Lens Tests
The process of testing is a major component of the cinematographer’s trade. It is a neverending task of examining the latest equipment and revisiting older tools with a fresh perspective.
Can always ask rental houses when a good day would be to come in and do some testing, as long as you don't leave the premises and are flexible on timing, it will be easy to get them to agree.
Keep notes of each of the variables you are tracking in the frame, it can be written on a slate or printed beforehand and held up in the shot. Hold these up throughout the duration of the frame so that even if you take a still or cut a portion of it, the notes are still visible.
Testing a lens at multiple stops is essential because the behavior of many aberrations changes at different iris settings. Testing a lens at every stop to gain as much information as possible is great, but that method can be tedious and time-consuming if you know you’ll only be using the lens within a particular range of stops.
Standard Test
Cameras
It’s best to test with a camera that you will be using — or if you’re inclined to add iterations to your testing, you can use multiple cameras and widen your knowledge base.
Foreground
To put all of these in the frame you can rig an 8' piece of square tube across the frame and clip all of these charts to the square tube. Then place the model in the center in the same plane as all of the cards so they can all be sharp at once, even wide open.
- Model
Since human faces are a key focus for cinematographers, it’s good to have one in frame.
- Slate
- Gray Card
- Color Chart
- Resolution Card
Background
Rig a 20x solid across the back wall. This should be a fair distance away from the model — about 6'-10'. Right in front of the black, stretch three or four nets of holiday twinkle lights. Ideally, the black and twinkle lights should span the entire field of view of the lens you’re testing, allowing for enough width for the widest lens that will be tested.
- Tungsten Holiday Nets
Do not use LED's as you'll end up with flicker. Nets are much better as well as it will allow you to cover the entire field of view of the lens with a few of them. Bokeh changes from center to edges, so this will allow you to see this easily.
- 150w Fresnel
Placed off to one side near the top of the frame, but still in the shot. This lamp should be pointed directly at the lens, in order to reveal how the lens will handle spot/ghosting flare. It will be switched on and off during the test. Additionally, another 150-watt Fresnel is positioned just outside the field of view of the lens, off to one side. Both of these lamps are cabled back to switches near the camera.
Lighting
Use a large, broad light as a key source to evenly illuminate the charts and the model. The fixture can be anything you have available, but tungsten is preferable as it will minimize variables with color rendition from lens test to lens test on the same camera. The light should be soft, it should evenly light the model and all the charts, and it should provide a minimum of 50fc (538 lux) of illumination on the model’s face and the charts (that’s a 5.6 at 800 ISO at 24 fps).
Focus
Focus is set on the model with the lens at its widest aperture and is not changed during the main body of the test. If you set focus with the lens at any deeper stop, it’s possible you will not have critical focus on the model, as depth of field makes it difficult to see the critical-focus plane; then when you open up the aperture, the model may be soft, which will bias the test. Once the focus is set, don’t refocus for each iteration, since that can also bias the test.
Process
- Start rolling with only the key light and holiday lights on.
- Turn on the rear flare light, directly striking the lens at an oblique angle. It remains on for a second or two and then is turned off.
- Turn on the side (veiling) flare light, striking the lens from out of frame. It remains on for a few seconds and then is turned off.
- Cut camera
- Change the iris.
- Change shutter speed and frame rate accordingly.
- Change the slate.
During the final iteration, before you cut, roll from close focus to infinity to check focus breathing. This tactic is generally best kept to the end of the process so that focus variation between iterations is not a factor.
Note that after each iteration, you change the iris as well as the shutter angle and/or frame rate.
To adjust exposure to test all of the different apertures it is best to change the shutter angle or frame rate as this will affect the least amount of variables. As long as the model remains stationary there should be no other visible changes.
This chart is provided as a guideline for exposure compensation via shutter angle and frame rate, with a base starting at T5.6, with a 180-degree shutter at 24 fps (noted in red). This base is chosen so that you may reach all of the common stops on most lenses without altering lighting. Bolded apertures are full common stops; non-bolded are fractional stops, allowing for lenses whose maximum speed is a fractional stop (i.e., T1.3 or T1.9 as opposed to T1.4 or T2). When testing on some cameras, not all of these shutter angles or frame rates may be possible. In those circumstances, you may need to compensate exposure via lighting.
On the chart, note that when you change frame rate, the duration of time and speed of movement also changes. If you switch to 12 fps, you should roll for twice as long at each stage of the test — including the length of time that you turn on your spot-flare and veiling-glare lamps. When you switch to 6 fps, you need to roll for four times as long. Inversely, if you switch to 48 fps, you should roll for half the amount of time.
Zooms
For zoom lenses, at the very minimum, you should test the widest and longest focal lengths possible within the testing environment, along with one or two positions between those points. For the most exacting results, test the zoom lens at every marked focal length on the barrel.
Taken from The Generic Lens Test
Specific Testing
Bokeh
One of the best methods for revealing the bokeh of a lens is using incandescent filament, net variety of holiday twinkle lights. These small, bright point sources are excellent for providing the highlights that, when rendered out-of-focus, can reveal much about the character of a lens. This gets you more spread out sources so you can see the changes over field of view. LED varieties may cause issues with sensor clipping and flicker.
- Place the holiday twinkle lights in front of a black background, roughly 8' away from the lens.
This will give you the cleanest representation of the bokeh.
- Test both focusing at minimum focus, then at infinity.
This will help you examine different qualities of spherical aberration, chromatic aberration, etc.
- Start with lens wide open and compensate for exposure by adjusting shutter angle.
Expose the holiday lights between middle gray and a stop and a half above. If you overexpose them too much you'll lose detail and the ability to see the character you're looking for.
- Begin with Minimum Focus, pan the camera so they move past the edge of frame, then slowly pan back in the other direction so they cross the entire frame to the other side. Repeat this for tilt if the lights don’t cover the full frame.
At the end of the iteration, you want to make sure that you’ve seen holiday lights over every inch of the field of view.
- Then rack focus to infinity in the same iteration and repeat the pan/tilt actions. Cut, stop down 1 stop, and repeat the test.
You may have to alter the focus of the lens to closer-than-infinity or farther-than-minimum object distance to adjust the size of the bokeh so they don’t overlap.
Characteristics of Bokeh
If there is color fringing around the bokeh balls, for instance, that is a sign of chromatic aberration in the lens. This color fringing can change depending on the focus distance, and it can change from the center of the image to the edges. I’m always looking at the pupil occlusion, or what is also known as the “cat’s eye” or “Petzval swirl” effect, where oblique angles get cut off and the bokeh take on a more elliptical or “football” shape. This is lessened by stopping down the lens. An aspherical element in the optical design may create a bullseye pattern of concentric circles in the bokeh — aka onion rings. Some of the bokeh may distort at the edges of the frame, illustrating aspects of geometric distortion and coma.
Flares
Flare testing happens best with a single small point source. Use a small 150w-300w tungsten Fresnel.
- Position the fixture about 8' away from the camera, in front of a black background.
- Pan the camera 90 degrees to one side, so that the light is actually shining on the body of the camera and not hitting any glass in the lens at all.
It’s best if this can be done with the camera pointing at another black background such as a 4'x4' floppy.
- Slowly pan toward the light, going 180 degrees past the fixture to the other side of the room
Ideally, you’ll have yet another floppy. This slow pan across the light source will reveal how soon the lens picks up veiling glare — and also the nature of the spot flare and “ghost” flare — as the lamp comes into and out of the shot.
As the source passes through the frame, you can also determine the lens’ quality of diffraction, especially as you stop down the iris in later iterations. Light diffracting around the edges of the iris can create a starburst pattern around bright highlights in the image. Odd-numbered iris blades will create twice as many burst rays as an iris with an even number of blades. (For example, an iris with eight blades will produce eight burst rays, while a nine-bladed iris will produce 18 burst rays.) Rounder irises will have softer, less-defined starbursts (or none at all).
Slating for Flares and Bokeh
For both of the tests above, slating can be a bit of a challenge. Without having a backlit slate, it’s not possible to have it in shot the whole time, so I generally place it on a stand in the corner of the frame, such that it’s captured at the camera’s starting pan position. At the beginning of the take, I’ll shine a flashlight on the slate to make it readable and then turn it off for the rest of the take.
Charts
Just some of the things that become readily apparent from shooting charts are contrast control, color bias, geometric distortion, resolution, edge-to-center performance, falloff (vignetting or shading), and chromatic aberration.
Charts are designed to be photographed at a specific magnification, which means that each focal length needs to be a certain distance from the chart in order for the full array of the chart’s elements to be properly recorded. Often, this equates to merely filling the field of view with the chart, but that is not always the case.
If you purchase a chart, be sure to check the instructions for the proper distance per focal length. My Century Precision chart requires a distance of 51 focal lengths — which means you multiply the focal length by 51. So, if I’m shooting with a 50mm lens, it needs to be 2,550mm (8.37') away from the chart, but a 25mm lens would only need to be 1,275mm (4.18') away.
Be sure to focus on the chart with the lens at its widest aperture, as focusing when stopping down may cause difficulty in critical focus, and the chart may fall out of focus as you open up the aperture.
The chart needs to be evenly lit with no hot spots, glare or shadows. Generally, use two fixtures set at about 45 degrees from the face of the chart.
Taken from Specified Lens Test