Using unity values for white balance and RAW recording-Q for Mitch Gross

[cuervo]

Specifically, a question for Mitch Gross.... Mitch, I've been getting familiar with the logic behind maximizing dynamic range on a DSLR by setting all the White Balance multipliers to unity. This will produce an image with a green tint, since the R and B sensors don't have the sensitivity of the G sensors. Manufacturers zero out this difference in gain by applying different multipliers to each color channel. As I am understanding this process, the result is that one can blow out a red or blue channel before the green channel, usually it's the blue channel that goes first. Setting gain multipliers to unity keeps all channels exposing with the same unity value, but, the result is a green tinted image which needs to be corrected in post. The process for forcing all the color channel multipliers to 1.0 (unity) is described for a DSLR( see http://www.guillermoluijk.com/tutorial/uniwb/index_en.htm) The FS700 has no such ability to control the white balance multipliers. I can use DCRAW to look at the multipliers being used and get the following: R:2.388916 G:1.000000 B:1.784440 This tells me that the RAW data for the red channel has a 138% multiplier and the blue channel has a 78% multiplier. So, the question is whether dynamic range on the FS700 can be increased by setting these values all to unity? Of course, a director might ask about the greenish image on the monitor, but, we're already living with the washed out look of SLog2 before LUT. Again, I'm curious if dynamic range can be increased by using unity values on the color channel multipliers? If so, are these multipliers applied to Sony's RAW data before it's fed to the O7Q, or does the O7Q apply these multipliers to the RAW data stream. If C-D has access to these multipliers, does it make sense to allow the user to set these to unity to increase dynamic range? An increase in dynamic range of up to 1 f/stop has been reported for exposing DSLR images with unity color channel multiplier values. It's worth noting that the histograms in a DSLR are based on a JPEG image of the RAW data, so can be pretty inaccurate. I assume the C-D histogram is calculated on the RAW data stream, yes?

 

[Mitch Gross]

Our Histogram is based on the RAW signal after we turn it into an S-Log2 video equivalent. That's why when you apply a LUT you can see the effect on the histogram.

The multipliers all happen inside the FS700. We accept the data, but do not control it.

While in theory the unity concept can expand dynamic range, it really only does so in very specific shooting scenarios, and these are almost always in daylight high color temperature conditions. All sensors are blue-starved, so they like daylight more than anything else. It's the blue channel where you will see noise first. ut I will tell you that it has been my experience that while one may technically gain up to one stop on certain cameras using unity control, this extra stop is almost always a useless stop of noise information. It ends up being thrown away in color timing. So effectively no improvement.

Sony does have a bit of a clue as to what it is doing, and those relative values for the color channels are pretty standard.

 

[cuervo]

Our Histogram is based on the RAW signal after we turn it into an S-Log2 video equivalent. That's why when you apply a LUT you can see the effect on the histogram. The multipliers all happen inside the FS700. We accept the data, but do not control it. While in theory the unity concept can expand dynamic range, it really only does so in very specific shooting scenarios, and these are almost always in daylight high color temperature conditions. All sensors are blue-starved, so they like daylight more than anything else. It's the blue channel where you will see noise first. ut I will tell you that it has been my experience that while one may technically gain up to one stop on certain cameras using unity control, this extra stop is almost always a useless stop of noise information. It ends up being thrown away in color timing. So effectively no improvement. Sony does have a bit of a clue as to what it is doing, and those relative values for the color channels are pretty standard.

Thanx for taking a moment to answer, Mitch. I appreciate the feedback.