Cinematography Playground July 27th
by Angus McLellan
Back on March 31st, 2018, we ran an event called "Cinematography Playground". The idea for the session, which was proposed and partially coordinated by Katarzyna Kochany, was to compare three of Ed Video's production cameras: the Canon C100, the Panasonic GH5 and the Blackmagic Cinema Camera. We arranged for three actors with different looks to come into Ed Video's studio space, where they patiently waited for us to experiment with a variety of camera settings, lenses, and lighting.
I'm actually not going to get much into the results of the tests in this 'blog', but I will post the results below. For me, the results aren't quite scientific enough to be definitive. I also have to admit that the results were largely in line with what I suspected, having used all three cameras a number of times. Each camera performed well in certain conditions, and picking a winner really comes down to what is most important to your eyes. Some cameras have more vibrant colour, others are better with skin, or general detail, or shading. Few do it all well.
For me, the biggest learning that came out of the session was just how difficult it is to conduct a test like this, and how meticulous you have to be in order to capture footage you can really look to for clear answers. For those of us who were there, I think we were willing to just experiment somewhat loosely. We tried so many variations that day, and I heard a number of remarks about what people were seeing in that moment. We all learned things as a direct result of the process, and there were a number of conversations that came up through the process that led to knowledge transfer and idea exchange, and that's awesome.
On the other hand, I don't think we generated footage that could give clear indications as to which camera or lens might perform best in which situations. I do think we created some material that gives a clearer picture of what settings might work well for each camera in the setting we created (so, pitting each camera against itself). That's certainly valuable.
Here are some things I'd do differently if I ever conduct one of these tests again:
- I would create a more specific 'scene' to film. The truth about these tests is that you can't determine what camera or setting is best, you can only really determine which camera or setting worked best in that given situation. And so, I would try to match a scene I was planning to actually film.
- I wouldn't cycle through so many settings on the cameras. I would focus on a much more narrow range of settings and lens combinations that would be the most likely to capture the results I wanted. For example, if you know you want to film in a flat picture profile, there's no good reason to test out the 'vivid' picture profile. I would dedicate the same amount of time to fewer shots, and make sure each of them was executed to the best of my ability.
- I would record, review on a big screen, and then record again after correcting any issues. Focus, exposure and framing all have to be perfect and matched for a test like this (unless you're testing the ability to manipulate an imperfect shot in post). It's too difficult to really analyze and compare images afterward if some of them have focus or exposure issues. It's just too hard to get past the flaws that aren't actually related to the setting or camera you're trying to compare.
- I would connect all three cameras to the same monitor, and compare the sources before recording. Are they really matched as closely as they can be?
- I would shoot it all in such a way that I had as little work to do in post to compare the footage as possible. Katarzyna put together the files for the C100 and GH5 pretty quickly, whereas I couldn't find the time to convert the RAW BM cinema camera files, and it took me months to get to it. I would go in with a very clear and simple test plan, take my time to nail each shot before hitting record, record only a few seconds of video (unless more was necessary to the test), and I'd have a thought-out plan for the post production and comparison.
- I would actually plan for the still that would be pulled from each shot, not the video. Unless you're comparing the effects of motion on an image, stills are actually much easier to compare than video. Video playback takes longer, and you have to skip back and forth or play with scaling in your presentation. Stills can be more easily compiled side by side.
- This would be a very different test, but I think a functionality test might be just as important as an image test. As important as the image is, the functionality and ease of use of the camera is just as important (sometimes more so).
- I would clearly determine if I was trying to compare images that would look their best 'out of the box' or images that would look their best after post-production work. These are two very different approaches, and require slightly different methods.
Overall, I'm glad we did the session. It was fun and informative. I hope the others who participate did as well.
Here are the results. The C100 and GH5 summaries were prepared by Katarzyna Kochany, who did a much more thorough job than I did with the BMCC footage. The BMCC results aren't very informative and don't do the camera justice, to be frank.