Infrared Images with Canon 20D Conversion

[David]

My Canon 20D has now now been converted by Advanced Camera Services of Norfolk (Google and ye shall find). The filter is the standard 720 - ca. 1200 nm, equivalent to the R72 from Hoya.

The system appears to work very well with normal exposure settings instead of tens of seconds with a screw on filter. There does seem to be some erratic variability about exposures whereby a setting that was giving a good exposure suddenly under or over exposes (in Av mode).

White balance is the one of the key issues in obtaining reasonable images as the raw images have a reddish cast as expected with the 720 nm filter. I've experimented with this using Canon's DPP, but use a custom white balance setting in camera obtained from photographing the grey matt side of aluminium foil (Apparently IR light is well reflected by metallic surfaces).

There are enormous and unexpected variations in reflectivity of IR light from dye coloured surfaces and I'll post on the results of some studies later. Meanwhile here are three shots of flower heads, post-processed using the white balance, contrast, highlight, and shadow settings in DPP to give interesting images, at least to my eyes.







Comments and crits are welcome as always

Cheers

David

[peterb533]

Hi David,

Welcome to the wonderful world of IR! Good first shots. I've noticed the variable eposure settings too, thought it was my camera playing up but more likely to be the differences in the way IR light is reflected. For white balance I've set mine from a verdant patch of grass, shot out of focus, as this tends to be white in IR. Lets see some more of your IR work. Maybe we could start an IR shots post?

Peter

[David]

Thanks for the encouragement, Peter. I agree that a centralised general post for IR shots would be a good idea, along with references to the other IR posts. If anyone else comments and agrees I suppose we could set that up.

Anyway, the sun came out briefly, so I trotted down to my local park in Bramhall. Here are some attempts. I was using my Canon EF 50 mm 1.4 prime lens and I noticed the return of the hot spot problem whenever I used small apertures; at f22 the shots were unusable, even at f8 I can see a lightening over the central area - very annoying. Thus, these are at largish aperture (5.6 I think).



I've called this "The Girl with the Red Jacket" for the obvious reason.



Careful adjustment of white balance, contrast etc. can give dramatic effects. Although the eye may tell you different, there is no white clipping in this shot.



As for the previous shot, no clipping, but that park bench will have to go!

Cheers

David

[rpcrowe]

Having a camera converted to IR is a great way to shoot because of the near normal exposure times, rather then the eons of time it takes when shooting with a filter in front of the lens. I use a converted Canon D60 as my full-time IR camera. I shoot in AV because I like to use an f/11 or f/16 aperture. However, I tend to shoot using exposure compensation in one stop intervals. Shooting in 3-shot bursts usually gives me at least one very usable exposure.

Some thoughts on full-time infrared.

I have not tried to do an HDR composite with IR and I don't really know why. Has anyone ever tried this?

Additionally, has anyone ever tried a night cityscape using IR?

Finally, although I really like the results from my converted D60, I probably should have converted a 350D which I owned so I could use my EFS lenses. The 350D would also have been a bit lighter in weight.

OR

I have often thought that it would be nice to have a small, but good quality, P&S converted to IR. That way, I could just slip it in a pocket of my shooting vest and always have IR capability without the weight and bulk of the larger camera. I wanted to shoot some IR during my April trip to China but, since I was carrying 30D AND a 40D cameras and the extra weight of the D60 would have been, IMO, just too much.

[David]

Hi Richard - Thank you for the comments. I like the idea of bracketed exposures as routine because there does appear to be some unpredictability about exposures, even a slight change of angle may cause a 2 stop difference, perhaps due to a variation in reflectance from foliage.

I have tried HDR using IR, but not very successfully so far. You can, of course, take 3 or more shots with different exposures and treat them with Photomatix or whatever to produce your image. However, there is a possible problem in you do not know what is a shadow area in the IR world and what is a highlight. Well, highlights are probably easy to guess, e.g. certain types of foliage, but shadows are definitely difficult to determine. I took a self portrait yesterday of me standing against a dark curtain. The image showed the curtain brighter than me (come to think about that, almost everything is brighter than me!!!). Then you have the problem that the camera is translating the IR image into its own image space. Thinking cap required. Regarding nightshots, these are on my list to try.

Cheers

David

[Dave Humphries]

Keep going chaps; I, and probably several others, are following this with interest

David, you've made a good start with the info Peter had already supplied here at CiC, I guess that must have saved a few 'dead ends'.

Thank you both,

[David]

This is definitely my last post (cornet calling) before I disappear to my Scottish retreat.

The IR wavelengths cannot be directly perceived and thus have no colour as far as human perception is concerned. Consequently, there is a strong argument that IR shots should only be seen as monochrome images. I don't fully hold with this, but here are two shots from my expedition to my local park yesterday, rendered in B/W, again processed via Canon's DPP, using contrast, highlight and shadow sliders, along with sharpening.



This first one, of a weeping willow, illustrates the deception in what is seen as the whitish area in the foreground is not grass but pond weed.



This second with its variety of reflectance from different foliage also has some interesting ripples that IMO add to the feeling of depth. I wonder whether they would have been so clear cut in the visible spectrum. I really don't know.

I'm looking forward to photographing the area surrounding my cottage. Be warned! There will be lots more to come.

Cheers

David

[David]

Having just read the threads re HDR and realism I'm in two minds about posting anything that isn't spot-on true to the eye. Anyway, I'm back from Scotland and thought that I would share some images that I took with my recently converted Canon 20D. I've posted several shots at http://www.pbase.com/david_ws/irauchencairn

but highlight a few here showing some of the insights and pitfalls in IR photography.



This first image is a classic shot seen on most IR photography websites and in IR photography books showing the dramatic effects of "white" foliage" against "black" sky, absolutely typical for IR work. It also shows one of the pitfalls - unexpected apparitions in the sky - con trails in this case. No doubt, they were there when I took the shot, but I had not visually noted them. The black sky and white cloud then increase the visual contrast. So, check the sky before taking the shot.



I took scores of shots of plants and trees and flowers, almost all of which show the "white foliage" effect. However, in this second shot the plants have very dark sepals (the bits that enclose the flower heads). The plants are I think hawkweed or hawkbeard, related to the daisy and dandelion group, neither of which show this effect. Thus the second lesson is to watch out for the unexpected; clearly the sepals contain an IR absorbing substance. More investigation is needed.

The next shot is maybe a first - it's an IR photo of a rainbow.



I didn't have much time for composition so the image is what it is. Nevertheless, you can see the bow. The question is what exactly are we looking at? Is this the camera's attempt to show the visible rainbow via an infrared shot or is it diffraction of the infrared wavelengths? As you know, red is on the outside of a primary rainbow, blue on the inside. The normal angle from the horizon is about 42 degrees. If this were an IR bow, the angle would be about 45 degrees - too close to call by eye. Note that there are two lighter bands on the inside of the bow - one should be easily visible in this rendered shot, but on my screen two can be seen. What are those? All very puzzling. Thus, lesson three is that you need your thinking cap on with IR work.

Update 240810 - The bow seen above is very probably due to IR radiation. - There are several photographic examples on the web (Google: IR rainbow). What is interesting is that those examples are very like the one above. The two inner bands are known as "supernumenaries", sometimes also seen with visible bows



The fourth image below was taken while messing about with a telephoto lens (Sigma APO 70-300mm). Again, I'm not so worried here about composition, but simply with testing the lens.



What is remarkable is the clarity over distance. The mountains in the background show detail and they are in the Lake District about 50 kilometers from where I was standing. I wasn't even aware that they were visible as the object of interest was the island in the bay. I have photographed over the Solway Firth many times and have never seen the peaks in the Lake District so clearly. So lesson four is that it is true that IR imaging can cut out visible haze and then some.

The fifth image illustrates another pitfall or positive possibility with IR photography: white balance.



IR texts usually refer to using grass as a white balance reference. This can be used because grass is normally "white" in IR photography. However, having now shot scores of images, I think that using a white balance based on a suitable point in the image itself is a more flexible and artistically useful method. Try clicking on various features until a tone emerges that augments the image. Alternatively, use the software, in this case Canon's DPP, to adjust white balance. Try not to go to monochrome settings if you want to preserve or enhance richness of tones.

This sixth shot provides another example of where IR shots may score over conventional ones.



The shot was taken at a small lochen near Mossdale on a grey day with a fine smir and little contrast. To help matters there was a compliment of vicious midges intent on doing me harm. All in all not a promising venue. However, IR seems to bring out the best in water reflections and the pale luminosity of the lily leaves and grass stem might be hard to achieve by other means. Thus, the lesson is that unpromising light may be more attractive in IR.

The final image I wish to post again illustrates the curious reflective power of water.



Some authors claim that water absorbs IR light and hence is black in an image. This is not so. At the wavelengths of near IR (720 - ca. 1200 nm) water is totally transparent to IR. However, there may be something going on at the very surface layer that may cause some optical effects. Anyway, the images can be quite startling. (BTW, I cheated - the ripples - I threw a stone in!!!)

Comments and crits are welcome as always.

Cheers

David

[Dave Humphries]

Fascinating David, thanks for sharing with us and taking the time to explain what we're looking at.

I'll be coming back to this later when on my own monitor for a more careful look.

Initial observations are;
#3 might benefit from some perspective correction, it looks like you set the pole vertical, but the church/chapel and house on right are, for me, disturbing/distracting from the bow.
#4 I find the lack of detail in the water doesn't help, but that could be my current less than ideal viewing conditions (at work). The island seems to 'float' somewhere above where my brain tells me it should be, it looks to be closer than far shore and yet is alomst half way up the dark hole that is the sea.

Are the overall 'toning' colours of #4 and #5 something applied in PP?

Cheers,

[David]

Lenses and Hot Spots

I'll reply to comments in due course, but I thought I'd add some further information about my experiences of using the converted Canon 20D. In other recent threads there has been discussion about hot spots arising almost certainly from internal reflections in lenses that have "anti-IR" coatings. I thought I would test the lenses in my kit on the Canon 20D to see if the conversion had had any effect on hot spots.

I set up a view with a wide dynamic range, at least in the visible, with the camera on a tripod and the sun at right angles to the lens direction. I shot at each f number from maximum to minimum. I found the following:

Canon EFS 18-55mm - no hot spot from f11 and down to lower f numbers, over the whole range of 18-55 mm. Also f22 at 55 mm did not produce a hot spot. Thus, good overall coverage.

Canon EF 50 mm 1.4 prime lens - only good at f5.6 and below. Otherwise bad hot spots. Also appears more IR light sensitive than the EFS lens by about 2 stops.

Canon EF 24-105 zoom - good at f8 and below over the range 24 - 55mm, possibly good at f5.6 and lower at 105 mm. There may be a strong directional effect with this lens with light from behind causing less of a problem. No use on high f numbers.

Canon EF 75-300mm zoom - good at f8 and below at 75mm, possibly OK at f8 and below on 300mm. Slightly disappointing results.

Sigma 70-300 mm APO - good at f5.6 and f4 at 70 mm and good at f11 and below at 300mm.

Overall - EFS 18-55mm good for general purpose shots; Sigma 70-300mm telephoto/zoom good at f11 and 300mm; EF 24-105 good at f8 or below.

David


030910 - I have acquired some Pentax Super Takumar lenses and have carried out a similar test on those to establish whether or not hot spots occur. This is in response to previous observations, occasioned by comments from Stephany, that older lenses without so many elements and coatings should not display hot spots.

The lenses are: Super Takumar 1:3.5 28mm f22-f3.5; 1:3.5 135mm f22-f3.5; and 1:1.8 55mm f11-f1.8. In short the results are that none of these three lenses from their highest to lowest f numbers showed any hotspot. This is clearly good news for taking IR photographs with greater depth of field associated with larger f numbers.

Here are two examples from the test files.






The first shot was taken with the 28mm lens and the second with the 135 mm lens, both at f22.

David

[David]

Exposure Monitoring, Neutral Density Filters

Keen to explore the possibilities of IR photography, I thought I would try some experiments with long exposure times. I used a ND4 filter (Conkin) to bring the Tv to about 1/5 second. However, no matter what I did, my shots came out overexposed by about 2 stops. I'm ashamed to say that it took me the better part of an hour to figure what was going wrong. Neutral density filters are completely transparent to near IR light. However, the conversion of the Canon 20D to IR does not, apparently, touch the metering system. Thus, put on a neutral density filter and the camera's metering of visible light tells the camera to change the Tv (or whatever). However, IR light is not affected.

The question then arises as to whether there are filters that might act as neutral density filters in the IR range. I found that a green PO1 filter (acquired for a couple of quid from Dixons a few years ago in a sale "there's no call for these any more since the digital age") fits the bill. It reduces IR light by about 6 - 7 stops! When time permits I'm going to experiment with waterfalls and the like to see what I can produce.

Cheers

David

[Dave Humphries]

This is going to be very useful information David.

So much so, that I have added a link to this thread here:
Links to Useful & Informative Threads

However, I know others have written on the topic and their threads should also be linked from there - Colin, Rob, if you spot any, please add to the post, thanks.

(I started a new IR section under the HDR one)

Cheers,

[David]

Dave - Thanks for your comments and the addition to "Informative Threads".

Regarding your observations on my images, I've not really done much PP except for using Canon's DPP to enhance contrast etc and sharpening.
Re #3 I have not put any perspective correction in, as it's not meant to be a composed photograph in our usual sense. It's a snap taken quickly to get the rainbow (see update in the main post re IR bows).
Re #4 - Yes the island floats but again this is an experimental shot - simply seeing what happens with this particular lens. What you see through the view finder and what you get are normally quite different! When I looked at the shot on screen I was startled to find how clear the distant mountains were. I had read about this effect but this is an example of it.

Re #4, 5, 6 - These toning effects are all a result of changing the white balance via DPP. The subjects of white balance and channel mixing are complex at least for me to get my head round. With visible photography the ideas are well discussed but with IR you have the problem of defining what a white balance is. In my view it's easier not to have a preconceived or rigorous definition and just test out the possibilities - it' easy enough to do.

Cheers

David

[Dave Humphries]

Hi David,

Thanks for the feedback, I mis-understood the purpose when I saw the standard phrase "Comments and crits are welcome as always"

I have now added a link to Peter's Canary Wharf IR thread, as that also contains a lot of useful information.

Cheers,

[Shadowman]

Having a camera converted to IR is a great way to shoot because of the near normal exposure times, rather then the eons of time it takes when shooting with a filter in front of the lens. I use a converted Canon D60 as my full-time IR camera. I shoot in AV because I like to use an f/11 or f/16 aperture. However, I tend to shoot using exposure compensation in one stop intervals. Shooting in 3-shot bursts usually gives me at least one very usable exposure.

Some thoughts on full-time infrared.

I have not tried to do an HDR composite with IR and I don't really know why. Has anyone ever tried this?

Additionally, has anyone ever tried a night cityscape using IR?

Finally, although I really like the results from my converted D60, I probably should have converted a 350D which I owned so I could use my EFS lenses. The 350D would also have been a bit lighter in weight.

OR

I have often thought that it would be nice to have a small, but good quality, P&S converted to IR. That way, I could just slip it in a pocket of my shooting vest and always have IR capability without the weight and bulk of the larger camera. I wanted to shoot some IR during my April trip to China but, since I was carrying 30D AND a 40D cameras and the extra weight of the D60 would have been, IMO, just too much.

I understand that once you convert the camera, that's it, you can''t convert back.

[David]

Hi John - As far as I know you could convert a Canon 20D back by re-replacing the filter in front of the sensor. However, Steaphany describes her system with Sigma SD 14(?) in this post:

IR converted camera vs IR filter

That sounds simpler to do, but requires a different camera altogether.

Cheers

David

[David]

Here are a couple more shots from my recent Scottish sojourn.



This first one kind of blurs the line between IR and visible photography.

The second below makes use of the absorption of IR light by the sepals of the hawkweed. In addition, the normally grey rocks become "white" because of the seaweed.



Finally, a high contrast image, taken to see if IR could pick up the faint outline of a path across the field. Allegedly, this is one of IR photography's capabilities. It did, but not really to any great extent.



Ciao

David

[bleys]

I freakin' love these IR shots. The world revealed by IR is so alien and just plain intriguing.

Keep them coming!

[David]

I've updated this thread regarding lenses and hot spots as of 3rd September. Basically, I've tested three Pentax Super Takumar prime lenses, 3.5 28mm, 3.5 135mm, and 1.8 55mm, over the range of their f numbers and found no hot spotting.

David

[rpcrowe]

In my film days, there were dual focusing marks on the focus scale of most manual focus lenses. The first mark was for full spectrum photography and the second was for infrared focusing. The standard procedure when shooting infrared film using a filter, was to focus normally and note where the lens was focused for normal light and then refocus the lens so that the infrared focus mark was on the original focus mark. Let me rephrase this... Say you were focused at 10 feet for normal photography, you would simply move the 10 feet indicator on the focus scale over to the infrared focusing mark.

None of my EOS lenses has an additional IR mark on the focusing scale. The technician who converted my Canon D60 (not 60D) into an IR version stated that the best way to get around this focus problem was to shoot aperture priority at f/11 and let the DOF take care of the focusing anomally between full spectrum light and IR light.

It seems to work O.K. Does anyone else have any other solutions to the difference in focus point between IR and full spectrum light. I wonder if a person could adjust the focus using micro focusing (on the 50D and 7D as well as some 1D cameras). All of these images were shot at f/11.