Earlier this month, Canon announced the C300 as the first product in their new Cinema EOS series, which they see as their next step in the world of digital filmmaking. They stumbled into the forefront of DSLR film-making entirely by accident, and are trying to figure out how to capitalize on that position. Their new camera is intended to bring some of the benefits of DSLR type filmmaking to the high end market, without the limitations presented by the existing 5D and 7D cameras.
Canon advertises the C300 as having a 4K imager, although that is misleading for two reasons. QuadHD (3840×2160) is not technically 4K, and regardless of imager resolution, it only records 1920×1080. This seems like a wasteful oversampling until you compare it to other cameras at a technical level. A traditional 3-chip video camera with full resolution sensors is capturing 3 times the output resolution, with the extra data used to generate color information. A DSLR type camera has a single large sensor, and doesn’t use a prism to isolate the colors, so a bayer pattern filter is used instead. This usually requires a complex demosaicing process to extract the color information from the RAW image, like an R3D or CR2 file, and this involves a lot of interpolation. If a full resolution decode is not required, then it is simpler to assign the values from a square of four pixels to the three channels of a new single pixel. This is how a fast 2K decode works from a 4K RED file. Canon took advantage of this concept to create a camera with a single large imager, that doesn’t require a processing intensive demosaicing because they are oversampling. This allows them to compress the resulting 1920×1080 three channel image in an existing standard codec instead of having to create something new, and then supporting the required new workflow that would entail. They chose to use their existing XF codec, which is 4:2:2 MPEG-2 at 50Mb/s and is already compatible with most systems and software. This method of processing the image from the sensor will alleviate many of the moiré and aliasing issues caused by the line skipping method used on existing DSLRs to pull an HD image from an 18-Megapixel sensor. Unfortunately this will not necessarily solve all other rolling shutter issues that are prevalent in DSLR footage, but those are also found in footage from RED cameras, and nearly every other large sensor imager. All in all it is a very effective design, but Canon’s 4K marketing is misleading. It would be more accurate to call it a full color single sensor HD camera.
Canon’s existing video DSLRs have two major things going for them. One is that their small lightweight form-factor is easy to use and allows the cameras to be put into places that any other high quality camera wouldn’t fit. The second is that with such a large imager, they give a unique style to the images they shoot, that many people prefer, even when compared to the results of cameras that cost 10 times as much.
When compared to existing HDV and AVCHD camcorders, the idea of a video DSLR totally revolutionized the low end market. It had so many advantages in that space, that it even caught the attention of the high end market, in places the Canon never dreamed that it would go. For the money, nothing else even comes close in regards to image quality, so DSLRs are an obvious choice for students and other independent filmmakers, and freelance videographers.
But people who are familiar with larger professional cameras look at a DSLR and see a huge stack of limitations. (SDI output, 15min record limit, remote control, excessive moiré and aliasing, etc.) They are looking for the image style and quality of a DSLR, but without those limitations, regardless of the cost and size of the resulting product. The C300 seems to be aimed at that market, and instead of revolutionizing it, it is just going to offer itself as one more option, which is a disappointment to those who were expecting more. It offers nothing to further the interests of the independent videographer or film-maker, because that is not who it is targeting. It is designed to tie into the existing equipment that larger organizations have invested millions of dollars into, the benefits of which, they are not willing to forego, in favor of the image style and depth of field offered by existing DSLR sensor systems. Whether it will be successful in its target market remains to be seen, but I am confident that many of its key features will eventually make it into future products that are more similar to the existing 5D and 7D DSLRs.