Forum Replies Created
-
Replies
-
Telecine transfers of movies were primarily from a color-timed low-contrast interpositive struck as also a protection master after answer-printing. Projection prints are a bit too high in contrast to make a good element for transfer though they will be used if nothing better exists. Sometimes a cheaper option than an interpositive was to make a low-con print, a stock that Kodak discontinued eventually. I believe that interpositives were made on a step contact printer which was more stable than the high speed continuous contact printer used for release prints.
Sometimes, at some risk, the original negative was transferred for the final video master… but since it is not a color-corrected element, and sometimes has A and B rolls, plus has splices at every cut, it was not the first option though the quality was the highest.
It’s just my opinion, but the early daylight-balanced Kodak EXR stocks (50D and 250D) were definitely higher in contrast than the tungsten-balanced ones, but by the time of Vision-3, Kodak made them all pretty similar (almost too similar). The main change between Vision-2 and Vision-3 was the introduction of “micro grains” (very slow in speed) which increased detail by 1-stop in the extreme overexposure areas.
I use daylight stocks sometimes but whether I do a show that is, let’s say, 250D and 500T versus 200T and 500T, depends on a number of factors. As Roger says, you can pull the 85 filter on tungsten stocks for cooler shadows or a blue dusk effect, or let the scene itself be timed on the cool side (i.e. only partially correct for the missing 85 filter.)
So I tend to base the decision on whether it’s a movie that will lean cool or lean warm — if I’m doing a wintery movie in the woods, I’m more likely to stick to just tungsten stocks, if I’m doing a summer movie in the desert, I’m more likely to get the daylight stocks partly to avoid the 85 filter.
I worked 2nd Unit on Season 1 of “Westworld” matching footage from the first half of the shoot before they shutdown, and one DP used 50D whenever he could and the other DP shot everything on 500T — I’m not sure viewers saw a big difference…
Roger has switched to slower stocks in day exteriors in the past, like the use of 100T 5248 outside in “No Country for Old Men”.
There are DPs who use 500T for everything — John Seale for example, or Emmanuel Lubezki in “Tree of Life”.
There are arguments both ways. Yes, using 500T for everything in theory is more consistent… but often grain is easier to see in scenes with large areas of midtones (a blue sky for example) and day scenes can have more midtones than night scenes. You can also argue that our eyes perceive daytime exteriors as sharper and more clear, with more depth of field, than we do in interiors and night scenes.
On the other hand, 250D is so similar to 500T in look (and is just slightly finer-grained) that it’s a toss-up whether you need to switch. But I think 50D does tend to stand out when mixed with 500T — it’s SO clean. 200T is somewhere in between as you can imagine.
Just from a stock management angle, I’d try and limit yourself to two stocks though. Unless you really want that 50D look for select scenes.
Why? What difference does it make? If you want to learn good handheld staging and operating, why does it matter if only 50% of the movie was handheld versus 80% or 90% or 100%?
“Jarhead”!
The boat sequence in “Jaws”, the running around the submarine shots in “Das Boot”… Most of “Man of Steel” for some reason is handheld.
“Chinatown” has some good use of handheld (John Alonzo was good at that technique.)
If your wide shots clearly show the cold flashlight beams bouncing off of warm surfaces and picking up that color, then normally you would try and match that in tighter coverage — it’s not disorienting nor unrealistic if it is clear what causes the colors.
I don’t think it has to be exact when going in tighter, just depends on whether you think the viewer’s eye will catch a shift. You might, for example, use unbleached muslin for a slightly warmed-up bounce when the wood walls are off-camera if you less of a difference.
Of course, a room is made up of more than one color surface so once the actor moves closer and deeper into the room into a tighter shot in a different area where there are more off-camera surfaces, you might be more free to let the bounce just be the same color as the source light.
Tends to be a practical issue involving the space restrictions. Sometimes you don’t have the room to back up a light to fill a bounce surface so you’re better off with multiple smaller lights closer. But that’s the nice thing about a big diffused light, you can always add another light into the mix to increase the output if you’ve fallen short.
The cutters shouldn’t be reducing light intensity on the subject, just the background, but if they are, then you might need to increase the light intensity to compensate (add an additional light to the bounce or use a stronger light). You can also use negative fill opposite the bounce to add some contrast.
Generally no, I shoot on the Alexa and stick to a ISO 500 to 1000 range, mostly to control depth of field more than anything. Otherwise I’m mostly at ISO 800. I control my highlights or shadows by exposure and lighting. I control contrast through lighting and by adjusting the LUT using ASC-CDL, mainly for dailies (an adjustment might be to match two cameras or lenses for example, or to deal with a hazy set, etc.)
Either way you are designing a shot that requires the creation of a traveling matte — a chroma key / color difference key (using a green or blue screen usually) is the simplest post method compared to rotoscoping. Now if the actor became a near silhouette against the new background, you can sometimes get away with a white screen and pulling a luminance key rather than a chroma key, but even that may take some clean-up work.
Generally it’s better to take the time in shooting to get something that works well for post than to hand footage to post that needs a lot of work, especially if you don’t know how good your vfx people are or what sort of budget/time constraints they will be under.
Or put the actor in front of a big LED screen where the screen image itself rotates from the old to the new and spin the actor on a platform…
As to whether you heavily push the ISO setting or create a stylized LUT (for monitors and dailies) / or stylized post color-correction, that is entirely project dependent. Your reasons may be practical or aesthetic.
But there are some tried-and-true generalizations:
(1) The simplest approach is often the best.
(2) Garbage In / Garbage Out — the shots you create in camera that come closest to the final look will go through the post process the easiest and come out closest to your intent — whereas the shots that are far from the look you wanted and need a lot of post to fix might not end up close to what you wanted, or you may run out of time and budget in post to get what you wanted.
It’s not as binary as “using natural light = not manipulating the scene” vs. “using artificial light = manipulating the scene” — natural light can be altered in ways (flagging, diffusion frames, etc.) and often artificial light is mixed with natural light if on location.
Cinematographers use light — natural, artificial, or a mix — to create mood and to tell a story. If the goal is a completely realistic and “honest” visualization, the truth is that the lighting could still be entirely artificial.
And what defines “artificial”? If you are shooting a scene in a auto repair garage and you ask the art department to hang a fluorescent shop light over the car being repaired, is that a natural or an artificial light? And if the light is above the frame line so instead you hang a LiteMat LED light instead of a fluorescent, if the effect is more or less the same, how artificial is it? And if the location already had fluorescent shop lights but you switched the tubes from Warm White to Cool White or to color-balanced 3200K or 5600K tubes, is that “manipulating” the scene? Maybe the original tubes would have made the scene look less “natural”, creating a more stylized effect?
Roger generally doesn’t use a separate eye light, except in some of his early works — if you watch “1984”, you can see some use of an eye light when the key is a soft top light in that Gordon Willis style, I don’t know if Roger did this because he was using a silver retention process for the prints and wanted to make sure the eye sockets didn’t get too dark, or if was just an earlier technique of his.
If you study Roger’s work, what you see is that his key is often placed to create a good catchlight in the eye at the same time, i.e. the eyeball is reflecting the key.
In the case of when other cinematographers use a separate eye light right next to the lens, how much contrast you plan on using in the final grade or LUT will affect how much that eye light also adds to the fill in the shadows — if the contrast is a bit higher, with deeper blacks, then the shadow detail drops out a bit more… but the reflection of the light in the eyeball can pop a little more.
Also sometimes “power windowing” the eyeball in post color-correcting and adding some sharpening to just the iris area will make the highlights in the eye brighter too. But that can be time-consuming and tedious for a colorist to track if the actor is moving much.
The size of the reflection of the eye light is a factor too and will not only be affected by the size of the light source but the distance to the subject, plus the size of the close-up — if you get extremely tight, you need to start thinking more about the shape of the eye light since it will be more obvious.
As I said, for a given ISO, the noise in the lower tonal values in the frame are the same whether it is a brightly-lit scene or dimly-lit scene, the only difference is that these lower tonal values take up a larger portion of the overall frame in the dim scene, which probably makes the same noise level now more obvious to the eye, and then once you start color-correcting the image, the changes in noise might be more visible just because so much of the frame has lower tonal values.
But in theory, the noise doesn’t get worse just because a scene is darker, anymore than a scene with normal noise level goes noisy the moment the character switches off the lights or the camera does an iris pull to create a fade-out or the camera pans over to a black curtain, etc. (assuming a fixed ISO setting and a recording with minimal compression and a high-enough bit depth.)
