DPanswers

essays | home

IR and Digital Cameras

by Gisle Hannemyr

This page is part of a series of articles about digital infrared photography. The complete set of segments in this series is:

Table of Contents

  1. Introduction
  2. Finding a Suitable Camera
  3. Table
  4. Exposure Value
  5. Camera Conversion Linkfarm

1. Introduction

All sensors used in digital cameras are (to some extent) sensitive to infrared light. This is simple to verify. Standard remote controls uses infrared light to signal the device. Take a TV or DVD remote control, point the remote at your digital camera and push one of the buttons. No visible light can be seen, but if you look at the preview or review of the remote on your camera's LCD screen, you should see the infrared light emitted by the device as a bright spot.

IR-light may cause image defects such as flare, false colours, hot spots and chromatic aberrations. It is not a good thing (in an ordinary photograh) to have the image affected by IR-light. As a measure against these defects, an IR-blocking filter in usually placed in the front of the camera's sensor.

Depending on the characteristics of the digital sensor, and the efficency of the IR-blocking filter, the sensitivity of different digital cameras to infrared light vary a great deal. The simple test with a remote described above will not reveal to what extent your camera is able to register infrared light.

A better test of your camera's infrared capabilities would be to borrow or buy a cheap IR-pass filter such as Hoya R72 or Wratten 89B and go out on a bright sunny day to photograph scenes with a lot of green foliage. Or, if you don't have access to a real IR-pass filter, do the test using an unexposed (black) frame of slide film taped in front of the lens. (Slide film has about the same transparancy to near-IR light as the Hoya R72, but it is probably not as good, optically speaking.) A look at the resulting images and the histogram should tell you a lot of how capable the camera is in the near infrared part of the spectrum.

Some time ago, I started to collect data about how sensitive the different cameras were to IR-light. My idea was to find a model that was sensitive enough to give good quality infrared images when used handheld. (I ended up buying an Olympus C2020Z, which has served me well.)

Around 2004, IR-blocking filters used in digital camears started to become much more effective. As a general rule, cameras released after 2004 need to have their IR-blocking filter removed to if you want to use them for infrared photography.

As a result, I've stopped collecting this data. I'll keep the table online for some time, as an assistance to those looking for an old, cheap camera that can be used unmodified for infrared photography.

2. Finding a Suitable Camera

In the table below, you will find a list of camera models and a column labled Avg. IS. This column gives the average computed IS (Infrared Sensitivity). This is number that is computed just like Exposure Value (EV), but with a filter in front of the sensor that cut off visible light so that only infrared light pass.

The average IS indicates how sensitive the camera is to infrared light. The higher the average IS, the more sensitive the sensor is to infrared light. For comparison, a camera's sensitivity to a scene lit by bright sunlight corresponds to a EV around 15. To find out how less sensitive the camera is to infrared light, you subtract the Avg. IS in the column below from EV 15, and the result is the difference in number of stops.

Example: The Epson 850Z infrared light sensitivity with the Hoya R72 filter is IS 9, so it is about 6 stops (15-9) less sensitive to infrared light than to visible light with this particular filter.

IS 8 corresponds to f/2 at 1/60th of a second at ISO 100. To use a digital camera for handheld infrared photography, I think it should have an IS of 8 or more if its maximum aperture is f/2.0, and 9 or more if its maximum aperture is f/2.8 (YMMV).

A basic problem with digital IR photography is that thermal noise increases and “hot” pixels appear due to the long exposure times involved. This means that EV is not the only thing that determines suitability of a particular sensor. If a less sensitive sensor has better noise characteristics, then it may be give an overall better IR-performance than one with a higher EV number, but worse noise characteristics.

3. Table

The table below indicates the relative IR-sensitivity of a number of popular digital cameras, with the average EV for each model indicated (some models that share the same sensor is lumped together.

Sony 1/2" 2.11 MPx
Camera Avg. IS Max f/ IS      T      f/ ISO Fltr. Link
Epson 850Z 9.0 2.0 8.9 1/60 2.8 100 R72 ddb #1
9.2 1/75 2.8 100 R72 ddb #2
Oly C20x0Z 8.1 2.0 8.6 1/100 2,0 100 R72 gd #1
7.9 1/30 2,8 100 R72 gd #2
7.9 1/60 2,0 100 RG715 jr #1
7.9 1/60 2.0 100 R72 gh #1
Nikon Coolpix 800 7.9 3.5 7.9 1/19 3.5 100 R72 rh #1
 
Sony 1/1.8" 3.34 MPx
Camera Avg. IS Max f/ IS     T     f/ ISO Fltr. Link
Canon G1 6.0 2.0 6.0 1.6 7.1 50 R72 cx #1
Minolta DiMAGE 5 6.0 2.8 6.0 1/8 2.8 100 R72 fm #1
 
Sony 1/1.8" 4.0 MPx
Camera Avg. IS Max f/ IS     T     f/ ISO Fltr. Link
Canon G3 5.3 2.0 5.3 1/5 2.0 50 ? ml #1
Canon A80 5.0 2.8 5.0 2 5.6 50 R72 cac #1
 
Sony 2/3" 5.24 MPx
Camera Avg. IS Max f/ IS     T     f/ ISO Fltr. Link
Minolta DiMAGE 7 7.3 2.8 7.5 1/30 3.5 200 ? rt #2
7.0 1/16 2.8 100 R72 dt #1
7.5 1/45 2.8 200 RG715 jr #2
 
Other compacts
Camera Avg. IS Max f/ IS     T     f/ ISO Fltr. Link
Panasonic DMC-TZ5 3.6 3.3 3.6 1/4 4.9 800 W89B dc #1
Canon G5 3.2 2.0 3.0 8.0 5.6 50 R72 ds #2
3.4 1 3.2 100 R72 gh #2
Panasonic DMC-FZ7 3.0 2.8 2.4 6 5.0 90 B+W092 grr #1
3.6 4 6.3 80 B+W092 grr #2
Oly 5050Z 2.4 1.8 1.0 3.2 2.0 64 ? kt #1
1.5 3 2.3 64 ? kt #2
2.6 1 2.0 64 ? kt #3
3.4 1 2.6 64 ? kt #4
3.3 1/2 1.8 64 R72 hc #1
Casio QV-4000 -0.6 2.0 -0.6 6 2.0 100 ? ca #1
 
Rangefinders
Camera Avg. IS Max f/ IS     T     f/ ISO Fltr. Link
Leica M8 7.8 - 7.8 1/45 8.0 1250 W89B mr #1
 
DSLRs
Camera Avg. IS Max f/ IS     T     f/ ISO Fltr. Link
Pentax K100D 7.9 - 7.9 1/30 5.6 400 R72 ma #1
Nikon D100 7.6 - 7.6 1/6 8.0 200 IR720 ss #1
Nikon D70 /
Nikon D50 /
Nikon D40
6.6 - 5.3 1/3 8.0 400 R72 rt #1
8.7 1/13 8.0 200 W89B br #1
6.3 1/8 4.5 200 R72 hcb #1
6.5 1/15 3.5 200 R72 wjc #1
7.3 1/25 3.5 200 W89B tm #1
5.8 3 18 200 R72 rs #1
6.4 1 13 200 R72 rs #2
Fuji S2 6.0 - 6.0 1/4 5.6 200 R72 ddb #3
Oly E-300 4.3 - 4.3 1/4 4.5 400 R72 sh #1
Canon D30 4.3 - 4.9 4 11 100 B+W092 cs #1
4.0 2 8.0 200 R72 cx #2
4.0 1 5.6 200 R72 ca #2
Sony DSLR-A100 1.8 - 2.3 25 11 100 R72 jk #1
1.3 13 8.0 200 R72 jk #2
1.7 5 8.0 400 R72 jk #3
Canon 350D 1.6 - 1.6 2.5 11 1600 R72 nn #1
2.7 10 8.0 100 R72 pn #1
0.7 0.8 4.5 1600 R72 fe #1
1.4 25 8.0 100 B+W093 je #2
Fuji S3 1.5 - 1.5 23 8.0 100 W89B cm #1
1.5 23 8.0 100 W89B cm #2
Canon 10D /
Canon 300D
1.1 - 1.5 0.71 1.4 100 R72 bc #1
1.4 6 8.0 400 R72 aa #1
0.4 6 4.0 200 R72 gs #1
Nikon D80 0.8 - 0.8 30 8.0 125 R72 eh #1
Canon 1D Mk2 0.7 - 0.3 4 9.0 1600 R72 nn #2
1.1 15 8.0 200 R72 nn #3
Nikon D300 -0.9 - -0.9 15 2.8 100 R72 wjc #2
Canon 400D -1.2 - -1.2 70 5.6 100 IR720 sk #1
Nikon D200 -1.3 - -1.3 25 4.5 200 R72 an #1
Canon 20D -2.0 - -1.0 15 11 1600 R72 nn #4
-3.0 20 4.5 800 R72 ck #1
 
IR-modified / without IR-blocking filter
Camera Avg. IS Max f/ IS     T     f/ ISO Fltr. Link
Nikon D1 ir 15.3 - 15.3 1/320 16.0 200 W89B  
Canon 300D ir 14.9 - 14.9 1/250 11 100 R72  
Canon 10D ir 13.8 - 14.0 1/1000 8.0 400 R72 je #1
13.6 1/400 11 400 R72 jwk #1
Oly 2040Z ir 13.3 1.8 13.3 1/500 4.5 100 RG715 jr #3
Sigma SD10 13.0 - 13.3 1/160 8.0 100 W87C ca #1
13.0 1/125 8.0 100 B+W093 spm #1
12.6 1/50 11 100 W87C rd #1
Canon D30 ir 11.9 - 11.9 1/60 8.0 100 W89B jrs #1
Nikon E990 ir 11.6 2.5 11.6 1/158 4.4 100 W87 rdh #1
Sony f828 ir 11.3 2.0 11.5 1/60 5.6 64 ? aw #1
11.0 1/200 4.5 200 ? aw #2
Minolta D7 ir 11.3 2.8 11.3 1/125 4.5 100 RG715 jr #4
Kodak DCS 460 8.4 - 8.4 1/16 3.3 80 R72 gh #3

How to read the table: The first column (Camera) list the camera model, the second lists the computed the average infrared sensitivity (Avg. IS) for the camera, and the third the maximum aperture (Max f/) for the particular camera. Then the next six columns list image specific data: The IS for the particular image (IS), the shutter time (T), the aperture (f/), the ISO setting (ISO), the filter(s) used (Fltr.), and finally, in the Link column, the photographers initials, with a a link back to the page with the original image

Methodology: The IS numbers listed in the table are computed from images taken by different photographers, at different times and under different conditions. It would obviously have been better to make comparisons by setting up all the different cameras under identical conditions and make fair and direct measurements. However, I don't have the resources to do such a controlled experiment, and those who have (e.g. DPreview) are not suffiscient interested in ir-photography to include ir-sensitivity in their standardized testing suite. I give priority to images taken under roughly the same conditions, depicting the same subject matter. This means that if possible, I use images taken with a filter with an IR pass point of 720 nm (Wratten 89B or equivalent) depicting a bright sunlit landscape (f/16, 1/250 sec. at ISO 200, equal to EV 15 in the visible spectrum) with plenty of foliage. While these constraints are not suffiscient to eliminate all errors, it is the best I can do. What I eventually hope, is to have a large number of samples from each camera. By computing an average IS I hope that the “law of large numbers” eventually will even things out.

Disclaimer: Many of the samples linked to has been extensively post-processed by skilled artists. Do not assume that this is how the image appears out of the camera. The samples indicate at most what can be expressed by a skilled craftsman or artist and a particular camera, lens and filter combination.

4. Exposure Value

The exposure value (EV) system was invented in the 1950s to give an absolute measure of expo­sure needed. You get an EV when you combine sensor sensitiv­ity, shutter speed and aperture. Sensor sensitivity settings, shutter speed and aperture combina­tions that results in the same exposure have the same EV (e.g. ISO 100, f/8 and 1/125 have the same EV as ISO 100, f/5.6 and 1/250 and ISO 200, f/8 and 1/250, and so on). EV is designated by integers such as ..., -2, -1, 0, 1, 2, 3, 4, 5, ... . Each increment of 1 EV corresponds to a increase in the light reaching the sensor by a factor of 2 (letting you use half the ISO value, double the shutter speed, or close the aperture down 1 stop).

Formally, at ISO 100, EV 0 corresponds to a shutter speed of 1 second and an aperture of f/1.0:

EV 0 = (ISO 100, f/1.0, 1 second)

EV 0 is very dark, e.g. a night scene with dim ambient light. By comparison EV 15 is f/16 at 1/125th second, at ISO 100 - this is what you would use for a landscape in bright sunlight (aka known as “sunny sixteen”).

See Fred Parker's Ultimate Exposure Computer if you are interested in learning more about EV.

Note: There is some confusion whether EV takes film speed into account or not. Some insists that EV is only valid for ISO 100, and use a different term, such as light value (LV), or actual EV (aEV) for a number that is a function of the film speed, aperture and shutter triplet. To cut a long story short, the most useful metric for digital, where film speed is just as variable as the other two, is one that takes film speed into account, so that is what I use and call EV. Please note that the calculator featured in the excellent glossary at DPreview is in agreement.

5. Camera Conversion Linkfarm

The links below leads to some pages outlining DIY procedures for IR camera conversion. I've also included some links to outfits that claim that they will do the conversion for you. Linking to these outfits is not an endorsement. I haven't used any of these companies or individuals, and don't know anything about the quality of their work.


If you want to comment, use the blog!

About DPanswers | Terms of Service | Privacy Policy | FAQ | Contact
Copyright © 2009 Hannemyr Nye Medier AS. All rights reserved.