Saturation

[Irri]

I realise I am being a little rude here by asking a “not specifically RED” question, but it is certainly a question that relates to RedAlert…

When I lift saturation, what exactly happens in an RGB colourspace? Doesn’t lifting all channels equally just increase luminance/brightness?

I had an interesting conversation with a director and he was wondering what gave video a certain “video look” in terms of colour. I could only guess at the idea of matrixing pushing the colour values, and that we could try using very saturated colours in the costumes etc to see if that made a difference.

[Anders Holck]

Im no expert, but I believe you covert to HSL or HSV in floats, apply the adjustment and convert back to RGB. The adjustment is simple as you can work directly on the Saturation channel.

[GlennChan]

Most saturation algorithms will boost one/two channels and make the other channel(s) darker.

2- Video look:
Film has a s-shaped transfer function to it. Combined with very high dynamic range, it tends to produce detail in highlights and shadow. (Whereas with video you'd get noise on the bottom end if you wanted a lot of compressed highlight detail.) The s-shaped curve will de-saturate highlights and increase saturation in other tonal ranges.
*I believe it gets a lot more complex than that.
In Redcine, duplicate this with curves.

A lot of video tends to have too much sharpening.

Depth of field.

No grain.

There's some HD stuff out there that is difficult to tell apart from film. But otherwise I'd just focus on making the image look good?