7
\$\begingroup\$

I made this little video. You can see a white flickering light coming out of my remote control.

When I look at my remote control with my naked eyes, there is nothing to see when I press a button of the remote control. Why can I see the radiation (infrared, I guess?) on my mobile phone?

\$\endgroup\$
  • 11
    \$\begingroup\$ This is not really an EE question, this would be better on the photography SE, basically our eyes can only see visible light frequencies and IR is outside of that band. Camera's are digital optical sensors that can see frequencies outside of the visible band, we often put UV & IR filters Infront of standard camera sensors to avoid UV & IR distorting the 'Natural' colour balance of the image. These IR filters are not perfect so you can still see light bleed from a direct IR source like that found in the remote. This is a basic answer to a complicated question but gives you the basics. \$\endgroup\$ – Jack Soldano Sep 17 at 16:08
  • 10
    \$\begingroup\$ Are you an elf or halfling? If you are not, that would explain why you do not have infravision. \$\endgroup\$ – Harper - Reinstate Monica Sep 18 at 4:53
  • \$\begingroup\$ @Harper-ReinstateMonica LOL! In fact, I'm the crawling king snake (J.M. gave me permission to use his words). :-) \$\endgroup\$ – Deschele Schilder Sep 18 at 7:00
  • 2
    \$\begingroup\$ For most remotes, one can in fact SEE the radiation, given it is dark enough for eyes to adapt. Most IR LEDs have some visible "tail" in their specters. It looks faint and deep red. \$\endgroup\$ – fraxinus Sep 18 at 8:21
  • \$\begingroup\$ @fraxinus When it's dark, I'll put some fresh batteries in the control and try seeing the tail. Thanx! \$\endgroup\$ – Deschele Schilder Sep 18 at 8:25
35
\$\begingroup\$

Building on the good answer by nanofarad, "But why white?" is a good follow up.

We don't know what the precise sensor in your phone is, or how good the IR filter in front of the lens is, but we can study the sensitivity plots for an example CMOS sensor.

enter image description here

Notice the three distinct humps, or increased sensitivity, occurring at different wavelengths for visible light. Notice also the increase in sensitivity towards infrared, for all three sensors equally.

By tuning the efficiency to different colours you get different light sensitivity. (More on colour selective sensitivity below)

In the curves you'll see the blue, green and red sensor sensitivity plotted, with peaks at 450 550 and 600nm wavelengths respectively.

White is for instance an equal mix of red, green and blue.

An area for the larger wavelengths on the right is supposed to be blocked by an IR filter. IR light is on the far right, at 940nm. All three sensors of this type, according to the graph, have the same efficiency at infrared frequencies.

This means that IR light will translate to the same amount of free electrons (current, voltage) for each sensor (R G and B) and will be stored as equal RGB pixel values (8-bit, 16-bit or 24-bit numbers) in a typical photo file.

When shown on screen, these equal numbers will appear "white" to the eye.

Thus, the 940 nm wavelength typically used by remote controls falls into the range where all three sensors have roughly the same response, giving a "white" result.

It should be noted that not all devices will have colour curves that raise at lower frequencies, as shown. Some will remain low, but usually all three will converge to exhibit about the same "efficiency". These devices would require a higher level of IR light intensity to translate to a "white" depiction in an image. This does not mean that the sensor has to be saturated with IR to appear white.

Now a bit more on colour selective sensitivity:

Technically, quantum efficiency refers to how well a CMOS semiconductor can transfer light energy into free charge. A sensor's colour selectivity is accomplished by placing a mosaic of tiny pixel-sized colour filters in front of the sensor array. This is pixel filter is called the "Bayer Filter", after its inventor, and it was developed by Kodak, already back in the 1970s.

Polymers are used to dye the filter, which will then absorb light (or reflect light - depending on the type) of specific wavelengths, and thus filter it.

So, the tuned efficiency is, more precisely, the quantum efficiency of the CMOS sensor cell's semiconductor times the transmittance of the optical filter in front of it. With the two together, we can speak of colour dependent quantum efficiency of a colour CMOS sensor.

And by combining different base colours (RGB) in the camera's photo sensor array, we can represent practically all colours for the full range of colour visibility of our eyes.

By the way, a CMOS sensor is not to be confused with a CCD sensor, which is an older technology now phased out for cameras.

There is of course much more to read about this, but I think a bit of an introduction is warranted here, as invited by the comments.

Image credit https://www.thorlabs.com/NewGroupPage9_PF.cfm?Guide=10&Category_ID=220&ObjectGroup_ID=4024

| improve this answer | |
\$\endgroup\$
  • 3
    \$\begingroup\$ What are those 3 curves? Not the actual sensors? But rather the combined response of the CCD sensor + the Bayer filter(?) \$\endgroup\$ – Peter Mortensen Sep 18 at 3:06
  • \$\begingroup\$ Great answer!!! I don't know what else to say. \$\endgroup\$ – Deschele Schilder Sep 18 at 7:03
  • \$\begingroup\$ Looking at the efficiency graph I’m surprised that cameras (especially tiny smartphone cameras) don’t have a removable IR filter for low light conditions. Wouldn’t it increase low light performance a lot? (at the cost of color saturation because IR just appears gray/white) \$\endgroup\$ – Michael Sep 18 at 7:57
  • 2
    \$\begingroup\$ @Michael Because the three sensors have different responses to IR, it would change the color, not just saturation. The goal of these cameras is to produce an image which convincingly fools the human eye. Bringing in spectra which is not part of the normal human vision would create images which don't feel quite right. \$\endgroup\$ – Cort Ammon Sep 18 at 8:50
  • \$\begingroup\$ You could elaborate that the 940 nm wavelength typically used by remote controls falls into the range where all three sensors have roughly the same response, giving a "white" result. \$\endgroup\$ – Ralf Kleberhoff Sep 18 at 11:39
21
\$\begingroup\$

The remote control emits infrared light, typically around 940 nm, while the human eye is receptive to ~400-700 nm. However, the CMOS optical elements in a phone camera have a response up to 1000 nm (source), and for cost/space reasons, the phone may omit an IR filter.

I recommend reading (and accepting) P2000's answer as it includes far more details about the actual sensitivities of CMOS sensor elements in the IR range.

| improve this answer | |
\$\endgroup\$
  • 1
    \$\begingroup\$ Of course! How else could it be? \$\endgroup\$ – Deschele Schilder Sep 17 at 16:08
  • 2
    \$\begingroup\$ But why is the IR-radiation visible in white? \$\endgroup\$ – Deschele Schilder Sep 17 at 16:12
  • 9
    \$\begingroup\$ This 'White' light means the IR light is overexposing the sensor resulting in a maximum readout. This will be displayed in the image as white (No Data/Max Data!) if you reduce the exposure the IR light will start to show up as a purple colour. The reason its purple is quite complicated but basically comes down to the way a sensor is made (With Red Green and Blue pixels) the IR interacts more with the Red & Blue pixels than the green so the sensor interprets IR as a purple sort of light. IR is out of the visible spectrum; Purple is not this is an artefact of our sensor design. \$\endgroup\$ – Jack Soldano Sep 17 at 16:19
  • 6
    \$\begingroup\$ @JackSoldano Interesting. I took some photos when I was heat-treating a tool-steel part I machined (inside a kiln, so reddish hot) and the images came out as very purple. The extra blue content is the unexpected part. \$\endgroup\$ – Spehro Pefhany Sep 17 at 17:41
  • 1
    \$\begingroup\$ Thank you for your shout out. Seems to be a popular topic! And I learned about mediocre or omitted IR filters. \$\endgroup\$ – P2000 Sep 18 at 14:41

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.