Welcome to our community

Be a part of something great, join today!

Register Log in
  • Hey all, just changed over the backend after 15 years I figured time to give it a bit of an update, its probably gonna be a bit weird for most of you and i am sure there is a few bugs to work out but it should kinda work the same as before... hopefully :)

Opinion POLL: EPIC native ISO = 800 or ISO = 320?

Opinion POLL: EPIC native ISO = 800 or ISO = 320?

  • Total voters
    159
Adrian, no matter how much you might want it the ISO is just not a constant value, on any digital camera. Despite what Arri, Canon or Sony might tell you they are all reccomended settings. My experience with the Alexa is consistent with my experience with the Epic in this regard, ISO 800 provides the best starting point but the actual exposue I use is based off the scene's DR and the intent of my lighting. When using a handheld meter or spot meter I'll set it to ISO 800 but I only ever use handheld meters for rough measurements and ratios. The actual iris setting is determined using the in camera histogram or a calibrated monitor.
 
Evin Grant said:
Adrian, no matter how much you might want it the ISO is just not a constant value, on any digital camera. Despite what Arri, Canon or Sony might tell you they are all reccomended settings. My experience with the Alexa is consistent with my experience with the Epic in this regard, ISO 800 provides the best starting point but the actual exposue I use is based off the scene's DR and the intent of my lighting. When using a handheld meter or spot meter I'll set it to ISO 800 but I only ever use handheld meters for rough measurements and ratios. The actual iris setting is determined using the in camera histogram or a calibrated monitor.
Click to expand...


So are you saying that the EPIC's sensor and any other digital camera's sensor for that matter has a "variable signal to noise (S/N) ratio" rather than a "constant S/N ratio"? IE: "variable" meaning that the sensor will automatically change it's sensitivity to light on a whim to give you the best image possible ('cause it's awesome!) or "constant" (sometimes referred to as native) meaning that the sensor has a fixed point at which it's sensitivity towards light produces the least amount of noise and the largest amount of dynamic range (and when one pushes or pulls the sensor's mid-point, by gain otherwise known as ISO, you add or subtract both noise and DR).

???
 
If we want specs, as in what does the manufacturer say well that would be 66db S/N ratio, 13.5 stops of dynamic range on the Epic. The M-X sensor did not change the sensitivity of the camera per se it changed the noise floor. How RED goes from point A (RAW data) to point B (an RGB image) is not summed up with one number or one answer for that matter. If one really wants to understand the nitty gritty details of how that happens, it takes more than a glance at a spec sheet. It requires testing and experience. Why? Because there are some things that can be measured scientifically and there are some things that are subjective and can only be judged by hands on experience.

Adrian, I want to believe you are only trying to drill down into the nitty gritty details so you have definitive answers to provide to people when they come to you for those answers. I want to believe you just want to make sure you understand exactly how the RED camera systems work, right down to the very last spec. I respect that. I don't see a problem with asking the questions. I will be honest though, there have been a couple of posts where it comes off as having an agenda that goes beyond specs. In may not be intentional but it reads that way. So, hopefully this thread doesn't deteriorate. There's some good info in here, some opposing viewpoints but that's okay, sometimes that can lead to really good things.
 
i set the camera to get the best image possible. For most of my shooting styles that means setting the ISO to 320 for tungsten lighting and lighting my scene to look good. Then make sure I don't clip the highlights.

In daylight balanced sources, i find 800 to be as good as 320 under tungsten and use that setting to get a low noise floor and protect my highlights. Tungsten at 800 is just too noisy for my taste.
 
There are tests floating all around the internet. For example here. Whether one relies on those tests as fact is up to the person looking for the answers but i have found that the answers are usually derived by means of accumulated data. So you start with existing data (manufacturer, others tests, etc) then you move on to confirmative testing (does your data match what the manufacturer says or that other tests have indicated) and then you move on to subjective testing (where you find what works best for you). I believe the reason you are seeing the disparity in the numbers you are seeking is that last stage. In the example with RED. They have said 800 is the best starting point in their opinion. But clearly some people have arrived at a different conclusion and I would think they probably followed the steps I outlined above and decided their way of shooting, the results they desired worked better when working with a different ISO. Therefore, it's hard to say they are wrong if they achieved the image they set out to capture. So, I don't think this is as black and white as it may seem, even if that does mean there might be some confusion that results in the outside world.
 
Steve Sherrick said:
If we want specs, as in what does the manufacturer say well that would be 66db S/N ratio, 13.5 stops of dynamic range on the Epic. The M-X sensor did not change the sensitivity of the camera per se it changed the noise floor. How RED goes from point A (RAW data) to point B (an RGB image) is not summed up with one number or one answer for that matter. If one really wants to understand the nitty gritty details of how that happens, it takes more than a glance at a spec sheet. It requires testing and experience. Why? Because there are some things that can be measured scientifically and there are some things that are subjective and can only be judged by hands on experience.

Adrian, I want to believe you are only trying to drill down into the nitty gritty details so you have definitive answers to provide to people when they come to you for those answers. I want to believe you just want to make sure you understand exactly how the RED camera systems work, right down to the very last spec. I respect that. I don't see a problem with asking the questions. I will be honest though, there have been a couple of posts where it comes off as having an agenda that goes beyond specs. In may not be intentional but it reads that way. So, hopefully this thread doesn't deteriorate. There's some good info in here, some opposing viewpoints but that's okay, sometimes that can lead to really good things.
Click to expand...


SS:

I'm with ya here, brother. Though some may see ulterior motives I am really just "drilling" for a better understanding, not necessarily the "truth". I agree that these things can get extremely complex, technical, and head-exploding. But that does not mean that we should turn away from reaching our goal when things get too difficult. The real issue for me is in being able to explain the concept/process of native sensor sensitivity to a DP (the artist). Though some have thrown up their arms and exclaimed, "RED recommends 800 and if that doesn't work for you then shoot 320 and make it look good", that's simply too vague of an answer. I could accept "RED recommends 500 but shoot 400 if you want a cleaner image". But the difference between 800 and 320 (or 200) is over a full f-stop.

I agree that the manufacturer in this case has defined the EPIC's sensor at 66db S/N. This data, unfortunately, will mean almost nothing to a DP. Hence the need for a more accepted, real-world metric (ISO).

And I believe if you change the noise floor of a digital camera you are in turn changing it's sensitivity (S/N) via a new sensor. But correct me if I'm wrong.


:beer:
 
Blair S. Paulsen said:
One of my favorite methods for getting the most usable code values into the RAW data set is to adjust the iris while watching the 800ISO RGB histogram. What I'm looking for is how the values stack up or spread out, particularly when I take it too far to either extreme - how does it behave? Early on I learned the hard way that there were times that exposing to the right as dogma made noiseless but flat images that lacked guts no matter how much I jacked them around in post. Bottom line, each scene and lighting condition has a natural distribution of values and in most cases maintaining an accurate representation of those relationships will yield the most useful starting point for grading.

Cheers - #19
Click to expand...
Great post Blair, this is excellent practical advice. Thanks :)
 
Why don't we make it easy and rate the camera like film stocks in both tungsten & daylight:

I.E. The MX sensor is rated:

ISO800D (Daylight)

ISO320T (Tungsten uncorrected)

This I believe would be a good starting point for everyone, less noise in tungsten and lowlight, and less clipping in bright daylight and in hmi (or tungsten corrected) lit scenarios. With this keep the understanding that you can push your low light footage and pull your daylight footage if need be.
 
Matt... that would work really bad. Try the oposite,,,

due to the whiteballancing you need to underexpose the tungsten a bit, while daylight will work better at 320, if you have to,

This thread OTOH...
 
BTW, I don't really subscribe to this debate... Hahahaha. But I'll vote 800, just to show my opinion on where it is usually smartest for most for the most pleasing imageresult under most conditions for most people, unless... of course, you wanna do it differently...

Hahahaha
 
Gunleik Groven said:
BTW, I don't really subscrube to this debate... Hahahaha. But I'll vote 800, just to show my opinion on where it ia usually smartest for most for the most pleasing imageresult under most conditions for most people, unless... of course, you wanna do it differently...

Hahahaha
Click to expand...



Debate ??? I'm just after a few beers with friends!


:cheers2:
 
Gunleik Groven said:
Matt... that would work really bad. Try the oposite,,,

due to the whiteballancing you need to underexpose the tungsten a bit, while daylight will work better at 320, if you have to,

This thread OTOH...
Click to expand...

shooting at 320 forces you to open up thus reducing noise. Exposing daylight for 800 causes you to close down and protect the highlights...Isn't that what you just said?
 
Take out Raw view... ? It makes no sense and spreads confusion

Take out Raw view... ? It makes no sense and spreads confusion

Gunleik Groven said:
It IS fun that this is an "opinion" poll, though, and when the chief engineer answers, noone seem to listen... :)
Click to expand...

I have no idea how G N can take this... People do not seems to take any time to read what has been explained again and again to them.

TO RED please pull out (or rename) raw view. It has confused people and you are suffering for it. Raw view makes no real sense and now since the Mx and other advancements is not needed.p

Even my partner is completely confused by Raw view (unlike GN i have explained 3times and that is it). He really will be better off without the option.

Michael L
 
Michael Lindsay said:
I have no idea how G N can take this... People do not seems to take any time to read what has been explained again and again to them.

TO RED please pull out (or rename) raw view. It has confused people and you are suffering for it. Raw view makes no real sense and now since the Mx and other advancements is not needed.p

Even my partner is completely confused by Raw view (unlike GN i have explained 3times and that is it). He really will be better off without the option.

Michael L
Click to expand...

+1! I've seen people viewing RAW view while on set shooting, not just for exposure checking. Please remove RAW view or name it highlight view....
 
You must log in or register to reply here.
Back
Top