This morning Zacuto released part 1 (of 3) of the highly anticipated and controversial Revenge Of The Great Camera Shootout 2012. Their shootouts have been one of the most frequently visited and discussed events of the online filmmaker communities for the last two years and I will be offering my take on things along the way. Check back often for updates!
Disclaimer: This section of the shootout was much more subjective and gave the cinematographers a chance to change the lighting, etc. Thus it is NOT a pure specs test and my comments are not to be used as an endorsement or condemnation of the RAW performance of the cameras, nor of how they would perform under other circumstances, unless specifically stated otherwise.
UPDATE 1: For my GH2 readers, note that 16:47 starts the GH2 section of the Shootout (the process they used) and it ends at 18:40.
Comparison footage starts shortly after 19:00. It ends at 23:22. There are 9 test clips.
Interviews end at 27:44. The rest is credits.
The interviews have some great stuff and many from DSLR proponents like Philip Bloom and Vincent Laforet to cinematographers that made their mark on cinema decades before digital film even existed share their thoughts and a bit of themselves.
I will try to address previous concerns expressed about the set, etc. by Francis Kenny in this comment on RedUser.net.
The music: I would have composed it differently (and if you ever watch anything I’ve scored, you will see I do) but I did not ever feel that the music distracted me from the music and if I were the producer, I would be happy to use music from that composer. It just would help to have them write the music for the production instead of using music from a library.
The set: I felt it did a good job of doing what the test set out to do, which was to address certain challenges, not to great a dramatically inspiring environment in and of itself. I have no issues with it in the context of the test.
I say that with deference. I enjoy Kenny’s work in the comedic genre and it definitely enhanced the comedy of films like A Night at the Roxbury and can recall the way he shot one of the car scenes in it without having reviewed the film in over half a decade.
UPDATE 2: Starting my observations and guesses. This round focuses entirely on the last part of the shot, with the two actors sitting down in the foreground.
Which one is the iPhone? D. If D is not the iPhone, somebody made a big mistake creatively with it because there is no reason for the DOF to have to be that deep on any of the other cameras.
UPDATE 3: Why are the highlights so yellow on the exterior for camera C? Seems an odd creative choice to keep them that way, whether or not the camera rendered them that way to start with..
I like the choice of blue on the shirt in the foreground in E and H. It is an interesting variation from that used in the others.
UPDATE 4: If you haven’t checked out EOSHD’s article on the setup used for the GH2 and FS100 yet, here is a link for that, that can add to your perception. http://www.eoshd.com/content/8213/how-the-gh2-and-fs100-were-setup-for-revenge-of-the-zacuto-shootout
UPDATE 5: DVXUser, Keylight, compiled comparison screenshots that have all the cameras in rows. Note that the individual screenshots have been downscaled and are not shown at full resolution. http://www.dvxuser.com/V6/showthread.php?282641-Hacked-GH2-Impressed-in-the-new-Zacuto-shootout&p=1986160004&viewfull=1#post1986160004
UPDATE 6: Here are some of my thoughts on first part of the opening, before the door opens when the camera is still at a high angle.
- D and G shot (or were graded to deliver) a visibly much lower resolution than the other cameras. This can be seen in the zebra pattern on the woman in yellow’s top, as well as in the tile pattern near the bottom.
- F had a lot more resolution than D and G but still seemed like it might have less than the others.
- B did fine in most areas but does more poorly than F in the detail work near the upper left corner of the frame.
- D and G also did poorly in the upper left corner.
- When it comes to the yellow contents of the vase with flowers, B and E look almost as if they had additional sharpening applied or something compared to the others. Not entirely sure what caused the look, but the contours are more sharply defined – which could either be good or bad depending on the context.
- A, F, I, H and G have a noticeably redder tinge to the right curtain than B, F, C, D.
And thoughts on the section when the guest first opens the door:
- D and G show much lower resolution on the leaves outside as well, as compared to the other cameras. The consistency of this issue throughout the shot leads me to believe that it is not just a question of approach, but an actual reflection of the camera sensor resolution. My working guesses (subject to change as I review more footage) are that D is the iPhone 4S and G is the Canon 7D (which records less resolution in video than it does in stills or than the C300 or 500 have in video).
UPDATE 7: Re-watched it and the motion rendering in D was so much poorer than in A, B and C. It looks visibly less fluid and is further argument that the clip may have been shot with the iPhone 4S, since it is the only camera with a shutter not set to 180 degrees or 1/50. The iPhone 4S was set to 1/60 and the narrower shutter should cause more stacatto motion and less motion blur, consistent with what is displayed here.
UPDATE 8: Whoever made the lighting and color decisions for H may have created harsh highlights on some of the guests, but boy did they bring out some pleasant colors in the spread on the back table and on the flowers behind them. There is such a pleasant glow to the approach used and I would have loved to see what the other cameras looked like with an approach more approximating it, for that section of the scene.
You can’t really judge the camera in most respects from this test, except for the ones that are below the baseline in terms of resolution. I could see myself being happy to use all but D or G for the right project. D or G suffer from too many issues to ignore, though. That is to say, if I did not have access to one of the other cameras, I could still do good things with G or D – but if any of the others were around, G and D simply would not get used.
UPDATE 9: If you haven’t checked out Ryan E. Walter’s discussion of the way he setup the RED Epic yet, it is featured at around 13 minutes 22 seconds. Ryan had, arguably, the most difficult job there – participating in a comparison that RED themselves had opted out of – and based on my limited experience with RED cameras (about a week) I agree with every comment he made about how approached the setup.. On top of that, he is a nice guy. He also helped me get my hands on a Promote Control last year that has been used for the majority of the landscape photography featured on this site. In other words, he both understands and appreciates dynamic range. If you see him on DVXUser, feel free to give him some kudos.
The other thing about shooting with a RED Epic or Scarlet (or another camera with similar RAW functionality) is that you really are not shooting a pre-baked look, at all. With a GH2, you want to consider what you are going to do in post when you shoot, because the further you get away from the look you bake into the footage, the more difficult it becomes to work with. On a GH2 the white balance, ISO, saturation, sharpness, noise reduction are all set before you hit record and the de.bayering is performed in-camera. In other words, you have defined a huge part of what the look will be when you shoot it. With a RED camera like the Epic or Scarlet, none of those things are locked – every one of them can be modified in post without a visible loss in quality, meaning that the footage can be used equally well for several very contrasting looks.
So Ryan had to make a huge number of choices and deal with a lot of backseat drivers. I just want to say that I will not be one of them and MANY kudos for taking on this challenge. And as long as you didn’t come up with D or G, I think you did great!
UPDATE 10: Andrew Reid has posted his impressions at EOSHD. Andrew and I do not always see things the same way, but I think it is helpful on something like this to get several perspectives and he does some great testing that others sites just don’t even try, so i am happy to support him.
Here is another one from Joe Marine at NoFilmSchool.com. I have never read any of Joe’s articles in the past so I cannot comment on his tastes, but NoFilmSchool.com tends to put up some good articles and I trust Ryan Koo to surround himself with good people.