RED Build 30 is a step backwards

[Brian D. Goff]

You are talking about the new exposure tool that only shows what's actually clipping and what's under exposed?

Not to be mean Stuart, but I feel like you all had a great tool in the previous FC, and now you have replaced it with a far less useful tool because it simply tells us far less information.

Matthew

I'm absolutely with you here. FC gave me far more information on what is happening over the entire dynamic range of the image - B/W just does not

 

[Stuart English]

False Colors in Build 30

False Colors in Build 30

I'm absolutely with you here. FC gave me far more information on what is happening over the entire dynamic range of the image - B/W just does not

Actually that isn't true. The multi-false color shows you the dynamic range of the ISO and Gamma adjusted RGB video monitoring space.

In a film / lab / telecine analogy, it's showing you "exposure" of the video output of the telecine. And yes that is important at that stage of the workflow.

What you really NEED to know when in the field is the exposure range of the film stock in the camera. Or in our case the RAW data levels from the sensor.

If you want to check the RGB levels as well, then select the Video Level false color.


If you wish you can put both on the same User Key ( the MULT selection) so you can use one button to cycle between them.....

 

[Brian D. Goff]

Actually that isn't true. The multi-false color shows you the dynamic range of the ISO and Gamma adjusted RGB video monitoring space.

In a film / lab / telecine analogy, it's showing you "exposure" of the video output of the telecine. And yes that is important at that stage of the workflow.

What you really NEED to know when in the field is the exposure range of the film stock in the camera. Or in our case the RAW data levels from the sensor.

If you want to check the RGB levels as well, then select the Video Level false color.


If you wish you can put both on the same User Key ( the MULT selection) so you can use one button to cycle between them.....

Ok, that might work for me. The old FC in RAW mode at 320iso told me all I needed - and helped me determine were in the image I would add some more light to get it say in the "green" region. I think "video FC" and adjusting the exposure will give me this too.
With all the improvements, sometimes good things get lost (like zoom in function on playback) - or maybe it will come back again:001_wub:

 

[Mark Wilkinson]

It seems what's missing (and what I was certainly used to doing) is the ability to light to the RAW false color - not just set the exposure.

What I would like is a RAW FC with ISO and FLUT settings visible so I know what I'm looking at.

I would also like a FC "key" sitting just below the histogram so I can see in a glance exactly what is going on. I have that FC gradient image on my iphone, but I'd rather see it in the camera.

 

[Matthew Rogers]

Actually that isn't true. The multi-false color shows you the dynamic range of the ISO and Gamma adjusted RGB video monitoring space.

In a film / lab / telecine analogy, it's showing you "exposure" of the video output of the telecine. And yes that is important at that stage of the workflow.

What you really NEED to know when in the field is the exposure range of the film stock in the camera. Or in our case the RAW data levels from the sensor.

If you want to check the RGB levels as well, then select the Video Level false color.


If you wish you can put both on the same User Key ( the MULT selection) so you can use one button to cycle between them.....

Okay, I think I understand what you are saying, but it still doesn't really make sense when I think about it (that's an interesting statement, isn't it!) You say that the Video FC is showing us what we'd get out of the end process in a telecine (aka: a render out with REDGamma/REDColor.) However, Video FC still doesn't show me when something is clipping..even though in the render out, the footage will be clipped in certain spots.

All I want to be able to do is know, hey I am within 10% of clipping on this spot, now I am at 5%, etc. The old FC told me that. The new "exposure" tool does not. It just tells me, <20% under and 110%-clipping.

The more I think about it the more I don't understand it. Maybe you should make a video showing us exactly what's happening, because written words don't seem to be working.

Matthew

 

[Scott Webster]

Stuart,

If you are shooting RedColor on camera, monitoring picture based on RedColor and processing only RedColor RedGamma, what is the best tool for exposing on the camera for a RC/RG workflow?

I ask because a lot of rental TVC work is based around this type of shooting.
The DP is giving the client a picture they are happy with (RC) and exposing for it.
The client is going to post asking for the picture that was on set. They process as RC/RG because it's the least labour intensive way of processing the files. The DP is unlikely (budget) to be in the grade to say "Lets go back to the RAW for that scene"

It doesn't matter if the RAW image has the information if no one has any intention of grading from it.

I know the above is in total contradiction to using a RAW camera but I'm sure it happens a lot and not to the benefit of the camera.

May be it's an Australasian thing

 

[John Brawley]

It's because FLUT is trying it's hardest to keep the RGB image from over-exposing.

BTW - just tried same experiment on M, and 99 works there too for me.

Graeme

I was working with one of my non MX cameras yesterday and I could get zebras to work past 95%. I'll try with another camera, but clearly something is going on...

jb

 

[John Brawley]

As John notes, a zebra reference at 100% can be an excellent way to ride your iris during a shot when required.

Ultimately the real answer is the individually adjusted viewing options anticipated for each monitor path on the DSMC line. Of course that's not soon enough to help JB. Good on Graeme to make the effort. In theory you could make a variant of the "E" screen that put the red fizzle over a color image but is it worth the effort? Perhaps for John.

The beauty of Zebra is that they are not distracting to the operator AND we're used to seeing them. And as I've already mentioned, I don't mind and can accommodate the NEW metering modes, but not at the expense of a feature most DOPs that aren't owner operators (like me) will expect to see AND have work in a logical way.

I started thinking that the new exposure mode would work OK if it was colour (redcolour) instead of black and white. Even that might be too distracting to an operator. Zebras are elegant and we're used to seeing them.

jb

 

[John Brawley]

John, all that Zebras are doing is using a value in the RGB domain to indicate a clip point in the RAW domain (or SMPTE range HD-SDI output peak white)

Under FLUT that means a value of about 99 IRE, but the actual value that you use as your clip indicator is up to you, so if you want to use 95 that is fine too.

Stuart I'm saying that zebra's DON"T WORK unless they are set at 95% or BELOW.

If I set a zebra to 100-108% and point it at a very hot reference then it doesn't even come on unless I view in RAW.

Redcolour is the mode operators almost have to use when rolling on a take because that's what feeds the split. So now, you've broken the most reliable way of looking out for clipping as you go through a take.

jb

 

[John Brawley]

The histogram and the warning goalposts and the traffic lights all tell you if there's over-exposure.

Graeme

Except....not WHICH part of the image dynamically as you shoot....

 

[Michael Epple]

Man. This discussion is just one big mess. I don't see why it's bad to be able to see the RGB values in False Color for what your settings are. At least that way you have a standard to work off of. That's how it was and it worked fine. You knew that if you were just clipping in the RGB space that you had a little headroom, which could be validated if you quickly went to raw view mode and went off of those RGB clipping values. You don't have that level of control now.

 

[Deanan]

Stuart I'm saying that zebra's DON"T WORK unless they are set at 95% or BELOW.

If I set a zebra to 100-108% and point it at a very hot reference then it doesn't even come on unless I view in RAW.

Redcolour is the mode operators almost have to use when rolling on a take because that's what feeds the split. So now, you've broken the most reliable way of looking out for clipping as you go through a take.

jb

If you didn't black shade after upgrading, that might affect the image.
Black shade again to be safe.
What ISO are you at?

 

[Tim Schumann]

Hi Deanan. He is shooting on 4 of our Lemac Red cameras.

We always do the upgrades by the book and black shade accordingly.

The problem he is talking about here is to do with the fact that 100-108% Zebras @ 320 ISO no longer work if you are viewing in RedCLR because the Video out RGB values never hit 100 as the FLUT is feathering the values off as it approaches 100.

This is great for post. But is bad for judging exposure whilst rolling.
Zebras at 100-108 only work in RAW now - which most directors don't want to look at.

Zebras in camera are really an exposure aid tool, as is a histogram.

We have been suggesting for a long time that there should be the ability to set the histogram and zebras so that they only reflect the RAW values. Maybe just a setting in the menus that enables the DP to choose to view Hist & Zebras as RAW or Relative (to your LUT) regardless of what LUT you are looking at. What's hitting the sensor & clipping is really the main concern.

The new False Colour (exposure) mode is good but not usable whilst rolling on a take.

I am going to have to spend a LOT of time informing clients to be VERY careful using the False Colour (video) mode from here on in because its purpose has changed fundamentally from previous builds. So much so that it could be disastrous for a DP who is expecting to see Red and Orange values (which you can no longer hit at all when set to RedCLR 320ISO)

Steve & Greame, thanks for your swift replies. This is one that we need to try to get sorted ASAP so your input is greatly appreciated. There are lots of really great things about Build 30 but the zebras are a problem.

 

[Stuart English]

Zebras

Zebras

I'll say once again, you should not only be careful trying to judge sensor clip with the Video false color, you shouldn't do that at all.

Look at the traffic lights, look at the bar to the right of the histogram, these will indicate clip.

Then if you need to see where that is, look at the Exposure false color.

And if you check (and adjust the LO if necessary) the Zebra 1 does overlay the red sizzle of the Exposure false color.... then where is the issue?

 

[Vico Martin]

I'll say once again, you should not only be careful trying to judge sensor clip with the Video false color, you shouldn't do that at all.

Look at the traffic lights, look at the bar to the right of the histogram, these will indicate clip.

Then if you need to see where that is, look at the Exposure false color.

And if you check (and adjust the LO if necessary) the Zebra 1 does overlay the red sizzle of the Exposure false color.... then where is the issue?

then...

1. Check false color to see where are you (right iris where you want)
2. Check Exposure false color
3. Shoot.

 

[Matthew Rogers]

I'll say once again, you should not only be careful trying to judge sensor clip with the Video false color, you shouldn't do that at all.

Then what exactly is the point of having the "video" FC mode? To me, it seems pretty worthless.

Look at the traffic lights, look at the bar to the right of the histogram, these will indicate clip.

Then if you need to see where that is, look at the Exposure false color.

The issue here is we want to know more than just what's clipping. In both Exposure AND video mode, I want to know what's under exposed, what's at 18%, what's at skin color, and the different levels right before clipping. The stoplights don't tell me any of that, except for clipping, and neither does the new exposure FC.

Is it impossible to get a FC mode that displays all the colors for their respective IRE's? If you switched into RAW, it would show the values in the RAW image. Heck, also a option that allowed you to monitor in REDColor, but was showing you the FC values of the RAW image. I firmly believe this is what we all want, because FC kinda used to be that way (seems like it changed every build) and it's extremely useful.

If it's not physically possible (which I don't see why not since you are doing colors in the video FC), then I understand. However, if you all are choosing not to do this, what is your logic? The waveform does help me expose without a meter, but it really still doesn't give me detailed info that I have enjoyed with the old FC.

Matthew

 

[John Brawley]

I'll say once again, you should not only be careful trying to judge sensor clip with the Video false color, you shouldn't do that at all.

Look at the traffic lights, look at the bar to the right of the histogram, these will indicate clip.

Then if you need to see where that is, look at the Exposure false color.

And if you check (and adjust the LO if necessary) the Zebra 1 does overlay the red sizzle of the Exposure false color.... then where is the issue?

I'll say it once again cause you're not hearing my issue.

Zebra is an indicator of exposure.

You're telling me that you now roll off the highlights when viewing in REDcolour so that they never clip on the monitor output incase I was recording that output.

You're telling me that I can't judge exposure in redcolour because it's now a floating LUT.

You're telling me that I should use the histograms, traffic lights and goal posts.

You're telling me to use the multi-colour to judge exposure.

You're telling me to judge sensor clipping using exposure check.

I really do hear WHY you want me to work this way so stop trying to explain it to me.

I'm telling you right now that you are still overlooking the critical advantage that zebra offers. The ability to have an exposure guide that can be used during a take with a minimal interruption to the onset viewing experience for those that need to see the image as close to normal as possible.

It is not good enough to keep telling me to use the histograms, traffic light and goal posts because none of these tools do what zebra does. Zebra indicates exactly WHERE in the image there is clipping.

PLease stop telling me how good the new exposure methodology is. Stop telling me I need to adopt my exposure methodologies. It makes sense to you and it works a long as you aren't ROLLING.

You also are flying in the face of the cumulative experience of every other DOP and operator out there that is used to seeing Zebras work as they USED to work with RED. RED now has a zebra function that is completely counter intuitive when compared to every other camera that has a zebra.

Why not just make it impossible to use zebra with REDcolour ? Because if you allow me to set a zebra at 100% I expect that thing will come on at 100%. That's how it works with every other camera on the planet, and that's also how it used to work on RED.

Don't you see how dumb this is ?

then...

1. Check false color to see where are you (right iris where you want)
2. Check Exposure false color
3. Shoot.

4. Try and work out if you need to go again cause the Sun came out during your take and MIGHT have lifted your bounced key a stop on someones face....How would you know if the skin tone was still safe ??????

Dude. You're not getting it. It's about monitoring while rolling. Histo's, goal posts and traffic lights are useless at judging WHERE in the picture it's overexposing.

jb

 

[Steve Sherrick]

On M-X sensor

ISO 800 = Set Zebra LO to 99
ISO 320 = Set Zebra LO to 96

Now, you've matched the Exposure Mode clipping indicator and the Zebra. You have your indicator while live shooting and you don't have to go into black and white mode.

 

[Steve Sherrick]

I also find that when you use the EVF, which many DPs/camera ops do, you can map those user keys (defaults are pretty good) so it's very easy to flip through these various modes. But to John's original point that all that matters to him is seeing a Zebra that represents clipping, following the above guidelines works pretty well, just different than what you are used to.

 

[Robert Trim]

Wow. Well put John. I'm in a different place in all this. I teach cinematography. Today, we started our discussion and labs on exposure, light meters, ASA/ISO and such. After the class I had the students install B30 no knowing all this. It's interesting enough teaching exposure concepts with and on the 4 different HD cameras we use... now I lose the one touch stone that ties them together (other than the histogram on some). Ugh. One student piped up "So is RED getting like Apple, if they don't like Flash because the head guy has a beef with Adobe, they just leave it out of their hot new product despite the huge amount of Flash driven sites on the web?"

This should be interesting to create a learning module that makes sense, based on this new way of looking at the image. Wish me luck and hope this gets worked out.