The C100 is doing well, but it could be better

I haven’t posted many updates lately so sorry for that. I’ve been working on a project involving a lot of two camera interviews using my Canon C100 and C300. This has been a good test for the profiles I created to match the camera and so far it’s going really well, the images from each camera are really close and I’m finding that with careful setup of exposure the results cut together well with the same grade applied.

The one thing that seems to let the the C100 down slightly is of course it’s internal 4:2:0 codec writing at 24MB/s, this doesn’t hold up as well as the C300‘s 4:2:2 codec at 50 MB/s. Below are some frame grabs showing one of the interviews I’ve shot and you will notice that although the C100 close up looks very similar to the wider C300 shot, closer inspection reveals compression artefacts that become more apparent when the footage is graded.

Both cameras were shooting in log with a minimal grade applied as an adjustment layer in Premiere CS6. I used Premieres RGB curves effect to add a shallow S curve and added some saturation using the fast color corrector tool. I’ll need to spend more time on the grade later in production but for now this gives a rough idea of how it effects the compression artefacts. Both the ungraded and graded screen grabs are shown below.

The wider shot was recorded with the C300 fitted with a Canon 50mm f/1.2 lens with the closeup recorded on the C100 with the Canon 70-200 f2.8 L. Both lenses were at f/2.8.

c300log c100logc300g c100g

I’m going to be reviewing the Atomos Ninja-2 external recorder over the next few weeks to see how that improves the images from the C100. Because the camera outputs 4:2:2 over the HDMI port that should in theory produce results much closer to the C300. Watch this space for more on that very soon.



You may also like...

6 Responses

  1. Oleg says:

    c100 outputs 8bit stream anyway, can u tell whether there is difference between c100 with ninja and plain c300?

  2. Matt Davis says:

    Paul, will you be checking the three different ProRes codecs?

    If I remember correctly (and this may be with tests on EX1R ‘way back’), one had to use ProRes HQ before motion aliasing visibly improved.

    • Paul Joy says:

      Yes I’ll do that for sure Matt. The Ninja arrived yesterday and I’m hoping to have time to look at it early next week.

  3. Andree says:

    Hi Paul, thanks for sharing.

    What are we looking at here? Can you point out where you feel the C100 breaks up and where you see compression artifacts?


    • Paul Joy says:

      Hi Andree

      The AVCHD from the C100 contains blocky artefacts which although not obvious do result in an overall texture on the images that the C300 doesn’t exhibit. This texture is generally well masked but can become more apparent with grading as the blocking becomes more visible.

      If shooting with just the C100 it’s not something that would likely cause too many concerns, but when shown side by side with the C300 footage it’s clear that the C300 produces a cleaner image that does not exhibit this texture created by the C100’s internal codec.

      • Andree says:


        OK, I think I understand what you mean.

        Do you see those artifacts in the images you posted?

        I’ve done some pretty serious pushing and pulling on my C100 footage, and I think it holds up beautifully.

        I’m having a hard time motivating buying a Ninja 2, even if that would put it firmly above the C300 internal codec.

        What is your take?

Leave a Reply

Your email address will not be published. Required fields are marked *