“Why would I need a camera test? I’ve shot with this camera so many times before!”
This is something I have heard more frequently than I care to remember—not just from my film students at UCLA, but also from professional shooters who have attended my camera training seminars around the country.
When shooting a project on film, camera tests were a given. They were not something a DP would even have to request. Camera tests were budgeted and scheduled into the preparation period of a movie, even if a DP knew a film stock like the back of their hand. Why bother test them again and again?
The answer is that you are not testing the acquisition medium; you are testing the interrelationship of three separate elements: your imaging device, your subject and your targeted look. Knowing your imaging device is only a small part of being able to achieve your desired look. You need to understand how your source imager will interact with your subjects in trying to achieve said look.
In the days of film, several film stocks (some known to you, and some new to your vocabulary) would be tested through several lenses, with several filters, through several lab processes, to see how they reacted when photographing your actors, in hair and make-up, in their costumes, and possibly even on the actual sets of your final shooting environments. You would then sit with a color timer and evaluate which stocks, lenses, filters and lab processes would get you closest to your desired final look. These principles have not disappeared simply because we are transitioning away from film onto digital acquisition. If anything, the interplay has grown exponentially more important, since our list of variables has easily quadrupled.
Sure, you agree that testing this interrelationship is important. But you also feel that it is simply out of your budget. This feeling is not without foundation—with digital, shooting schedules have shrunk down to sometimes-unreasonable levels. To make matters worse, on a low-budget indie film you will most likely not have access to the actors prior to principal photography, will most certainly not have their wardrobe already fully picked out, and will most likely be shooting in practical locations inaccessible until the day of the shoot.
Does this mean that you should give up on pursuing tests to achieve a specific look? Absolutely not, but if your resources are very limited, you need to be smart about what you test, how you test it and why—and you need to eliminate candidates early on. Don’t go into a camera test having to test 20 cameras. You will quickly get overwhelmed. Narrow your selection of potential cameras to a maximum of three. How? These questions will whittle down your pool of suitors.
- Is this project destined for theaters? Television? Gaming consoles? Web? Portable devices? Usually, the larger the venue, the higher the resolution and color depth you will need to capture, at the lowest compression ratios possible.
- Are there any specific deliverable requirements? Some networks, like Netflix and Amazon Studios mandate ultra-high-definition (UHD) acquisition at the source. Others, like Discovery Channel, will have certain specifications on codecs and color space. Make sure your camera can deliver on these requirements if these are your targeted outlets.
- How much storage/bandwidth/computing power can you afford? The larger the raster size (i.e. the actual pixel dimensions of the frame you are recording), and the lesser the compression of your codec, the more storage you will need (both near-line, off-line and archival), the faster that storage will have to be, and the more computing power you will need to handle that media all the way through post (especially important if you will be handling the editorial and finishing process yourself).
- Are you buying the camera and accessories or renting? Answering this question alone should cut your selection in half.
- Are you shooting mostly handheld and in tight spaces? Or is this more of a studio/dolly-and-track shoot? Ergonomics play a huge factor in choosing the right camera.
The more specific and realistic you can be about your project, the easier it will be to narrow down the possible candidates. Once you have no more than three, it’s on to the testing phase.
Since money and time are of the essence, test only elements of true concern to achieving a specific look for your project. Here are three big questions and testing methodologies that should serve you well.
1. The Color Test
Does your look rely on a specific color palette? Is your film very saturated and punchy, or does it have a dryer, desaturated vibe? What color tones are you aiming for: pastels? Steely blue? Or perhaps a warm amber look? If a distinct color palette is vital, you must set a lookbook with the help of all your designers (production and costumes) to make sure everyone is on the same page. Armed with this lookbook and palette, you can prepare your color test.
My color tests look very much like skin tone tests, but don’t be fooled, I’m looking for much more than skin tones. I have seen many people only shoot faces, with the excuse that faces are the most important part of a movie, and that’s all you really need to know about the camera.
The dirty secret is that you can get any camera to produce decent skin tones. That’s right, any camera. The issue is not really the skin tones; it’s what happens to the surrounding colors when you balance your skin tones. Every camera has different color filter arrays on their sensors, and by extension, records into a different color gamut. What this means is that every camera will record the spectrum of available colors slightly different. Can they be color timed? Sure. But it’d be intensely expensive and laborious to pull a secondary color correction on every frame of your movie just to get your desired color palette. Why not just shoot with a camera that has a color palette that already closely matches the visual intent of your story?
Armed with your lookbook, bring in friends, family, PAs and stand-ins that have mostly similar skin pigmentation to your lead actors and actresses. Surround them with as many color swatches that match your chosen color palette as you can find. These are quite literally loose pieces of fabric that you can find in the garment district for a few dollars each. Don’t worry about how the costumes will look (since you are not fitting the real actors, anyway) but rather how the skin tones will fare against the color palette you have chosen, given the camera.
With all three cameras shot, it’s time for some lab work. My favorite digital DIY Lab is Blackmagic’s free DaVinci Resolve Lite Software for Mac or Windows. It runs on reasonably cheap hardware and is fairly easy to learn. I first import the footage from my three cameras and organize the timelines by shot size (wide/medium/close-ups). I then use only primary color correction controls to balance the skin tones and make them look healthy. If you cheat by making individual color adjustments via secondary controls (or windows), you’re committing to making those adjustments on every frame of every shot of your movie. Those controls are best left for final color.
Once the skin tones fall nicely into place I move my focus to comparing the surrounding color swatches, and how they differ from camera to camera. When I desaturate the skin tones, do the other colors look too muted? If I make the skin tones warm and healthy, are any of the other colors now overly saturated and losing detail? As I dial in a color look, how are my skin tones fairing? Are the people looking morbid and ashy? Are they looking like orange pumpkins? Looking for the path of least resistance here. Good-looking skin tones are your guiding light; try to get to as close as possible to your ideal color palette while only using primary color correction adjustments. If you can get 70 percent there, it should be a breeze getting the final 30 percent using the rest of the available toolset during final color.
2. The Dynamic Range Test
What kind of image dynamics does your look require? Does your film take place under mostly overcast exteriors, or in a hospital with white walls and costumes and very little contrast? Or is it a horror film with lots of shadows and bright shafts of light, requiring a high-contrast look? If achieving specific contrast is crucial to your look, you need to get a feel for your camera’s dynamic range, or the difference between the brightest and darkest shades it can see.
I have done very elaborate and complex dynamic range testing over the years. It’s easy to get carried away in this department. I don’t always shoot this test—if my film takes place entirely in controlled lighting within a limited tonal range, why waste my time? Most cameras today can safely record nine stops of dynamic range. If your movie calls for affinity in contrast and the lighting can be controlled throughout, I would skip this process. That said, most indie feature films don’t have the luxury of controlled lighting, and often have to deal with very harsh image dynamics.
Schedule this test for when it is brightest outside. Find a window in an apartment or office that looks out onto something with lots of detail. It can be another building, a storefront, mural, anything with fine detail. Bring a friend and ask them to stand next to that window. Not in front of the window, but next to it, so you can see outside the window clearly and have the person to the side in half of the frame.
At this point, try to recreate the dynamics you expect to have to support in your film. Is it 10 stops of dynamic range? 12? 15? Whatever you expect your biggest challenge to be, test it here. This can be done in two ways. My preferred way is using a spot meter. Get a reading outside the window, and get a reading on the person’s face inside with no lights on. See how many stops apart they are. If the face is way darker than the amount of stops you are looking to test, shine a light on the person. This can be a studio light if you have one, or it can be a scoop light from IKEA with a 500W CFL bulb ($7 at Home Depot).
You’re not looking to sculpt a beautiful image here; you’re simply looking to create a lighting ratio. If you don’t have a spot meter, you can count stops using either the zebra stripes function in your camera or viewfinder, a waveform monitor in your camera or monitor, or a dedicated waveform monitor like the Leader 5330.
I then bring that footage into DaVinci Resolve Lite and compare the amount of detail out the window to the detail in the person. I look at how each camera deals with extremes. Are the highlights completely blown out? If there is detail, are there any strange artifacts (aliasing, step-laddering, sensor smear, etc.)? Are the colors shifting as they clip? What about the shadows—are they buried in noise? These elements are crucial to understanding if your look will require fine manipulation of image dynamics. Does this mean that if a camera performed beautifully in the Color Test but has poor dynamics, it is automatically disqualified? No. If the color palette you desire is important enough to select that camera, then you need to work hard to limit the dynamics of that camera to an acceptable range, either by applying ND (neutral density) gels or nets to windows, or by simply shooting later or earlier in the day. This is particularly the case for DSLRs. You can achieve quite beautiful imagery if you can control the dynamics on those cameras.
3. The Low Light Test
Is your look reliant on low-light or dark environments? Does your film take place in the wilderness at night? In the desert at midnight? On the streets of a small town after dusk? In the basement of a building?
If your script calls for dark and moody scenes, lit by moonlight or candlelight, you know that a large portion of your frame will be dark. You need to know how your chosen camera performs in dimly lit environments.
For this test, find a small space with plenty of objects and detail in it. This can be an office, a broom closet, or even your bedroom. Light a couple of candles and maybe a very small practical desk lamp if the room is large enough to absorb that light. Have someone walk through this environment in dark clothing with a flashlight on. Have them occasionally shine the flashlight at short distance to certain objects, and then occasionally shine it right into the lens as if searching for something.
DaVinci Resolve again. What you’re looking for here is not whether you can see the person clearly in the dark, but how much noise you have to be willing to accept in order to see detail in the room. This is a measure of signal-to-noise ratio.
Many cameras allow you to boost gain to very high ISOs; some even advertise ISOs over 100,000. But at what cost to the noise floor? Incredible low-light performance is great and exciting, were it not for the issue of dynamic range. Bumping up the ISO to astronomical levels does not increase effective dynamic range. It simply raises the exposure floor—hence my inclusion of the flashlight.
Look at how the objects react to being lit by the flashlight in relation to the noise floor in the shadows. This is very telling, since most low-light scenes will have occasional specular highlights that need to be handled. This is especially true for nighttime exterior scenes (cars driving by with headlights on, signal lamps, storefronts). You cannot simply boost your ISO to infinity and hope to hold any detail in your specular highlights—it’s a balancing act between noise and dynamics.
Why shine the flashlight into the lens? I do this for two reasons. First, it gives me a sense of how the lens will flare for a given sensor. Every sensor has a different OLPF (optical low pass filter), which, in combination with the coating surrounding the sensor, will produce different flare effects given the exact same lens. The second reason is to check for any “black sun” effects. This happens when a bright highlight (like the sun) inverts and turns black on screen. This mostly happens on lower-grade cameras and is somewhat fixable in post, but shining a flashlight right into the lens will reveal this problem early and alert you to additional work in post.
Are these tests fully comprehensive? No. But if you perform them with diligence, you will be surprised by how much confidence you gain in the camera you eventually choose. Happy testing! MM
Suny Behar has shot the HBO Camera Assessment Series for four years.