Tag Archives: Ryan Walters

Decent Exposure

Some of my posts are emotional, others more prosaic, and this one, at the opposite end on the spectrum of the mundane, is so purely technical that I think it kind of loops around and becomes poetic. Of course there has always been a poetry to technical pursuits that appeals to me as an artist, but I understand if you don’t find the poetry here, hidden as it is. 

This post is all about calibrating my Sekonic L-758C light meter. Keep in mind I’m not a director of photography or a gaffer, and 90% of the time union regs don’t allow me to even touch the lights or the cameras at work.

That being said I think that light meters are cool and I follow my interests.

A few years ago I bought a trumpet.

I’m quixotic.

800_270_T2.0_ND1.2+.9

3 reasons to use a light meter…..

1. Proper exposure

Access to DSLRs like the 5D (not to mention Sony’s new A7s) have given us a terrible education on how much light it takes for most cameras to register a rich image. Scrappy shooters like me who cut their teeth shooting web videos with small crews are accustomed to walking into a room, ANY room, and thinking “of course there’s enough light to shoot.” This has spoiled us, and in many ways ruined us when it comes to using better cameras, which incidentally have smaller sensors (less light gathering) and don’t automatically apply their own noise reduction to the image (which is how DSLRs achieve those super high ISOs). I’ve read a lot of internet posts where people turn on their cameras, aim them at the sofa with only the room lights on, and then post those pictures to the web saying that the image looks grainy. Yeah, those images do look grainy. The Red Epic, and the Red Dragon are not low light cameras, especially compared to DSLRs. So just how much light is enough light? Well as it turns out, Red gives you a lot of tools to evaluate the light reaching your sensor. There’s a histogram, the goal posts, and the traffic lights. But this post will address NONE of those in-camera tools, because I went third party and got a Sekonic L-758C light meter. This is now my light measuring device, and it’s like having a year of grad school in my hand every time I turn it on. If you want those deep rich blacks that your camera is fully capable of, then you’re going to need more light, and if you can quantify those amounts it’s going to be a lot easier than just saying “add a bunch more.”

2. Know the usable latitude of your sensor

Latitude, or dynamic range, is an expression of how much your camera can see into the shadows on the low end, and into the highlights on the high end. Everything below the dynamic range is too dark, everything above is too bright. Everything in the middle makes a picture, and that’s your camera’s latitude. Whether you are shooting in full sun or deep in shadow, you want to consider where your imagery falls on the available latitude spread that your camera has. I was listening to Alex Vendler’s interview on The American Cinematographer’s podcast, and he talks about shooting the dark imagery for the horror flick The Woman (see scary woman). the-woman-stillHe explains shooting in the “toe of the curve.”  When I heard that expression I found it to be movingly poetic, and I had NO IDEA what it meant, although I instantly felt a need to be able to do this as well. Essentially what he meant was that he was exposing a lot of scenes in that subtle range of the camera’s latitude where it’s juuuuuusssst about to be too dark, but it’s NOT too dark. You can make out detail but everything is on the low end of the dynamic range, and within a stop or two of being underexposed. You can’t shoot in the toe of the curve without knowing exactly where the falloff happens, and for that you need a light meter.

3. Don’t trust monitors

Apart from the DIT’s monitors, which are usually calibrated, monitors on set or on the camera won’t give you an accurate sense of exposure, so I have a desire to be able to circumvent the monitor entirely through my own ability to quantify the amount of light in a scene, and to know how it’s going to show up in the footage, AND to able to do this without relying on a monitor. Does this mean that I am banning monitors? No. Monitors can stay. But for my own education, with my own camera, I want to know when they are lying.
 800_270_T2.0_ND.9
 
So I could have just bought the light meter and stopped there. But in my quest for truth, I discovered that light meters come from the factory with their own inherent deviations, and that the instruments themselves need to be hooked up to OTHER INSTRUMENTS to be calibrated. For this you use a clandestine service like Quality Light Metric, which involves dropping your meter at a mailbox store on the Northeast corner of LaBrea and Hollywood. Just walk in and flash your meter, the people behind the counter will give you instructions. It’s kind of like being a spy. Not that I’d know. Or would I…….(beat)…..but Sekonic also provides software that allows you to calibrate your meter according to the sensor of your particular camera, and that to me felt very sexy and intimate, so I set out to master the Sekonic Data Transfer software. If you are doing this for a cinema camera that shoots RAW, be prepared to go super deep, lock yourself in the office for a few nights, have a container of gorp and a pee bottle under the desk. And to instruct you in this endeavor I will send you to Ryan Walter’s indispensible video tutorial where he gives an in depth, click by click tutorial in how to do it. I watched this video 30 times last weekend, and it’s more like a textbook to me now. Without Ryan’s help I’d have been totally and completely lost.
 
Somewhere out there there must be a prelude to this video because Ryan starts with “Now that you’ve shot your exposures using the Sekonic Profile Chart…” But I couldn’t find the video that talks about how to shoot those exposures, and while you can largely infer how it’s done from Ryan’s video above, there are some basic practices I had to deduce by trial and error. Determining the usable latitude of your sensor is entirely dependent on how you plan to color correct your footage. Don’t forget that. Red, for instance, has 16 stops of latitude for the Red Dragon sensor. That may be the case, but from what I’ve found, the entire usable range is only apparent in uncorrected RAW footage, and no one will finish a project in the uncorrected RAW output. Maybe you have created a look that is super flat, and maintains all information, but for me, based on my own subjective taste, I always adjust shadow and highlight levels in a way that’s visually pleasing, and which reduces the visible latitude. Dramatically. You see in order to calibrate your sensor, you are going to export jpegs for the Sekonic software to analyze, and in order export those jpegs, you’re going to WANT to put a look on them that you like. Maybe it’s a look you’ve created for one specific shoot, and in that case you can create a Sekonic profile which is project specific. Or maybe you want to export jpegs which match the RAW look, so that you always know how much information is there to be played with. It’s entirely up to you. But spending time with this process will help you understand the complicated and subjective relationship between light levels, aperture settings, RAW footage, color correction, and the final latitude represented in your image. To begin to understand this chain of consequences will make you appreciate your camera a lot more, and it will bring you more into the mindset of how professional cinematographers evaluate light and exposure, EVEN if they don’t use meters themselves.
Ryan captured his chart images at various exposures but with a constant aperture. He did this using a series of stacked ND filters and by varying the shutter angle of his camera. There is some extremely complex addition and subtraction here, and I will leave you to puzzle over the particulars yourself. It’s like solving a riddle, and when you get it you’ll be happy. All the info you need is in Ryan’s video, just be prepared for some new forehead wrinkles.  20150201_174010
There’s only one key piece that you need to know going in that is not stated clearly enough. You will need a light powerful enough to give you an initial reading of somewhere between f32 to f45. The first two times I did this I used my old Mole Richardson soft light with a 1K bulb. And my initial light readings were too low for any of this shit to make sense. In fact, it wasn’t until I was all done that realized that I didn’t have enough light, so fortunately for you I plodded on and made another revealing mistake….You have to make sure that your profile chart is evenly lit. During setup I noticed that I got a slightly different reading at the outer edges of the chart than I got at the middle of the chart. These readings were off by 1/10 of a stop and I sort of figured that it might not matter. But it does.
unevenchart
This process involves looking at your footage on a waveform, and if you’ve exposed properly you’ll see a perfect staircase of exposures, representing the descending brightness of each chip. If you have uneven illumination, like I did, then your waveform will be bowed and slanted, and the software won’t give you accurate results.
Bad waveform
Bad Waveform
Better waveform.
Better waveform.
My first batch of charts were crap, and it’s because I was using the wrong light and the light was too close to the chart. Moving the chart further away from the light evened out my illumination a little bit, but not completely. And while moving the chart further away made my exposure more even, it also decreased the amount of light even further. Remember how I said you’ll need enough light to get somewhere between a 32/45? Well as you can see I was working with a 5.6. That’s not enough. You’ll see that the “better waveform” on the right still has slanted chips, and the amount of slant gets worse towards the edges. As of this writing I haven’t gotten a chart illumination perfect because I’m going to need more light.
Also before you start, you’ll need to go into the custom settings of your Sekonic meter and make sure that the meter is set to display aperture values in full stop increments, as opposed to 1/3 stop increments. If you don’t do this, then you’ll get meter readings that say things like 6.3 and 2/10. But when you go to enter your meter values into the Sekonic software, the dropdown menu jumps from 5.6 to 8. That’s because 6.3 is one third of a stop above 5.6 and the drop down menu does not allow for thirds. You’ll get it when you do it.
I’m totally not finished calibrating my meter, but I can see where it’s headed. Based on my how I shoot and the looks that I favor, I can expect around 8 usable stops from my Dragon. Having spoken to a few DP’s I trust, this is a totally acceptable shooting range. I know that the full dynamic range is there for me to use, should I get into a scrape but knowing how my standard curves look, I’m using only the part of it that I want.
800_270_T2.0