|
Over the years I've noticed what appears to be a loss in fine detail when I shoot warmer scenes. It was particularly noticeable when I first started shooting, and finished filming a scene in the late afternoon. The director and I went in to see dailies; everything looked fine with the shots we did in the middle of the day-the time we started the scene. As the shots progressed, the film warmed up considerably. (This was in the days before a "one light" print was timed scene to scene, the way it's done today.) I squirmed a bit in my chair, knowing that I hadn't compensated for the sun's shift in color temperature as it approached the horizon. But she was oohing and ahhing at what she saw. When the lights came up, her only comment was, "I wish everything looked like the shots at the end, all warm and fuzzy." I assured her the warmth could be timed in on the print, but the fuzzy was a problem. That instance was the first time I encountered the loss of fine detail with the shift in color temperature. That small production was shot on 16mm with an old 12-120mm Angenieux zoom. I don't remember the stock, but it was probably an early incarnation of 7247. Times have changed! Objective I've been curious over the years about how much detail was lost on the film, at what point the fine detail diminished, and if using color temperature (e.g., using filters like Corals and timing them out) as diffusion was practical. I know of one director of photography who uses Corals for diffusion. I use the #81 series of filters, the #85C, along with the #85 to correct Tungsten film for daylight, and to shift my artistic rendition of a scene to a specific color temperature. For this test, I wanted to test something similar to what I do so I could learn results that help me. To that end, I didn't test Coral filters to see if they decrease the film's acuity, thereby acting as diffusion filters. Corals have pink in them, and consequently affect the red layer more. In relation to simply shifting the color temperature, I discovered there is very little loss of acuity in the film, at least in relationship to how the test was shot. This was a big surprise. I assumed that as more of the red layer was exposed and providing density for the print, the resolution would drop. That wasn't the case, but what I did discover was more interesting than what I set out to determine. For the moment, let's take a moment to look at the methodology of the test and the lens used. Chromatic Aberration One of the first questions I needed to address was if the lens could focus all colors of the visible spectrum on one spot-the film plane. What I found out is that all modern lenses are chromatically correct. What this means is that the lens is able to focus all the colors onto one plane even though the colors are different wavelengths. The old lenses had chromatic aberrations. Those early uncoated lenses focused the image the best way the current technology allowed. As scientists learned more about the physics of color, they were able to modify the lenses to the point where a Primo, Zeiss, Cooke, or other contemporary lenses can focus the colors of the visible spectrum at the same point. An exception to this is in macro photography with a large T/Stop and the magnification factor. That's an issue of diffraction limitation. Simply put, there is a limitation of contrast and resolution that is affected by the magnification of an object, and the T/Stop. All lenses will fall apart beyond a certain relative point, and this issue is taken into their design. Another situation that occasionally happens, where the 1st AC may be blamed for soft focus, is an object or area that is very red in the scene appears "soft." If the object is bright, the red emulsion can be exposed to its Dmax with the resulting fuzziness of the edges than can be confused with soft focus. A simple solution is to cut down on the amount of red light on the object, or light it with a CalColor 15 or 30 Cyan to cut out some of the red. A red neon sign will always bloom if it's more than 3 stops over the exposure. A side note to this is that color saturation proportionately drops as the luminescence of the color exceeds the T/Stop of the exposure. A red neon sign shot +3 stops will be pink with fuzzy red edges looking slightly out of focus, not the saturated red the director probably wants. This problem is not a lens issue (or 1st AC issue), but an exposure and film issue. The Ins & Outs of Resolution The resolution chart I used is a modified Century Precision Resolution Chart. The distance the chart was photographed from the film plane was the recommended 51 focal lengths. For the newly designed and manufactured 75mm Cooke prime I used at Clairmont Camera, that was 153 inches from the chart to the film plane. This relationship is critical since it would indicate the resolution of the lens in a quantifiable manner. Looking at the chart, each section indicates the number of lines per millimeter. When I speak with a lens tech and ask about a lens's resolution, one of the first things I am told is the number of lines/millimeter a lens will resolve on a lens projector. The critical numbers I am interested in are the "A" section (112 lines/millimeter), and the "B" sections (80 lines/millimeter). Alan Albert, at Clairmont, told me the new Cooke primes could resolve to almost 200 lines/millimeter when placed on their projector, but that's beyond what the film can resolve. I know from conversations with Bob Boyce, that their minimum contact printing standard is 80 lines/millimeter. All of this leads back to my question regarding the film's resolution as the color temperature shifted. I needed to know what was acceptable, and the standard at each step. The following chart indicates the lines/millimeter in each section of Clairmont's resolution chart. The Chart of Negative Densities from Candlelight to Daylight As I said, I was surprised at the acuity of the film at 1660° K, but I was even more amazed at the loss of negative density as I approached 1660° K. I pondered this. I questioned myself about how I determined the stop and how I exposed the film. Finally I decided to plot the 18% gray card values we shot at each change in the color temperature. I guess that with all of the LAD Tests I've done, my approach is "When in doubt, read the density and plot the graph." The chart "From Candlelight to Daylight" shows the results. As I entered the data into my chart template, I watched the blue density drop below the green emulsion's density. I calculated a 1.4 stop loss in density, and when I looked at the print, it looked about a stop dark. Here's the calculation:
Here's the formula: Density - LAD Density The blue emulsion had a -3.3888 stop loss. The red emulsion gained
.7222 stop in density. This isn't an optimum negative for telecine,
and would be a big problem if I had to go to print. Film Reality v/s Meter Reality v/s Candlelighted Scene What I discovered was most interesting. Two issues: one my meter
wasn't reading correctly, but more about that later, and two,
there was a physical problem in the way the film reacts to the
shift in color. Density is achieved predominately with the green
layer. I've used the green layer to determine the film's EI since
I developed my LAD Testing method. But I looked at the chart and
could see the green layer decreasing along with the blue layer.
I expected the blue emulsion to lose density since there was very
little blue light on my "candlelighted" chart. Mitch
explained that the loss in green density is a byproduct of the
yellow dye coupler as it reacts with the blue emulsion's lack
of exposure and resulting loss in density. Daylight at the End of the Tunnel At the opposite end of my test spectrum you can see how the colors
spread apart. There is about an equal spread of color that you
can see on the chart. The blue emulsion has gained about 1.4 stop
of density, and the red emulsion has lost about -1.1 stop in density.
The green layer virtually remains constant. The net result is
the overall negative density remains close to optimum with just
a shift in color. |
