Welcome to our community

Be a part of something great, join today!

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 :)

False Colour and Exposure confusion

D

David Litchfield

Well-known member
Joined
Aug 14, 2007
Messages
161
Reaction score
0
Points
0
Can someone explain to me why I can get the RAW image to show clipping when using Exposure check, but when I then do False Colour in Redcolor nothing shows as clipping? If the sensor is clipping, Redcolor should definately show clipping. Its confusing and nobody I've worked with understands this either.

Cheers

Dave Litchfield
 
One possibility is that exposure check shows absolute clipping at the sensor level while false colour uses relative RGB values with the current gamma curve/ISO/etc. So if your RGB signal doesn't go to 100%, it won't show as clipping. Don't know if that's the cause, but at least to me it would sound like a plausible scenario.
 
Jaakko Rinne said:
One possibility is that exposure check shows absolute clipping at the sensor level while false colour uses relative RGB values with the current gamma curve/ISO/etc. So if your RGB signal doesn't go to 100%, it won't show as clipping. Don't know if that's the cause, but at least to me it would sound like a plausible scenario.
Click to expand...

Yup that's essentially it. The role of the FLUT color science is to map the RAW sensor data into legal RGB space... you see the same thing on the camera.
 
Thanks for the response guys. Lets see if I understand this correctly! If I've got a window thats blown out, the sensor is clipping but that is mapped to a legal RGB space so it shows it as 100% white. False Colour won't show it as clipping because its been brought within the legal levels?

If we're sticking with Redcolor for our workflow, surely this renders false colour useless as an exposure tool? The window is now within legal levels but its got f' all detail in it! That just doesn't make sense to me. It should still show that area as clipped. If i want to get a little bit of detail in that window and expose it just below clipping, false colour cannot help me in the same way that the exposure check in raw can.
 
The explanation of the above is the VIDEO (i.e. RGB) false color overlay is NOT an exposure tool. It's a video output level indicator, valuable when you have broadcast oriented work - such as live HD-SDI feeds and / or when you need to exposure primarily for skin tones, and the FLUT color science rolls off the highlights so that you don't create hard clipped whites. On the other hand the EXPOSURE (i.e. RAW) false color is overlay IS an exposure tool. Simply stated if you don't see red in the frame when you have this selected you are not clipped at the sensor level and therefore everything in that scene can be mapped into RGB space w/o clipping either.

Does that help ?
 
Sorry Stuart,

But I'm more confused about the VIDEO overlay now than I realized before reading this thread. It's not an exposure tool but it is useful "when you need to (check?) exposure primarily for skin tones..."? What are you saying here? It is not a general exposure tool but a task specific exposure tool for mid range values only?

Stuart English said:
The explanation of the above is the VIDEO (i.e. RGB) false color overlay is NOT an exposure tool. It's a video output level indicator, which valuable when you have broadcast oriented work - such as live HD-SDI feeds and / or when you need to exposure primarily for skin tones, and the FLUT color science rolls off the highlights so that you don't create hard clipped whites. On the other hand the EXPOSURE (i.e. RAW) false color is overlay IS an exposure tool. Simply stated if you don't see red in the frame when you have this selected you are not clipped at the sensor level and therefore everything in that scene can be mapped into RGB space w/o clipping either.

Does that help ?
Click to expand...
 
Thanks Stuart,

I understand all that, I just don't think its logical. You have 2 colour overlay modes each with a different logic. How can you say VIDEO false colour is NOT an exposure tool and then say its used for skin tone exposure, and presumably the other colours that are represented. Every Epic/One manual I've read would suggest it is an exposure tool. What would be the effect or harm of adding the red clipped colour to the VIDEO false colour overlay so that they match up with the raw?

Cheers

Dave
 
I guess what I’m experiencing is counter to what I expect.

If I was shooting on Alexa I could look at the rec 709 view and think its really contrasty and the whites are horribly clipped, but I’d be safe in the knowledge (and could check) that on the LogC recording the highlights are looking fine. So I’d feel like my LogC is getting more of the sensor data and is a bit of safety net for those highlights.

That's how I feel about a RAW image. Its getting me more than the baked in rgb look does. But the overlay modes can suggest the opposite. If I was just going on the false colour alone, which a lot of dops are used to, its perfectly reasonable to think your exposure is fine because there are no Red warning colours, but the reality is you could be clipping without realising it. Therefore getting less of the sensor data than you think you are.

The solution then is you have to use 2 different overlays for one goal. Please just add red to false colour :)
 
It's been this way ever since FLUT was introduced, so if it changes, it needs to be announced prominently, otherwise lots of people will be confused all over again!

Rather than add it by default, I'd suggest an additional option tick box in False Colour menu. Red VIDEO (RGB) "clipping" would actually be indicating that FLUT was coming into play rather than clipping, as by definition FLUT makes hard video clipping impossible.
 
There already is RED in the VIDEO false color overlay.

Even the Epic Manual shows red as a color for Super white at 108-109 IRE
 
Martin Stevens said:
There already is RED in the VIDEO false color overlay.

Even the Epic Manual shows red as a color for Super white at 108-109 IRE
Click to expand...

It shows yellow, but I should have added that I wanted it consistent with the Raw red overlay. If its clipping the sensor, it should definately be showing up as super white on the VIDEO false colour overlay, which it doesn't.
 
I just assumed the reason RED assigned a switch button on the Redmote and side handle was for the purpose of quickly checking for clipping or offering a visual of the director's feed. When I'm shooting with no one looking over my shoulder I use zebras and the exposure setting. I'm not finishing the footage in camera. I'll wait until post to do that. All I'm interested in is getting the most data I can from the shot.

Works well for me.
 
I thought I'd do a quick test just to make sure i'm not taking crazy pills. We've been blessed with a rare spell of sunny weather here in the uk so I'm pointing my camera out the window and at ISO 800 the sky is half Red with Exposure check and completely Yellow in False Colour. This is as I would expect and it makes perfect sense with the Yellow being 100IRE and Raw sensor being able to see a bit more than this. Adjusting my stop and shooting at ISO 640 I can have slightly more Yellow in False colour than I can have Red in the exposure check. At ISO 500 and lower there is no Yellow in False colour whatsoever. The screen goes white with over exposure before any yellow 100 IRE comes in. That doesn't seem right to me.
 
David: Seeing how at lower ISOs the RGB signal doesn't go to 100IRE, I don't really see the problem here. Like it was said, false color works with the RGB values and those come from the gamma curve you're using. You can make your own curves adjustment if you want it to work differently. If not, you should look at the traffic lights and goalposts and use the exposure tool if you want to know if you're blowing the highlights.
 
there is something funny going on with the false color imho, if I raise the iso to 1600 or above I get RED in the video false color - like in the old red1 days - false color showed all of the colors from purple to red.
- Also what happend to the 3rd option for raw exposure check the vertical color bars? that was super useful.
 
Jaakko, I don't actually have a problem getting good exposures and neither do the dops I work with mainly because if it looks good to the eye it usually is, and we use the histograms and exposure check and then the DIT deals with it after that. I just have an issue with a feature of my camera which is confusing me and I wanted clarification on it. I like the false colour approach as a quick reference and use it to get my mid-grey and skin tone and its great for green screen. But as I've said it should show clipping. Regarding the lower ISOs, if the sensor is clipped and that is then converted to an RGB space resulting in my screen being white with overexposure in Redcolor, I must have white values of 100 + IRE. In which case they should appear yellow.
 
David Litchfield said:
Regarding the lower ISOs, if the sensor is clipped and that is then converted to an RGB space resulting in my screen being white with overexposure in Redcolor, I must have white values of 100 + IRE. In which case they should appear yellow.
Click to expand...
But the thing is, that at lower ISOs you're not getting anything at 100 IRE in the RGB signal.

Open any clip with blown highlights in RedCine-X Pro. View it with REDgamma2 or RG3. Look at it with a low ISO and a high ISO. What you'll see is that at lower ISOs the blown highlights are further away from 100 IRE than at higher ISOs. That's just how the gamma curve that is being used works. It doesn't go to 100 IRE at low ISOs, no matter how much you're clipping at sensor level. If you want to reach 100 IRE with something like 320 ISO you need to apply a custom curve on top of RG2/RG3 or use RG1. And with RLF the RGB signal will obviously be even further from IRE 100.
 
I can't check this right now, perhaps someone else could. If you shoot at 320 and blow your exposure would a waveform monitor not show that White level to be around 100 IRE? I think it would in which case false colour should be showing yellow.
 
To measure the exposure you have to concider two separate issues. You are shooting RAW, the camera is recording RAW but what you see in the viewer is an interpretation of the RAW image overlayed with metadata. To measure RAW you only have to be aware about the goalposts as well as the stop lights. The histogram and the false colour are dependent of the interpretation you have made in the metadata settings.
For instance changing f-stop will actually affect how much light will reach the camera sensor, while changing ISO will not, since ISO is only metadata. (It's like using a certain film stock while using a different ISO. Both the film stock and RAW format have a set ISO/sensivity to light. This doesn't change.)
On the other hand both the histogram and false color are used for making an interpretation. This is probably the interpretation the DP want's to convey to the director and the colorist. This is also the way the signal is played out of the camera on the HDSDI and the HDMI ports, as well as in the viewer, and an editing system will default to this look when playing back the media.
Hope this helps
Andi
 
Shooting with low ISO is like: shooting at 800 ASA and then turning the luminance down. So in practcality you could clip the sensor (and lose detail) and then turn it down to "legal" levels. The clipping has to be noticed by watching the goalposts and the stop light. Even here the stop lights can show clipping while the goal post doesn't. This is because the goal post will show what percentage of the image is clipping while the stop light shows that there are pixels that actually are clipping.
Andi
 
You must log in or register to reply here.
Back
Top