C100: Is the high iso performance really that good on C series cameras?

R

RZE

Active member
I saw some comments on nofilmschool.com that inspired me to test how the c100 low light compared to what is understood by at least the online community as a low sensitivity camera, the red scarlet x. I exposed middle grey at 3200 ISO on both cameras, I used clog avchd on the c100 and added contrast to remove some of the green magenta artifacts and to make the image similar to Red gamma 3. On the scarlet I only recorded through the hdsdi to an atomos samurai in prores 422 hq no raw. I denoised the second shot in neat video will only 10% of luminance noise reduction. Results are not what you might expect.

http://www.mediafire.com/download.php?3qyw9j094tc5d3u
 
You...denoised using Neatvideo? 10% luminance noise reduction does still does quite a lot because it does temporal noise reduction too unless you turn it off.

You should've went to ISO 20 000 to really test that low light. C100 is a bit noisy in the shadows but it's noisy in the shadows also at iso 850. The midtones and highlights will be cleaner. When DPreview.com or another photography site does a lowlight test, they test all the isos. Because the real differences are quite easy to spot when you go high.

Same thing with the Sony FS100 which is great in lowlight.
 
Ryan, thanks for doing another test.

Like I said on Nofilmschool, it's fine using NeatVideo since it's the end result that counts. If Scarlet somehow can put its resolution advantage in play through downscaling to 1080p then that's fine as well.

Looking at your images I have the following remarks:

Scarlet is clearly noisier pushed to ISO3200 and this harmonizes with "internet consensus" I would say. Scarlet at this levels needs NeatVideo, while the C100 is totally usable.

Unfortunately the C100 was quite shaky, but I managed to extract one of the sharper frames. I then compared this to the first Scarlet sequence that didn't have any noise reduction. Resolution wise the C100 is at a disadvantage since you have framed its shot wider and the card is smaller in the frame. But still, even under these circumstances, the C100 outresolves Scarlet. Text is crisper on the card and the same goes for the texture/scratches on the clamp.

It's pretty much what I expected and now I'm really curious about how these two cameras compare at lower ISOs. Also remember.... You are comparing Scarlet ProRes to internal AVCHD on the C100.

Exiting times!!

AS ALWAYS: open images in new tab to look at 100%. It's more fun that way =)



Scarlet.jpg

C100.jpg
 
There was no way I was going to put the scarlet on the ground in my basement so I had to do the c100 handheld. What you pointed out is interesting on the sharpness on the c100. You definitely are getting more out of the box, but one thing to keep in mind is that the scarlet has no sharpening at all applied in camera. My c100 does. I have to do another test with proper processing and higher isos. It certainly interesting to see the differences between c100 avchd, c100 prores, scarlet prores, and scarlet raw. They all have unique pluses and minuses. I'll shoot some stuff at lower Isos but probably not test chart stuff maybe a small lighting set up.
 
maarek said:
You...denoised using Neatvideo? 10% luminance noise reduction does still does quite a lot because it does temporal noise reduction too unless you turn it off.

You should've went to ISO 20 000 to really test that low light. C100 is a bit noisy in the shadows but it's noisy in the shadows also at iso 850. The midtones and highlights will be cleaner. When DPreview.com or another photography site does a lowlight test, they test all the isos. Because the real differences are quite easy to spot when you go high.

Same thing with the Sony FS100 which is great in lowlight.
Click to expand...

Based on your comments on other threads, can you try to leave this one alone
 
By all means keep the tests coming! If possible keep the setups as similar as possible, but I'm sure you're doing that anyway.

We shouldn't forget that Scarlet and the C100 has similar (recording) pixel dimensions. C100 is 3840x2160 and I think Scarlet is shooting true 4k at 4096x2160?

It's also possible that Scarlet suffered a bit since it was the internal downscaling to 1080p? It's possible that a full 4k output from Scarlet, noise reduction on that, and then export to ProRes would be crisper? But it also adds to the workflow and storage/time equation of course...

Anyway, for 1080p delivery there is no denying that the 4k-no-debayer-sensor is quite successful for Canon at the $6k mark.

But we should also note—perhaps contrary to what people are saying—that Scarlet can be made to look good at ISO3200 as well.
 
My conclusions were exactly the same. I've had the scarlet for a year and have never gone above 800 iso because of rumors of low sensitivity. But at the same time I had to attach a bridgeplate, screw in a noga arm, mount the monitor, mount the external recorder, attach the battery plate and attach the battery, for the c100, I just took it out of its case. It hard to describe the feeling of prepping those two cameras back to back. Its pristine quality vs really good quality at the cost of slow set up and operation. c100 will still be chosen on most projects for me.

Keeping the tests completely controlled is difficult when I'm filming them myself. If I can get some studio time I would rather do it there but its up in the air.
 
You must log in or register to reply here.
Back
Top