Purple shirts & IR contamination

R

rvaubel

Well-known
Local time
3:21 AM
Joined
Feb 11, 2006
Messages
789
Fellow RD1ers

Just to confirm the fact that our precious RD1 suffers from the same kind of IR contamination as the M8, I submit these samples for you peruasal. The lighting in this particular sample, highlights the probably, as it is normally a lot harder to spot.

Notice that the Heliopan filter is totally effective at mitigating the problem. These samples have not been post processed at all.

inquisiting minds want to know

Rex

:eek:
 

Attachments

  • RD1_HeliopanTest_Filter.jpg
    RD1_HeliopanTest_Filter.jpg
    133.9 KB · Views: 0
  • RD1_HeliopanTest_NoFilter.jpg
    RD1_HeliopanTest_NoFilter.jpg
    152.4 KB · Views: 0
In stead of my buying filters to eliminate this effect people should be wearing natural fabrics, dyed in natural dyes. :)
 
Corrected in Photoshop <1 min vs buying IR filter and then still having to correct cyan corners. IMO the black shirt is a lot more natural than in the filter shot.
 
Last edited:
we should put all these threads together. The bottom line is Rex uses weird lighting ;)

In "normal" lighting situations, the effect is much less dramatic, see my blog, also Rex should consider using the original firmware and posting his wb settings.

Pro photographers have dealt with this same issue with film, and then with early digital cameras. It is nothing that can't be fixed with lighting, picasa, printing, without resorting to image tone sucking filters.
 
barjohn said:
Tell us what you did to correct it. Thanks.
Click to expand...

The fastest thing to do is to remove all the Magenta from the Black color using "Select Color" in Photoshop.

All the steps have been mentioned a number of times; using the Search function you may be able to find them.
 
Gabriel M.A. said:
The fastest thing to do is to remove all the Magenta from the Black color using "Select Color" in Photoshop.
Click to expand...

Not in this case because he's holding a "magenta" filter. Select Color selects it along with anything else in the shot that's really supposed to be magenta. What I do is use the "Magnetic Lasso" tool to outline the IR-magenta areas and desaturate them. In this case I was able to do the entire shirt and go around his hand and the filter in one loop. I had to go back a second time for the little bit of shirt that shows below the filter as he's holding it. I also find that it's better to desaturate the IR-magenta and adjust the contrast to get the right shade of black, because messing in the color adjustment it takes too long to get a neutral black.
 
Anyone have the sensativity curves for the R, G, and B pixels that make up the sensor? I'm figuring that the red pixels are the ones that are contributing to magenta by extending sensitivity past the red and into the IR.

Mark
 
anselwannab said:
Anyone have the sensativity curves for the R, G, and B pixels that make up the sensor? I'm figuring that the red pixels are the ones that are contributing to magenta by extending sensitivity past the red and into the IR.

Mark
Click to expand...
The sensor itself sees in monochrome. There is an RGB Bayer filter over the sensor and I'm not sure we can make any assumption about how the Bayer filter handles infrared, since its design parameters likely included the assumption that the light reaching it would already have passed through a robust IR filter.
 
Ben Z said:
Not in this case because he's holding a "magenta" filter. Select Color selects it along with anything else in the shot that's really supposed to be magenta. What I do is use the "Magnetic Lasso" tool to outline the IR-magenta areas and desaturate them. In this case I was able to do the entire shirt and go around his hand and the filter in one loop. I had to go back a second time for the little bit of shirt that shows below the filter as he's holding it. I also find that it's better to desaturate the IR-magenta and adjust the contrast to get the right shade of black, because messing in the color adjustment it takes too long to get a neutral black.
Click to expand...

All right Hotshots, try this one. Notice that Laya is holding the Heliopan IR cut filter, so there is obviously no filter on the lens. What color do you think the shirt really is? I have a picture taken with the filter, but I won't reveal it until you guys take a swack at it.

Rex :p

BTW, the light is not that weird. The main light is normal incandesent can lights. The undercounter are floresencnts.
 

Attachments

  • Laya_purpleshirt- copy.jpg
    Laya_purpleshirt- copy.jpg
    234.4 KB · Views: 0
rvaubel said:
All right Hotshots, try this one. Notice that Laya is holding the Heliopan IR cut filter, so there is obviously no filter on the lens. What color do you think the shirt really is? I have a picture taken with the filter, but I won't reveal it until you guys take a swack at it.

Rex :p

BTW, the light is not that weird. The main light is normal incandesent can lights. The undercounter are floresencnts.
Click to expand...
Alrighty, then; I'm "game".

This is an experiment, and I say that, because I've never handled an RD-1 file before, so you let me know if this close to "reality", whatever Descartes may have thought that may have been if he wanted to be buddies with Sartre and not get into arguments:
 

Attachments

  • Laya_purpleshirt-GMA.jpg
    Laya_purpleshirt-GMA.jpg
    326.8 KB · Views: 0
Gabriel M.A. said:
Alrighty, then; I'm "game".

This is an experiment, and I say that, because I've never handled an RD-1 file before, so you let me know if this close to "reality", whatever Descartes may have thought that may have been if he wanted to be buddies with Sartre and not get into arguments:
Click to expand...

Well, good try. However the actual stripes were magenta! They almost matched Laya's boyfriends "magenta AKA black" shirt. But I can't prove it because no digital camera I had handy could reproduce what the eye would see.

Rex
 
Actually there are plenty, maybe even the majority, of times when an IR filter would be the only practical solution. Last night I was in a furniture store and saw some chairs upholstered in a floral pattern, gold foliage on a black background. The resulting photo with gold foliage on a magenta background would have been impossible to correct with Select Color, and would have taken hours using my method of Magnetic Lasso + Desaturate. Where the RD-1 has it over the M8 is that it is less IR sensitive so that it isn't necessary to always have an IR filter on every lens.
 
Ben Z said:
Actually there are plenty, maybe even the majority, of times when an IR filter would be the only practical solution. Last night I was in a furniture store and saw some chairs upholstered in a floral pattern, gold foliage on a black background. The resulting photo with gold foliage on a magenta background would have been impossible to correct with Select Color, and would have taken hours using my method of Magnetic Lasso + Desaturate. Where the RD-1 has it over the M8 is that it is less IR sensitive so that it isn't necessary to always have an IR filter on every lens.
Click to expand...

Yes, the RD1 exanple I posted above is an extreme example. The RD1 normally doesn't show that much IR contamination, even from photos taking in the same room. The IR filter that I do have (77mm shown in pictures) doesn't even fit any of my "m" lenses. I purchased that filter for my Canon 20Da (the astrophotography model) which as the IR filter modified. That is a whole other story. But in brief, the 20Da has IR contanination that is of a total different character than the RD1 or the M8.

Rex
 
Ben Z said:
Corrected in Photoshop <1 min vs buying IR filter and then still having to correct cyan corners. IMO the black shirt is a lot more natural than in the filter shot.
Click to expand...

Ben, I disagree. The black shirt in your example is just a shirt with desaturated colors. It looks very fake to me. The filtered shot shows a black shirt with color cast that you would expect in the room the picture was taken in.
 
Last edited:
I agree, but I was only trying to correct the magenta as that was the subject of the thread. Had the shot been mine I would have either corrected the color balance of the entire shot first, or else adjusted the shirt with color balance controls rather than desaturating, so that it had the same yellowish tinge as the rest of the shot if that was what I was after.
 
The thing for me is I bought my R-D1 for 1299. At that price a little IR contamination is no big deal. The M8 is three times that price. I expect perfection. Nice kitchen man. I think there is some banding on her sweater!
 
you know those color checker charts? Has anyone tried this on those? Don't they show the IR sensitivity?

And what to pro's do who HAVE to get accurate color? Like they're photographing paintings for a catalog?
 
pfogle said:
And what to pro's do who HAVE to get accurate color? Like they're photographing paintings for a catalog?
Click to expand...

X-rite color densitometers, gretag macbeth color checkers, calibrated imacon scanners, calibrated monitors, and custom ICC profiles for every paper they run through their printers.

The pros keep everything in control every step of the way through.
 
You must log in or register to reply here.
Back
Top Bottom