Removing the IR filter

[Rodrigo Prieto]

In the General Discussion Forum, I posted a question about Night Vision for a film I am prepping which has a scene in a desert at night. I am looking for ways to shoot with the least amount of lighting possible (or even in moonlight). I am exploring the Genesis camera or other HD cameras pushed to the limit. Luke Prendergast suggested removing the IR filter, but I don't understand what this means, and if it indeed would help with capturing extremely low light levels. Has anyone done this, and what effect does it have on the image? Please explain it to me as if I were a 4 year old. Thanks!

[David Mullen ASC]

In the General Discussion Forum, I posted a question about Night Vision for a film I am prepping which has a scene in a desert at night.  I am looking for ways to shoot with the least amount of lighting possible (or even in moonlight).  I am exploring the Genesis camera or other HD cameras pushed to the limit.  Luke Prendergast suggested removing the IR filter, but I don't understand what this means, and if it indeed would help with capturing extremely low light levels.  Has anyone done this, and what effect does it have on the image?  Please explain it to me as if I were a 4 year old.  Thanks!

<{POST_SNAPBACK}>

 

The CCD's in digital cameras are overly sensitive to the infrared end of the spectrum so that light is filtered out internally. I don't know what removing the filter would look like at night in terms of extra information, but it would only be info at the IR end of the spectrum (heat for one thing).

 

It would be something to test in combination with boosting the gain and turning off the shutter, the other two techniques of boosting a camera's sensitivity. You figure that a +12 db boost would get most cameras in the 2000 ASA range, and then going to 1/24th of a second on the shutter would get you in the 4000 ASA range, not that you can shoot a scene under real moonlight necessarily but it would be close if you combined that with shooting wide-open on a T/1.6 HD prime lens.

 

Here's a bizarre idea that probably won't work: find two HD cameras that can do 12 fps at 1/12th of a second per frame (like the Varicam can), put them into a 3D rig but lined up so that are shooting exactly the same frame, then trigger them somehow to be 1/6th of a second offset from each other (half a frame), and then in post somehow overlay them to smooth out the jumpy motion from only shooting at 12 fps... Of course, doing all of that only gets you another stop of speed so it may not be worth it...

 

I'm not sure how you can avoid SOME lighting unless using a Night Vision rig and shooting by starlight. Even if with an HD rig, you somehow had enough exposure to shoot by the light of a full moon, for example, the moon is only fullish a couple of days per month, and what if it is overcast that night? So once you're committed to adding some minimal amount of light, perhaps getting an HD camera into the 4000 ASA or 8000 ASA range would be sufficient.

[Ilmari Reitmaa]

The CCD's in digital cameras are overly sensitive to the infrared end of the spectrum so that light is filtered out internally. I don't know what removing the filter would look like at night in terms of extra information, but it would only be info at the IR end of the spectrum (heat for one thing).

<{POST_SNAPBACK}>

 

For what it's worth, CCD spectral sensitivity does not go much beyond 1200nm (which is well into the infrared zone though), usually being most sensitive in the 600-800nm area. Human body heat radiation peaks at 9300-9400nm, so one shouldn't expect to be able to pick it up much, even with IR-filter removed.

 

The Genesis uses, I would think, a Sony HAD CCD, so probably either Sony or Panavision will be able, if willing, to provide spectral sensitivity curves for purposes of comparison with, say, film stock sensitivity curves.

 

(Still, probably wouldn't try to explain it quite like this to a four-year old...)

[Phil Rhodes]

Hi,

 

I still think it'd be worth a try. When you're fighting for whatever exposure you can possibly get, if your camera can see IR then imaging the IR might be a perfectly reasonable way to go.

 

For that matter, you can triple the sensitivity of any video camera by using a prism block without the filters, which would get you three times the sensitivity and a monochrome image, but I should stress: I don't think it would be practical to remove the IR filtering, and even less the RGB filtering, from most video cameras. It's all fused onto the prism block. If you're doing a particularly upscale production you might be able to get them to make you one without the filters, but get ready for a very large bill.

 

Phil

[Ilmari Reitmaa]

For that matter, you can triple the sensitivity of any video camera by using a prism block without the filters, which would get you three times the sensitivity and a monochrome image

<{POST_SNAPBACK}>

 

Hate to correct you there, Phil, but that's actually incorrect in the general case; the prism block colour separation isn't based on subtractive filtering but on reflective coatings, refractive indices and whatnot, thus preserving, theoretically, all of the incident light. There are certainly variations and combinations, though.

 

Would work with a single chip configuration, though, i.e. replacing a Bayer-pattern (or which-ever-pattern) chip with a monochrome one. That, actually, might be feasible with the Genesis (might), or better yet, Kinetta, yielding about three times more incident light on the CCD.

[Sam Wells]

Would work with a single chip configuration, though, i.e. replacing a Bayer-pattern (or which-ever-pattern) chip with a monochrome one. That, actually, might be feasible with the Genesis (might), or better yet, Kinetta, yielding about three times more incident light on the CCD.

<{POST_SNAPBACK}>

 

My recollection is that Jeff, when I asked said that a chip not IR filtered would be possible but it would, as you suggest, be a monochrome chip, not the same as using the AltaSens sans the filter.

 

-Sam

[Elhanan Matos]

This doesn't really have anything to do with HD, but it does have some relevance to this post.

 

I got the link off of a website called hackaday.com

 

Make your own IR webcam

[Kevin Zanit]

The page Elhanan linked to had a picture of an area illuminated with infrared LEDs (from a remote control).

 

What if you were to create large panels of IR LEDs (or any other IR source) to "light" your scene? Then, obviously, record the scene with an IR camera (or one with no IR filter).

 

The light from an IR LED is invisible to the naked eye, thus it would not intrude on your actor's performances. It would still be pitch black to the eye.

 

 

 

Kevin Zanit

[Eric Steelberg ASC]

I've inquired with Panavision to remove the IR filter. According to them it's not possible to get a "workable" image. "Bad things" happen. I didn't push the matter or get additional information.