C300MKIII vs. C70 - An Unsolved Mystery

[chris f]

I'm working through all of my side by side tests I ran the other week, and right out of the gate I was surprised at how different the color & curve/lift seems to be between the two cameras. I shot BTS of the whole test so I can confirm that both cameras had the exact same settings across the board, but they look so different to me that I'm wondering if there could be some other user error on my part? Both cameras had been on for about 30 minutes before I black balanced both cameras. Both cameras had the same picture settings, recording settings, and the exact same lens.

Picture Profile settings for both cams are below, only difference is the line "Look File - Off" is an option only on the C70, but as best I can tell from the manual this only for turning custom loaded LUT files on or off, and doesn't apply to the built in CP files (can someone with a C70 confirm this?)
Screen Shot 2021-02-05 at 2.12.51 PM.jpg

The differences between the images and scopes won't be as noticeable on here because DVXUSER compresses the heck out of photos, but you can download the two 4K .png exports from my Premiere timeline & the Scopes here:
https://drive.google.com/drive/folde...Ku?usp=sharing

C70
C70.jpg
C70_Scopes.jpg

C300MKIII
C300MKIII.jpg
C300MKIII_Scopes.jpg

There's an obvious color shift between the cameras, but I also find it strange that there's such a difference in the scopes. It's almost as if the Wide DR in the C70 is a different "Wide DR" that's punchier....

[ozmorphasis]

Hi Chris,

What kind of color shift are you seeing, by eye and on the scopes? Can you describe it? I don't see it.

What I definitely do see is brighter overall exposure on the C300mk3. Since every part of the waveform looks uniformly higher on the C300 shot than on the C70, I'm not sure that the C70 is punchier, if by punchier you mean more contrasty.

Not sure what would account for different exposure though.

[chris f]

Hi Chris,

What kind of color shift are you seeing, by eye and on the scopes? Can you describe it? I don't see it.

What I definitely do see is brighter overall exposer on the C300mk3. Since every part of the waveform looks uniformly higher on the C300 shot than on the C70, I'm not sure that the C70 is punchier, if by punchier you mean more contrasty.

Not sure what would account for different exposure though.

If you download the originals from the Google Drive link you can see (especially in the scopes and waveform) that the C300 is leaning more yellow & green and the shadows aren't as crushed (coming in at about 4 IRE, whereas the C70 is touching 0 IRE)

[ozmorphasis]

If you download the originals from the Google Drive link you can see (especially in the scopes and waveform) that the C300 is leaning more yellow & green and the shadows aren't as crushed (coming in at about 4 IRE, whereas the C70 is touching 0 IRE)

Ok, I just looked on a different monitor and I definitely see a more magenta lean on the C70. However, regarding exposure, it's not just the shadows that are crushed. The entire C70 waveform is lower by those 4 IRE that you are mentioning. The peak of the highlights on the forehead for example, or the highlights from the lamp. If you take any of those as a reference point, that difference of 4-ish IRE is present between the two, so I'm not sure that it is a different tonality spread. It seems to be uniformly lower exposure on the C70. The color difference is definitely more concerning and notable. It could probably be easily corrected, but that defeats the point of buying two cameras that are supposedly using the same sensor for multicam shoots so I definitely understand the concern.

[chris f]

I have a friend that owns two C70s. I might see if I can borrow both and do another test with a color chart and gray card as well to get a better idea of what's going on and to make sure it wasn't a fluke or just the specific C70 that I used last time.

I expected to see some sharpness/noise differences in some of the different modes of the two cameras, but I didn't expect to see such a big color difference. Definitely disappointing...

[Jaime Valles]

Fascinating. I wonder if you'd get the same color shift if you shot in C-Log 2 on both cameras, or some other picture profile.

[boosh]

I just run a three camera setup last week with 2 x C300MKIII's and a C70 and can confirm the C70 colours do not perfectly match the C300MKIII's. I was slightly shocked as its the same sensor. Not a huge difference between them but still they should just match out of the can. Not sure if this is something Canon should really address in a firmware update. In my opinion all of the modern Canon cameras should match perfectly in terms of colour.

[Ben_Ericson]

It’s almost like the C70 has a color cast across the whole image. The image doesn't look as neutral as the C300. I actually ran some tests last week with some IR cut filters because I was wondering if the C70 suffered some sort of minor IR pollution (it doesn’t look like it.)

I haven’t done a test to see if the NDs themselves are creating a non IR related color cast or color shift. Might be worth testing? If it is truly an identical sensor, the only variable would either be the NDs used for the RF mount or the color profiles themselves.

A lot of the footage from the C70 I see online has these similar color qualities. The skin has kind of a reddish/brown look, rather than a pink or golden look you’d see from the C300ii or the C100. If you dial out the red, you end up with a really cool toned image.

I’ve been digging into the Gamma Shift settings in the color matrix settings. I think I was able to improve it but I definitely don’t like the image straight out of the camera as much as I did with the C300ii.

[digitalinnovations]

I’m following this because I find it interesting. It must be extremely frustrating because these cameras are marketed in such a way that would lead you to believe you could use the C70 as a B-Cam to C300 MkIII or even the C500 MkII with their speed booster. So much so that I had considered picking up the C70 as third camera angle for multi-person interviews but I would also want the ability to do live feeds that just match (with the same settings of course) and/or I also wouldn’t want to spend a bunch of time in post matching LOG footage.

[chris f]

It’s almost like the C70 has a color cast across the whole image.

I haven’t done a test to see if the NDs themselves are creating a non IR related color cast or color shift. Might be worth testing? If it is truly an identical sensor, the only variable would either be the NDs used for the RF mount or the color profiles themselves.

That is a good thought, but unfortunately in my testing I did lots of tests with and without the ND filters and it is the exact same difference in all of the clips.

If I can get the C70s again I’ll test all of the cams in Log 2 as well and try a different copy of the lenses, but I will say that I did a two camera shoot earlier that week with those cams with different lenses and was noticing the same types of differences...