2

I'm very new to photography and got attracted to the infrared photos I keep seeing online.

I decided to try experimenting with something cheap to see if I have any hope in this hobby or not. I got an old Sigma SD14 because it doesn't require conversion, as the internal filter can be simply popped out. I also got a Hoya r72 filter. Using the kit lens, for the moment.

Now, for some reason, all my photos end up either horrendously overexposed but with at least some information in them, or total garbage with literally no information in the photo.

I've tried what I feel is every combination of ISO, F and shutter speed with different, but all equally terrible results. I know I can't expect too much from the cheap old camera I got, but it should be able to produce something usable, at least according to other experiences I find online.

I've uploaded a couple of RAW samples and would be endlessly grateful to anyone who'd be able to give me any pointer at all about what I am doing so deeply wrong. How come even on the best of my photos, if it's blueish - everything is equally blueish, or if it's reddish - everything is reddish (see the linked samples). On other photos I see, trees look white or pinkish, but the buildings do not.

Could it be that I simply somehow damaged the sensor so everything appears too bright?

Terribly sorry for the broadness of my questions, but I still lack the vocabulary to express myself better. I'm learning, but need direction...

  • 1
    Your raw files appear to be ~5-stops underexposed, not overexposed. – xiota Aug 9 at 22:38
2

The most likely cause and solution to your problem is:

  • You need to tweak settings when processing RAW files.Leon demonstrates this to likely be the case. I used RawTherapee to process one of your images. The following settings are particularly relevant:

    • Color management must be turned off because it is unable to handle IR correctly.

      Color Management
      Input Profile: None,
      Working Profile: sRGB,
      Output Profile: No ICM.

    • Increase exposure compensation to 5.5. (Consider increasing EV when taking IR photos.)

    • White Balance temperature 1500, tint 10.

    • RGB curves to visually "normalize" histogram.
    • Adjust tone curve to mild s-shape.

    After exporting the image, you can make further adjustments in an image editor, such as GIMP. In this case, I extracted the red channel because it has much more detail and less noise than the other channels. There is no significant color information in this image, so there is no benefit to keep the other channels.

    img1

    Here is a 100% crop to compare the RGB channels to show the difference in noise among the channels. Note that other cameras, because of the Bayer filter, may have different noise characteristics across the channels.

    img1-noise

Other possible causes that do not necessarily apply to your situation include:

  • Your camera settings may not be optimal for IR. You may need to use a different metering setting, add exposure compensation, or even use full manual mode. Since camera settings for IR tend to be idiosyncratic, you will need to experiment.

  • Your camera may not be able to handle white balance for IR. This is very likely because most cameras have this problem. This issue affects your ability to preview images in the field and your post-processing workflow. It should not hinder your ability to process the raw files.

  • You could have a poor-quality infrared filter. Based on examination of your raw files, the filter appears to be fine.

  • The camera could be damaged. This is unlikely if your camera works as expected for visible light. Based on examination of your raw files, your camera seems to be working appropriately.

Other points to consider for near-infrared photography:

  • Your raw files are monochromatic because you used a 720nm filter. Longer wavelength filters, such as 850nm or 950nm, are likely to be "more" monochromatic. The subtle color effects some people get with 720nm filters comes from the quality of the IR filter. Lower quality IR filters let more visible light through, creating false colors.

  • If you want to create false-color IR images, you will need to use a 680nm filter. Swapping the R-B channels would create dark-blue skies and golden leaves.

  • As for why IR images tend to be reddish with unbalanced color channels, consider the reduced blue/green sensitivity in the infrared range of the Foveon sensor (source).

    Foveon sensitivity

    The drop in blue/green sensitivity of Bayer filters is more extreme.

  • Any clue on how to achieve that beautiful pink foliage? – kaqqao Sep 28 '18 at 14:08
  • The raws were underexposed over 5 stops. Maybe some color will be captured if you increase exposure. If not, it might be a limitation of the sensor. – xiota Sep 28 '18 at 19:28
  • @kaqqao Edited to show noise difference among channels. For images that are essentially monochromatic, it may be worthwhile to simply extract the "best" channel (red). – xiota Aug 14 at 5:31
4

This is standard for digital IR photography.

You will get better results if you set the camera to a B&W mode (if there's such an option on your camera) or if you manually convert your images to B&W and adjust in post-production just like you would any other image from a digital camera.

This is a necessary post-production process for digital IR photography.

You can also do color manipulations (See: https://petapixel.com/2016/10/27/introduction-digital-infrared-photography/)

B&W Conversion

  • Hey, thanks for the answer and the great work you've done with the photo! Much appreciated! But... how do people get the beautiful pink and yellow foliage in their photos? Perhaps different filters? – kaqqao Sep 26 '18 at 16:47
  • 2
    There is IR with different spectrums - 750nm, 850nm, which produce slightly different looks. but ultimately all of that is done in post-production. One technique is called "channel swapping" (slrlounge.com/…) – Leon Sep 26 '18 at 16:57
  • Ah, the photo in the article you linked looks absolutely dreamy 😍That's what I'm ultimately after. – kaqqao Sep 26 '18 at 17:03

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.