Canon C100: Ninja 2 ProRes v AVCHD (hi ISO)
I carried out a quick experiment tonight to see if the extra bit rate and colour space of ProRes 422 recorded by the Atomos Ninja 2 was any better than the internal AVCHD 420 codec recorded by the Canon EOS C100. I use the medium setting 95% of the time so this was the setting I used on the Ninja 2. This is a totally unscientific test and I can in no way claim that the results are definitive. The test is intended to allow me to decide whether or not to continue using the Ninja 2 with my C100 on a regular basis.
Here is a video carried out in a room in my house using available light only. The ISO varied from 12800 to 20000 so there is some noise present:
I was slightly supprosed that there is more noise in the ProRes (left) than the AVCHD, but I have read somewhere that the AVCHD is cleaned, compressed and processed in camera whereas the HDMI signal is sent straight from the sensor in the C100 to the Ninja. I also can see the that the ProRes footage is slightly sharper when viewed on my FCPx timeline and this may be exacerbated by the shaky unstablised handheld footage shot using my Canon 70-200 F4 L non OS lens. The sharpness advantage is less noticeable when Vimeo compression is applied however.
I can see a definite sharpness advantage in the ProRes stills above, and a slightly softer look to the AVCHD stills. Perhaps its due to the higher ISO used or the shaky fast movement. However, when the picture is moving (i.e. video) its hard to tell the difference in sharpness, however, the noise in the AVCHD is ugly! Its more organic in the ProRes footage. There are a whole bunch of different factors that need to be taken in to consideration when analysing the results of tests like this. Maybe you have your theories and opinions as to which is best but for me the ProRes recording Ninja 2 has the edge on the stills test.
So whats my conclusion? The Ninja is a worthy device for broadcast use where higher bit rates are required. Its also a great device where fast movement and high detail is present and where heavy grading is to be carried out. Personally for me the disadvantages are its extra weight on the camera, the “hard to bend” HDMI cables (they’re always in the way!), and of course the larger file sizes it produces. For example, the test video above shot at ProRes 150mb/s was 682.7mb whereas the AVCHD 24mb/s file was only 139.9mb. Added up over a large shoot and the differences become very apparent.
Will I keep using the Ninja? Time will tell… I want to carry out a lower 850 ISO test in daylight to see if the results are as clear cut.