Comparing style and technique. They are intertwined.

I can shoot available light as well as anyone but on important assignments (or when I just want stuff to look really good) I bring along lights and I've learned to use them effectively. I have several friends who are full time videographers and they are sometimes at odds with the way I shoot video. They have bought into the idea that "good" video work should be cinematic. But like every term tossed about these days nearly every person with a camera has their own definition of "cinematic."

What I see most often from my "cinematic" friends are very, very flat files with no well defined blacks, not much structure in the shadows, milky and opaque mid-tones, endless highlight tonality and very low color saturation. Much of their work is "available light" which means: "Hey, these cameras are so noise-free I can use them without having to set up lighting. No, I am not looking at their ungraded S-Log files I am looking at their samples and their finished work.

The friends whose work I know best are both fairly accomplished videographers and I understand that they prefer a flatter contrast than I do. They also seem to enjoy files with about half the saturation I want. They also have, in the past, dreaded setting up lighting that I would routinely consider mandatory.  That's all okay. It's style versus style. Taste for taste.

But I recently swapped files with a good friend and he said he struggled to get the same kind of color and contrast out of his dedicated video camera that I've been getting out of my cheaper, all purpose camera. We compared technique and we more or less concluded that lighting a scene, well, made all the difference in the world. Controlling variables is the magic bullet.

So, let's talk about light. Here's the first fact: digital image files look best and are measurably better (meaning they have the highest potential sharpness, lowest overall noise, best dynamic range and most saturation) at their lowest ISO. This is almost always the "native" ISO of a device. You can extend the ISO range of many cameras but the lowest ISO, without extension, is the native ISO of your device. While an A7Sii may be useable at 12,000 or 96,000 ISO there will be a deterioration of all the desirable attributes listed above. It is true of every digital camera sensor.

Here's a second, related fact: Getting good exposure is more than just getting enough photons on a sensor to make an image, it also means getting good balance, from light to dark, in the image file. Yes, most cameras with Sony sensors generate files that can be "pushed" or "pulled" in post processing. But, if you are underexposing because you live in fear of overexposing your highlight tones you will be degrading the tones and information in the shadows of your image and, when you "pull up" your exposure via curves, levels, or just straightforward exposure sliders you are shifting crappy, low information parts of your tonal curves from the shadow areas into the mid-tones. By having fewer bits of information to work with in your mid-tones you are sentencing your mid-range tones to a life in which optimum separation of details and tones never meets its potential.  Want a "better" camera? Buy a meter and learn how to use it. Or really pay attention to all those exposure indications that come bundled with your camera. A balanced file is our initial aim point.

What this all means is that in order to realize your vision of a file full of good information; from shadows to the highest highlights, you need to expose correctly in the first place. If you are "saving" highlights by pulling down curves, or "opening" shadows by pulling up curves, you are losing vital information that is essential to the optimum file integrity you pursue.

If you are videotaping or photographing in a location with low light you have only one real choice if you are to pursue ultimate technical quality (and by extension garner the artistic control you desire) and that is to add light to your scene in order to get closer to the optimum exposure at the native ISO, or to use wider apertures and slower frame rates, or shutter speeds, to get the same exposure results. Since depth of field is generally determined by what you need to keep in focus it makes sense to have lights handy and to know how to use them.....well. Changing depth of field to compensate for not lighting isn't always a workable compromise.

There is a reason that Hollywood blockbuster productions don't scrimp on lighting, or attempt to shoot feature movie files at nosebleed high ISOs. They get the best bang for their production bucks the closer they hover near native ISO for their cameras and the closer they hover to the idea of balanced files. We may have different tastes in the look for files but we can generally agree that most feature productions aim for uniform lighting styles which means most film are well lit and lit well enough to use the cameras the way in which they were designed to excel.

While it's true that there are many fast lenses and that fast lenses will save you if you need the speed and don't mind a very, very narrow depth of field the fact is fast lenses can't do much to help you if the lighting contrast of your scene is out of whack. If you are shooting interior locations and showing some exterior scenery you are playing around with a set of variables. To max the image quality of your still or video camera you have limited options. You must either bring up the level of interior light on your interior subject or bring down the level of exterior light which your camera can see outside --- nothing else is an optimum image quality solution, no matter how badly you want to think you can always fix the problem in post.

If you opt not to light your interior subject you may have to let your exterior burn to white ---- if you want the light on your interior subject to be correct and ample enough not to use up your accurate shadow detail. Just for reference, most of the information in digital files is engineered to be in the mid-range to highlight areas. The nature of the way files are created takes advantage of the fact that our eyes see less tonal variation in shadows and so camera makers don't put as much information in the areas of shadow. The eye is much more capable of detail discrimination in highlights, and when the files are written detail in shadow areas of the files is somewhat compromised, by design, in order to maximize detail in the medium toned and highlight areas; areas where your eyes and brain are most sensitive (it's a "rods and cones" thang). If you choose to "hold" the highlights in your exposure (under expose) and sacrifice the mid-tones with the idea of fixing it all by lifting the shadows in post you'll be pushing up the part of the file with the least bit depth and tonal information and placing it onto the visual curve at a part of the curve where the human eye has much more ability to discern and value the opposite kinds of visual parameters. Pulling shadow mud up into the mid-ranges only gives you mid-ranged mud. Achieving acuity in visually vital parts to the tonal curve requires balanced files.

Let's move on to one more vital part of the whole quality chain. This is color balance. This is most important to still shooters who like to shoot and stay in Jpeg, and to filmmakers who are using any camera file except raw files (and, in the case of Canon, many aspects of what are traditionally expected of raw files are already "baked in" to their cine cameras...).

If you are shooting any of the popular video camera profiles (or Jpeg in still cameras) it really doesn't matter (much) what your camera's bit depth is or how color is written. The basic balance of the color spectrum is locked in when you shoot the file. You may be able to adjust the colors somewhat but making any significant changes to the color results in throwing away valuable spectral information that's a necessary part of a full information file. Shifting color balance after exposure is like working with a light that only puts out partial colors ---- something is missing that needs to be there to make sense to the human brain when looking at a finished representation of a scene. After the fact color balancing on anything but a raw file always involves tossing away information that you need in order to have color that reads as "real."

If you have a file that's too yellow and you change your RGB curves to "add" more blue you may find that what you are really doing is tossing the yellow information altogether which will affect your greens, reds, oranges and their sub-hues. There is no free lunch in post exposure color correction. Except, sometimes, in raw files.

You have choices. You always have choices. If you decide to rescue a scene by placing a subject that should be in the mid-tone range down into the shadow range to later bring it "back to life" in post you'll be "reviving" an image that's already been damaged and lost information. If you add to that the degradation of also having to throw away pixels to color correct you will have thrown away even more valuable information. At some point, without actually going in and painting back color with a brush, you will never be able to get to the aim points that constitute a "good" file, or a file that looks good. There's just not enough magic in the software to save the results of low information files.

Raw is a different story. But raw isn't magic either, the same basic constraints apply, the only difference is that in raw there is a lot more leeway for slop because the raw files start with more native information. You can throw more away before the files look as crappy as baked in files because there is more to through away to begin with. But even raw files have their limits.

I'm a plodding photographer and a more or less novice film maker but I am a creature of habits and most of the habits have survived from the days when files were very fragile and there was very little information in the files to compromise with. I am a stickler for getting things right before I push the shutter button to make a photograph, or push the red button to start making video.

Step one is always exposure. If the exposure is off one way or another then it's not always practical or practicable to get a good, solid custom white balance. There are logical steps to doing the process correctly. In film making and in photography the first step is always exposure. After the right amount of light is falling on the sensor only then can you make a perfectly accurate custom white balance. Anything less than a custom  white balance will require changes in post and those changes will require you to throw away information. Even if you are shooting video in S-Log the two steps to proper file creation are still carved in stone. Exposure and then color balance. Then everything else will fall into place.

Finally, to my friends who are working with inexpensive (+/- $1,000 cameras) to mid-priced (+/- $12,000 cameras) I would council shying away from using S-Log ( or V-Log or C-Log) profiles in any environment except full, high contrast sunlight. The reason is simple, Log files work by shooting super flat and then mapping the resulting files into a fixed gamut. It's hard to do well and requires shoehorning wider data into a smaller gamut space. Some aspect has to be interpolated in order to fit. You may use curves to bring down shadows into a usable and range and visually desirable contrast range but it's rare for me to see files color managed this way that don't have milky, almost opaque mid-range tones and lifeless near shadow tones. Just doesn't look convincing. The only way it really works well is with cameras that shoot at higher bit depth and write color as 4:4:4:4. Those cameras? Sure, use all the Log profiles you want...

I'm a proponent of getting as close to your desired "final look" as you can get while operating your camera and only depending on post production for things that absolutely can't be changed any other way. Is shooting and doing post (laboriously) in S-Log 3 really more desirable than taking time to put up a diffusion scrim on top of your subject to mediate the collimated rays the sun? I think the change in the quality of light speaks for itself, and the scrim gives you added control and uniformity. Sure, you might be able to get a decent file by presenting a flatter file in post but is it in line with our preferred visual references? Don't we want to see deep, rich black? Don't we want to see contrast in our mid-tones? Don't we want to see open and airy highlights?  I contend that we do and that the current style of flat light, vague saturation, and muddy shadows and lower mid-tones is a short-lived style that will soon recede from general practice.

Much like the mania for shooting everything with super narrow depth of field is already looking like a "last century" style because of it's massive overuse. In fact, one could say that these styles are analogous to the constant zooming that amateur film makers did in the 1960's when inexpensive zoom lenses first flooded the market on the front of consumer-accessible Super8 movie cameras. And we tired pretty quickly of that kind of "tromboning."

The bottom line is that if you start your process with a file that needs rescued you will never squeeze out all the imaging potential your camera is capable of. I may make files that are a bit too saturated but it's easier to get rid of excess saturation information in post than it is to add it to a file in which the color was absent because non-linearities in color curves change the look of the file as you add saturation. And the color spectrum shifts. I can start with a perfectly exposed file and throw away information to make the file conform to my vision but it's much harder to take a dark file and make it lighter with the same quality as an optimum file. It's true in reverse as well. If you start too light you might be able to get to a workable file by pulling down the curve to recover highlights but if you start with a file that is right on the money you can shift it with total control to a lighter version.

Pianists practice scales and keyboard exercises so that when the opportunity comes to play real music they will have already mastered the techniques and muscle memory required to play well. Sloppy technique is not a sign or great artistic promise or merit. It is only a signifier that you still have some technique to master and some skills that need practice. In digital imaging style and technique go hand in hand. You master technique in order to master your style. Anything else is just wishful thinking or playtime.


Wally said...

I shoot landscapes photos for still images with a Sigma Merrill DP2. Any ISO above 200 is pure garbage with noise and color banding due to the Foveon sensor. If your workflow is as you suggest the images are phenomenal.

I picked up a Nikon 1 in the fire sale and occasionally shoot birds. When well lit in direct sunlight the images are also phenomenal. Same goes for my D7200 when using my Okta Box and 1600 watt Bluff Monobloc.

What’s the not so secret sause for all this gear? Its light on the scene, exposure, lens, and camera in that order. Need more light you need f1.4 or crank up the gain -another way to think of ISO. More gain = more distortion just like in your stero speakers.

Want better images add light.

Otherwise its show up at dawn, big heavy tripod, cable release, mirror lock up shutter delay, pray for no wind, ISO 100.....stupid.

Sherwood McLernon said...

Kirk, reading this post just made the lights go on in my mind. I can't think of a post that has pointed out some of the shortcomings in my finished work in such a succinct manner and how to improve what I produce. An invaluable post. Thank you

MO said...

Nice post. I even brought cameras with No log on. 1/3 The Price and 2 % lost due to same argument your Making. I shoot The profile that hit The Colours I like best and try to get it right before I hit the shutter.

Edward Richards said...

Wow - I though cinematic implied complete control of the light to SIMULATE natural lighting. Like practicing a card trick for a decade so it looks effortless.

Frank Grygier said...

The following excerpts from the famed Zacuto Camera shootout in 2012 showed the film making community at the time a $700.00 camera could produce an image that great film directors chose over cameras costing over $15000.

"The results are in from Zacuto’s Revenge Of The Great Camera Shootout 2012. The majority of those at the cinema screenings – including Francis Ford Coppola preferred Colt Seaman’s lighting and the capturing of it by the Panasonic GH2."

“The thing I was most impressed with is that some guys or gals with something to prove did better at lighting than the established cinematographers with a good camera” – Bruce Lundeen"


You should write that book!

Anonymous said...

OWEN ROIZMAN: ...Everybody thought I shot The French Connection in available light. I always joke and say yes, I shot in whatever light was available from the truck. The goal was to make it look like it wasn’t lit, which was a radical notion in those days...

...My question is whether future directors and cinematographers will need to be concerned about lighting?

OWEN ROIZMAN: My simple answer to that is FOREVER. https://edistocam.blogspot.com/2008/06/dp-owen-roizman-asc-french-connection.html

Teheimar said...

Very good blog post.
:) Really very educational.
I second that the Nikon 1 sensors are in good light...well phenomenal. Habe a V1.