F35: Make it more filmic

Haha I've never shot it above 450 so I feel slightly ridiculous thinking it would crank that high

Mac, I've kept it in Cine-EI too- I figured in regular CINE or Custom mode when you navigate outside of native ISO the DR % might start to drop or become remapped as with other cams.

As with SJX I'm also mainly wondering what kind of adjustments you made in resolve and how you measured stops gained- perceived is certainly fine in this context but I know there's probably a more scientific method too.

Yes. Can somebody please explain?
 
In Cine EI mode, the ISO means pretty much nothing, as all you're doing is mapping the mid gray into a different zone of the system zone. The sensor is capturing the same image but a different gamma curve is applied on the monitor output. Higher ISO means setting mid grey into a lower zone.

Alexa-Dynamic-Range-Spread.jpg

This graph is from the Alexa and illustrates this concept clearly.

On Cine EI in the F35 however, the ISO setting does not change the gamma curve (which happens in the Alexa) of the HDSDI outputs (what you should be recording from) and that process should be done in post by the colorist.
 
Here's the 4K version, better compressed for youtube.


So basically I'm out there exposing the camera as if it was a 2000ISO with a lightmeter.
I'm also shooting in tungsten modo, because I find it very easy to dial the WB in Davinci afterwards.
For this test I used the old Kodak film print emulation LUT posted on the main F35 thread.
The grading has a few nodes but nothing too crazy.
Neat video comes at the end of the chain, and it's not a specially built profile or anything. I just use temporal noise reduction, no spatial noise reduction, and I blend it with the original at 60%.
The results are undoubtedly very filmic, basically because we are taking advantage of the great colorimetry of this camera and adding two extra stops on detail in the highlights while reducing the shadows, which as explained earlier, it's closer to what film does.

I will shoot my next project with this technique for sure.
 
Cool, thanks for the explanation! Can't wait to try this all out in my F35 (long story but we are separated at the moment). In the meantime I have been devouring everything I can find about the F35 till I get my hands on mine in July. Needless to say I look forward to everything everyone can post about the cam esp. Footage, and workflow tips and techniques. The stuff you have done with yours is Awesome, and you inspired me to get one for myself. All the best! I hope to contribute more to this community very soon.
 
FWIW some time ago I did a project and to recreate "home movie" footage I shot with a lot of gain in the F35, a small F-stop (about 16) and an off-frame rate, about 16-18 or so. Really did a good job of mimicking old home movie stuff.
 
In Cine EI mode, the ISO means pretty much nothing, as all you're doing is mapping the mid gray into a different zone of the system zone. The sensor is capturing the same image but a different gamma curve is applied on the monitor output. Higher ISO means setting mid grey into a lower zone.

Alexa-Dynamic-Range-Spread.jpg

This graph is from the Alexa and illustrates this concept clearly.

On Cine EI in the F35 however, the ISO setting does not change the gamma curve (which happens in the Alexa)

Exactly. This will not work with F35 to shift it like above. Because iso don’t change internally. So information is already recorded no matter what you do in post. It is not RAW. So you can’t shift like above in post like with Alexa. Alexa do it in camera. So if iso don’t change in F35 you are not shifting or gaining anything. What I’m missing?
 
You can't shift it in camera but you can shift it in post. That's what color grading is about, right? It's the exact same thing.
 
You can't shift it in camera but you can shift it in post. That's what color grading is about, right? It's the exact same thing.

How is this exact same thing? After recorded it is recorded. You can't shift like this. F35 is not RAW. So I don't understand you.
 
OK let me explain a bit.

Here's a simplified scheme of what happens in a camera (if there's an engineer in the room please correct me): light hits the sensor > AD converter converts the electric signal into digital 0s and 1s > camera boards creates a image (linear) > other boards make that image pretty (ISO boost, gamma curves such as rec709 or Slog or whatever) > pretty video output (clean output to be recorded either on board or external).

On a Red camera, which captures raw files, the chain ends right after the AD conversion basically, and that is what it gets recorded. The rest of the process is done in your PC.
On a Sony F35, you get the full chain until the pretty output.
On an Alexa it's the same thing, but a bit more flexible because you have the possibility of recording both RAW or the the video output.

Since these cameras do not seem to apply analogue gain to the sensor to boost the signal, the ISO/sensitivity increment is done by adjusting the gamma curve when going from linear to rec709 or log gamma.
This needs to be done at a high bit depth (10-16bit) to get decent quality. In a Red camera this is done in your computer (in redcine or Scratch or whatever you like to use), in the Alexa this is done in the camera if you're recording Log, and in the case of the F35 this ISO boost is done in the grading (if you are shooting in CINE EI mode). This is how Cine-EI is supposed to work.
In any of the above cases the cameras recorded the same image, it's just in the processing and what curve is applied (which dicatates where mid grey will be mapped) what makes the image brighter or darker.
 
Last edited:
Back
Top