When you read the economic headlines, the first question has a lot of weight. Me, I’m following the sage wisdom of Brooks Jensen in the Lenswork Podcast #478 — I’m refusing to participate in the recession. But it sure doesn’t mean that I’m ignoring it.
When I decided to make photography my full-time profession I had already committed to including video in my technology mix. The shingle says “photography”, but I consider what I do to be “visual storytelling” and the technology to do that includes moving and still imagery. Video is quickly becoming as pervasive as still photography and an important component in any advocacy or documentary project.
I’ve had a small video camera for a few years. I purchased it hoping to do some video during my Wild Mustang photo shoot back in 2007. That particular subject was far beyond the capabilities of the camera. I quickly realized that doing video well, particularly nature video, was going to require a lot of hours of practice.
So about this time last year I purchased a Canon XH-A1 video camera. I don’t have nearly as many hours with that camera as I should, but coming from a photography background I was struck on how difficult it was to compose scenes the way I wanted to. I also, obviously, do a lot of nature photography and the limited focal range and depth-of-field of the video camera irked me considering the large investment of glass I have for my cameras. Within a month or two of owning the camera I was looking at gizmos like the Redrock Micro, but that looked to be impossible to use in the field.
When Canon announced the 5D Mark II (hereafter referred to as the 5DMk2) would include HD video it seemed like a perfect fit: a full-frame, high-resolution DSLR that could also shoot video (just like my 3-year old Elph point-and-shoot … only good) using all those lenses I’ve spent years acquiring — a solid camera that would stand up to the elements.
Then came Vincent Laforet’s Reverie short, shot entirely on the 5DMk2. Here was a video, shot by one of the best photographers out there, and it looked pretty darn good — particularly in low light. I talked it over with my financial advisor (Betsy) and put a deposit down for one the next day.
What works.
The 5DMk2 hits a particular sweet spot in my workflow. I now have a camera that readily switches between very high quality stills (and with low-light performance that greatly increases the quantity and final quality of possible subjects) and can very quickly change to shoot SD or HD video.
There are three reasons why this is of value to me: 1) video allows me to capture certain types of behaviors and scenes that simply don’t work well as still images; 2) even relatively static scenics have a certain subtle verisimilitude which works better in video than stills; and 3) when I was purchasing the XH-A1 I asked about “environmental seals” on the camera and I got a puzzled look — at least the 5DMk2 can get a little damp and operate in the -40 to 100+F range. OK, four reasons: 4) I think there is a business window of opportunity for stock video and I’d like to see if some money can be made there.
Having a separate video camera, and believe me I’ve tried, just doesn’t work. The question was whether or not this “combination camera” would.
From a workflow perspective my answer is an emphatic “yes”. Granted I’ve not spent months with the camera, so check back here in May to see if my opinions change – but I’m pretty confident about this. The ease by which I was able to switch between shooting video and stills was nothing short of exhilarating: compose the shot, focus the shot, take the shot, would this make a good video?, yes?, hit the button, recompose slightly for 16:9, if the shot still works, set the exposure, and shoot (the focus is already done). If the shot doesn’t work you’ve lost 4 seconds.
If I’m shooting primarily video, I can compose and focus scenes many times faster and more accurately than I could with the video camera. As a still photographer I’m already used to standing on my head for odd shooting angles, so the lack of a movable view panel isn’t something I’m going to whine about.
A 7-second HD video clip roughly equates to 1 still. Most clips are in the 7-15 second range, with some behavior clips running longer.
When I get back to the laptop, I stick the card in and my custom import workflow copies the stills to an “Inbox” folder for subsequent Lightroom processing while sending the movie clips to an “Inbin” — with all files renamed to my personal scheme.
I can review the video clips with the new “Quick Look” feature in Mac OS X Leopard: Cmd-Y on the file and it plays a large version of it. If it was garbage, delete the file. Bam.
If I need to put together a quick video with the clips, iMovie imports the H.264 files and you can do a simple cut.
But normally the clips come back to the office, transferred to the clip library. Those that look to be worth using for a project are dragged to a droplet that runs Compressor, transcoding the H.264 to ProRes 422 HQ (and to 720p if I know that I’m just going to make a web video anyway and want to save some space).
Edit, grade, compress, publish…
So, what’s wrong?
I previously mentioned the Vincent Laforet Reverie video. Unfortunately, Vincent’s chosen subject inadvertently (or perhaps intentionally) made the camera’s video capability appear much better than it was — unless you were always shooting at night. It was known that the 5DMk2’s video would not have full manual control, but I didn’t count on how fully automatic Canon made it.
From my rather limited perspective there are two significant things wrong with the Canon 5DMk2 video mode: it only shoots 30p (no 24p) and the draconian automatic exposure control. The sad thing is that both are fixable in firmware. The really sad thing is that they probably won’t fix either. Deep down I’m hoping that Canon takes advantage (so to speak) of the economic slowdown and actually updates the firmware instead of spending hordes of money developing a whole new camera before the potential of the one they’ve just shipped is fully realized.
The lack of 24p isn’t the show-stopper for me that it is for other people, but it is a powerful creative option that seems easy enough to have included (obviously there’s enough bandwidth for 30p, a 24p rate should let the camera run cooler and have better storage usage).
Many folks lament the lack of full manual control. I do too. But I’d be willing to make a compromise that Canon already knows how to do and is rather easy to implement control-wise: move “program shift” into the video mode. One of Canon’s automatic still modes, “P”, tries to do the right thing, but you can rotate the shutter wheel to get different shutter/aperture options. I’d be happy if they would just play aperture vs. ISO against each other. That would give me about 80% of what I need about 80% of the time. I can play games with ND filters for the other 10-15%.
Oh, there are plenty of other annoying aspects to the video shooting mode (like it resetting the exposure compensation) and no histogram or “blinkies” during composition – but I can learn to live with those.
Bottom line: it is close to the ultimate “b-roll” camera for a photographer. I’ve used it for nature and business shoots. In a pinch I might use it for interviews, but I’d rather use the XH-A1, plug the mics into the XLR jacks, set my audio levels, and let the tape roll for 15+ minute interviews. But after the interviews are done, I can shoot b-roll with the 5DMk2 about 10x faster than I could with the XH-A1 using anything from telephoto to wide-angle to macro lenses.
Did I make the right decision? Overall, I think so. Purchasing the first generation of any new technology (in this case a combination DSLR and HD video camera) is fraught with danger and we’re certainly bumping into some disappointing limits here and there. (Made more disappointing as they would appear to be purely parochial marketing decisions rather than engineering or technical limitations.) But my hunch that it fit well into my normal shooting and processing workflow has, so far, proven correct and I expect to start producing video at a much higher rate than I could with the XH-A1 alone.
What’s next: I need to come up with a good “frankenstein” rig that will take will work well in the field. It will need to support both a regular video camera as well as DSLRs, have a place for audio levels and mics, and have a place for a fill light. I’m also going to rig up a nice “rebel” follow-focus (at minimum to make precision rack focus moves). This will take a few months of experimentation and I’ll report on these changes as I go.
