The challenge of exposing the Helium sensor

[Audy Erel]

I thought ISO is used for a DOP to determine the lighting power as a guide to correctly exposing the image and lighting design even before the camera is arriving on set.. So one could set all the lights with predetermined values with confidence to work with ANY camera that adheres to the ISO standards. Not to be confused with a camera's DR which is a variable between cameras, ISO values should be an exact science for every camera IMHO, what's the point of a "standard" then?

 

[Josiah Duncan]

That's interesting- did not know that. Do you have a source for that?
BTW I may be wrong but the posts above of the tests and comparisons seem to contradict it- meters seem to be matched to ARRI rating. Correct me if I'm wrong please.

I might be incorrect about that, I thought I read about it here from one of the big dogs, but it could be some BS I heard somewhere else. Don't take this for fact unless an expert like Graeme or Phil can chime in, I couldn't find the quote myself.

I'm not for one side or the other, I just expose based on the in camera metering tools and a calibrated display. I'll likely do some testing for green shadows later this week to see if Manual/Auto black shade yields different results, because this thread piques my interest.

 

[Anton Shavlik]

ISO is somewhat arbitrarily defined for digital sensors, and it's especially arbitrary for a sensor that records in raw and doesn't apply any analogue gain when you change the ISO setting... It's somewhat presumptive that arri's gamma is correct.

ISO's namesake, International Standard Organization, has definitions of ISO that relate middle grey values as ratios to sensor clipping points in linear light, or average value of pixel luminosity in an sRBG image, or leave definitions to the whims of camera manufacturers trusting that they know best what their clientele need from their sensors.

To the person who made their own LUT to push the exposure, why not use FLUT -1 since that would make the displayed ISO pushed a stop as well... seems easier.

 

[Hrvoje Simic]

You are the expert (unless you imply you are a sorcerer ;-) and I'm not a camera operator. Or an electronic camera designer for that matter...
Look you aking for my opinion....it is completely irrelevant. Only Red knows the core of the matter. I'm sure they want to address it more than anyone.
I only pointed out one obvious issue (ISO discrepancy) that may contribute to the exposure difficulties the OP is having ...
Hope that's clear...

Ok, thanks for the clarification. We are talking about different things.
I haven't negated the ISO discrepancy. I'm saying it is not the cause of this issue and that published numbers are insufficient to get a best possible results out of any camera.

ISO value doesn't tell the full story as signal properties vary through types of light and its spectral properties. DR numbers do not tell anything about specifics of shadow performance and highlight roll off and every camera has particular offsets in its response. That all requires testing and getting to know your creative tool to be able to get the most of it. There are too many variables in image acquisition and raw signal chain to be able to take every test published online as referent.

 

[Hrvoje Simic]

The shadows have always tended green in my footage. They dramatically shifted green in the test I posted (shot by someone else). They shifted green in John Marchant's (from Kipper Tie) dynamic range tests. Hell, they even shift green in Phil Holland's extensive Dragon v. Helium tests.

I mean, have you been watching any of this stuff? It's not just a one-off fluke.

I don't have time finding and watching every test published online, and cannot comment results of something I have zero information about. I do my own tests where I know the scene information, camera setup and image development process so I can take many variables into account to form conclusions based on known premises.

I have a bunch of Helium material and have yet to see that dramatic green shift so if my logical process is correct the issue is not consistent and there are some variables causing this.

No matter what exposure strategy you use, no matter how you rate the sensor, you're going to end up shooting a scene that has shadows. If those shadows shift green when they shouldn't, then there's a problem--and that seems to happen a lot on this camera.

The reason I posted that -4 stop image in the OP was to show how dramatic that shift is.

Yes, and by that you showed that if you push an image 4 stops (for some reason) you enhance the signal's color offsets.

If it helps, I'll be happy to look at a problematic r3d grab with a green shift.
Shot with a properly black shaded camera, no NDs or Motion Mount, lit by Sun, plasma, HMI or tungsten with lamps in good condition.
LED panels, CCFL, streetlamps, club/stage/party/working lights don't qualify.

Even if you don't push your footage in post (which I really never do anyway), you will have parts of your image that are 4+ stops underexposed (shadows), and that pushed image in the OP shows JUST HOW GREEN THOSE SHADOWS ARE.

If you don't push the image 4 stops you don't enhance the color offsets which may happen for any of multiple possible reasons, and those offsets remain a) smaller and b) buried in the shadows.

 

[Jake Wilganowski]

Yes light meters match Alexa. When using with red I add in a -1 stop exposure compensation, so if my meter says 800 it's actually giving me a reading for 400, with Alexa 800 is 800

 

[David Mullen ASC]

If you underexposed by 1-stop from your meter reading and your meter is set to 800, then aren't you actually exposing at 1600 not 400?

 

[Jacek Zakowicz]

I don't think so David.

 

[jpmcabe]

I don't think so David.

Wow!

 

[David Battistella]

Yes light meters match Alexa. When using with red I add in a -1 stop exposure compensation, so if my meter says 800 it's actually giving me a reading for 400, with Alexa 800 is 800

Or in other words by setting your meter in this way you Rate the RED at 400 ISO? I think that is what you mean?

I agree that using an 800 ISO as a basis (number) for exposure, images will end up about 2/3 to one stop under exposed.

d

 

[Nick Morrison]

Or in other words by setting your meter in this way you Rate the RED at 400 ISO? I think that is what you mean?

I agree that using an 800 ISO as a basis (number) for exposure, images will end up about 2/3 to one stop under exposed.

d

We threw a grey card up on our Epic Dragon, previewed in RAW, and basically realized the camera was 320, period. We light for that.

 

[Jake Wilganowski]

If you underexposed by 1-stop from your meter reading and your meter is set to 800, then aren't you actually exposing at 1600 not 400?

What I mean is since i keep the meter iso set to 800, but internally in the meter set a -1 stop filter, the meter is effectively giving me a reading for 400 iso, which f-stop reading then can be used to expose the sensor properly in my opinion (regardless of camera ISO, based on what you want from the scene of course)

 

[David Battistella]

We threw a grey card up on our Epic Dragon, previewed in RAW, and basically realized the camera was 320, period. We light for that.

I'm pretty much at 400, so agree completely. Nice thing is that exposing this way and processing at 800 RWG/RG10 is a really decent combo.

d

 

[David Battistella]

What I mean is since i keep the meter iso set to 800, but internally in the meter set a -1 stop filter, the meter is effectively giving me a reading for 400 iso, which f-stop reading then can be used to expose the sensor properly in my opinion (regardless of camera ISO, based on what you want from the scene of course)

Yup. Totally get this strategy the way it is being explained here.

David

 

[Kenneth C Merrill]

If you don't push the image 4 stops you don't enhance the color offsets which may happen for any of multiple possible reasons, and those offsets remain a) smaller and b) buried in the shadows.

What you're ignoring here is that the Dragon screen capture in the OP was shot under the exact same conditions (presumably), and it shows none of the green shift that the Helium does. So that would at least eliminate most of the external variables because it reveals that the scene isn't actually green, just the way the Helium is capturing it. As far as I can think, the only variables in capture and post would be black shade and the raw processing. You'll likely chalk it up to one of those, but since I've seen green shadows in other tests (and my own footage, when properly blackshaded), I'm not satisfied that that is the answer.

At face value you can't deny that it looks like the Dragon is capturing the scene accurately while the Helium is not. I mean, come on.

 

[Kenneth C Merrill]

For reference to other tests I've seen, here is a still from Phil Holland's Helium vs Dragan VV test. The Helium tends green again.

[/url][/IMG]

 

[Kenneth C Merrill]

I don't really see this conversation going anywhere, unfortunately. I'll be doing my own tests this month, and I'll come back when I have some R3D's.

 

[Jacek Zakowicz]

I don't really see this conversation going anywhere, unfortunately. I'll be doing my own tests this month, and I'll come back when I have some R3D's.

Trust me- this conversation is going straight to the RED HQs. They are the only ones with the real solution and they are on it but things like this just don't get resolved overnight unfortunately.
I know it's not much of a consolation but this is the world we live in now- hardware sales followed by software refinements.....

 

[Hrvoje Simic]

What you're ignoring here is that the Dragon screen capture in the OP was shot under the exact same conditions (presumably), and it shows none of the green shift that the Helium does. So that would at least eliminate most of the external variables because it reveals that the scene isn't actually green, just the way the Helium is capturing it. As far as I can think, the only variables in capture and post would be black shade and the raw processing. You'll likely chalk it up to one of those, but since I've seen green shadows in other tests (and my own footage, when properly blackshaded), I'm not satisfied that that is the answer.

At face value you can't deny that it looks like the Dragon is capturing the scene accurately while the Helium is not. I mean, come on.

I'm not ignoring anything, exactly the opposite. Exactly why I mentioned variables to avoid oversimplification and jumping to conclusions.

I have Helium material from various cameras shot in various conditions plus my own tests and it is without the mentioned green shift so my experience differs and attempting to compute why some may get it while shooting something under some conditions is an equation with too many unknowns.

I don't care about the destroyed image test shot under unknown conditions with unknown post process. You can post a raw grab of a proper shot with standard factors I listed so one can get more objective impression and I can offer a comment with any merit. This way the discussion is revolving around guesswork.

 

[Antony Newman]

Is the difference in metering caused by Dagon Colour Science on a Helium sensor?

When Helium CS (& IPP2) is released, it may well be this Anomaly evaporates.

AJ