OVERVIEW:
You may have read my post What Camera Settings Are Best? (Differs For Each Lens?) where I cover how I think every lens deserves its own settings (because each lens sends light to the sensor differently). If not, check out that post, but the idea is that I test every lens I get to figure out what I call lens-sensor specific camera settings (and LUTs) to make the footage look more "cinematic". I call them "REAL" settings because my goal is to not just to make each lens-sensor combo look as "cinematic" as possible, but to make it look REAL enough that it's not a distraction to the viewer. To do this, I bring each lens-sensor combination to what I call a "base" where the image looks realistic (like what your eye sees).
ABOVE: While testing a Canon FD 50 1.4 (on my Panasonic G7) I came to the conclusion that while good, there are some lenses that can't render a realistic look, no matter how hard you try. What that means for me, is they won't match other lenses. So, I decided to create a rating system that shows how close each lens-sensor combo can get to being "REAL".
WHAT'S THE BIG DEAL (ABOUT BEING "REAL")?
Well, the main reason I like to keep the footage "REAL" is because I don't like the technical aspects of filmmaking to overwhelm the artistic and/or storytelling aspects. I also don't like spending lots and lots of time shooting, only to discover some of my footage isn't going to be able to match other cameras. So, now let's talk about WHY I think my system can help make that process of shot matching (and other things) so much easier. We'll start with what type of rating/grading system I use (hint: it's letters) and then we'll go into more detail about how I use it to rate lenses.
WHAT RATING SYSTEM DO YOU USE?
The standard I use to express the ratings is simply the A, B, C, D, F letter grading system which all young Americans have grown to love (or hate) from years of formal education. It's not an exact science (like, where did the letter "E" go?) but I use the system because most people can relate to it.
ABOVE: There are a number of things that can either make and image look "REAL" or fake, including the contrast, sharpness, noise reduction, color and sometimes hue settings. It's important to note that these are lens-sensor specific settings.
WHAT ATTRIBUTES DO YOU RATE?
I know there are more complex systems out there to rate the "IQ" of lenses (and other systems to rate the digital resolving power of sensors, etc.) but with my system, I'll try to keep it simple:
CONTRAST (Too much can kill image details by blowing out the highlights and crushing the shadows.)
SHARPNESS (Excessive sharpness can reveal the pixel structure of a sensor which can make it look too "digital".)
NOISE (Too much "noise" is distracting, unless it looks like film grain?)
COLOR (Oversaturated color is another thing I consider to be distracting, and it can also emphasize colored artifacts in digital noise, which is even more of a distraction.)
RATING 1 CONTRAST
To me, if you get the contrast of a shot to look "REAL" (on all cameras) you're getting pretty close to matching all shots. I think varying degrees of sharpness (between cameras/lenses) can be acceptable, if the contrast of the two shots, matches. So, CONTRAST might just be the most important lens attribute (when it comes to not only "shot matching" but also to rendering what I call a "REAL" looking image). Yannick Khong has a really good article on "micro contrast" and I really love how he puts the emphasis on his goal of creating "more realistic images"! However, I have a slightly different opinion when it comes to which lenses are considered "good" and which are "bad" because I think it can depend on which SENSOR/CAMERA the lens is sending the light to, but this points back to my central theme that "It's all about (lens-sensor) relationships". Ansel Adams (the famous photographer who created the Zone System) was also well-known for being a master of controlling contrast.
"REAL" RATINGS FOR CONTRAST:
A = The contrast looks REAL.
B = The contrast looks ALMOST REAL (is not distracting).
C = The contrast looks HALF REAL, HALF FAKE (halfway distracting).
D = The contrast looks MOSTLY FAKE (is distracting).
F = The contrast looks FAKE.
RATING 2 SHARPNESS
In all of the testing I've done so far, I've only come across 3 lenses that could not render acceptable sharpness...and I'm pretty picky. What that means to me is the problem with most lenses is not having too LITTLE sharpness. So, that brings me to one of the worst problems of many modern lenses (besides excessive contrast) too MUCH sharpness. The problem is, when they're paired with the digital sensors of today, the sharpness often reveals the digital nature of the sensor, and this is where any "REAL" organic film look is lost.
"REAL" RATINGS FOR SHARPNESS:
A = The sharpness looks REAL.
B = The sharpness looks ALMOST REAL (is not distracting).
C = The sharpness looks HALF REAL, HALF FAKE (halfway distracting).
D = The sharpness looks MOSTLY FAKE (is distracting).
F = The sharpness looks FAKE.
RATING 3 NOISE
I'm in the process of doing more testing and research about this lens attribute, as it's something that is difficult to explain and quantify, but the short version of this story is that I think some noise is distracting, while other noise can look like "film grain" which some people really like (if you were born in the era of watching film-based films, I guess). That's why I say this one can be a bit difficult to explain or quantify, yet my overall goal is for the image to look "REAL" so I guess I would say that if the noise looks "organic" (i.e. like film grain) then it is better than if it looks "digital" (more like pixels).
"REAL" RATINGS FOR NOISE:
A = The noise looks REAL (looks "organic").
B = The noise looks ALMOST REAL (is not distracting).
C = The noise looks HALF REAL, HALF FAKE (halfway distracting).
D = The noise looks MOSTLY FAKE (is distracting).
F = The noise looks FAKE (looks "digital").
RATING 4 COLOR
The difficult thing about color, is that fashion trends change, and the appetites for too little or too much color sway back and forth. My solution is to keep it "REAL". Make color look like our eyes see. Not too much, not too little, just enough (to quote a book from my childhood). The result is a standard, one which may seem boring to some, but which is a good starting point for all. If we make "REAL" our base, all shots from all of our cameras should match.
"REAL" RATINGS FOR COLOR:
A = The color looks REAL.
B = The color looks ALMOST REAL (is not distracting).
C = The color looks HALF REAL, HALF FAKE (halfway distracting).
D = The color looks MOSTLY FAKE (is distracting).
F = The color looks FAKE.
WHAT'S UP WITH "HUE"?
You might wonder why I don't have "HUE" listed, when it is one of the variables that you can adjust on certain cameras (such as the GH4, GH5, etc.) HUE refers to the color accuracy, whereas the "COLOR" setting is focused on just the color SATURATION. I think the color SATURATION can make things look "FAKE" a lot faster than a slight shift in color HUE, but the main reason I don't use the "HUE" setting, is that some cameras don't have this option (especially the cheaper ones I have) so I'm sticking to the controls that exists in all of the budget filmmaking cameras I have (Contrast, Sharpness, Noise, Color).
MY SETTINGS DO NOT CORRECT FOR WRONG WHITE BALANCE
While both lenses the camera SENSOR can have a color bias, another reason I don't adjust the "hue" with camera settings, is that I think this is the job of proper white balance. Therefore, I think it's important to use a white balance tool #ad (or even a piece of paper) to correctly set your white balance.
Comments