Ask David Mullen ANYTHING

[Davide B.]

Hi David,

Thanks for your advice on the scene I'm planning. Once it's done I'll upload screen shots to show it went. I was also wondering if you had read Paul Schrader's book, "Schrader on Schrader." And if so what did you think of this claim he made in the chapter on film noir:

"Because film noir was first of all a style, because it worked out its conflicts visually rather than thematically, because it was aware of its own identity, it was able to create artistic solutions to sociological problems. And for these reasons films like Kiss me Deadly, Kiss Tomorrow Goddbye and Gun Crazy can be works of art in a way that gangster films like Scarface, Public Enemy and Little Caesar can never be."

Do you agree that the core meaning of film noir film lay in their cinematography just as much as, if not more than, in their narratives? Do you consider cinematography to be something that complements and emphasizes a film's story or shapes it and gives a it different meaning than the bare script could ever have had? When you shoot do you think of yourself as representing the story visually or trying to give, as Schrader put it, "artistic solutions" to the themes raised by the script?

I guess one of the movies that makes me wonder about this is: He Walked By Night. The story of that film is pretty much no different than something you would get on Dragnet. But the visual style of Anthony Mann and John Alton on that film is far different than any TV show. Do you think that gives the film's story a different message than if the script had been turned into an episode of Dragnet?

thanks,
davide

 

[David Mullen ASC]

David,
If (or back when) you face the situation of shooting near window(s), and needing to balance to tungsten for interior, what type of gel do you like on the windows? 1/2 CTO? 3/4 CTS? Full? Any ND?

Full CTO would convert daylight to tungsten. How much ND just depends on how much I had to darken the window.

I never find the color of any orange gel on the window to look right for matching to tungsten. I usually use HMI's. On my last movie, I did gel a workshop with Full CTO on the window but that was because I was lighting inside with Cool White fluorescents and I wanted them to look blue-green compared to the windows so I needed a 3200K base for "white" so the 4800K Cool Whites would look cooler in comparison.

 

[David Mullen ASC]

Hey David,
Our opinions on "The Diving Bell and the Butterfly" seem to correspond. What fantastic cinematography, especially in serving the story line! One of the things I really loved was the camera operation, in particular the car shots, and in particular the car shots when he was driving with his son near the end of the movie. It was an interesting mix of mechanical ability and human fluidity. Do you know anything about the rig they used to achieve that mix?

I don't know what rigs they used for driving scenes. I'm out of town about to shoot a feature, so I don't have access to the American Cinematographer article to see if they said anything about that.

 

[David Mullen ASC]

Do you agree that the core meaning of film noir film lay in their cinematography just as much as, if not more than, in their narratives? Do you consider cinematography to be something that complements and emphasizes a film's story or shapes it and gives a it different meaning than the bare script could ever have had? When you shoot do you think of yourself as representing the story visually or trying to give, as Schrader put it, "artistic solutions" to the themes raised by the script?

I guess one of the movies that makes me wonder about this is: He Walked By Night. The story of that film is pretty much no different than something you would get on Dragnet. But the visual style of Anthony Mann and John Alton on that film is far different than any TV show. Do you think that gives the film's story a different message than if the script had been turned into an episode of Dragnet?

The semi-Expressionist lighting in film noirs suggests hidden psychological states of mind that reflect the story and setting, so yes, I think it's a genre where the meaning is partially driven by visual style, or at least, it is enhanced and made more clear by the visual style. I'm not sure there is definitely a hard line between noirs and gangster films, nor that one can be art and the other not (were "The Godfather" movies gangster or noir films? Aren't they artistic?)

The Expressionist style suggests a broader meaning to the small crime story in a noir, that we are looking at primal human behaviors and perhaps even seeing elements of spiritual issues like Good vs. Evil. Maybe because to some degree, Expressionism reminds us of classic fairy tales, like by the Brothers Grimm. Or also that the darkness used is both obviously visual but also symbolic of the dark parts of the human soul.

 

[Jim McKinney]

I don't know what rigs they used for driving scenes. I'm out of town about to shoot a feature, so I don't have access to the American Cinematographer article to see if they said anything about that.

American Cinematographer had a lot of good stuff on the lenses and in camera techniques used, but very little about the camera movement - none about the car rig.

FWIW, I rarely even check my email when I'm on a feature. That you still take the time to answer our queries here, while you're working those kind of hours . . . . Well, I'm just blown away. Thanks for all that you do for the worldwide cinematography community.

 

[Mike S. M.]

Hey David, a question about the F900's HyperGamma curves. Between the 4 of them, which ones do you recommend for what applications, situations? Thanks.

 

[David Mullen ASC]

I haven't used them -- Hypergamma was a feature added to the F23 and the F900R, but I don't recall it being an option in any F900 I ever used.

On the F23, the Hypergamma feature allows extended dynamic range without getting into the flat LOG emulation, so how aggressive of a gamma manipulation you want to do depends on what look you want and whether it is for a film-out (but don't want to use Sony LOG) or broadcast.

I don't think there's a free lunch in video, so generally the more aggressive a gamma manipulation is in terms of squeezing more dynamic range out of the image, the more likely you'll start to see things like noise in the shadows. So if this is just for broadcast work, I'd probably use one of the less strong Hypergamma choices, though I haven't studied them in detail. It probably also depends on how you will be color-correcting in post, and in terms of scene content.

Like I said, I haven't used it myself yet.

 

[Luc Montpellier csc]

Hi David,

I'm a Toronto based DP. I'm about to start a round of tests on the camera and was wondering what your thoughts are on the fact that you're not able to monitor a live RAW image on set like the Viper and Genesis... I find this to be a fundamental flaw in working the RED into a traditional production environment. I've shot a series on the Viper, and I found myself constantly referring to the live FILMSTREAM image on set. Do you or anyone know if this is being considered in future firmware versions?

Thanks for your time.

 

[Bruce Allen]

Luc, the idea is that future firmware versions you can view using whatever LUT you want (including one that shows the whole RAW range). I think you would create the LUT in RedCine or RedAlert and then upload it to the camera. Hasn't been implemented yet though.

David, thank you so much for this thread BTW.

Bruce Allen
www.boacinema.com

 

[David Mullen ASC]

Hi David,

I'm a Toronto based DP. I'm about to start a round of tests on the camera and was wondering what your thoughts are on the fact that you're not able to monitor a live RAW image on set like the Viper and Genesis... I find this to be a fundamental flaw in working the RED into a traditional production environment. I've shot a series on the Viper, and I found myself constantly referring to the live FILMSTREAM image on set. Do you or anyone know if this is being considered in future firmware versions?

To be accurate, the Viper and Genesis do not produce a RAW image -- RAW means an image from a Bayer-filtered single-sensor pre-conversion to RGB. The Viper is a 3-CCD camera and the Genesis has an RGB striped filter. Besides, you can't really look at a live RAW image anyway, by definition, it needs to be converted to RGB to be a viewable color image.

And the natural output of the Viper's three CCD's is somewhat linear -- what you get with the Viper in Filmstream mode is a minimally-processed HD signal (no options like gain or Rec 709 color matrix or white balance) that has been gamma encoded to produce a LOG emulation, in the tradition of a Cineon Log scan of film. The Genesis also produces its own LOG emulation (Panalog) and the Sony F23 has their own version. All of those LOG emulations are 4:4:4 10-bit 1080P.

The Arri D20 is a 2.8K RAW camera but can produce a 4:4:4 1080P signal with a couple of choices for LOG emulations.

Since all of these cameras are designed to create a 10-bit LOG signal (and again, LOG is not the same as RAW), you have the option of viewing this flat-looking LOG image on a set monitor, or running it through a LUT for proper monitor viewing in gamma-encoded Rec 709.

The RED records RAW but creates a gamma-encoded Rec 709 8-bit 720P 4:2:2 signal for monitoring purposes.

I think what some people have been asking for is a LOG output for the RED, with the notion that this milky, flat-looking image would be somehow more indicative of the dynamic range being recorded in RAW (and then have the option of applying whatever LUT's they want for manipulating the live monitor image.)

While I'm not sure that it has to necessarily be in a LOG format, perhaps some sort of monitor output that is very low in contrast so that the recorded detail at the extreme ends of the RAW capture are visible would be useful.

 

[Charles Angus]

David,
How do you feel about new DP's who may never have used film learning to light by looking at a monitor/wvfm? Do you think that lighting to a monitor can impair the development of the eye required for the job? Or is it just the new way, with metering the key and lighting be eye going the way of the dodo? Just looking for your thoughts on the differences in learning to light differently.

Sincerely,
Charles Taylor

 

[David Mullen ASC]

I don't think it matters much whether you learn by using a light meter or a waveform (another type of light meter) or just a monitor, though any DP has to be comfortable with a meter -- they may have to scout a location or pre-light a set without the camera being there.

Learning to expose is a subset of learning to light, and ultimately you are learning how to use the best tool you have, your eyes. A meter can't tell you whether to light soft or hard or from high or low or from the side or make it warm or cold... Ultimately a meter reading should just be a confirmation of what your eyes are telling you.

Now, if a process is particularly poor in dynamic range (like cross-processed reversal) then metering becomes more important because you have such a limited range to work in.

 

[Policar]

David,

I'm shooting a short film that contrasts a horror movie aesthetic (slasher films) with a teen comedy aesthetic (Clueless). In terms of mis-en-scene and camera movement/framing I've figured out mostly what I want to do, but I need some help with lighting.

My kit (which I can expand if necessary) is going to be a couple 650s, a few 300s, a few inkies, china lanterns, a couple 1ks (with medium softboxes), two kinoflo 4-banks, a firestarter or two if necessary (never used them, though), and a 1.2kw HMI. The shoot is 95% interiors, a mix of day and night. I may rent a Honda generator, too.

Anyhow, for the horror scenes (95% night interiors) I figure I'll do a grainy/de-saturated warm look with a nearly nonexistent (but very cool) fill. I am thinking of using undiffused fresnels for key and backlight and a 4800K china lantern for a very little bit of fill? (One character retains a "comedy" look so I'm going to hit him straight on with a 650w softbox.) But I prefer soft light. Can I get away with using big soft sources and relying on a lack of fill for a creepy dark look? Also, should I bring my backlights a stop or two higher? Are fog machines good for fill? Any tips for creating "pools" of light? Any movies that really embody the slasher movie aesthetic that I should watch first?

For the high key scenes, how do you light that? Movies like clueless have no shadows whatsoever. Is it just a wash of kinoflos facing straight on? Or is it a big top source? Can I emulate this by bouncing an HMI off the ceiling? But wouldn't this result in shadows under the eyes? Also, how much should I mix color temperature in each setting?

And quickly: Do you know if Bringing out the Dead used ENR or bleach bypass? I thought that movie looked great and much better than Casino, though maybe I just saw one on DVD and one on laserdisc.

Also, I think (but am not sure) that CTO3/4 corrects as much as CTB and that CTO (full) corrects for a bluer-than-5600K source.

And thanks in advance for the help. This thread is great.

 

[Tom Lowe]

David, did you see Michael Clayton? Thoughts?

 

[David Mullen ASC]

"Bringing Out the Dead" was shot in 35mm anamorphic with net diffusion and used an ENR process on the prints; "Casino" was shot in Super-35 with spherical lenses, blown-up to anamorphic -- I don't know about the prints. ProMist diffusion was used.

Full CTO is designed to convert daylight to 3200K tungsten but it all depends on how you define daylight, which is reality is all over the map. "Photographic" daylight, according to Kodak, is 5500K.

Full CTB is designed to convert tungsten to daylight.

If you look at MIRED shifts, 5500K is 182 (1,000,000 divided by 5500) and 3200K is 313.

So to shift daylight to tungsten or the other way around, you need a MIRED shift value of 131 (+131 for orange gel and -131 for blue gel.)

According to the Lee swatchbook (and Rosco doesn't necessarily use the same values) Full CTO has a MIRED shift value of +159 and 3/4 CTO has a value of +124. Full CTB has a MIRED value of -137 and 3/4 CTB has a value of -113.

So you could say that Full CTO is too much and 3/4 CTO is not enough to shift 5500K to 3200K. But the truth is that daylight is a mix of warm and colder light, and HMI lighting can often drift much bluer than 5500K, often more like 5600K to 6000K, so they made Full CTO more orange than +131.

As for horror lighting (about to start shooting a horror film myself) you have to be more specific about the light sources you are emulating -- hard moonlight coming through a window, etc. Unless you are shooting b&w and are not so concerned about creating realistic source lighting. But in color these days, you need to start with a grounding in what the natural light would be before you start adding to it. In terms of soft lighting, it's all about flagging it down off of the walls to create more contrast, and using less fill. And coming from a side or 3/4 angle rather than frontally.

Fill is a case-by-case situation as to where it can come from, whether it's a bounce or a diffused light or a light on the lens, etc. You're not really supposed to notice fill.

"High key" just means that there is a predominance of highlights in the frame whereas "low key" means there is a predominance of shadows, underexposed areas, in the frame. Either style can follow natural source lighting -- for example, a bright fluorescent-lit supermarket is naturally high-key.

I liked the photography of "Michael Clayton" quite a bit; moody but natural, low-key.

 

[Deanan]

Full CTO is designed to convert daylight to 3200K tungsten but it all depends on how you define daylight, which is reality is all over the map.

Literally all over the map because daylight varies depending on location.
5600k correlates to midsummer day at high noon in Washington, DC.

 

[Policar]

Thanks for the extremely informative reply. FWIW, Rosco CTO3/4 and CTB (full) both have MIRED values of 131 so while I was mistaken in this being the reason for CTO looking funny or overcompensating, you can see why I'd bring it up.

I should have been more specific re: horror lighting. The fill is from moonlight through a window, the key lights are motivated by lights around a kitchen, the main location in the film. What you say re: soft sources needing flags makes a lot of sense and since I don't have the budget/crew to set up multiple flags quickly, I'll stick to undiffused fresnels, I think.

But just reiterating: in Clueless and movies like that, do cinematographers generally use a strong frontal source (kino wash) or a top source (chimera on ceiling; hmi bounced off ceiling). I have the motivated light sources already pretty well defined and can probably justify either look. The film "quotes" specific genres so I want to nail the look of each.

 

[David Mullen ASC]

Soft sources like from Chimeras and Kinoflos naturally contain some spill and can use egg creates to control spill further, so are good when you don't have a lot of flags.

But I'd say that carrying some flags and c-stands is sort of mandatory in the case of creates interesting areas of bright and dark on a set.

Unless it is a really low-level source like candlelight or other dim sources, it's a bit odd for me when moonlight is bright enough to register as fill on a room with the lights on, unless the lights go out in the scene so you are transitioning from a lamp-lit to moonlit space.

There is no standard approach for a movie like "Clueless" -- generally the DP creates a big soft key light, how big depends on the scene and the space and the camera move, same for whether it comes from overhead or the sides. Even on a movie like that, there is some logical justification -- is the source a soft overhead ceiling light or sky light? Is it from a floor lamp? A window? Could be a 12'x12', 6'x6', 4'x4' soft key, just depends. How much fill they use, how they do the fill, whether they use a Kino or not, it just depends, but fill is also generally soft. The only time fill is not soft is when it's a weak, small eyelight over the lens. Otherwise, you generally don't want to notice the fill light casting a shadow. The fill is sort of a mechanical light designed to control contrast and should be subtle. Otherwise, if you key is soft enough and wraps around the face enough, fill may not be necessary. Also, in some light-toned rooms, there is enough ambience from the keys bouncing back to fill in the shadows. Sometimes all you need to do for fill is put a white card opposite the key to bounce a little of it back in.

Rather than throwing a lot of light around in a room, it's better to start out thinking about the space and where the light may be coming from, and start to simplify everything. Too many sources, too many backlights, too many ways of filling, and it all becomes an overlit mess. I feel that in any scene, there should be a dominant source and all other lights are secondary in feeling, if noticeable at all.

 

[Policar]

Thanks again for the help. Due to this being a small student film, I think I'll only have two flags and two c-stands and I am shooting on location, so that was my concern about avoiding the use of too many flags. I don't have egg crates but I'll build some.

And this is all great advice: I'm going to re-read all this a few times before starting on light diagrams.

 

[Rudi Herbert]

Once upon a time in the West

Once upon a time in the West

David,

I just saw a screening of a restored version of the 1968 classic "Once upon a time in the West", showing here at the Miami International Film Festival. Print looked good, a bit of CA on some edges here and there, but for the most part, tack sharp as was Leonne's style. Amazing what a difference it made to watch that film on a big screen (I'm not old enough to have seen projected it originally, so it was all DVD and VHS until now), and suffice it to say that the 2:45 hours went by a lot quicker than in the "comfort" of home. Anyway, my question to you was this: I understand this film was shot on the "techniscope" (or is it technoscope?) format, where basically two frames were squeezed onto a single 35 mm frame, making the frames the equivalent of a 2K resolution in digital terms? Or pretty close actually to the S16 resolution as well? With that in mind, the film didn't exhibit any of the grain inherent to something close to 16mm origination, and in terms of sharpness, I could not tell it was shot using only half the resolution of a 35mm frame. Is this due to the better 35mm stock as compared to 16mm, and also, was the DOF achieved by the lenses the same as it would have been at full 35mm frame or more like 16mm?