light, but with lens detection UV/IR on. That will give you a "red cornered" reference image. Generate a profile from that, use it to correct, and use the chroma slider to fine tune....

 

Sandy

[Albert4321]

Thank you all for the detail and helpful response.

 

Actually, connerFix is the tool I have in mind when I saw the OP image. Shooting a profile for each location make sense as I do that with my whiteBal anyway. In time, I will have enough profiles to handle most lighting situation without shooting new profile each time.

 

Lecia community owns Sandy, Sean, JLM etc big time.

[marknorton]

This might of course be an example of a subject where it's best to remove the IR filter since as Sandy points out, there will not be a huge amount of IR around. Interesting also that the camera has some baseline cyan correction for the sensor cover glass even before you select "UV+IR filter" so it would be interesting to see if there was some red even without a filter.

[JeffreyTotaro]

Thanks for all of the responses. Some more info:

Yes, it's the latest firmware

Yes, it's the Leica filter

Yes, it's the Milich adapter

[sean_reid]

A few comments:

 

One issue here is that judging from the image, you were under fluorescent lights - fluorescent lights have a line spectrum, not a continuous spectrum, so are a problem to correct in a "one size fits all" form such as in the firmware. The required correction varies with the type of lamp - different lamps have very different spectra. Combined with the firmware correction which isn't too good for the WATE even in natural light, you can get results like this.

 

CornerFix can help here in a number of ways. Firstly, you can shoot a reference image under the fluorescent lights in question, which will give you close to 100% correction. But even if you were using a profile generated under natural light, Cornerfix would probably give better results: (a) for the WATE, the basic correction will be better especially if you generate the profile specifically for your lens/filter/camera combination, so the corners won't be so far out, but ( you can use the chroma correction slider to manually get rid of any remaining red corners.

 

If you need to correct these images, there are two things you can do with CornerFix:

 

1. Go back to the location, and shoot a profile under the same conditions that generated the red corners - CornerFix can use this to correct. This will get you close to 100% right.

 

2. If you don't have access to the location, you can try the following - shoot a reference image using the WATE without the IR filter under natural light, but with lens detection UV/IR on. That will give you a "red cornered" reference image. Generate a profile from that, use it to correct, and use the chroma slider to fine tune....

 

Sandy

 

 

And this is a *free* program that Sandy has devoted much time to simply because its useful.

 

Cheers,

 

Sean

[sean_reid]

This might of course be an example of a subject where it's best to remove the IR filter since as Sandy points out, there will not be a huge amount of IR around. Interesting also that the camera has some baseline cyan correction for the sensor cover glass even before you select "UV+IR filter" so it would be interesting to see if there was some red even without a filter.

 

Under some circumstances, the color drift in the corners will be off even with no filter and the camera set to detection "on". As Sandy and I have said above, there are limits to what a generalized correction in firmware can do. Mostly it works fine, sometimes it won't.