builds/dynamic range

[A.R KAIN47]

i was just wondering in the official presentation of the red one he said that in one of the shots with the car that the light is 14 stops hotter than the car, but i ve been told by a few people on these threads that red one can only do 8 to 9 stops is this true. plus if it is will this build 16 be able to rectify that or is that a problem that cant be fixed! coz to be told one thing and find out it isnt right is frustrating!

[Andrew M.]

Here it is:
http://www.reduser.net/forum/showthread.php?t=13403

You have 10 smooth levels of light (power of 2 increments) if you make sure that histograms of each color (RGB) is aligned, (color temperature of the sensor is used)
I guess, in the film nomenclature it mean:
starting from stop 0 you have 9 stops increments if the temperature of the scene is matching the optimum temperature of the sensor with no tolerance to the noise or grain in the picture. (10 levels of light intensity) translate to stops because I still don't know how to do it.

[Graeme Nattress]

You're miss-quoting. It was said that, with respect to that early shot with the car, that some cinematographers, upon seeing it, thought they could see dynamic range in the region of 14 stops. However, even to an experienced cinematographers, it's very hard to judge dynamic range by eye, and even harder to judge from an image when you were not actually at the shoot itself.

We have a very linear sensor over a 12 stop range. The darkest stops are noisier, but there can be detail in them. I've seen, in controlled tests, detail at the limit of 11.3 stops. As for how many of the 11.3 you'd want to use for the best quality work, that's up to you after performing tests.

I have seen outdoor, real world shots, that show the full range of the sensor has detail, and you've got to dig really deep into the shadows to see it all.

Graeme

[Michael Lindsay]

Hi Graeme

What's the chance we might see the ability to load a 1D LUT into the camera for its output... (Ideally in the future with RED/EPIC separate 3D luts per output would be marvellous)

Keep up the good work!

thanks Michael

PS Do you think there will ever be option for a non-linear encode of the sensor output?

[Andrew M.]

I think the biggest misconception coming from the jeremyn test is that we are trying to evaluate all colors and large surfaces in the lower and lower exposure.
In the real scenes the low level of light is associated with the blacks or shadows, that do not have much of color in it. Dark doorways, texture on the black jacket or pants etc.

Sure in low light all colors performance count.
But then we should rather talk about low light or low exposure performance instead of dynamic range.