|
The first clue that it's a rec709 transform is the fact thats it's called The Blackmagic 4.6K Film to Rec709 V3 LUT. It's a color space/gamma transform with a soft clip curve. It's not meant to be "cinematic". The function is to correctly map one gamma/gamut space into another. Most are probably looking for the Blackmagic 4.6K Film to IhavenoideahowmycameraorResolveworkbutIwantsomethi ngthatlookslikecinemawhateverthatmeans LUT.
while appreciating what you re saying CaptainHook (and I certainly respect your superior expertise here!) why do you think there is an impression that one camera makers 'reference' rec709 LUT is 'worse' or 'better' than another?
is this just inbuilt bias? I am assuming your inference is that as the gamma is defined there is no real room for making it respond in a 'nicer' way. but that doesn't seem to 100% jibe when (some) people compare to other systems... (and we do like to compare!)
Because the 4.6K cameras include information above the standard range, I think it causes a lot of problems with LUTs. If it's a finishing LUT or a gamma transform, this isn't a big deal because that goes at the end of the chain. You can correct under it and salvage those highlights and massage the look all you want.
A bigger issue, at least to me, is that the 4.6K to Linear LUT clips all highlights stored above the normal code values. You can't correct under it because then it changes the math that makes the whole thing work in the first place. The Color Transform node is a work around but requires the footage to go through Resolve.
If you speak of Arri, their colour scientist said their 709 LUT is more his preference for a gamma curve on a 709 display for their footage. Our 709 and Video settings/LUTs are the standard gamma curve with a small bit of highlight roll off. Both may have "709" (or Video) in their name as that's the target they're designed for, but that doesn't necessarily mean either are a "technical" LUT/transform - just that for BMD the decision was made (when the first camera was released) for it to be a very close technical transform to the Rec.709 standard to fit into existing broadcast workflows, and offer BMDFilm for those who wanted to create their own curves/looks/transforms.
Hey Captain,
I don't know if you're still monitoring this thread but I've found what I believe to be a glitch in the LUT that's carried over from DR 14.2. I was trying to figure out exactly what the LUT was doing so I used a CST applied to 709 color bars to transform into the V3 gamma/gamut then transformed back to 709 and all was good in the hood, everything was exactly where it should be. When I applied the LUT to the bars in V3 space the green channel jumped out of alignment significantly. I couldn't figure out why until I tried a soft clip on the CST-CST version to try and match the roll off. As soon as I touched the soft clip curve control the green channel jumped out of alignment. The soft clip behavior has been fixed on v14.2.1 but the behavior is still burned into the LUT. Obviously the saturation is going to be different because of the effects of the curve but I don't think the change in vector is by design.
Glitch.jpg
Hey, which LUT exactly are you talking about? The ones in Resolve are created by the Resolve team but i can look into it and pass it on to them if there's an issue. Thanks!
« Previous Thread | Next Thread » |
Tags for this Thread |