Sat vs Sat curve compared to Hue vs Sat explain please.

I watched a tutorial the other day that made me aware of a curve I had not thought to use previously, and I thought it could solve part of the intense Magenta/red saturation that I get in some shots from the BMPCC4k.
However, in practice, I don't really like the results.
The tutorial used the Sat vs Sat curve to tone down any colour that was over saturated. I quite often get colours in a shot that are way more saturated than the rest of the image. I have a shot here, where the red part of the bike is taking over the image, but I don't want to loose saturation in the skin tones. Previously I've used Hue vs Sat to tone it down, but this always brings the skintones to a muddy dark beige. The bike is clearly more saturated than the skin, so I figure Sat vs Sat would help tone it down, and it does without effecting the skintones too much. But the effect that this curve has on the bike is clearly different to the Hue vs Sat. With Sat vs Sat, the gradation between what is affected and what is not, seems a lot harsher. It creates unnatural bands within the red of the bike. Hue vs Sat looks a lot more natural on the bike, but ruins my skin tones.
I've exaggerated the Sat vs Sat adjustment a little to show what I'm talking about here. You can see the darker parts of the red are heavily contrasting with the lighter parts. This does not happen with Hue vs Sat.

Screen Shot 2022-08-17 at 8.47.35 AM.jpg (395.07 KiB) Viewed 1425 times

Screen Shot 2022-08-17 at 8.49.38 AM.jpg (379.16 KiB) Viewed 1425 times

Is it 8-bit 4:2:0 source video? That would explain the banding, perhaps. Could be the Sat vs Sat can break footage with less color info more easily than Hue vs Sat (I don't know, just guessing).

Is it 8-bit 4:2:0 source video? That would explain the banding, perhaps. Could be the Sat vs Sat can break footage with less color info more easily than Hue vs Sat (I don't know, just guessing).

It's BRAW
Thinking about it a bit more I realise that it is the darker areas that hold more saturation, naturally. The Sat vs Sat adjustment is targeting based on saturation levels. The darker areas are more saturated. It's bringing them down more. Hue vs Sat is targeting the hue and bringing all of that hue (light and dark) down.

But I would be interested to hear any ideas for smoothing this banding issue out, the curve does not quite cut it.

I wouldn’t have expected banding with BRAW. Maybe someone else will have a suggestion.

I wouldn’t have expected banding with BRAW. Maybe someone else will have a suggestion.

Yeah, it's definitely not the old 8 bit banding.

If just the brighter part of that magenta area is oversaturated, I would try to use the Luma vs Sat curve as well, in addition to the other curves. Maybe increase the Sat of the darker stuff and decrease the Sat of the brighter stuff, in addition to the Hue vs Sat fine-tuning.

As there are multiple curve parameters involved here, I would also try the Color Warper instead of the Curves. (Maybe with max amount of warp grid points and with Auto Lock set to 1 or 2 points.)

Are you using Gen 4 or Gen 5 color science?

Convert RGB to HSV or some other model where saturation is separated and look at that channel. It could be pretty far from being smooth.

aha, yes, these suggestions make sense, I'll give them a try, thanks

try keeping more of the sat vs sat curve straight across so all the values move down the same amount see if that helps. Not sure if it will but give it a try.

Are you using any colour management systems?

Could always reach for the qualifier if the curves are not working.