|
|
|
|
(IR) Infrared Cut-Off filter for digital cameras
This topic is one which comes up frequently and the reason seems logical - digital cameras contain a Silicon sensor (usually) and Si is not only transparent in Infrared but also IR can penetrate optics and reach the sensor and such theoretically can contribute to the formation of the image signal. Usually, one can read that especially the older cameras are subjected to this problem and therefore it is highly advised to use an IR cut-off filter in front of the camera lens to improve the overall image quality. It is usually further stated that due to the IR contribution in the formation of the image, the image will become sharper and the color cleared if the IR cut-off filter is applied. I tried it out - with a rather old digital SLR, a Kodak DCS 460 (6 Megapixel) and a brand new one, again my favorite, the Nikon D2X (12 MPx). What I was looking for was not the resolution difference of course but any significant differences in color rendition, sharpness and other changes due to the usage of such an IR cut-off filter. As B+W is often regarded as the top of the line, I used a B+W UV-IR CUT (486) filter (price about $ 300.00!!) and my 17-35mm ED Nikon lens on both cameras. The FOV of these two cameras is different, so I had to use slightly different focal lengths to get to comparable shots. To get a significant amount of IR, I shot all images on a very sunny day ( which also was rather hot..). At my computer, I processed all images the exactly same way and to the same amount to avoid any alterations due to different procedures. The results were a bit surprising at the first moment. Here they are: 1. Nikon D2X - the specs claim to have already an IR block filter built-in so that an additional IR cut-off filter should not have much visible influence. And it did not as expected. The upper part of the image below was shot withOUT the filter, the lower one WITH the filter - hardly any difference visible : If you look at the very fine details, there is no visible difference regarding sharpness of the images but the lower one, taken with the IR cut-off filter, appears to be a tiny bit softer, a tiny bit less contrasted compared to the upper one. I have no ready-mixed explanation on hand other than the IR filter cuts off also a part of the darkest red part of the visible spectrum, which could contribute to enhancing slightly the overall image contrast. You can see that small difference in the contrast (and perceived sharpness) on the screenshot below which shows the crops of the raw images - the upper one without IR cut-off filter and the lower part with IR cut-off filter applied ( no additional sharpening applied)
The next set of images have been taken with the old Kodak DCS 460 - don't focus on resolution, that's clear, it is a 6MPx camera...but the difference with and without the IR filter is much more visible but still less than one could conclude from reading the camera manual and various publications of manufacturers.
The maybe best location to see that difference is in the clouds - the IR cut-off filter reduces the reddish halo of the extremely bright clouds visibly but the other parts of the images are more or less very similar. Here I could say, that the overall impression of the image taken with the IR filter is slightly more appealing than the one without the IR filter. As a side remark - I am positively surprised how good the overall image quality of that camera is - keeping in mind that it is an over 10 years old design (at the time of market presence, its retail price was over $ 20,000.00!) Is it possible to quantify these minute differences - well I used a program, which becomes more and more popular and which was written by my friend Norman Koren - IMATEST, and took images of the Gretag Macbeth Color Chart with and without the IR filter applied. Here the results of the two evaluations and calculations - the upper chart is withOUT the IR filter and the lower one WITH the IR filter:
The deviations of the resulting colors from the color checker chart are shown in the two charts below - here the output of the calculations withOUT the IR filter:
And here the same WITH the IR filter in front of the camera lens:
To complete the test results, here the final two charts, again the one withOUT the IR filter first:
And there the data for the shots WITH the IR filter applied
I am not finished and have asked another filter manufacturer for feedback and so, as time goes on, I hope to get to a clearer picture about the importance of an IR cut-off filter for digital photography. One thought crossed also my mind - using highly corrected ED lenses means also a lot of glass in front of the sensor - this could reduce the IR sensitivity of the camera significantly as well. And of course the modern multilayer lens coatings might also have an IR-reducing effect. Anyway, the search for truth goes on! Update per August 7,2005: No response at all from Tiffen, none from Hoya and a short, very general response from B+W, telling me that these filters were designed at a time when digital cameras were still sensitive to IR and needed such a filter. Included in their response were also two transmission curves. Let me know, if you have suggestions, advise, feedback etc !
|
|
Send mail to
Webmaster with
questions or comments about this web site.
|
