Difference between Kodak print film and movie stock

[harrykauf]

Does anybody know if there is a Kodak negative film available that matches
the Vision2 movie stock ?
Especially its claimed latitude? I heard somewhere that it has over 18 stops
and now read an article in American Cinematographer about the movie
"Sunshine" and the Director of Photography says this:

"The blessing and curse is that today`s stocks are really good; I could
overexpose a scene by 5 or 6 stops and easily bring it back to normal
in the digital grade. When we wanted the image to really burn out,
I had to overexpose by 8,9 even 10stops to really burn the information
away to the point where we couldn`t get it back."

So does anyone know if you could overexpose a scene by 5 to 6 stops
on Portra NC for example and drum scan it and make it look normally
exposed?

I have only limited experience with Vision2 but I have shot some heavily
overexposed stuff and got everything back in the grade and thats without
compensating at the development stage.

[Ash]

Can't help with your question, but I watched the movie the other night. Loved it :)

[David A. Goldfarb]

The new Portras use 2-electron sensitization like the Vision 2 stocks, but I don't think they have the same latitude, since they are designed to have more contrast for printing to RA-4 papers. Cine stock is less contrasty, because it is designed for printing to higher contrast release films for projection.

The new Portra 160NC, though is beautiful stuff--very fine grained with good latitude.

Here's a thread on APUG with some tests of the new Portra films--

http://www.apug.org/forums/forum40/33887-now-new-portras-whos-used-them.html

[Bruce Watson]

So does anyone know if you could overexpose a scene by 5 to 6 stops on Portra NC for example and drum scan it and make it look normally exposed?

Graininess is directly related to density. If you over expose by five or six stops (just makes me cringe to think about it) you will have an extreme amount of graininess which a scanner will not correct.

Also as a consequence of the extreme density you will lose quite a bit of sharpness.

You'll almost surely generate some interesting "density halos" where light and dark things in the scene are in close juxaposition. I don't remember what causes this, but I've seen it in severely overexposed Tri-X, so I'm assuming that you'd get it with a C-41 process as well. Maybe not, but it's a risk.

I would expect some color shifting as you push the image up the film's characteristic curve. The Portra films have nice long straight line sections to their characteristic curves, but everything shoulders somewhere.

And if it doesn't shoulder, as 100Tmax is rumored not to shoulder, you can easily put the density even beyond the reach of a first class drum scanner. I've scanned some 160PortraVC on my drum scanner and seen densities from the blue channel of more than 2.5. Five more stops would push it to around 4.0. I wouldn't guarantee to be able to read it, but I would guarantee that if I could read it, I'd also return some scanner noise.

If you add it all together, the grain, loss of sharpness, color shifting, scanner noise, weird density related artifacts, etc. I think it's safe to say that no, you can't make a severely over exposed negative look "normal" in post.

[Ben R]

I'm missing something here, why would you want to overexpose film that much just to bring it back, is this for some sort of effect or to shoot scenes with a huge difference between the highlights and the shadows with some sort of HDR effect?

[Mark Woods]

Hello Harry,

Motion picture film is an ECN-2 process -- not C41. And slightly contrary to what Bruce mentioned, the Vision stocks are known for having a tighter grain structure with a bit of added density. Also, the Vision stocks' color emulsions track very parallel, i.e., there is very little color shift anywhere on the H&D Curve. I believe you mentioned that the Cinematographer said he could bring back an image "5 or 6 stops overexposed" and get it back to a "normal" look. I think that's a bit over the top since my tests show about 13 stops of total latitude. That said, the most expensive drum scanner made is pocket change compared to the equipment in a top of the line Post Production house that creates the Digital Intermediate for a film out. Those bays run between $1.5-5 million/bay -- maybe more. Needless to say, the bells and whistles they have to make things better are not only more robust, but they do film corrections at 24 frames/second -- real time in 2K and they're getting close to real time rendering in 4K. I'm looking forward to 6K real time, but those files are huge for each frame. :-)

Hope this helps.

MW

[Helen Bach]

I'm also surprised by the '5 or 6 stops overexposed 5218 and still normal' comment, but 'overexposed' is a vague term and if he meant 'over the metered value' I could believe it easily. 500T Expression 5229 (which I like very much) and HD Scan Film 5299 have even more dynamic range than 5218.

If you are scanning rather than printing optically you can easily use the full dynamic range of still colour negative film, which I find to be at least 12 stops for Portra 160NC for example (the exact number of stops depends on what you can accept at the limits).

Bruce,

Though silver-image negative film gets more grainy with density, dye-image negative film tends not to. In fact it can get less grainy as the density increases - rapidly at the toe, then more gradually. It does lose definition though, as you say. ECN-2 is a continuous agitation process (naturally) so there isn't the same opportunity for edge effects as there is with intermittent agitation.

Colour negative film doesn't have a 1:1 exposure:density relationship. Six stops of extra exposure doesn't add 2.0 extra density - more like 1.0, varying from film to film.

Ben,

I think the idea in the case of Sunshine was to get a blown-out look - they didn't want to overexpose and bring back to normal. On the other hand, with films that roll off a little at the very top end of their curve (towards the shoulder, not always shown on the manufacturer's published curves) generous exposure puts the highlights into that lower contrast zone, and it softens the definition a little - effects that can be used, especially if the film won't be post-processed digitally.

Best,
Helen

[Bruce Watson]

Though silver-image negative film gets more grainy with density, dye-image negative film tends not to. In fact it can get less grainy as the density increases - rapidly at the toe, then more gradually. It does lose definition though, as you say. ECN-2 is a continuous agitation process (naturally) so there isn't the same opportunity for edge effects as there is with intermittent agitation.

I'm not a film engineer and don't even play one on TV. But this isn't my experience, nor does it fit with what I've learned over the years about image formation. I'm not saying you are wrong. I'm just saying that yours is the first voice I've heard say this. I didn't see this in Haist, or Henry, or Morgan, etc. If it was there I must have missed it.

My understanding is that the image is formed with silver halide crystals. These crystals establish the graininess and sharpness of the image. During development, dye couplers form dye clouds based on the sliver halides. Dye inhibitor releasing (DIR) couplers are used to control the growth of the dye clouds.

In image areas that are struck with low light levels graininess is low because the silver halide development is low and therefore dye cloud development is low and the clouds small. In areas that receive more light, more silver halide grains are developed and therefore more dye clouds are created.

Each of the many image forming layers in color films do their own latent image capturing, silver halide development, and dye cloud formation. When a scanner (or enlarger) is used to "read" the negative, the layers add together to give the total density and graininess of the film.

What I see in scanning 160PortraVC is a fairly linear increase in graininess as the negative goes from less dense to more dense. White fluffy clouds in a print at high enlargement ratios start to show this graininess as a sort of "color confetti." I start to see this around 10-12x and therefore limit my enlargements to that level.

That said, my education is far from perfect. If you could point me toward further reading on this idea that increasing density can result in decreasing graininess from a color negative film I'd love to learn more. I'm serious. The more I learn about how it all works the better my prints seem to get, and I'm happy to be proven wrong if it makes my prints better.


Colour negative film doesn't have a 1:1 exposure:density relationship. Six stops of extra exposure doesn't add 2.0 extra density - more like 1.0, varying from film to film.

True enough. The slope of the characteristic curves can be just about anything the engineers designed into the film. Since I don't know and can't know (no lab to play in) what the actual slope is for 160PortraVC I was just picking on the definition of a "stop" as a doubling or halving of luminance and translating that to a doubling or halving of density, which would be about 0.3D per stop IIRC. Many B&W films are actually fairly close to this IIRC. But I doubt that color films are.

[Mark Woods]

Hello Bruce,

Here is a paper presented by Kodak regarding graininess:
http://www.kodak.com/global/en/professional/support/techPubs/e58/e58.jhtml

It is about color negative film for print on paper, not motion picture film. Your statement about graininess showing up in the highlights may be your experience in stills, but it's quite the opposite with motion picture film. "Noise" may show up in the highlight in a film to tape transfer on older machines, but it's rare to see it currently with new equipment. In the photo-chemical arena, the highlights rarely have any grain, but in a "normally" exposed negative, there can be grain (a lot of grain) in the shadows. This is why many Cinematographers subtract 1/3 stop from the EI to add 1/3 stop density to the negative. By printing down the image, the blacks are printed down to Dmax on the print.

Hope this helps. Motion picture film is a different animal from still negative film. Different development process, different H&D curve, and a different presentation.

Kind Regards,

MW

[Bruce Watson]

Hello Bruce,

Here is a paper presented by Kodak regarding graininess:
http://www.kodak.com/global/en/professional/support/techPubs/e58/e58.jhtml

It is about color negative film for print on paper, not motion picture film.

Yes it is, because that is what the OP asked about. He said "So does anyone know if you could overexpose a scene by 5 to 6 stops on Portra NC for example and drum scan it and make it look normally exposed?"

[Helen Bach]

Bruce,

I think that there is a perfectly logical explanation for the apparent paradox, and I probably should have mentioned it already.

Two factors are at play:

the measurable granularity of the film; and

our ability to perceive it.

We can perceive much smaller changes in density in light tones than in dark tones. Therefore it is possible for the light tones in a print to appear more grainy than the darker tones, even though the measured granularity is lower in the highlights than the midtones and shadows.

A good test would be to compare the graininess of two otherwise identical prints (matching in terms of density and contrast): one with two stops more exposure than the other, though one stop should do. Far from being a lone voice, I thought that I was just following the herd - the fashion herd that gives colour negative film a deliberately generous amount of exposure to make skin tones smoother and softer.

Oh. Why is colour film less grainy in the highlights? One reason could be that colour film is usually made up of multiple layers in each spectral sensitivity, and the high densities tend to be dominated by the finest grains.

The other reason (I think) is the difference between the behaviour of opaque silver grains and necessarily transparent dye clouds. As an aside, in the good old days I used to think that Fuji got its reputation for less grainy films by making the dye clouds less distinct, while Kodak went for more distinct, and hence sharper, dye clouds.

How does that sound?

Best,
Helen

[harrykauf]

Thank you all for those very informative answers. Now I know that
the still and movie stocks are different and why.

I can also confirm Helens experience that overexposing reduces grain
on movie stock. I work in a post production house and shoot on super16
with a friend and we can use our spirit datacine for grading on weekends.
So my experience is limited but I have overexposed by 3 stops atleast
(3 stops above normal reading) and got everything back, even detail
in a white suit or newspaper. Without pulling in development.

After the first positive tests I started to overexpose everything by 1 stop
and it really helps keep the grain down which is a bigger problem in 16mm
than 35mm.

Another reason to overexpose can be on a shoot that will get elements
added in post so you want to keep detail in the shadows for 3d tracking.
Sometimes you just cant control the contrast when the camera is moving
and ND grads are useless. So you expose for the darkest area and know
that the film will hold all the highlight detail to bring it back in post.

[C. D. Keth]

It's my undertanding that some overexposure always makes film images appear less grainy. Large grains of silver halide are the fastest and smaller grains are slower and need more light to hit their exposure threshold to be developable. Overexposure exposes more of these smaller grains and "fills in" the spaces between larger grains with them.