Nonlinear optical filter

[Mitch Deoudes]

Here's one for the guys out there with outsized, egg-shaped heads:

Is there such an animal as an optical ND filter with a nonlinear response, brightness-wise? I.e., something that would attenuate highlights, while leaving mids / shadows alone (or attenuating them less)?

I'm guessing no (though a pola is a similar thing, I suppose) - but it's an interesting thought exercise. (It should be obvious where this came from - compressing dynamic range, etc.)

[Graeme Nattress]

Nearest thing I know of is an ultra-contrast filter. They sorta-work to raise just blacks, so you can expose more for the highlights and capture a bit more shadow data. I've not tried one on a RED yet.

Graeme

[David Mullen ASC]

Well, there are filters like the Schneider DigiCon and Tiffen SoftCon that have an ND element that you are not supposed to compensate for, that push down the exposure slightly to help with highlight detail. They may involve using black specks rather than overall ND but I'm not sure that really makes a difference (why you can't just stop down a little while using a regular UltraCon or Low-Con, I don't know).

They also have a low-con effect that lifts shadow detail, somewhat to compensate for the lowering of exposure for the highlights, so you get a somewhat milkier image to color-correct back to the contrast you want later.

Flashing does a similar thing -- adds light that has more of an obvious effect in the shadow regions than in the highlights.

But the reverse effect seems to only happen when you flash a positive piece of film like a print, the added light darkens the bright areas (since technically a print is still a type of negative stock where more light makes areas darker, just that it is a negative copy of a negative image and thus a positive image.)

I did recall seeing a post somewhere about a new product being developed that reacts only to bright areas, creating an ND effect for the highlights only -- sort of an electronic filter device. The main problem with such an idea is response time -- as you pan the camera, you don't want a "lag" where you see a darkened afterimage where the bright object was.

[jbeale]

There are many nonlinear optical materials. Most people have probably seen the kind of glasses that become tinted when you go outside into bright light, but they work very slowly. The ones I know about that are really fast are only significantly nonlinear at extreme brightness levels. For example, you might need to be within inches of a several-kilowatt bulb for it to start working.

If you did have such a filter, it would have to be in the image plane, of course. If you put it in front of the lens it would darken uniformly, just like the sunglasses.

[Mitch Deoudes]

If you did have such a filter, it would have to be in the image plane, of course. If you put in on front of the lens it would darken uniformly, just like the sunglasses.

Yah - this was the thinking behind my "I'm guessing no" comment... though I'm a bit too woolly-headed today to have really thought it out.

w.r.t. to Tiffen: According to [this page], the difference is that the Ultra-Con filters don't flare as much as the other two types. Though I'm still a bit confused as to what's physically going on here - it sounds like a flashing sort of effect, as david mentioned, or a diffusing of the highlights into the rest of the image... That seems like it might perceptually affect the contrast, but not necessarily allow you to re-construct the original information later. (Again, excuse my woolen head.)

[David Mullen ASC]

Lifting the blacks with flashing or Ultracons, etc. has a very mild effect on improving shadow detail, bringing barely recordable information into a more visible range -- but by the time you reset the blacks to normal in post color-correction, most of the extra information is buried again, though you could you say that you've increased some flexibility in adjusting the shadow detail in post. But like I said, the improvement in actual shadow detail is minimal compared to the amount you are just lifting the blacks.

[Policar]

http://www.cinematography.net/digicon-test.htm

Actually kind of impressive. FWIW, the ultracon 3 I used recently was too hard to control for my tastes: it would flare up when hit by sunlight and do nothing when no light was hitting it. But it did help "soften" the image a bit in terms of contrast.

[Anders Holck]

Doesn't flashing only give benefit to a non linear medium like film, where it can move the shadow detail from the compression of the toe up to the linear part of the response curve?

[David Mullen ASC]

Doesn't flashing only give benefit to a non linear medium like film, where it can move the shadow detail from the compression of the toe up to the linear part of the response curve?

It probably has more of a benefit to film but not for that reason. Basically a silver halide crystal needs enough energy from photons to become "developable" into silver, so there are some barely exposed crystals on the edge of being developable, so some extra exposure from flashing allows some of that threshold detail to become developable, plus the lifting of the blacks makes some of that information more visible. Whether or not it actually moves from the toe to the straight-line, I don't know.

Similar to latensification, though that works better with very long exposures to very low light levels.

I have heard that in early tube video cameras, there was something called a bias light -- a weak light shining onto the sensor -- used to increase the signal enough to reduce lag artifacts. Anyway, Jeff Krienes put a bias light into his Kinetta prototype and showed me how it could increase the shadow detail of the camera in low-light, though at the expense of the blacks.

So I think low-cons and whatnot have similar effects in video as they do on film -- lifting the blacks and bringing out a little shadow detail.

[marasco ivan]

more on digicon test:
http://www.alfonsoparra.com/pages/press/e18.html