Dual Sensitivity/ ISO

[Richie Allen]

A questions i'm hoping the more experienced can answer. The Gemini has dual ISO for low light. As i understand it, this it down to how the raw image is processed by the firmware. So its software based correct? If this infact the case, why would this not be possible with a Monstro or Helium sensor but a firmware update. OR is the sensor fundamentally different that allows this to be possible?

 

[Adam Brian Dada]

I believe it's a totally different CMOS sensor -- basically it's already a low light CMOS sensor by default but software tweaks the data to likely provide for the "normal" ISO such as in other RED sensors.

I don't believe you can take a normal light sensor with smaller "pixels" and adapt it to higher ISO.

Red's video on it: https://www.youtube.com/watch?v=xcKPupoLtKM

 

[Fabricio Morato]

I have no idea how Reds implemented this, but in this PDF it shows the idea on how Panasonic did with varicam https://pro-av.panasonic.net/en/sales_o/broch_pdf/varicam35.pdf
From what I understand it's kind of like what Alexa has, but instead of using for the highlight protection, it uses to have another ISO base. Meaning its a hardware design...

 

[Bob Gundu]

Definitely a combo of software and dedicated hardware.

 

[Noah Yuan-Vogel]

This is almost definitely not software it's sensor design. Dual conversion gain modes is a very common sensor feature these days and exists in a lot of cameras that are good in low light even if they don't advertise with marketing terms like dual native ISO or dual conversion gain. Most of Sony's still camera sensors use this and some of Panasonic's do as well.

That said in most of the cameras that implement it, the high conversion gain mode is not free, it tends to reduce highlights by 2 stops while reducing noise by only about 1 stop, meaning about a 1 stop loss of DR. That is what I've found in my tests of Gemini and what I've heard from varicam usere (in that case about 2.6 stops of highlights and 1.3 stops of noise)