Join the discussion about work/life balance in relationships:

June 20, 2012

The Visual Anatomy of a Scene: 'One Click Away' Part 3

This article is a continuation of the series “Anatomy of a Scene.” See Part 1 and Part 2. In this post, I’ll discuss the use of some non-traditional lighting sources that we used for the “One Click Away” project. These low light level sources would require a very light sensitive camera.  First, I’ll discuss our choice of camera (Sony PMW-F3) overall.

My intention is not to spark a debate over what camera is best in all cases – but to illustrate the aesthetic, technical, and financial factors that influenced our decision. I’ve worked with the RED MX, the RED Epic, Arri Alexa, Canon C300, and Canon DSLR cameras. They are all good tools to choose from, depending on the specific needs of a project.  I'm not suggesting to anyone what camera they should use.

Camera Choice

We chose the F3 based on a combination of low light sensitivity, dynamic range, and rental cost. On this project, I knew that there would be situations where working in low light would be advantageous financially and aesthetically.  The F3 has a native ISO of 800 and it is amazingly noise free at ISO’s as high as 3200. Note: the Canon C300 had not been released at the time of our filming. I would now consider it as a viable alternative.  The project was originally budgeted to shoot with an Arri Alexa. This would have been an excellent choice for low light and dynamic range.  However, the Alexa is an expensive camera to rent. As the project progressed in pre-production, some of the original parameters expanded to include a larger cast. This increased costs, but the budget remained the same. As a DP, it’s only natural to want to use the best tools available. However, it’s important to look at the big picture for the sake of the film as a whole.

The Alexa’s 14 stop dynamic range is unbeatable, but I found myself arguing for the less expensive Sony F3 in order to help balance the budget in other areas. For me, the Alexa really excels when shooting sunny day exteriors or day interiors with bright windows that are difficult to control. Two of our project days would be filmed in studio interiors. The location day interiors had dark trees outside the windows and we could adjust our schedule for the best light.  I knew that I would be able to control the lighting well enough to work comfortably within the F3’s 12+ stops of dynamic range.  We also chose not to use the Sony S-log gamma. One factor was that Jason Satterlund, our director, had not graded with it before and we ran out of pre-production time for him to test it.

Sony S-log

I love the extra dynamic range possible with S-log, but it does require more effort with color grading.  Another factor for not choosing S-log was the additional cost and hassle to rent an external recording deck such as an AJA Ki Pro Mini.  While it’s true that one can record to S x S cards in S-log, I had doubts about the color fidelity going that direction.  In addition to creating a flat contrast image, S-log severely de-saturates the color.  XDCAM only records in 4:2:0 color space to S x S cards.  I was concerned about restoring that color loss accurately.  With more time, testing would be the best way to prove or disprove this theory.  Based on our location scout, I knew that I would be able to control my lighting ratios fairly well.  I find S-log most helpful in extreme contrast scenarios such as bright sunny exteriors seen through un-gelled windows.  When lacking these extreme conditions, I’ve had very good luck working with the various Cine Gammas options (Cine 1 - 4) found in the F3 Picture Profile menu. They require less severe adjustments when grading.

One of the scenes in “One Click” called for a little boy (actor Fred Pitman) to appear in front of a giant TV with a channel clicker (3:25).  Jason and I discussed several ways of approaching this challenge including using green screen and a virtual TV.  In the end, we settled on filming it “live” with a projector to help the kid's performance and to see the “real” light reflection on his face.  We rented a portable screen that was literally 16' by 9' in dimension.  We used a HD video projector to rear project the pre-edited video material on the screen.  To make the young boy appear dwarfed by the screen, we shot at a low angle with an extremely wide-angle lens: a Duclos cine modified Tokina 11-16mm zoom lens.  This is where the sensitivity and dynamic range of the F3 worked really well.  By boosting the gain to +9 db (ISO 2000) and selecting Cine Gamma 3, I was able to hold detail on the bright subject areas of the screen and still see the reflected light from the screen flicker on the boys face - even though he had dark skin.

Cine Gamma 3 is fairly balanced between highlights and shadows.  Mid-tones fall at about 50 IRE.  But it offers more dynamic range on both ends than the Standard Rec 709 Gamma setting.  For comparison, Cine 1 setting holds more details in the highlights, but less in the shadows.  Middle gray mid-tones fall lower at about 35 37 IRE.  I use this setting for sunny day exteriors and day interiors with bright windows.  I don't use Cine 2, which is similar to Cine 1 except it clips highlights at 100 IRE.  Cine 1 records highlights up to 108 IRE.  Cine 4 is biased toward more detail in the shadow region, but less in the highlights.  I use this gamma setting more for night interiors or whenever I need more shadow fill.

When we photographed the boy directly in front of the video screen, I liked the starkness of having him silhouetted against the screen with no light on his back.  However, when he walked away, his profile felt too dark. To resolve this, we added a 1200W HMI par “edge light” or “kicker” just to the side of the screen in the direction that he would exit.  This light separated him nicely while maintaining a clean minimalist look.  The final visual step that really made this scene beautiful was done with a mop!  The floor of the stage was painted black, but it was very dusty and dirty from lots of previous foot traffic.  When we mopped a thick layer of water over it, the dingy floor magically became a black mirror – reflecting the video screen and the little boy.  If you look carefully, you can see a few air bubbles in the floor.

When it came time to shoot the boy's reverse shot, we slightly cheated his body to the side to avoid a camera shadow (from the screen) on his face.  With the reverse shot, we radically changed our view to a high angle to look down on the boy.  This of course diminished him and made the screen feel more powerful.  We turned off the kicker light because it would be too frontal and compete with the beautiful shifting pattern of light from the screen.  The boy was blending into the background of the dark floor, so we added a soft backlight using a Kino Flo Diva 400.  I often like using this particular instrument for backlighting because it is lightweight for rigging and can be dimmed to a desired output.  The challenge with positioning this backlight became hiding the reflection of the light in the shinny floor.  Since our camera angle was very steep, the backlight had to also be steep in order to hide the reflection of the light behind the boy's body.  You can still see a little light reflection in the floor, but it doesn't appear to be an obvious “movie light.”

No Movie Lights

One of my favorite scenes was the little girl (actress Deja Fitzwater) sitting in the “dollhouse” (4:07).  Our still photographer, Levy Moroshan, loaned us a dollhouse set that was originally built for a fashion shoot.  The size of the room was large enough to accommodate our little girl, but the claustrophobic scale made her feel trapped.  We wanted to enhance this feeling and symbolize the threat to children from unsupervised exposure to media.  To create a “big brother” feeling, we placed two TV's outside the windows.  With the brightness set at full level, we ran a pre-recorded pattern of analogue video “snow.”  The creepy flickering light from the TV's was augmented by the low angle light glow emanating from the girl's video tablet.  The camera gain was set to +12 db (ISO 3200).  The T stop was 2.6 on an 18 – 80mm Arri-Fujinon Alura zoom lens.  This type of shot (using video displays as light sources) would not have been possible a few years ago without the latest generation of large sensor cameras with clean low light capabilities.

In Part 4 of Anatomy of a Scene, I'll discuss some “in camera” visual illusions that we employed.

[First photo: Levy Moroshan]


Randolph Sellars, Director of Photography and Filmmaker, has over 30 years of experience photographing a variety of projects in 11 countries around the world. He has shot 23 feature films, including The Juniper Tree, which was a Grand Jury finalist at the Sundance Film Festival and was singer/actress Bjork’s first feature film.

Your Comment

18 Comments

Good stuff, as always. I'm really digging this series of posts. Very informative despite the somewhat goofy subject matter of the video.

June 20, 2012

0
Reply
Robert

Robert, glad you're getting something useful from the series. I appreciate your comments.

June 20, 2012

0
Reply

Hi. Nice work. I have a question. do you work with canon 5d mark II? My question it's if it's posible add Dynamic range to the Canon 5d Mark II apart of the cinestyle, and other like styles.

June 21, 2012

0
Reply

Raguel, yes I've worked with the Canon 5D Mark II many times. I'm sorry, I don't know of any secret way to increase the dynamic range that hasn't already been posted on many forums - such as the Technicolor cinestyle. The Technicolor cinestyle does a good job of flattening the image. However, it sometimes introduces more noise than I would like. Also, it requires a very skilled colorist to grade properly. Often, I prefer to use the in camera flat neutral settings. The 5D works very well in low light and has good shadow detail. The challenge is usually holding details in very bright highlights. I could have used a 5D for this project and gotten a good result, but it would have taken more time and effort for some scenes and probably more window gel. When shooting day interiors with a 5D, I try to lower highlights (by putting ND gel on windows) or avoid shooting toward bright windows when gelling isn't possible. Scouting locations and trying to schedule for good light and avoiding bad light is always helpful - but even more important when shooting with any camera with lower dynamic range. It's all relative and doable with planning and creative problem solving. I remember when film only had about 10+ stops of usable dynamic range.

June 21, 2012

0
Reply

Great Job, and little do people know how this subject is undermining this country. The grabbing hands was very well done. Hiding the rest of the actors was seamless. I didn't realize the color grading was simulating "in front of the screen" until I watched it over after reading your description (compliment). This article was VERY informative. Thanks so much for sharing.

June 21, 2012

0
Reply

Daniel Lee, thanks for the compliments. Glad you liked it.

June 21, 2012

0
Reply

Wow, superb Randolph. You are very generous. Thank you, really, I mean it. I have been looking for a real deep analisis of a piece on direction of photography for quite a while.
I have 3 questions:
1.- What place (institution, university, school, program, course, etc) would you recommend for having a really really good eductation on direction of photography? I know that it is quite ironic asking here (nofilmschool) but to me really make sense. I have learned a lot here. I´m from Chile (South Amercica) and I really want to improve my skills on that matter, also in color grading. I am looking for something really good, I don´t want to go anywhere just to waste time and have fun, fun is good though.
2.- Do you know about any other web site where I can find this kind of material, analisis, resources on direction of photography, etc?
3.- The whole triangle lightning, with the foam core, I guess that finding that solution and building it was pre-production time, wasn´t? So, in production time you just shoot, am I wrong?
Please excuse my english (not a native speaker)
Thank you

June 22, 2012

0
Reply
Matías

Matias, I'm not up on all of the latest film schools and I only know of them by general reputation - not personal experience. The following schools usually end up on many people's list: University of Southern California (USC), University of California, Los Angeles (UCLA), New York University (NYU), American Film Institute (AFI - post undergrad only). As I understand it, these schools all have strong cinematography training as part of their film production programs. Brooks Institute in Santa Barbara specializes in photography and that includes cinematography. Also, North Carolina School of the Arts has a strong cinematography program. I have two friends who teach there. I'm sure there are many others - do your research carefully. You should know that most of these schools are very expensive. Also, they are very competitive and not easy to get accepted into. Another option to consider is attending a less expensive school and supplementing your "cinematography education" with high quality professional workshops. I believe that formal training can be very helpful - but ultimately it is about shooting a lot of footage yourself. There is no substitute for experience. I also recommend working on professional sets in any crew position possible so that you can observe and learn how pros work. Ideally, working as a grip or electrician will contribute greatly to your understanding of lighting techniques.

I don't know of any other web sites that deconstruct or analyze specific scenes. I'm planning on developing more of this type of material that includes more video of the process behind the scenes. In the meantime, there are lots of free video tutorials online - although the quality varies greatly. Even some of the paid tutorials are not very good. Check out Ryan Walter's website and blog: http://www.ryanewalters.com/ He has some excellent information including videos. His material is excellent quality and he is a very skilled DP - and I admit a friend. Also, check out Art Adams: http://artadamsdp.com/. Select the "student" heading to check out his excellent blog. He has great articles about lighting. The brilliant Roger Deakins has a forum (http://www.deakinsonline.com/forum2/) where he actually answers questions that he thinks are relevant. Great info there.

Yes, during production you just shoot. I built the foam core triangle on my own time prior to the shooting - so I guess you would call that pre-production. It would have taken too long to construct on the set - and taken a crew member away from other duties. During my "official pre-production time" I was mostly scouting locations and conferring with the producer and director. BTW, your English is just fine! Good luck!

June 22, 2012

0
Reply

Again, Thank you so much!!

June 26, 2012

0
Reply
Matías

Thanks Randolph very informative
I have three questions ( Thanks in advanced for Answering )

_Is dynamic range stop different in photography and cinematography ?
As you know according to " zone system "
( http://en.wikipedia.org/wiki/Zone_System )
apparently we have 11 stops from Pure black to pure white so i confused when you said Arri Alexa has 14 stop dynamic range ?

_ " While it’s true that one can record to S x S cards in S-log "
do you mean for recording on different Gamma curve we need Specific memory card ? ( does different logs has different output sizes ? )

and the last one ;-)
using different gamma curve as you used for different lightning set up didn't make matching Scene hard in post ?

June 23, 2012

0
Reply

Hamid, excellent questions. I'll do my best to answer clearly and thoroughly. Responses will be separate posts.
First, the concept of dynamic range is the same for photography & cinematography. When Ansel Adams developed the zone system, it was based on the tonal reproduction capacity of black & white film negative and printing techniques. In his era, one could only reproduce 11 distinct b&w tones (including pure black and pure white) in full stop increments. Today, some color film stocks (Kodak 5219) and some digital cameras (Alexa) can record 14 stops of dynamic range. You can still use a "zone system" to previsualize, but the zone needs expanding to 14. Keep in mind that middle gray (18% gray) is not always in the middle of the zone anymore. Digital sensors tend to capture more zones below middle gray (more shadow detail) and fewer zones above (highlight detail). Some camera manufacturers are starting to publish this specific info. Otherwise testing is the only way to know. Also, keep in mind that we are still limited by the tonal reproduction capacity of our video displays. Most monitors are usually limited to approx 7 stops of dynamic range. Even quality projection maxes out at about 9 stops. That is why color grading is a necessary tool with extended range sensors. A camera might be able to "capture" 14 stops, but if we desire extended detail in shadow or highlight areas, we still need to "squeeze" them in post into the limited range reproduced by our displays. OLED technology is exciting because those monitors can display much more dynamic range. The blacks are beautiful and much more pure.

June 23, 2012

0
Reply

thanks Randolph very useful
another question came to my mind and that's more about color grading. if video displays can show approx 7 stops of dynamic range with which display we can see 14 stops of dynamic range footage for process in post?

and if I'm not wrong, dynamic range in negative is more in white stops and in digital as you said in dark stops, but beside which portion( highlights or shadows) have more dynamic range,
A 35 mm negative have more stops dynamic range or 35mm digital sensor ?

June 24, 2012

0
Reply

Hamid, you should always use the best monitor (in terms of color fidelity and contrast range) for color grading that you can afford or have access to. I'm not aware of any monitor that can display 14 stops of dynamic range. When you view your footage while color grading on a monitor (with approx 7 stops of dynamic range), it may appear as if some of your highlights are blown out. However, if you are able to preserve highlight detail with camera exposure by not clipping anything, then when you are color grading - you should be able to bring some detail back in the highlights by lowering them with your grading tools. How much detail to bring back is subjective based on taste. Sometimes its desirable to have some bits of pure white in a scene. The same is true of shadow detail. If you try to lift shadow detail up too much, you'll start to see excessive noise. Again, it's usually desirable to have some strong blacks - not milky shadows with noise.

You are correct, film has better dynamic range in the highlight areas, while digital is better in the shadows. Kodak 5219 has around 14 stops total of dynamic range and that exceeds most digital cameras (except the Alexa) by approx a stop. However, film is rather disappointing these days when it comes to low light shooting and shadow detail. If film stock is pushed much beyond 500ISO it begins to look grainy and muddy in the blacks. Look at the Zacuto shoot out with the large sensor cameras from last year to see what I'm talking about. To greatly generalize dynamic range advantages if money were no object - if I was shooting a costume period drama with mostly sunny day exteriors and interiors, I would consider shooting film. On the other hand, if I was shooting a horror film with mostly night exteriors and interiors - then I would probably favor shooting with an Arri Alexa.

June 25, 2012

0
Reply

Hamid, responding to question about S-log and S x S cards. The Sony F3 camera can only record internally to S x S cards using a 4:2:0 color subsampling compression. It doesn't matter what type of gamma setting is used including S-log. As far as I know, different gamma selections have no effect on file size. For many people 4:2:0 color space is adequate color quality for most uses excluding green screen work. However, one can use an external recorder with the F3 to record with less color compression (4:2:2) or no color compression 4:4:4 - but file sizes are much larger. For a detailed explanation of color subsampling see: http://en.wikipedia.org/wiki/Chroma_subsampling. The S-log gamma was optimized to work best with 4:4:4. Many have reported that the S-log works well at 4:2:2. I would agree. However, my research indicates that using the S-log with 4:2:0, yields sub-standard color quality when graded. Of course, the luminance tonal range is not affected.

June 23, 2012

0
Reply

Hamid, responding to question about changing gamma curves. Generally, I would not change gamma curves in the middle of a specific scene. That might make it difficult to keep the lighting contrast ratios consistent. However, different scenes in a film may require different looks including the lighting. For instance, some scenes may be more moody than others. For me, gamma curves work hand in hand with existing lighting conditions and the lighting package available to work with. If I have bright windows, I may want to help control the brightness by selecting a gamma curve like Cine 1. Of course if I do that, I may need slightly more fill light in the shadow areas. When using Cine Gammas, I'm counting on color grading to achieve the final tonal contrast that I desire because all of the Cine Gammas will tend to flatten the image in different areas of the gamma curve. I still want to control as much of the contrast range as possible with my lighting choices - but the flattening effect gives me some extra protection for more detail in either the shadows, highlights, or some of both. This doesn't make it any harder to match scenes, it just gives me more options to achieve the final look in grading. FYI, for this project, we didn't make many big changes in the grading - mostly just increasing the contrast slightly and balancing out the color shot to shot.

June 23, 2012

0
Reply

gostei do filme tem uma temática muito forte, fará sucesso. qual das câmaras é melhor canon 5d ou sony fs100?para trabalhar com pouca luz e atendendo que o meu projecto será filmado a noite com cenas com efeitos de luz e sombra e com actores negros e o orçamento é baixissimo.

June 24, 2012

0
Reply
domingos geovete

Thank you for these posts. It's one thing to read articles about cinematographer's choices and (before or after) seeing the films...but its another to get so in depth about it...with more to come, still.

June 28, 2012

0
Reply
Daniel Mimura

Randolph, thank you for going writing up these four posts. It was a great introduction to real cinematography and how it's done.

July 27, 2012

0
Reply
Mike W