Welcome to our community

Be a part of something great, join today!

  • 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 :)

Access to RAW data in secondaries?

Kendal Miller

New member
Joined
Jul 9, 2009
Messages
3
Reaction score
0
Points
0
It seems to me that a lot of times prioritizing a grade towards highlight recovery etc...happens in the secondary room. For example an interior shot with a window. Lets say in primary I balance the shot toward the interior say push 320 to 640ISO but in so doing I now push the window too high and its now over. I would love to apply a secondary that had access to the original raw data to now pull the secondary back down into range. Is there currently a way to address that? Or a suggested solution?
 
Apple needs to make Color's 'floating point' renderer actually be floating point in the sense most people use the term.
 
So Apple Color doesn't use the RAW data in the secondaries? What is it using then?
 
Color allways use Raw, but in secundaries you can´t edit metadata.
 
DRX

DRX

Are you using the DRX in the RED program, it seems to be using the over-white-clip point data to get some more highlight detail.

I have been trying that with the de-Bayer program I am working on, it seems that the red sensors are exposed less than the green and blue for daylight, so when you overlay that data you get about 1/2 stop more highlight detail, for 3200K the blue might make a better overlay.

There is no reason to loose highlight detail if you change the ISO since whatever program is being used should NOT rase the tones close to white when you pull up the shadow and mid-tones, if that is the case use another program, white and black should stand still, only mid-tones should go up and down for ISO adjustments.

Never use brightness or contrast as they will both clip data, only use gamma and S-curves.

Hue (sometimes mis labled Tint) is not a Cinema control and has no place in filmmaking, it is a phase control only useful for video sources that are out of phase, it is not useful for RGB film sources that are NEVER out of phase.

You may be able to make two ISO versions of the image and merge them to avoid software clipping issues in the programs you are using, but try DRX first.
 
Once the image leaves the RED tab in Color it's locked into RGB. Sucks, but I'm guessing that having it access FLUT et al through the entire image chain would entail a total software rewrite.

If I've got a shot that needs highlight recovery, I usually do a secondary *before* any of the rest of my grading. So do the various secondaries, using one of them to pull down the highlight, then use the Primary Out to push the rest of the scene up to where it needs to be. It's kinda kludgy, and not good for more than about a stop of recovery, but it seems like the best option so far in Color.
 
Ben,
I thought of that the problem is if you want to push the scene using ISO in Red tab it isnt available in Primary out. Basically we need access to all that great information in Red Tab, in secondaries. That where I find myself trying to use power windows ets to dial stuff back in, but I cant get access to that data because if its clipped in Red Tab its gone.
 
I use the secondaryes in Scratch to pull down the higlights and it works pretty good.
If that dosen´t work in Color a good place to deal with that is in Red Cine X, where the
FLUT control deals with the midtones without crushing the highlights. I´m looking forward to have those controls in Color and Scratchs as well.

Best
 
Once the image leaves the RED tab in Color it's locked into RGB. Sucks, but I'm guessing that having it access FLUT™ et al through the entire image chain would entail a total software rewrite.

Well, if a gamma in floating point mode can access all the pixels, there is no need to specifically support FLUT, you just need to do more work to make the roll off right (which is what makes FLUT magic right?). That also means you don't actually "need" ISO or other settings in secondaries as you should be able to access any possible pixel values you need. I can see for example very focal hue changes *before* burning in the white point could be useful in secondaries.

But as Nick infers, perhaps the floating point support in Apple Color is not capable of accessing all the data effectively, Nick do you have details on what the problem is with Color's floating point support?
 
Nick do you have details on what the problem is with Color's floating point support?

Most floating point renderers (Nuke and Baselight are two that I use) do not clip values at 0 or 1, or give the user the option to do so of required. Color appears to use the term 'floating point' only to indicate float precision.
 
I've been pondering this for most of the day, and I've got some thoughts, and maybe a suggestion as to a possible solution/workaround for the OP.

First, thoughts: I think Color is capable of 32-bit processing in the traditional sense. (Super-Whites, Sub-Blacks) At least up to 110 IRE and -10 IRE, according to a very brief unscientific test. It's possible that's just the limits of the built-in scopes, though.

The issue at hand, though, is that the RED plugin locks footage into traditional RGB space, clipping/compressing at >100/<0. In a traditional grading environment, what I'd probably do is jack the ISO up to expose for the room, sending the windows into super-white, then bring them down in secondary. Obviously, though, that's not possible given that RED won't let us go super.

There are a couple of possible workarounds here:
1) Render to DPX or something that won't mind super-whites, set your exposure in the RED tools for the windows, then bring that into Color and do what you've got to do to brighten up the room.

2) Start your grade in REDLog. That might flatten the image enough that you can pick an ISO that brings both the window and the room into a usable range. Stretch the image out to taste.

Obviously neither of those are totally optimal solutions. The best solution would, of course, be to have access to the extra data, but since that's not currently an option, maybe one of those suggestions will help.
 
Back
Top