Xbox 360 WebCam IR filter

My first blog post was about mounting the Xbox Live Vision camera on a telescope for astrophotography. Last night I removed the built in red glass filter, which makes the webcam much more light sensitive - especially for infra-red (IR). I actually bought a second Xbox 360 camera for this experiment, so I can show photos with and without this modification.

Removing the red glass filter

For my second attempt I found opening the Xbox webcam case was much easier - and again I removed the green LEDs and cut off the nose cone for use with a telescope adapter. As before, this is based on the advice and photos from Cliff Tate on the Star Gazers Lounge forum.

To get at the IR filter you also need to unscrew the PCB from the back of the case, and then unscrew the black plastic threaded lens mount. For this a small cross-head screwdriver would be advised - I just about managed with a flat headed jeweller's screwdriver but it was quite a struggle. (Update - a Phillips head PH 000 screwdriver works beautifully.)

Two PCB screws removed

Note the internal USB connector

You don't actually have to remove the board completely, I did it to take these photos.

Back of PCB

Front of PCB, with mount

Front of PCB, without mount, showing CCD

Lens mount showing red glass

The red glass filter can now be removed by gentle pressure on the edge with a small screwdriver from the front (through the lens hole). Then reattach the threaded mount, making sure the screws are done up tightly, replace the PCB, and put the case back on. If you have suitable small crosshead screw drivers, it should be quite easy.

Testing in low light

Since I now have two Xbox Live Vision cameras, I can compare with and without the red glass filter side by side. In normal light, both work fine but of course the colour balance is a bit strange without the filter.

I tried this in the middle of the night, with moderate ceiling light, indirect light from the hall, and indirect light from a desk light - I didn't note down exactly which was which.

All the images are taken with Photo Booth, a simple webcam application which comes with Mac OS X. This offers no control over the raw image capture settings at all.

The photos on the left are from the infra red (IR) sensitive camera with the red glass filter removed (slightly leaning to the left), those on the right are with the original filter in place:

In the photos above, in normal lighting levels, the pictures are not so different - the colour is more natural for the unmodified camera (right), especially in the green leaves of the plant and the brown of the picture frames.

Once the light starts to get lower, the improvement from removing the red glass filter is very apparent (left) - although there is still some details in the nearly black images from the original camera (right):

There is now quite noticeable image noise in the white walls from the IR sensitive camera (left), while only a hint of the picture frames and mantlepiece present with the original camera (right):

Then I tried some experiments by candle light, here with a candle on the mantlepiece shelf itself - notice the unmodified camera captures this as a yellow light (right) while it appears more white on the IR sensitive modified camera (left):

Here I've moved the candle away:

And even further away - but there is still a recognisable picture from the IR sensitive modified camera (left):

Here the candle is on the floor, giving a slight reflection from the wood of the mantlepiece, which just shows up on the unmodified camera as a faint smudge (right):

Here the candle is positioned to give a reflection from the fire place tiles - giving just one spot of light on the photo from the unmodified camera (right), but a pretty good image on the IR sensitive modified camera (left):

The results are pretty impressive. I found in low light conditions like this my eyes were still more sensitive than these single images captured with default settings - however, I'm sure that with fine control of gain/exposure or stacking multiple images this can be improved. That is encouraging for my intended astronomy usage.

I'm looking forward to trying this out with the telescope - prior to removing the red glass filter I'd not managed to capture any stars at all. The moon, Jupiter and Saturn are all relatively bright - the dimmest objects I'd managed were faint smudges for some of Jupiter's moons. Hopefully I can do a three-way comparison: Original Xbox Live Vision with its red glass filter, without the red glass, and without the red glass but using an astronomy IR filter threaded onto the adapter.

Note - an astronomy IR (and UV) cut filter is recommended if your telescope has any lenses, since these may only consistently focus visible frequency light. If you have a reflecting telescope (e.g. a Newtonian) this shouldn't be a problem unless also using a Barlow lens for magnification.

Update

Sadly it was cloudy tonight, so I couldn't try this with astronomical targets - but I did have a go with some city lights. This does have the advantage that I don't have to worry about sidereal tracking!

I picked my target using the IR sensitive camera. To switch cameras, I gently unscrewed the camera from the mounted nosepiece, and then screwed the other camera on in its place. This usually disturbed the view point a little, but the focus point seemed to match nicely. Next time I'll try removing the nosepiece while switching the cameras - which will avoid twisting the USB cable and causing tangles (update - I bought a second nosepiece which makes switching cameras a lot easier).

I used the IR modified camera first (left photo), then the unmodified camera (right photo), then back to the IR modified camera to check the pattern of lights hadn't changed. It did with my first target - someone had turned a light off while I was switching equipment!

The square light (bottom left) is probably a large lit window, while the central and top right lights are probably street lights or something. As expected, the modification to remove the red glass filter makes a noticeable improvement in sensitivity - but is isn't as dramatic as with the indoors tests.

Update: I've also tried a DIY infra-red pass filter with some indoor targets.


Update: I've also tried adding IR LEDs to make a night vision webcam.