Testing the Sony F3
Prior to shooting Pas De Restes, my first project with the Sony PMW-F3, I decided to do some tests to figure out how the camera responded to light and color, how flexible the footage was in grading, and what picture profile settings would best suit the look I was trying to achieve.
It was fascinating to see how different picture profiles affected the image, especially the way colors were handled, as well as the camera’s ability to hold over and under-exposure. I hadn’t been paying much attention to this camera as it seemed to get kind of lost somewhere in between DSLRs and the high-end digital cinema cameras such as the ALEXA. These tests, however, drastically changed my view of the camera.
For reasons outside of my control, production made the choice to record onto the standard SxS cards as opposed to an external recorder. As this limited us to the 35mbps MPEG2 4:2:0 XDCAM codec, one of my concerns was that the grading required to make something like S-log watchable might be too much for the footage to handle.
For this test, I created a lighting setup similar to what I planned for the biggest scene in the movie. I bounced a slightly dimmed 100W pepper off of a 4’ square of bleached muslin from below and created a soft, wrapping backlight with a 420W pepper through a small chimera, diffused further with a 4’ frame of 216. I put an accent on the back wall with a 250W light with a doubled up blue party gel to emulate a feature that existed in our filming location. A spot reading on the gray card gave me T2.8 at ISO800, the highlights on the face were around T4, and the blue highlight was a little over T5.6. I left the diffusion and bounce slightly in frame to see how the camera reacted to extreme highlights.
Then, with the help of my lovely wife, I made a series of exposures using the following settings (the camera was constantly set to 3200K, 180° shutter, 24fps and for the lens I used a 35mm Zeiss Super Speed MkII):
- S-log with LUT turned off. 0db gain/ISO800. Exposed normally, as well as 1 and 2 stops over, and 1, 2, and 3 stops under.
- S-log with the REC709 800% LUT baked in. 0db gain/ISO800. Exposed normally, as well as 1 and 2 stops over, and 1, 2, and 3 stops under.
- S-log off, and using the Abelcine Range picture profile. 0db gain, which is approximately ISO400 on this camera when S-log is turned off. Exposed normally, as well as 1 stop over (the lens wasn’t fast enough to go 2 stops over), and 1 and 2 stops under.
- I was planning to light the film with an ISO800 rating in mind, so I tested the Abelcine Range picture profile again at 6db gain/ISO800 to see how much noise this introduced. Exposed normally, as well as 1 and 2 stops over, and 1 and 2 stops under.
- Finally, I chose a standard picture profile, set the matrix to Cinema, then, at 6db gain/ISO800, made normal exposures using the following gamma settings: Cine 1, Cine 2, Cine 3, and Cine 4.
I then took all the footage through Apple Color to see how it responded to grading. My adjustments consisted mainly of tweaking the luma curve for contrast and bringing the offset exposures to a normal brightness, tweaking saturation, and slight color tweaks, mostly consisting of a slight increase of blue in the highlights and decrease of blue in the shadows. Certain picture profiles created some nasty clipping in the blue channel as a result of the background light, in these instances I used a secondary correction to tone down the saturation in that part of the image. I was surprised by how easy it it was to bring down the over-exposed footage. With the under-exposed footage, however, if it began to get unacceptably noisy, then I would leave it dark. The footage was rendered to ProRes 422 (HQ) from Color.
You can download a high quality .mov from the Vimeo page.
Based on these tests, I decided to shoot the project in S-log, with no LUT, and rated at ISO800. Because I was confident in the camera’s latitude, I felt free to expose it much like I do film, relying on my meter, and not being afraid to let characters move in and out of hot spots or shadows.
This test was very illuminating, and helped me to become much more comfortable with the F3. I know it’s not comprehensive by any means, but I hope this footage can give you a good idea of what this camera is capable of, and the effect that different settings have on the way it reacts to different material.
Colton Davie, Cinematographer