CallaghanFilms
Veteran
:beer:
"Film is Finished" article
- Thread starter vanvideo
- Start date
Didnt read the article as I'm not that interested after seeing the same headline 1000 times since 2007. If someone handed me a budget - I'd shoot film.
Why?
I wouldn't.
Actually, Nolan has said, "We save a lot of money shooting on film and projecting film and not doing digital intermediates. In fact, I've never done a digital intermediate." That was from this article back in Spring of 2012. I get the impression that he wouldn't even use any digital scan of the film if he could get away with it. But obviously, he's had to have parts of his films digitized for special effects.
As for the cookie comment he made, I seem to remember that he said more than the quotation in another response includes. The impression I got was that digital is like those cookies that say "chewy" on the package, but upon tasting them, you discover that they're just a bunch of synthetic garbage that will probably kill you and doesn't actually taste very good.
visugeek
Well-known member
I think it depends on the person looking at the painting. Artists know, the difference between Acrylic and Oil, but many in the general public don't. Most digital is not on par with film, only the higher end cameras that have enough dynamic range can substitute for film in every situation. The public may not say "That's film, or that's digital" but they may prefer the look of one over the other and not know why.
visugeek
Well-known member
Why?
combatentropy
Veteran
You can make digital look just like film. It's two steps:
1. When shooting, don't overexpose.
2. When grading, give it an S-curve.
For all you lawyers out there who can't absorb the gist of something unless it's qualified and hemmed and hawed into oblivion, yes, the devil is in the details and there is a great deal of finesse needed in each of the two steps.
1. When shooting, don't overexpose
I've come to believe that all these years almost all video has been overexposed. To understand why, you need to understand the second step, about the S-curve. But what I mean is that if you think of the entire dynamic range of a video sensor, we have been setting middle gray too high up the scale when shooting. We have been doing this because we have been shooting on cameras that bake in a look that is optimized to reduce noise, or really have none at all. They make the lower f-stops really dark to hide the noise. So naturally we open the iris more to expose faces well. But then fewer f-stops are left for highlights. Hence two of the complaints about the video look: (1) really, really clean, no grain, (2) hard-clipped highlights. These both spring from the same setting.
The look of film depends a lot on the S-curve, which eats a lot of f-stops. So it helps to use a camera with a wide dynamic range, like the Sony F3, with its 13 stops. But a lot of the f-stops that the S-curve eats are the ones above middle gray. Creating that long tail fade into highlights is expensive, dynamically speaking. It lifts the rest of the levels. So you actually want to start with an exposure that puts the mid range below the half-way point, maybe the one-third point, because like I said when you apply the S-curve in post it's going to lift everything. If you had the midtones already at the halfway point, then they would become too bright.
A final result of applying curve to the highlights and lifting the shadows is that you will see a bit more noise. But sensor noise looks quite a bit like film grain, especially when it's mild, as would be the case with modern sensors using this technique. It doesn't add a lot of noise, but more than is typical for the default settings of most digital cameras. Since grain is one of the things film lovers say they miss with video, maybe it's not such a bad thing. Since some film filters are even adding grain, maybe it's even killing two birds with one stone!
2. When grading, give it an S-curve
Film's S-curve is a happy accident. It turns out it's a perfect way, in my opinion, to tranfer a real-life image with real-world brightness ranges (very, very wide) to a television or movie screen which, for the sake of the audience's comfort, has a more limited brightness range. I believe you actually want to boil down a 14-f-stop real-life scene to a screen dynamic range of 6 or 7. It's so that the audience doesn't squint. Anyway, there are several ways to reduce a scene from 14 to 7. One would be to simply cut off the upper or lower stops. This would yield a hard clip, which we all know and hate. Another is a very low-contrast image. It would look like the subject is bathed in milk. Finally, you could do an S-curve, which means you have low-contrast lows, slightly-high-contrast-mids, and low-contrast highs. The genius is that the midtones, where faces and other focii of attention are --- the content of the scene --- is rendered in a little bit higher than normal contrast, causing it to pop. It's one of the main reasons why film looks so good, because what we want to see is easy to see. And the tapering lows and highs are a graceful way to show detail in those areas without requiring your TV to be as bright as the noonday sun. So, I like the S-curve, and I think all those people trying to make themselves like the way video renders life, calling it the New Aesthetic, are misguided and causing themselves needless pain
1. When shooting, don't overexpose.
2. When grading, give it an S-curve.
For all you lawyers out there who can't absorb the gist of something unless it's qualified and hemmed and hawed into oblivion, yes, the devil is in the details and there is a great deal of finesse needed in each of the two steps.
1. When shooting, don't overexpose
I've come to believe that all these years almost all video has been overexposed. To understand why, you need to understand the second step, about the S-curve. But what I mean is that if you think of the entire dynamic range of a video sensor, we have been setting middle gray too high up the scale when shooting. We have been doing this because we have been shooting on cameras that bake in a look that is optimized to reduce noise, or really have none at all. They make the lower f-stops really dark to hide the noise. So naturally we open the iris more to expose faces well. But then fewer f-stops are left for highlights. Hence two of the complaints about the video look: (1) really, really clean, no grain, (2) hard-clipped highlights. These both spring from the same setting.
The look of film depends a lot on the S-curve, which eats a lot of f-stops. So it helps to use a camera with a wide dynamic range, like the Sony F3, with its 13 stops. But a lot of the f-stops that the S-curve eats are the ones above middle gray. Creating that long tail fade into highlights is expensive, dynamically speaking. It lifts the rest of the levels. So you actually want to start with an exposure that puts the mid range below the half-way point, maybe the one-third point, because like I said when you apply the S-curve in post it's going to lift everything. If you had the midtones already at the halfway point, then they would become too bright.
A final result of applying curve to the highlights and lifting the shadows is that you will see a bit more noise. But sensor noise looks quite a bit like film grain, especially when it's mild, as would be the case with modern sensors using this technique. It doesn't add a lot of noise, but more than is typical for the default settings of most digital cameras. Since grain is one of the things film lovers say they miss with video, maybe it's not such a bad thing. Since some film filters are even adding grain, maybe it's even killing two birds with one stone!
2. When grading, give it an S-curve
Film's S-curve is a happy accident. It turns out it's a perfect way, in my opinion, to tranfer a real-life image with real-world brightness ranges (very, very wide) to a television or movie screen which, for the sake of the audience's comfort, has a more limited brightness range. I believe you actually want to boil down a 14-f-stop real-life scene to a screen dynamic range of 6 or 7. It's so that the audience doesn't squint. Anyway, there are several ways to reduce a scene from 14 to 7. One would be to simply cut off the upper or lower stops. This would yield a hard clip, which we all know and hate. Another is a very low-contrast image. It would look like the subject is bathed in milk. Finally, you could do an S-curve, which means you have low-contrast lows, slightly-high-contrast-mids, and low-contrast highs. The genius is that the midtones, where faces and other focii of attention are --- the content of the scene --- is rendered in a little bit higher than normal contrast, causing it to pop. It's one of the main reasons why film looks so good, because what we want to see is easy to see. And the tapering lows and highs are a graceful way to show detail in those areas without requiring your TV to be as bright as the noonday sun. So, I like the S-curve, and I think all those people trying to make themselves like the way video renders life, calling it the New Aesthetic, are misguided and causing themselves needless pain
Last edited:
Because digital cinema has come of age. I've worked in video my entire career, with some forays into film. I am comfortable with the medium. I would be quite happy working with Alexas or Epics, or with the new F55, or with whatever new digital cinema camera that comes out. They are more than capable of making successful motion pictures, producing excellent images.
If I were given the big budget necessary to produce a major film, I would, without hesitation, choose digital cinema.
Last edited:
Or, more likely, they won't notice and won't care.
I've heard this same mantra before, and the success of digital cinema movies says otherwise. Nobody I know made a single complaint about how The Avengers or Skyfall looked because it wasn't shot on film, as an example. This includes people who work in the industry. And for those who did notice - well, it simply didn't matter. Both of these movies were wildly successful, each making over a billion dollars. What sold these movies to the public wasn't which medium was or wasn't used, but story and content. As it always has been and always will be.
You may believe the general public has some sort of subconscious perception as to which medium they prefer, but I do not.
Digital cinema is here now and is the future of movie making. It is inevitable and final. And I personally couldn't be happier.
If audiences prefer one over the other they are keeping very quiet about it. They will ***** about actors, lack of a good plot, it was to long or to short, poor special effects, the script didn't match the book, but I haven't seen any discussions by mainstream audiences outside of film maker blogs about the look being digital vs film, or just not pleasing in some way, on Hollywood productions like the Social Network or Girl with the Dragon Tattoo. No comments like I don't know why, it just didn't look right, or as good as X movie, etc., Nothing.
If the public didn't like it believe me they'd let you know. It would be all over Rotten Tomatoes and other sites.
2. When grading, give it an S-curve
Film's S-curve is a happy accident. It turns out it's a perfect way, in my opinion, to tranfer a real-life image with real-world brightness ranges (very, very wide) to a television or movie screen which, for the sake of the audience's comfort, has a more limited brightness range. I believe you actually want to boil down a 14-f-stop real-life scene to a screen dynamic range of 6 or 7. It's so that the audience doesn't squint. Anyway, there are several ways to reduce a scene from 14 to 7. One would be to simply cut off the upper or lower stops. This would yield a hard clip, which we all know and hate. Another is a very low-contrast image. It would look like the subject is bathed in milk. Finally, you could do an S-curve, which means you have low-contrast lows, slightly-high-contrast-mids, and low-contrast highs. The genius is that the midtones, where faces and other focii of attention are --- the content of the scene --- is rendered in a little bit higher than normal contrast, causing it to pop. It's one of the main reasons why film looks so good, because what we want to see is easy to see. And the tapering lows and highs are a graceful way to show detail in those areas without requiring your TV to be as bright as the noonday sun. So, I like the S-curve, and I think all those people trying to make themselves like the way video renders life, calling it the New Aesthetic, are misguided and causing themselves needless pain
Maybe this is why the scene in teh Zactuto shootout looks so bad. It's lit for 14 stops, and all the camera operators testing their cameras are TRYING to get 14 stops to show off their cameras capabilities. It's not a pleasing look at all.
I think the same goes for HFR, projecting at 48 frames. We've had that look for many years, it was called SD video. No one liked it then, I have no idea why we'd suddenly like it if it's done on a RED. That's why everyone was trying so hard to get 24p video, because it's very pleasing to the eye unlike 60i. Projecting at 48fps just makes it look like really crisp 60i video.
androbot2084
Carbonite Member
4K progressive 48 FPS does not look like standard definition interlace video.
Well neither did 24fps at SD resolutions with a 3:2 pulldown so your point is moot.
Kenneth Fisher
Active member
We'll know when film is finished, because people will stop declaring it.
androbot2084
Carbonite Member
It's the film look that audiences are after. If 35mm film were shot and projected at 60 frames per second audiences would complain about it looking like video.