Experts

The film look: what is it, and why are there two of them?

There are two basic techniques for making digital images look like film. Here’s what they are, and why.

The film look is an elusive thing. Everyone has their favorite flavor, but there are two basic, yet opposite, techniques in vogue right now—and I think I’ve figured out why.

When the film-to-digital transition commenced in earnest, back in the early 2000s, the biggest complaint about digital was that it had a “video look.” Over time the general consensus was that it boiled down to two things: over sharpened detail, and very saturated and hue-distorted highlights. Early HD cameras (I’m looking at you, Sony F900) only gave us about two stops of overexposure latitude above middle gray. Highlights clipped in the most hideous manner, often showing hue shifts (cyan and pink were not uncommon). And the rule of thumb for those of us with film experience was to keep the detail way down when shooting for broadcast (-40 in SAU, or “Sony Arbitrary Units”) or off when shooting for a film print.

The original Panasonic Varicam was gorgeous. Its color was very delicate, accurate, and somehow filmic, and even though it only captured 720p it outpaced the F900 among owner/operators because cinematographers were drawn to this “non-Sony” look. It had several gamma curves that aggressively compressed highlights, and while they were meant to capture images for post grading we often used them for WYSIWYG imagery because they desaturated highlights and rolled them off in a more filmic manner.

As cameras got better and better, we gained more dynamic range but the image didn’t always look filmic. It wasn’t until the Arri Alexa Classic debuted in 2010 that DPs like myself, who started out in film, felt we had access to a true digital cinema camera. The amazing thing about this camera was that the entire dynamic range was usable: rather than having to protect highlights, or keep flesh tones below 60% on a waveform to prevent hue distortions, we could overexpose or underexpose and it all looked great. What had been technical considerations evolved into artistic choices.

I was fascinated by Alexa’s color, and I became determined to figure out what was going on. I set up a Sony F55 (pre Cine-EI mode) next to an Alexa, aimed them both at a color chart, and rolled through as much exposure range as I could, observing what happened to color saturation as exposure increased.

The Sony F55, in custom mode, saturated colors right up until luma hit 70%, at which point highlights showed dramatic color distortions and the diamond pattern on the vectorscope resembled a shrunken and very unhappy starfish. In Alexa’s case, the diamond pattern on the vectorscope grew until luma hit 35% or so, and then the pattern locked. As exposure increased, colors got brighter but they didn’t get more saturated.

That’s what film does. Unless you talk to a colorist.

A common colorist trick is to desaturate shadows for a film look, which seems backwards to someone like myself who has actually shot film and thinks Alexa does a great job of emulating the “film look.” I like this look, but I didn’t understand for the longest time why I focused on desaturated highlights and they focused on desaturated shadows.

When I realized what was going on, I felt like a bit of an idiot.

I’d been looking at highlight quality because I transitioned to digital at a time when that was its huge weakness. In film I could overexpose highlights, blow out windows, use practicals to light the shot, all because the highlights rolled off gently to a desaturated white. Digital held more shadow detail than film, by far, but it couldn’t handle highlights. Film held highlights, but shadows were often a bit crunchy.

Well… print stock shadows were crunchy. When I started out I worked mostly on features, so I focused on what the print looked like. Print stocks don’t have nearly as much shadow dynamic range as film negative does because it is trying to block as much light from the projector as possible to create rich blacks. Colorists transfer directly from the negative, so they have a very different experience.

I found this 65mm negative from Lawrence of Arabia at the Widescreen Museum. (This is a great website, and if you don’t know about it you owe yourself a long visit.) I brought it into Adobe Lightroom and played around with it a bit:

In film negative, the dark areas are highlights and the bright areas are shadows. It’s interesting to note that most of the color information is in the mid-tones. It drops off pretty quickly in the shadows, because those drop off to transparency.

 

Transparency in a film print represents highlight information. Transparency in a film negative represents shadow information. In both cases, increasing transparency means decreasing color information.

As a film DP shooting for print, I knew that I had less dynamic range in the shadows because print stocks dropped off before they reached deep into the toe of the film curve. They had to, otherwise they couldn’t reproduce a deep black. Colorists don’t have that barrier: they create imagery directly from the negative, and they can shape the bottom of that curve to pull out more detail than I ever had access to in a print. In the same way that a digital camera captures more shadow detail than can be reproduced on a print stock, a telecine can pull more out of a film negative than a print stock can reproduce.

The upshot is this:

  • If you’re used to grading from a film negative, and you’re trying to make a digital image look like what a telecine would see in the deepest shadows of a film negative, you’re going to desaturate digital shadows because color saturation decreases as the negative becomes transparent.
  • If you’re used to viewing print film, you’ll be more interested in desaturating digital highlights, as color saturation decreases as the print stock becomes transparent.

I suspect most colorists use a combination of these tricks. Most cameras, though, focus on highlight desaturation over shadow desaturation. Alexa certainly does this, Panavision’s DXL2 with Light Iron color science does this, and modern REDs do this as well. It makes sense to emulate what a film print stock does because the only way to see a truly filmic look is to view projected film. As this has become somewhat uncommon with the advent of digital projection, it’s less important to emulate film than it is to differentiate between beautifully shot digital images and the traditional video look. From that perspective, either technique meets that need.

I have worked as a paid consultant to both Sony and Arri.

I’ll be teaching two Arri Academy classes in mid-March, 2018 in Brooklyn and Chicago, with another class in Vancouver, B.C. in April. If you’d like to learn the ins-and-outs of Arri cameras, accessories, color processing, LUTs, recording formats, etc., this is worth checking out. For more information about upcoming classes, or to request a class in your area, please contact the Arri Academy coordinator.

Art Adams
Director of Photography

 

 

 


Support ProVideo Coalition
Shop with Filmtools Logo

Share Our Article

Art Adams is a cinema lens specialist at ARRI, Inc. Before that, he was a freelance cinematographer for 26 years. You can see his work at http://www.artadamsdp.com. Art has been published in HD Video Pro,…

19
Leave a Reply

avatar
8 Comment threads
11 Thread replies
4 Followers
 
Most reacted comment
Hottest comment thread
11 Comment authors
GaetanoChris Wrightbobby movie for pcNeco result 2018Quentin Brown Recent comment authors
  Subscribe  
newest oldest most voted
Notify of
Tim Guy
Guest
Tim Guy

Great stuff – thank you Art! That has conclusively answered my original question. 🙂

Art Adams
Guest

Well, thanks for getting me thinking about it. 🙂 It surprised me to realize what was going on. I’ve never heard anyone else talk about this. People get hung up on techniques sometimes without asking where they came from. I’m always curious in the “Why?” as much as the “How?”.

bobby movie for pc
Guest

yea he really addresses problem with this article

Jamie LeJeune
Guest

Great article Art! It would be amazing to have one of those classes in San Francisco. There is a complete lack of advanced training in the Bay Area for some reason, even when compared to what I would assume are smaller markets. I sent an email to they address you linked to, but if you yourself have any impact on where they plan them, please let them know we need some Bay Areas classes : )

Art Adams
Guest

I’m guessing that Arri would encourage people to come down to Burbank for training, but if there was enough demand they’d do one locally. It would be nice to do a class so close to home.

Jamie LeJeune
Guest

Yes, that’s basically what they replied with when I contacted them. I wish they would hold a class here, but it doesn’t ever look likely. They’re certainly not alone in the industry for treating SF as if it’s just a suburb of LA instead of a market in its own right. I just need to do what everyone else has done and up and move to LA myself. Even most clients prefer to hire from there, even when the shoot is in SF. It’s crazy.

Bruce Schultz
Guest
Bruce Schultz

Art, I always saw this film vs digital as the difference between shooting on film positive (reversal) and film negative stocks – film being a negative (for the most part) stock and digital being a positive/reversal profile. I used to get 35mm strips rolled onto 35mm stills rolls from a lab in LA and get back from the lab both positive (slides) and negatives. I could see the difference you describe above when I later brought those negatives into Photoshop for archival process. There is a simple trick in DaVinci Resolve to take saturation out of the shadows using Lum… Read more »

Art Adams
Guest

Actually, Arri seems to emulate print film more. The shadows are really saturated, and it’s the highlights that roll off. That’s why I had to figure out why colorists are always talking about “the film look” being desaturated lowlights, when print film and Alexa do the opposite. I don’t think the A-Team used reversal. As best I know, none of the episodic shows from that era used reversal. They shot on negative, but the difference is that they transferred from a print or interpositive, as the producers didn’t want their negative to run through a machine with sprockets. Once sprocketless… Read more »

cpc
Guest

Shadows ARE saturated on a negative! The reason is exactly the same as the reason highlights are desaturated. Highlights desaturation is a product of the shoulder, shadows saturation is a product of the toe. After you print, you increase both effects — negative toe combines with print film shoulder and increases saturation in the blacks further; negative shoulder combines with print toe and decreases saturation in the highlights further. The main reason colorists desaturate shadows in the grade is to achieve deeper blacks, which means a perceptually denser image. This is literally the only way to pull shadows down to… Read more »

Art Adams
Guest

>Shadows ARE saturated on a negative! I don’t believe I said shadows weren’t saturated on a negative. I believe I said they lost saturation as they approached the toe (transparent negative), just the same as in a print—or exactly as you describe in your comment. >The main reason colorists desaturate shadows in the grade is to achieve deeper blacks, which means a perceptually denser image. Interesting. I’ve never heard a colorist actually articulate that. >This is literally the only way to pull shadows down to the black level without crushing a channel. I don’t think that’s as much of an… Read more »

cpc
Guest

Well, technically any color that has a channel landing above the toe, with at least one of the other channels below the toe will be (over)saturated — non-dominant channels are pulled down by the toe which increases the difference between channels. Similarly, any color that has a channel above the shoulder with a least one other channel below the shoulder will be desaturated — dominant channels are pulled down by the shoulder which decreases the difference between channels. This is a simplification, and “channel” is a bit of a misnomer in the film case, but this is the basic principle… Read more »

Art Adams
Guest

This is great information, thank you. Let me see if I have this right: In a film negative, a “color channel” that extends into the toe will desaturated as density diminishes, leaving the other colors dominant. Similarly, a “color channel” that extends into the shoulder will be desaturated as density increases, leaving the other colors dominant. It’s very interesting that you still see unbalanced shadows in lower-end cameras. I guess I’m working with the right cameras. 🙂 I rate a lot of cameras slower than their stated native ISO as I find many too noisy, and I suspect that might… Read more »

cpc
Guest

It is probably better to avoid using “saturate”/”desaturate” when talking about individual channels. I mean, these terms do have engineering (and mathematical) meaning when applied to any signal, but in trichromatic color “saturation” is really about the relations between channels. Maybe a more intuitive way to think about what toe and shoulder do to saturation is by considering contrast. The toe increases contrast in the shadows, the shoulder decreases contrast in the highlights. Increasing contrast increases saturation as the differences between channels are exaggerated, decreasing contrast reduces saturation as the differences between channels are diminished. Does that make better sense?… Read more »

Quentin Brown
Guest
Quentin Brown

I for one would be very interested in a deeper article here – I’m a DIT and slowly learning more as a Colourist. Anything that furthers our understanding and can inform our creative practice is very useful – for Cinematographers, DIT’s, colourists and all those in between. Art’s ability to illustrate and summarise the highly technical geekery in terms that the wider community of image makers can relate to would be really helpful here. There seem to be few easily traversed bridges into this highly technical and abstract world of deep image science that you are touching on here cpc,… Read more »

petrus
Guest

Wow it feels good to be am satisfied with the answer i got… Thanks for this article

mile
Guest

Great article Art! It would be amazing to have one of those classes in San Francisco. There is a complete lack of advanced training in the Bay Area for some reason…

Neco result 2018
Guest

yea the article is great

Chris Wright
Guest
Chris Wright

unfortunately, this also means that a color science lut won’t work to match arri to other cameras. its an electronic non-linear color compression.

Gaetano
Guest
Gaetano

>> The shadows are really saturated, and it’s the highlights that roll off.

If I shoot raw can I mimic that “perceptual attribute” with a master curve (RGB) by pushing density in the shadows while leaving midtones and highlights as is? The curve steepness in the shadows should push saturation in that area. I can then bring down saturation globally as that will affect highlights the most.

Thanks