Infrared Images with Canon 40D

[David]

Hi All - I've been playing with an infrared R72 filter from Hoya fitted to a Canon 50mm 1.4 prime lens. As the sensors in digital cameras are protected by a filter to remove IR light, using an IR filter to record IR images requires considerable exposure times, perhaps as much as 30 - 60 seconds. However, using the Canon 1.4 lens and a high ISO rating reduces times to about less than a second in good light. As described elsewhere on the Web, there are problems with focusing and with "hotspots" on some lens (including the Canon 50mm 1.4). To compound matters, the resulting images show a strong red or magenta bias which causes problems for post-processing. It's a wonder anyone bothers with the technique. Nevertheless, there are some fine and intriguing images to be seen and knowing how to capture them is always worthwhile.

I'm particularly indebted to Steaphany on this site for posts about a year ago on how to go about post-processing captured images. Until reading those posts I was struggling to get anywhere. The following image is the best of my first attempts and for anyone with a Canon DSLR and its associated DPP software here's a workflow.



Because the captured images have a strong red bias, it is very tempting to try and correct the white balance to give something more palatable. I spent a great deal of time on this and basically got nowhere. Having I read Steaphany's posts, I scrapped any idea of altering white balance. Instead, using DPP, switch the Picture Style to monochrome, adjust either in RAW and/or in RGB modes the exposure, dynamic range, contrast, highlights and shadows to give a workable image. Use the histogram to check the exposure. Increase Sharpness and check Noise Reduction, with values of perhaps 3 to 6 being appropriate for the level of noise encountered in these images. When satisfied, process via Convert and Save to JPEG or TIFF formats and open in your favourite editor. Further processing to enhance contrast seems to be necessary and you can use Levels and Curves to that effect. Some more sharpening may be appropriate. Finally, as an option and as I have done in this case, you can use an Orton effect to add a touch of glow to the image.

I'd be happy to hear from anyone else with a Canon DSLR who has tried this aspect of photography, and thanks again to Steaphany.

Cheers

David

[Steaphany]

Hi David,

I'm glad my work and posts helped you out, that's a great image.

Here is another suggestion for you to try:

Shoot two images, one color in natural light another in IR. Make sure the camera is on a good tripod and steady between in exposures.

In post, process the IR into monochrome and then split the channels, yes these will all be identical, but the fun comes with the natural light color image. Split it as well, now recombine as follows:

IR => Red

Green => Green

Blue => Blue

This will give you a false color IR image.

Normally, healthy foliage in natural light leaves the Red channel dark, but the monochrome IR in place of the natural light red makes this now quite bright. When combined with natural light green and blue, the foliage becomes a bright orange.

Another advantage of this method is that you do not have to worry as much about noise or IR focus. Just have the camera focus by natural light and then set your lens to manual. Remember that the green and blue channels will provide the overall sense of clarity. The now obviously out of focus IR will combine with the sharp green and blue yielding a natural looking aura.

To get an idea of the result, here is an experiment I shot several years ago. To make it a real challenge I used my antique Sony Mavica FD88 renowned for it's high noise, low light sensitivity, JPG only capture, and poor IR performance, but I wanted to see what was possible:



By any account, the IR image was under exposed, very dark, incredibly noisy, and pretty much worthless on it's own, but when combined with green and blue from the natural light image, it turned out to be good enough to get a nice effect. There is some remaining noise, but consider the camera, and some ghosting visible around the trees which was due to a breeze.

[peterb533]

Hi David,

IR is something I have in interest in although via a different route. I bough an old Canon 10D and had it converted so that it only takes IR, reduces the exposure to normal levels. I like the effect in your shot, very impressionist. The shot below was taken at Canary Wharf at little while ago.I intend to visit again very shortly.

[David]

Hi Guys - Thanks for the replies. Peter - that's the kind of shot I should like to take. Where did you get the conversion of your 10 D carried out? A company called ACR is mentioned on the Web and by Steaphany.

Steaphany - Thanks for the extra info. I tried the technique with the forest glade scene above but the effects were very subtle. The idea that noise and focus problems can reduced is intriguing.

Meanwhile I've been experimenting in my highly equipped optical laboratory, viz. on my study bookshelf. The following images are, respectively, the original colour image of books with coloured spines, shadows etc, the monochrome image produced via DPP, and the IR image via the R72 filter and processed as described above.







What is fascinating is not only the change in tonal (colour) information, but also the appearance of detail previously hidden in the shadows. For example, on the LHS, Lynn Trusses' "Eats Shoots and Leaves" has disappeared but Michael Swan's title for "Practical English Usage" has appeared! On the RHS Kenneth Clarke's title "The Nude" has disappeared. I can start to see the forensic uses of this technique. My next experiment will be to take some shots of subjects with a lot of visible red in them and see what happens.

Thanks again.

David

[peterb533]

Hi David,

The conversion was indeed done by ACR, very quickly and efficiently too. You would have to check with them on prices as it was a couple of years ago that mine was converted. Also they have some new more sensitive sensors now. If you are intrested in IR and have a suitable body then its well worth it.

Peter

[David]

Hi All - I may be getting the hang of this. The following shot was taken in bright sunlight, processed by selecting a white point (fluffy cloud), desaturating in RAW and RGB modes, enhancing contrast etc. I did some tweaking, like sharpening and cropping.



Overall, I think I shall approach ACS or ACR to convert my redundant 20D. There's a lot of scope here.

Many thanks again for your help.

David

[David]

Hi All - Further questions: I notice that most of my shots in IR have what is described as a hotspot, an extensive area normally around the centre of the image that is lighter than the rest of the image. From investigations on the Web there appear to be three different explanations put forward for such hotspots. First, some say that they arise from conventional camera lens flare; second, there is the view that they come from an internal back flare from the hot mirror; and third, that they occur because of anti IR coatings on the lenses. I would welcome any comment about the origin of these hotspots and, more importantly, on any techniques that can be used to get rid of them. In post-processing, I have tried creating a mask on a duplicate layer and then blending with Multiply. This is partially successful, but I should appreciate any other methods.

My Canon 20D has been sent off for conversion, but the turn round time is 8 weeks, such is the popularity of the technique at the moment.

Cheers

David

[peterb533]

I've heard about the 'hot spots' but so far not experienced them. I was told when my 10D was converted that it had been optimised for 50mm focal length, although not sure how (or why). I have also heard that some lens are particularly prone to hot spot.

Peter

[Dave Humphries]

Hi All - Further questions: I notice that most of my shots in IR have what is described as a hotspot, an extensive area normally around the centre of the image that is lighter than the rest of the image.

Are you saying the IR 'light' received is less nearer the edges of the lens than it is in the centre?
If so, I would tend to think of it more as a 'pronounced vignette' from this bit of the description, rather than say, a flare.

I can't come up with an explanation for it I truly believe in though here's some I discounted:
- Can it be that the lens edges are cooler (or warmer?) because in contact with barrel conducting heat out (or in?)
- Focus/refraction errors due to the longer wavelength of IR is beyond what the glass can cope with near the edges (causing the IR to hit the sides rather than go down the barrel to the sensor)

HTH,

[Steaphany]

Are you saying the IR 'light' received is less nearer the edges of the lens than it is in the centre?

Dave Humphries,

Often, a properly exposed digital IR image will contain a bright region in the center of the image, the situation is often repeatable, and specific to particular lenses.

Dave,

The Wikipedia page on Infrared Photography states:

Some lenses will also show a 'hot spot' in the centre of the image as their coatings are optimised for visible light and not for IR.

Which I tend to agree with.

If the IR Hot Spot was the result of a light source out side the field of view, but still shining light onto the lens, i.e. Lens Flare, then the effect would be an offset flare pretty much identical to the effect in visible light photography. A back reflection from the camera's hot mirror would also exhibit an axial asymmetry if a IR light source within the field of view is not centered on the optical axis. You can check for this by photographing a candle flame.

The characteristic of a IR Hot Spot is a precise optical axis centering and consistent irregardless of the orientation of the camera to the light source.

If you think about the optical geometry of a modern lens along it's optical axis, you will find that the lens consists of a series of flat parallel surfaces. Apart from the dielectric visible light anti reflective coatings, light is free to reflect repeatedly between these, which results in a Hot Spot. The Hot Spot gradient or fall off from the center corresponds to the over all optical element curvature.

The problem is, few lenses today are manufactured and tested for operation outside the visible spectrum. (Hey, no Sigma lens I have has an IR focus Mark, back in the 1970's and earlier just about all lenses did.)

My recommendation is to check each of your lenses for the number of optical elements and try choosing your lens with the least. Zoom lenses are particularly bad and they characteristically have a greater number of optical elements. The fewer number of elements translates to fewer surfaces to cause the Hot Spot. If you have any antique prime lenses that can be fitted to your 40D, you may want to give them a try.

Another alternative would be to shoot when there is less IR light. I have found that I can reduce the prominence of the Hot Spot when I shoot in the early morning or late afternoon when the Sun is not high in the sky and the scene is not flooded in IR. (Remember, the majority of Sunlight is IR )

[David]

Steaphany - Thank you for the explanation. What I had not realised until I read your post was that all the optical components will have been coated. By chance (or sentiment) I still have the pride and joy of my early photographic years - a Pentax Super-Takumar 1.4 50mm lens. I tried it out a few years ago on the Canon 20D having obtained an appropriate coupling ring. Apart from having to manually set exposure and check focus, it worked well. I have now looked it out again and shall experiment as soon as can fit an R72 to the front. Incidentally, did you know that a batch of these lenses was found to be radioactive? The coatings used in those days (1960's) sometimes contained inorganic salts of an element called thorium, which is radioactive. There are still some out there and probably worth a lot of money if their owners haven't died of radiation poisoning. (I'm still feeling fine and have only lost a couple of toes apart from my hair and nails.)

Dave and Peter - we have learned a great deal today.

Cheers

David

[peterb533]

Thanks Steaphany, explains why the efect is lens specific.

[Steaphany]

Don't worry about Thorium, apart from traces of short lived isotopes, Thorium 232 naturally amounts to 100% and has a half life of 14,050,000,000 years through alpha decay. Alpha particles, helium nuclei, can only travel short distances in air and are easily blocked by a sheet of paper.

(Yeah, I know, this shows how much of a techi I am, Yes, atomic physics was a hobby when I was in grade school.)

I'd say your loss of toes, hair, and nails is probably due to something else.

[David]

Hi Steaphany - The following images bear out your explanation. The first is taken with my Canon 50 mm 1.4 lens, the second with the 50 mm Super Takumar. Both were shot at f8 with 15 seconds exposure. Apart from enhancing the general brightness via DPP to make sure the tonal range looks more or less the same no other post-processing has been done.

Clearly, the Canon lens causes a hotspot, whereas the Takumar does not.





As regards being a techi, I used to know a great deal about such things and probably still do if I can be bothered. But I reckon someone who uses a picture of a nuclear explosion for an avatar has to be watched very carefully!

Cheers and thanks again for your help in these matters.

David

[Steaphany]

I reckon someone who uses a picture of a nuclear explosion for an avatar has to be watched very carefully!

LOL

The reason I chose that avatar image is it's an atomic test that was conducted out in Nevada in the year I was born. I also like the aesthetic of the image and it demonstrates our civilization's technological advancement. The Sputnik launch would have also been an appropriate avatar as that too was in 1957, but I have never found any quality photos.

It's good to see you're getting a handle on IR.

[David]

I thought I'd try what appears to be called the "blue sky" effect in IR photographic circles. Here's a couple of examples:



This was taken with very even illumination all around. The blue sky effect was achieved via the technique described at: http://layersmagazine.com/digital-ca...d-imaging.html

This second shot also had very even illumination, but was slightly underexposed according to the histogram. Nevertheless, the graininess seems to add to the effect. The young couple seem not to have moved during the few seconds exposure. I've no idea what they were up to!



David

[peterb533]

David,

Good effect, especially the second, very other worldly.

This is straight out of the camera

[Leighh]

Hi, sorry to revive an old thread but it seemed appropriate as I have a 40D
I bought an ir720 filter and thought id have a go at this myself, my camera is not converted so I am just using the filter and Photoshop, shooting in RAW I import in to Photoshop with Adobe camera raw, use the white balance tool and select an area then in Photoshop use a Channel mixer adjustment layer I Select Red and drag the Red setting to 0% and drag the Blue setting to 100% next I Select Blue and drag the Blue setting to 0% and drag the Red setting to 100% and below is the best I can come up with so far, is it possible to do better with my image?



Its here if any one can show me how

http://leighhogben.co.uk/IMG_9301.CR2

Thanks all