Jump to content

M8 Pics with IR cutoff filter


gseitz

Recommended Posts

Advertisement (gone after registration)

Hi,

 

Just got my M8 today and happen to have a tiffen standard hot mirror filter laying around. Enclosed are some pictures of a couple of household items I had that show the IR problem. The first picture is from the Canon 5D, the second is the M8 without filter, and the third is with the hot mirror filter attached. You can see significant improvement in the blacks although the purple is not completely gone (notice the lining of the upper boot)...

 

Thanks,

 

Greg

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

Link to post
Share on other sites

Hi,

Just got my M8 today and happen to have a tiffen standard hot mirror filter laying around. Enclosed are some pictures of a couple of household items I had that show the IR problem. The first picture is from the Canon 5D, the second is the M8 without filter, and the third is with the hot mirror filter attached. You can see significant improvement in the blacks although the purple is not completely gone (notice the lining of the upper boot)...

 

Thank you for posting these. I would say they are a "smoking gun" in terms of confirming that excess IR sensitivity (as opposed to white balance or whatever) is the cause of the magenta-blacks issue.

 

Your Tiffen filter seems to have done a pretty good job of reducing the magenta cast, and presumably a filter optimized specifically to the M8's sensor would do an even better job.

 

So, not a perfect solution, not necessarily a convenient solution, but at least it is a solution that can be applied immediately by people who encounter this type of problem.

Link to post
Share on other sites

THanks for these...

 

From eyeballing the wood on the deck which would you say is the most accurate rendition? The IR filter seems to change the color of the wood a lot. This could be a huge problem in trying to balance the non IR targeted colors by using these filters.

Link to post
Share on other sites

Greg,

 

Welcome to the Forum!

 

The filter definately improves things. I can still see a little purple in the "Tiffen" case and (as you pointed out) the boot lining, but that's a world better than what you started out with.

 

Thanks for sharing these. They're encouraging.

 

Allan

Link to post
Share on other sites

It's an improvement, but as I toggle back and forth between the first and third images I can still see a fairly significant difference. The blacks in the dart board in the 5D image are vastly superior to the purplish blacks from the filtered M8.

 

This makes me feel like crying. Even the "fix" doesn't get it there. Thanks for doing the test, even though it dashed my hopes.

Link to post
Share on other sites

Advertisement (gone after registration)

THanks for these...

 

From eyeballing the wood on the deck which would you say is the most accurate rendition? The IR filter seems to change the color of the wood a lot. This could be a huge problem in trying to balance the non IR targeted colors by using these filters.

 

The 5D image is pretty much spot on in terms of the actual color. I'm guessing it would be quite a bit easier to correct the yellowish/green vs the pretty much hopeless purple cast since it's pretty uniform across the image.

Link to post
Share on other sites

Greg,

 

Welcome to the Forum!

 

The filter definately improves things. I can still see a little purple in the "Tiffen" case and (as you pointed out) the boot lining, but that's a world better than what you started out with.

 

Thanks for sharing these. They're encouraging.

 

Allan

 

Thanks for the welcome. Lots of good info on this forum.

Link to post
Share on other sites

It's an improvement, but as I toggle back and forth between the first and third images I can still see a fairly significant difference. The blacks in the dart board in the 5D image are vastly superior to the purplish blacks from the filtered M8.

 

This makes me feel like crying. Even the "fix" doesn't get it there. Thanks for doing the test, even though it dashed my hopes.

 

Yeah, I tend to agree, but at least it gives a glimer of hope and certainly confirms the IR role in the problem (not that it needed confirmation with the pretty overwhelming evidence already posted).

 

Greg

Link to post
Share on other sites

I did some approach with PS2 to remove purple casting from the picture, however, it will change the whole picture's color attribute.

 

Actually, I just got an idea, tried it, and it seemed to work fairly well. Looking at the three samples, it appears that the problem with the "purple" one is an excess of red in the dark areas only.

 

Being a professional Photoshop user, I thought this sounded like a job for the curves palette. So, I downloaded a copy of the image, opened it in Photoshop CS2, opened the curves palette, and used the dropper to display where the dark areas of the dartboard fell on the red, green and blue curves.

 

Sure enough, where the green and blue channels read 30, the red channel (which also should have read 30 to give a neutral black) read almost 70!

 

So, I "pinned" the middle and upper regions of the red curve so they wouldn't move (preserving the balance of the lighter tones of the image) and lowered the toe of the curve until I got a reasonably neutral black. The result still wasn't as good as the 5D image, but it was a big improvement over the original -- and unlike the hot-mirror solution, it didn't change the tones of the wood deck. (And if I had had an original raw file to work on, rather than an sRGB JPEG, I might have been able to get closer.)

 

One good thing about this approach is that once you'd worked out the curve that gives you pretty good neutralization of blacks with your particular camera, you probably could save it and then use an action to run it on the images that needed it.

 

I'm not saying this lets Leica off the hook in terms of coming up with a real solution. But it does suggest that a software approach (maybe implemented as a firmware update) could provide at least a decent "Band-Aid" until a real solution is forthcoming, or for people who only encounter this problem rarely.

 

And in the meantime, if you're wondering how to salvage that great shot you took yesterday of the Supreme Court justices having a snowball fight in their robes, give the curves palette a try and see if it doesn't help.

Link to post
Share on other sites

Greg, did you shoot raw or jpeg? also what was the iso, TIA

 

The ones posted were shot raw at iso 160 and converted in Capture One, but I also shot all the way up to 2500 with pretty much the same outcome.

Link to post
Share on other sites

Actually, I just got an idea, tried it, and it seemed to work fairly well. Looking at the three samples, it appears that the problem with the "purple" one is an excess of red in the dark areas only.

 

Being a professional Photoshop user, I thought this sounded like a job for the curves palette. So, I downloaded a copy of the image, opened it in Photoshop CS2, opened the curves palette, and used the dropper to display where the dark areas of the dartboard fell on the red, green and blue curves.

 

Sure enough, where the green and blue channels read 30, the red channel (which also should have read 30 to give a neutral black) read almost 70!

 

So, I "pinned" the middle and upper regions of the red curve so they wouldn't move (preserving the balance of the lighter tones of the image) and lowered the toe of the curve until I got a reasonably neutral black. The result still wasn't as good as the 5D image, but it was a big improvement over the original -- and unlike the hot-mirror solution, it didn't change the tones of the wood deck. (And if I had had an original raw file to work on, rather than an sRGB JPEG, I might have been able to get closer.)

 

One good thing about this approach is that once you'd worked out the curve that gives you pretty good neutralization of blacks with your particular camera, you probably could save it and then use an action to run it on the images that needed it.

 

I'm not saying this lets Leica off the hook in terms of coming up with a real solution. But it does suggest that a software approach (maybe implemented as a firmware update) could provide at least a decent "Band-Aid" until a real solution is forthcoming, or for people who only encounter this problem rarely.

 

And in the meantime, if you're wondering how to salvage that great shot you took yesterday of the Supreme Court justices having a snowball fight in their robes, give the curves palette a try and see if it doesn't help.

 

AJ

 

I'm happy to see this software approach has the potential to work as well as it did for you. I have a special astro model of the Canon 20D with the infra-red filter removed. For regular use a could have elected to mount a infrared filter on the lens but instead I created a special color profile in CS2's Camera RAW. It worked perfectly well, in fact so well I never us the filter. I hope a similar approach could work for the M8. Your success with lowly JPEG files would indicate it could.

 

Could someone post a file of a GretaMacbeth color chart with a piece of black velvet that turns purple, in the frame? It needs to be a DNG file to build a profile. My email will accept 10MP attachments. Daylight would be the best light source.

 

Rex

Sean, are you listening?

Link to post
Share on other sites

AJ

 

Could someone post a file of a GretaMacbeth color chart with a piece of black velvet in the frame? It needs to be a DNG file to build a profile. My email will accept 10MP attachments. Daylight would be the best light source.

It has to be synthetic black material doesn't it? I don't think all blacks behave this way?

 

Asher

 

http://www.openphotographyforums.com

Link to post
Share on other sites

Are there different "strengths" of IR cut off filters?

 

While an improvement is shown in the 3rd shot, it's just not enough.

 

To put it in perspective ... I'd be dealing with 300 to 400 wedding shots a week ... 80% of which would involve various degrees of synthetic black materials juxtaposed with a woman wearing a white dress. No color shift can happen in one area without affecting the other.

 

Formal wear now comes in a wide variety of materials from wool blends to totally synthetic materials. In addition, the Bridesmaids dresses can range in color and material that has to be accurately rendered also.

Link to post
Share on other sites

As to software fixes, I found:

 

1. Select offending tinted black color with eyedropper

2. Select "color range" from PS "Select" menu - fuzziness varying, but 20-40 somewhere - this selects that particular dark magenta everywhere in the image.

3. Hue/Saturation/Lightness - desaturate overall about halfway, with additional half-desaturation for magenta, blue, red and green individually (to avoid taking neutrals too far to the green side), plus a slight lowering of lightness because the IR is raising the shadow exposure artificially.

 

But as fotografz says - not something one can do with 300-400 shots per week.

 

Wiht Greg's permission I'll attempt it on his unfiltered M8 shot and post it here.

Link to post
Share on other sites

Archived

This topic is now archived and is closed to further replies.

  • Recently Browsing   0 members

    • No registered users viewing this page.
×
×
  • Create New...