No way to preserve super-whites?

[Mike Jackson]

Yup, as I said above, I *am* working in 32bit. I came across that suggestion a few places, but it still doesn’t seem to change anything. It’s as though AE throws out the superwhites before it does any other processing / interpretation of the footage…

[Chris Wright]

post a project file with a single image. 32bpc has nothing to do with importing 8 bit footage. what does your color management look like? interpret, view output, simulate, monitor profile, AE preferences-gamma adjustments etc…

[Mike Jackson]

Not sure I see the point in posting a project file, when anyone with DV, HDV, or XD-Cam footage on their drives can test it for themselves in seconds.

Regarding my preferences and settings – Video Preview set to ‘More Accurate’ through a Blackmagic board (though I had the same issue with a Matrox board, and it really shouldn’t affect it anyway). Project color is set to 32 bits float (though the results are the same in 8-bit and 16-bit). Signal goes out through SDI to a Flanders Scientific monitor.

I have tried every combination of settings I can think of for color management in the project and clip interpretations. Most of the clips are embedded as Rec709 and can’t be changed, though even with one clip I *could* assign other profiles to, the results were the same. I’ve tried setting the project to Rec709, Re709 (16-235), Wide Gamut RGB, and many others. I’ve even fooled around with switching to linearized working space. I’ve tried all the same settings on the profile of the clip I *could* change, in various combinations with different project settings. The problem remains. The gamma will change, the white point and black point will rise and fall, but always the data above 100IRE in the original source is clipped, even if the new white point goes down as low as 70IRE.

So to my eye, it seems as though any of these color management options are applied only AFTER AE has thrown out the super-whites.

I would LOVE to bend the ear of someone at Adobe to see if there’s any way to address this.

-M

[Jon Bagge]

I usually try and keep the overbrights too, and I don’t seem to have this problem. My stuff is usually from Avid. I export using the 709 option from Avid to keep the full range, then use standard Rec.709 (but NOT 16-235) in AE. This will keep the superbrights, but also put blacks at 16. Then I add a levels effect to move the blacks to 0.

I use Rec.709 working colour space, and render to Rec.709 (16-235) to go back to Avid.

Occasionally I encounter files where I can’t use the Interpret Footage, and I’m not sure how to get around that. I believe this happens when I use Avid DV codec, which is only for temporary use anyway. Any 1:1 or DNxHD is always fine.

————–
https://www.jonbagge.net
Jon Bagge – Editor – London, UK
Avid – FCP – After Effects

[Mike Jackson]

That combination of color settings (Rec709 all the way) was the very first thing I tried. Mind you, I didn’t *export* anything from Final Cut – I sent an XML to Premiere, and from there cut-and-pasted the sequence right into AE – Thus linking AE directly to the original source footage.

But you have made me wonder – Are you on a PC, or a Mac? Could this be a Mac or Quicktime-specific issue…?

[Jon Bagge]

I use both. PC at home, usually Mac at client. This works fine either way.
I haven’t used the XML workflow so I don’t know what trouble that may be causing.

But the basic idea is to import a file with Rec.709 16-235 (or whatever your editing app would call that, and then interpret it in AE as Rec. 709. You have to deliberately misinterpret it in other words.

I almost always work in 16-bit by the way, not 32-bit. That shouldn’t actually matter for this to work.

Another option is to use a levels effect in Premiere/FCP and take the superbrights down to 235 first. As long as you’re working in 10-bit this should be fine. In 8-bit you’d be losing some information.

————–
https://www.jonbagge.net
Jon Bagge – Editor – London, UK
Avid – FCP – After Effects