Do RED cameras have 16.5+ stops of DR over all resolutions?

[Tom S]

Do RED cameras have 16.5+ stops of DR over all resolutions? Or does the dynamic range change from 8k to 7k to 6k to 5k, etc?

 

[Björn Benckert]

In One way no, it does not change. As the sensitivity does not vary with the number of pixels.

But in another way yes, what you will consider as "usable DR" becomes bigger the more pixels you got at capture. As that brings down the noise levels in you final image.

You can also turn that around and say for HD mastering the camera has X amount of DR but for 8k finnishing you cant only use part of X and shave of a few stops in the low end or image would have to much noise.

 

[Nick Morrison]

Do RED cameras have 16.5+ stops of DR over all resolutions? Or does the dynamic range change from 8k to 7k to 6k to 5k, etc?

What Bjorn said. On each sensor, as you go down in resolution, you have a higher chance of seeing noise (as you magnify closer to the pixels), so in effect you're losing some usable DR in the blacks. So YES.

To get the max DR of each sensor (whether its Monstro, Gemini, Helium, Dragon, etc) it's best to shoot the highest resolution you can.

If this isn't possible, then try and compensate by adding more light, or using less compression (two other variables that affect noise/signal chain)

 

[Kenneth C Merrill]

I haven't seen one that has 16.5 stops of DR period haha

MAYBE the monstro.

 

[Phil Holland]

I haven't seen one that has 16.5 stops of DR period haha

MAYBE the monstro.

I guess it's that time of the year.

- Monstro Dynamic Range

Dragon and Helium

There's a lot of "perspective" regarding Dynamic Range and what is eventually a usable stop in regards to how noise becomes visible in darker values.

Empirical data for me is Total Captured Dynamic Range as what is a visible stop. Then from there people can have the choice/preference/decide on how many stops are usable. For me that typically comes in exporting out various wedges at different ISO on proper exposed value charts so I know where my preferences and limits are. I do this literally for every camera I shoot with.

I should do a full blown Gemini test at some point I guess.

Generally, while we are always looking for more captured exposure data, all of the current sensors are pretty damn good in my opinion. Basically from Dragon on it's been pretty nice. Mysterium-X when used within it's playground is also nice and most people figured that out a long time ago. Beyond that color has been one of the biggest enhancements from MX to Dragon and forward from there.

And yep, Monstro at the moment is the king in terms of Total Captured Dynamic Range.

 

[Hrvoje Simic]

Do RED cameras have 16.5+ stops of DR over all resolutions? Or does the dynamic range change from 8k to 7k to 6k to 5k, etc?

Within a same sensor, dynamic range has zero relation to resolution.

Whether you crop the sensor or downscale, highest and lowest values captured are the same.

 

[Misha Engel]

I hope these guys will get a couple of recent RED camera's so they can test them under the same conditions as they tested the others.

https://www.cinema5d.com/reviews/test-lab/

Give them a test and callibration procedure because they forgot to blackshade the UMP4.6k after updating the firmware and before testing (and many of us know how important blackshading can be).

 

[luigivaltulini]

I tried to run imatest like from site Cinema5D.

ALexa - Monstro.

[/url][/IMG]

 

[rand thompson]

Luigi,

In the tests above it seems, In The Alexa, the Red and Blue channel Noise are linked closer together than the Red and Blue Channel noise in the Monstro. So that is probably why when you increase ISO on the Monstro you get more Blue channel noise than from any other color channel.

Also, all three channels in the Alexa are linked closer together than All three channels in the Monstro. I can only assume this was a decision Red made to somehow give the Monstro better lowlight handling capabilities than the Alexa.

I know when I was grading some of your Sony Venice Footage you posted a while back, that I didn't really notice any specific color channel noise when I increased the Iso on the clips. Maybe because the Noise from the three color channels (RGB) are so closely linked together like in the example below.

 

[Kenneth C Merrill]

There's a lot of "perspective" regarding Dynamic Range and what is eventually a usable stop in regards to how noise becomes visible in darker values.

"Usable" stops are the only stops that matter to me, and yes I will totally grant that it's a subjective issue. I've always appreciated your tests, and I think the image of the Xyla chart above supports my opinion that the Monstro has about 16 "usable" stops. It has the most of all the Red cameras so I think it's safe to say that none of them have more than 16.5 stops.

But like you say, it doesn't REALLY matter at this point. Pretty much every real cinema camera out there now produces a gorgeous image. The only thing that irks me is that Red has been throwing around the 16.5+ number for a long time, and they're only actually just achieving it with the Monstro. But whatever.