There is an awesome LinkedIn discussion happening on HD cameras for documentaries. I recommend you reading through the thread, regardless if you are in the market for a new camera or not.
In addition to these comments, I would ask you to think about some technical difference that might be important, depending on your production/post-production pipeline:
>> Subsampling Model (4:2:2, 4:2:0) – 4:2:0 does not do well in green/blue screen environment – too much noise. If you plan on doing a lot of post effects, then go with 4:2:2 or better.
SIDE BAR: For those interested in seeing the difference between 4:2:0 and 4:2:2, at least on a Canon 7D, check out this demo by Patrick Tong.
>> Chipsets (CMOS): DSLR CMOS chipsets (e.g., Canon 5D, 7D, etc) induce more noise over time since they are not designed to run continuously. Take a look a the first frame and compare it to 30 mins of on/off/on/off operations, significant visual difference in noise in mostly the blue channels.
>> Operations: Check to see if there are recording length limitations. For example, Canon has a 12 min (actually a 29min 59 sec bases on keeping it out of the video production class – different EU tax structure). In the 12 min case, it is based on reaching a 4GB storage limit.
>> Bit Depth: 8bit vs. 10bit vs. 12bit – All the cameras you picked are 8 bits. 10 bit (Cineon) and 12 bit (DPX) offer a lot of recording latitude (more stops).
>> Codecs: MPEG – not edit friendly since it does not store all the luminance/color data in a single frame (delta frame). ProRes – uses a different intra frame MPEG that is user friendly. MOV (H.264) – compression artifacts. RAW, CINERA, ARRI RAW – lots of data to work with, very edit friendly
>> Depth of Field: DOF is our friend and adds tremendous dynamics to a scene that can be capture in camera. The Canon 7D, with a fast lens, has great dof control. Other cameras may need external adaptors (Redrock Micro, etc.) or post-production processing.
For example, in think about whether a DSLR camera is right for recording a documentary, I would go back to the nature of the documentary itself – production (where will it be filmed), post-production (how will it be made), and venue (where will it be shown). If you need to post/distribute with lots of high quality color with lots of latitude (4:2:2 and 10/12 bits), then DSLRs are not a good fit (more the realm of a Red camera). If you can live with high quality color and modest latitude (4:2:2 and 8 bits), using a DSLR with a nanoFlash type transcode storage drive is great. On the other hand, if you can get by with modest color and latitude (4:2:0 and 8 bits), then you can shoot with just an DSLR or even an iPhone.
Once the color and latitude are thought through, other cinematographic characteristics should be reviewed. DOF, being one of that is often overlooked, is a core strength of the DSLR (through a fast lens). Being able to control what is being focused on in camera through DOF is preferable over doing adjustments (lens blurs) in post (it is cheaper and better quality). So, sticking with a pretty good, like weight, mobile rig with high quality color and great latitude could be a DSLR + nanoFlash. But, there are lots of combinations you can choose, once one gets out of the one camera vs another discussion. For example, grab a Canon Vixia HF20 and add a nanoFlash recorder => 4.2.2 color (great), still 8 bits (ok), 24p (awesome), aperture priority (quasi-dof, not bad, but not great).
So, the point I am trying to make through the last few exchanges is that you need to define the production pipeline (codec, chroma, etc.) and pick the camera/storage best suited for it, rather than the other way around. In the end, as a compositor, one doesn’t care what you shot the scene in. However, they do care about the quality of the color, range of latitude, and keeping them away from non I-frame codec so that I limit making cuts natively. Make sense?
