As a freelance camera guy, there are a number of events that I tend to photograph on an annual basis. In 2009 and 2010, I was paid to produce recap videos for Huntington University’s annual Olympiad (a miniature ”olympics” only with stranger events…) as a keepsake for the students.
This year, although I wasn’t officially on the job, I was in the area and couldn’t resist watching some of the festivities – and of course I had to bring the camera along for old times sake.
Unlike in past years where I was required to film a bit of everything (see video at end of post), I just stuck to Ultimate Frisbee this time around. I shot everything in 720/60P mainly because I’ve never really given the slow motion feature on my Canon 7D a fair shot before.
Take a look at the finished video below (embed SD, click here for an HD link) before reading my thoughts below.
Any camera with a proper cinema sensor would be a better choice for sports action than a CMOS-based DSLR. There’s just too much skew from the rolling shutter.
Certainly, on any high-end paid-project, I’d use a different camera for something like this, but when you’re doing something for fun on no budget, renting a Phantom is rarely an option.
That being said, I also wanted to test the limits on using a DSLR for video in the sports world. Again, I’d never given this a fair shot, mainly due to the rolling shutter, lack of autofocus, form factor, and other common issues that people have with DSLRs for video.
This being a personal project, I was able to experiment in some new ways and thought I’d use the rest of this post to share my findings.
I haven’t shot nearly as much in slow motion as I would have liked. As a cinematographer, I often see part of my goal as showing my audience something from life, captured in a way that makes it more beautiful and awe-inspiring than seeing it in real life.
If I’m photographing the Taj Mahal, I want to come away with at least a couple of shots that show something that 95% of the tourists didn’t see. Either a unique angle, perfect lighting, or beautiful use of depth of field not possible with the human eye.
In sports especially, slow motion is one of the most effective – and honestly, simplest – ways to get this unique perspective (see my post on why slow motion looks awesome).
The last time I did any significant work in slow motion was with a Panasonic HVX200 when I shot The Sketch.
I remember my producer, Joel Bain, came to me with a story about a girl drawing a picture who is shut out from the world both physically and mentally.
My first thought as he described the story was to show certain shots from her POV and film each of these in slow motion (overcranked at 60 fps) while leaving everything else at 24 fps. I was worried because we had one shoot date and if I changed my mind after shooting, it was too late to change it.
In some ways, this holds me back from slow motion. What if I’m shooting 60 fps and decide later I want 24 fps? On many of today’s cameras, this is a significant question because they allow for full 1080p at 24 fps and only 720p at 60 fps. I’m losing a significant amount of resolution each time I decide to overcrank – and I can’t get that back. I can either get a couple of shots in slow-mo blown up, or an entire project output at 720p.
Ultimately, this project allowed me to get past this dilemma by shooting a short subject sports piece where I could use entirely slow motion if I chose to do so.
Thus, I kept the following settings for everything:
For most shots, this meant I could use an f/5.6 with .6 ND to get proper exposure, but I used these two factors to adjust for how much sunlight I was getting during the afternoon. Heavy clouds made for an overcast day with very consistent and diffused sunlight.
I was pretty happy with the results. Being able to slow anything down to 40% in the edit bay was a real blessing for timing, and made for some real suspense on some of the more spectacular plays. I also double printed some shots where extreme 20% slow motion was warranted.
Although using a 7D for the fast-paced action (often including whip pans to follow the disc since I was so close to the sidelines) created some issues with the rolling shutter, I was able to edit around them easier than I had feared.
The higher shutter speed, combined with the low amount of light produced by an overcast November sky helped with the manual focus somewhat. I was worried that the depth of field would be too shallow to pull proper focus from the camera’s small LCD screen. In reality, this was one of those times when I should have reminded myself that “harder” is really “easier” when it comes to pulling focus, that is, when you keep it shallow, you can easily see that you’re either in or out of focus and adjust accordingly.
I downloaded this awesome piece of software about a month or two ago from Red Giant’s website after following a tip from a fellow blogger. Naturally, I started playing with it right away, but was waiting for a “test project” such as this to get a proper look at it.
I was pretty impressed at what I saw: quick, easy to use, and effective.
I’ve planned to do a full review since I downloaded the software, so I’m definitely going to save the bulk of my findings here for another post. I will share a before-and-after picture here showing the product in action:
I shot everything with a low contrast, neutral color profile in-camera to leave the greatest possible latitude to work with in post. You can see in the before-and-after above that I significantly pushed the blacks and gray’s toward black, and slightly pushed the saturation on most of the clips.
Colorista’s three wheel system (one for shadows, midtones, and highlights) made this a quick, simple process. So much so that I plan on making this my new go to plug-in for quick and dirty color edits. AfterEffects is still there when I need it.
Suffice it to say, when I showed this piece to people as a preview, the first comment I got was, “I really like the color.” Granted, I color correct pretty much everything I shoot to one extent or another (unless going for a very specific style that can be achieved solely in-camera), yet compliments on that part of the process are usually rare, even often among those in the industry.
As you probably noticed, I did some time remapping on a handful of clips. Usually, I just shoot the intended framerate and adjust the clip’s settings in Premiere to match the output framerate. Voila! Instant slow motion.
In this project, I wanted to have the ability to emphasis certain parts of each clip and effectively time things to the music by varying the speed within certain clips. I did this using the time remapping feature in Adobe Premiere CS5.
I was pretty happy with the results. This allowed me a lot of freedom with timing in the edit, something I usually don’t have the same amount of luxury with. Typically, I tend to focus on the cinematography and planning and let the editing take care of itself. Allowing myself the luxury of remapping clips (and shooting a high enough fps that it remained smooth) gave me a lot more leeway in post.
I generally don’t like to crop things after I shoot them. In cinema, this is now a rare practice. In television, it happens to anyone still unfortunate enough to have a 4:3 display.
I like to only make changes in post that are planned for during production. For example, I’ll usually crop my still photographs in post to make them fit common dimensions only if I plan it that way while shooting. I sometimes even go with strange dimensions. I once did an album of 3-by-1′s, but only because I went out planning to come back with an album of 3-by-1′s.
In the cinema world, there are fewer choices. Mostly 16:9, 21:9, and occasionally still 4:3. You pretty much plan to shoot what you want before ever framing an image, regardless of the cinematographer’s style.
For instance, I was the DP on a split-screen piece that required everything to be shot in roughly 5:4 so that when placed side-by-side with another image, each final frame would measure 16:9. We discussed the framing ahead of time and taped off our 16:9 monitors so that onset, we never saw anything other than the final 5:4 frame.
(Only while viewing dailies did we see the massive number of lights, crew members, dolly tracks, and other grip equipment that we recorded on the other half of the image! It was cropped out and entirely out of mind on set, since we knew it would also be gone in the end product.)
On this short, I shot full 16:9, and cropped off the top and bottom in post to produce a 21:9 final product. I didn’t use any guides while filming, simply because I wanted to save the option of whether to apply the 21:9 crop in post or leave the full 16:9 image.
In the end, applying the crop was beneficial, because it allowed me to reframe up or down as needed to better frame some of the shots in post where the person catching the disc reached higher than expected and such (hard to nail such things on-the-fly when you have a zoom lens designed for stills…). That way, I could get the image I wanted without punching in on some shots in post and losing vertical resolution as well (I did already have to settle for 720P…).
As such, some of the shots are too wide for my taste, as they weren’t composed strictly for 21:9, but rather as a compromise between that and 16:9. I do hope to do more pieces (probably landscapes and more cinematic stuff) in proper 21:9 in the near future.
Finally, I’ve included the main recap video from 2010 (embed SD, click here for HD).
By contrast, I edited this in 4 hours (had to be ready for the weekend’s closing ceremonies), from 5 hours of footage. That’s more than a bit of a rush edit…