LaForce
Active member
Ok. We've all been suffering with red.
Frank Glencairn has posted some good info on his blog and on the forums regarding this and the way the Colour/Color Depth setting actually works. In fact beyond just this particular setting; in pretty much all aspects of profiling, I must (although I've never even met or talked to the guy) give a huge thanks to Frank for all the information he has shared. It's really great to have such generous people who's love of the craft exceeds their desire to excel in it, and hence share the lessons they've learned for the benefit of all who share a passion for that craft.
Ive been doing some tests today which hopefully I will get time to post some stills of later in the week. Let me explain:
I've been shooting some test footage ofa scene which is basically a table set up with a tv (black); a microwave (white); some assorted colours of condiment jars in the foreground; next to an open doorway which has a semi-transluscent red drape hanging in it - concealing some white and black behind (a fan, a dress hanging). I have also shot the same scene with a face at bottom of shot to see the relationship between the skintones and whats been going on "behind the red curtain".
What I have discovered is that as has been discussed before as you move away from a red colour depth setting of 0 towards seven the red seems to become less intense and develops a richness. But the result of this has been that shooting through this transluscent screen the red despite becoming less saturated; seems to bleed out into the surrounding pixels...? - In effect as I up the colour depth into the positive the semi-transluscent curtain becomes less and less translucent until the black dress becomes almost imperceptible (really just a darker red shadow appearing like a stain in the curtain), and the white can almost disappears until only the highlights can be perceived.
However if I go the other way and shift the red colour depth into the negative, the red curtain becomes more and more transluscent until the footage begins to resemble what my eye is actually seeing (and what I desire to see). Now -7 tends to present it prefectly, but here is the catch - the lower I go, when I had a human face into the scene the more washed out the skintone becomes and becomes a little plasticy. In this particular scene (and I realise this will vary hugely from scene to scene) the best compromise between the skintone and the transluscency of the curtain has been to set the red colour depth to either -4 or -5.
From this experiment it appears that the following scale is accurate:
-7 - the further into the negative the more "tinted" the colours become : 0 - the most saturated the colours become : +7 the more "toned" the colours become
I'm going to experiment further with this in the next few days because it seems like there is a big correlation between how the black level and red colour depth interact in the rendition of skin tones: ie it seems like bu pulling down the red colour depth, and pulling down the black level, and sacrificing a little bit of top end info you can get the most dynamic range without sacrificing colour integrity as far as red is concerned. I feel like that is the choice I am being face with. Maximise dynamic range and over saturate red, or get red a little closer to "correct" but lose a bit of luminance info and have less room to move in post. I suppose in saying that I'll add that I m not one who favours a baked in look. I prefer to shoot as flat as posible and recover my low end in post...
I'd love to see what you guys think about this and see how this compares to your own experiments. I suggest using some kind of semi translucent red cloth to see exactly what I mean.
I"ll make some time to post some stills later in the week, and I look forward to your input. Thanks guys!!!
Frank Glencairn has posted some good info on his blog and on the forums regarding this and the way the Colour/Color Depth setting actually works. In fact beyond just this particular setting; in pretty much all aspects of profiling, I must (although I've never even met or talked to the guy) give a huge thanks to Frank for all the information he has shared. It's really great to have such generous people who's love of the craft exceeds their desire to excel in it, and hence share the lessons they've learned for the benefit of all who share a passion for that craft.
Ive been doing some tests today which hopefully I will get time to post some stills of later in the week. Let me explain:
I've been shooting some test footage ofa scene which is basically a table set up with a tv (black); a microwave (white); some assorted colours of condiment jars in the foreground; next to an open doorway which has a semi-transluscent red drape hanging in it - concealing some white and black behind (a fan, a dress hanging). I have also shot the same scene with a face at bottom of shot to see the relationship between the skintones and whats been going on "behind the red curtain".
What I have discovered is that as has been discussed before as you move away from a red colour depth setting of 0 towards seven the red seems to become less intense and develops a richness. But the result of this has been that shooting through this transluscent screen the red despite becoming less saturated; seems to bleed out into the surrounding pixels...? - In effect as I up the colour depth into the positive the semi-transluscent curtain becomes less and less translucent until the black dress becomes almost imperceptible (really just a darker red shadow appearing like a stain in the curtain), and the white can almost disappears until only the highlights can be perceived.
However if I go the other way and shift the red colour depth into the negative, the red curtain becomes more and more transluscent until the footage begins to resemble what my eye is actually seeing (and what I desire to see). Now -7 tends to present it prefectly, but here is the catch - the lower I go, when I had a human face into the scene the more washed out the skintone becomes and becomes a little plasticy. In this particular scene (and I realise this will vary hugely from scene to scene) the best compromise between the skintone and the transluscency of the curtain has been to set the red colour depth to either -4 or -5.
From this experiment it appears that the following scale is accurate:
-7 - the further into the negative the more "tinted" the colours become : 0 - the most saturated the colours become : +7 the more "toned" the colours become
I'm going to experiment further with this in the next few days because it seems like there is a big correlation between how the black level and red colour depth interact in the rendition of skin tones: ie it seems like bu pulling down the red colour depth, and pulling down the black level, and sacrificing a little bit of top end info you can get the most dynamic range without sacrificing colour integrity as far as red is concerned. I feel like that is the choice I am being face with. Maximise dynamic range and over saturate red, or get red a little closer to "correct" but lose a bit of luminance info and have less room to move in post. I suppose in saying that I'll add that I m not one who favours a baked in look. I prefer to shoot as flat as posible and recover my low end in post...
I'd love to see what you guys think about this and see how this compares to your own experiments. I suggest using some kind of semi translucent red cloth to see exactly what I mean.
I"ll make some time to post some stills later in the week, and I look forward to your input. Thanks guys!!!