Gen 5 vs Gen 4 Colour Science Grain/Noise

Can anyone confirm with me if you recognize that Gen 4 colour science on BMPCC4K is cleaner than Gen 5 Colour science in a darker environment?

I have avoided upgrading because of this for a while, but I am missing out on a lot of other good updates because of it.

The latest firmware I'm able to run with the cleaner darks is on 6.9.6.

I do understand that there is DUAL Native ISO and I have troubleshot this on both firmwares. Whenever I turn on the Focus assist and put a lens cap on, for the newer firmware (7.9.1) on Gen 5 Colour Science, there is so much noise. When I do this same test under firmware 6.9.6 with focus assist and the same 400 or 1250 ISO, there is no grain.

Can anyone explain to me why this is a thing? Is it because of the new parameters within the colour science?

Also, the LOG looks a lot more different within each Gen Sciences.

It is also pretty noticeable within a colour grade as well. Why do I feel like the lower firmware handles darks better than the newest firmware?

Are you seeing a difference [just] on the screen, or when the two shots are recorded (with identical exposures) and then graded?

The short answer is, there's no difference in noise between the two color sciences, although noise may be more apparent, before any grading, in one than the other.

As for log differences, it's true that Gen5 is more challenging to grade. You may want to use Resolve Color Management, to make grading easier and with a more accurate normalization than you can do yourself.

Why not just upgrade and try... and if not to satisfaction can always go back to earlier version.

I am seeing a difference on screen and also when I import into Davinci Resolve as well. I kept all the values the same on both firmwares when I was troubleshooting.

I thought it was just how Focus Assist is managed, but if I am looking at focus assist with just a black screen (my lens cap on), wouldn't it try to pull focus from grain?

I'll try to include an image of both what is on the camera screen for each firmware and also what they look like colour graded in post.

Why not just upgrade and try... and if not to satisfaction can always go back to earlier version.

That is my exact dilemma. There are other upgrades in the firmware I would be sandbagging myself with if I don't upgrade. I want to know if this is just what it is or can if I "fix" this particular difference within both firmwares.

You need to supply the original files, if you want to get to the bottom of it. How you grade the material could be the problem.

The sensor noise levels are not affected by the color science. Look for user error, somewhere or other.

It’s a much flatter log profile, so Gen 5 will have a lifted black compared to Gen 4 to start with in Resolve and the noise is more visible until you push it down. I often start with extended video or video instead of the film curve.

If you are viewing it in camera and the focus peaking is reading the sensor not the LUT then I imagine peaking would be affected right?

You can still use Gen 4 with Braw in post by selecting it in the raw decode in Resolve after upgrading the firmware.

gen 4/5 are a color science where you develop signal, noise born before color science.
Gen 5 give you better color, better skin tone, better rolloff on highlights, less halo around high saturated colored light and more.

Did you test it scientifically? did you de log by OFX? By wide gamut RCM?
you simply load file in resolve, setup delog, switch in camera raw tab between gen 4 and gen 5.
i can tell you that i saw exactly the opposite, less contrast during encoding give me less noise on high compression and allow denoiser and neat video to give me better picture.

could you post a sample of picture, a simple braw sequence to test it?

just have seen that gen 5 with last firmware you have different focus peaking value? may be that in old setup was low and now high?

Compare Gen 4 to Gen 5 in Resolve.