Recently, some friends asked me what dSLR to purchase if they want to make movies with it. They aren’t experts at doing SLR photography.
Nikon calls movie taking “D-movie.” It is currently the cheapest dSLR that can do video mode. It’s only one of three dSLRs that can do autofocus while taking video mode. This strikes me as the best balance between learning and using an entry level dSLR and being to take film-like movies. I’ll recommend some others below, but first I’d like to talk about the why and what of SLR movie-making (with the caveat that I’m a photographer, not a filmographer).
It’s about the sensor
First, I want to brag a bit. I predicted the emergence of dSLR moviemaking (to replace even videocameras for high end work) four and a half years ago. And the reason is “the film look.”
Over the years, various theories have been advanced over what encompasses a “film look” that makes film so much better than video. Some have claimed that it was the 24 frame per seconds (video did 30 or 60 interlaced), others had claimed it was the high-definition sharpness (video was limited to 480 vertical lines of resolution—film was analog but no more than 6000), still others have stated it is the processing (another future article). Each of those has fallen—to variable rate cameras, to high-definition, to plug-ins like After Effects—but still the “video-look” looks bad.
I’ve felt that while the above has contributed to the look that we recognize as film, that there are really two things that distinguish cheaply shot home movies (a la your average pr0n site), and a “film” movie: sound and image quality.
Sound is a production issue involving XLR inputs or a separate DAT or memory card digital recorder and an audio engineers time. It can be hidden by just putting a music track and going all montagé on things.
And image quality? It’s the sensor. The sensor in a video camera, even a high-definition video camera is 1/3″. The smallest sensor in the cameras I’m discussing is 16x bigger; the largest sensor is 32x bigger. A larger sensor means a smaller depth-of-field as I discussed here.
And that larger is immediately evident in the focus effect:
The limitations of the dSLR video mode
The most important limitation is the “jello effect” due to a rolling shutter. What is going on is that the bottom part of the sensor is recorded at a different time than the top part of the sensor. What you will see if you are panning or there is a lot of horizontal motion (for instance you have the dSLR video pointed sideways out of a moving car), is a jello artifact. Note that is a limitation based on the fact that a CMOS-based sensor designed for photography is being hacked in the image processing unit to handle video. This means that if movie-making on these cameras continues to be viable, the rolling shutter will disappear in future models. In the meantime, the solution is to simply avoid shots that exhibit this effect. This means not panning quickly (not recommended anyway), and doing shots where there isn’t too much side-to-side motion (front to back is okay).
The second limitation is that the dSLR camera design has exposure and autofocusing systems active only when the reflex-mirror is down. In order to record, this mirror comes up and you cannot get SLR quality metering (and especially) focusing. This means the bulk of these cameras don’t autofocus during exposure.
The third limitation is that since it is primarily a photographic camera, sound is an afterthought. This becomes especially annoying since the microphone can pick up any lens sounds.
Other limitations are increased optical cost (and weight), a body designed for handholding (instead of a shoulder-mount or steadycam), lack of a powered zoom—zoom is of little value anyway in anything outside a home movie, and lack of a RAW mode for heavy post-processing.
Having said this, there are four other cameras to consider at this point…
Reasoning for video: Panasonic is a mirrorless (non DSLR) system so it’ll have the best coupling between video and camera modes. This model also has a built-in jack for an external microphone. Panasonic is the only company with both video experience and dedication to video-in-large-sensor-camera. To see this comitment, look no further than the AG-AF100, which will use the same photographic lens system as the GH-2:
Also this model sells with the best video lens on the market, and the only lens specifically designed for video (totally silent operation)—the price linked above includes that lens which is very high quality.
Reasoning for photography: Better lenses for close focusing (macro and food photography). Articulating LCD display. Great live view mode.
Reasoning against: The sensor is smaller (2x multiplier). This means—especially with the small aperture general purpose lens above—that it has a video effect. On the other hand, it gives you insane zoom.
That video was taken from here:
Reasoning for video: It can take an external microphone. Some claim it has a smaller rolling shutter than the D3100 but I don’t believe it. (A rolling shutter is because when taking video the top of the image is recorded at a different time than the bottom—the effect can be mitigated by good camera work and preparation.)
Reasoning for photography: Higher megapixel and image quality; faster frame per second; 100% pentaprism viewfinder—it means the viewfinder is as bright as possible). It can autofocus older Nikon lenses.
Reasoning against: It costs more. It’s also harder for a beginner to use. You have to buy a lens as the price above is body-only.
Reasoning for video: It has a “full frame” when other cameras are “APS-C”. The sensor is 1.5x bigger (each dimension) than the APS-C class sensors. The big advantage of that is you get a smaller “depth-of-field” which is essential to the “film look.” It can take an external microphone.
Reasoning for photography: The larger sensor is a higher image quality and better dynamic range and low light photography.
Reasoning against: No autofocus while video. Very, very expensive.
Update: Sony Alpha A55
An earlier version of this article posted before I finished it. In it, I forgot to mention the Sony A55 and was rightly panned for the oversight. As I’ve discussed Nikon uses Sony-produced sensors, up to the image processing level, the Sony is the same as the Nikon only cheaper. However, the Sony uses sensor-shift image stabilization, which these others lack.
Reasoning for video: The system uses a semi-translucent mirror which we’ve seen before. This means this camera can autofocus (using the SLR’s AF system), while doing videotaking at a cost of some compromise to both systems in low lighting situations. That should make this camera the most responsive video preformer in its class. In addition, the in-body image stabilizer gives image stabilization during videotaking that is not dependent on the lens used. It also has an articulating screen like the Panasonic offering.
Reasoning for photography: latest generation Nikon-level image quality at a cheaper price with in-body image stabilization. Sony also has some pretty neat innovations from their Exmor pocket cameras trickled up: sweep panorama, high speed photography, multi-shot noise rduction, etc. This camera has a built in GPS. It also has an articulating screen for photographing in live view from intereting positions.
Reasoning against: Mostly, I’m still uncertain about Sony’s commitment to the alpha line. They seem to have a lot of misfires recently (the A230-380 design) and sales of their excellent full frame cameras have been anemic despite being the best value around. The semi-translucent mirror means that there will be serious compromises to the autofocus system in poor lighting.
Yes, other cameras do video, but these are the ones I recommend.
Update: Nikon D3100 video
My friend Sean, tested the Nikon D3100 to make this video (you can hear the autofocus motor also).