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 :)

LUTs, REDGamma LUTs, REDLogFilm LUTs,... free LUTs as far as the eye can see.

Could you post an r3d still to OneDrive/Mediafire/Dropbox? I'll take a look tomorrow afternoon in Nuke. Also a couple EXRs of the base debayer and what it "should" look like along with how I looks with a LUT. I wonder if it isn't just an offset issue.
 
The problem was actually overbrights not mapping properly.....nothing to do with HDRX, when you think about it if there are overbrights in an HDRX it will be in the A track not the X.

To get around this in fusion....and still do your CG in a linear workflow...

Bring in the footage as LINEAR----do your CG------convert to LOG-----and apply LOG to RG LUTS


Because going back to LOG normalizes all the values above 1.0...this maps the LUTS perfectly....if you just apply a LIN-RG LUT....it may not map the overbrights properly

LOG to RG LUTS can be found here

https://github.com/videovillage/RED-Conversion-LUTs

not sure if this is just Fusion specific....or if it the limitations of a LUT

At least contained within this thread is every possible conversion LUT for R3D in one place...so whatever your workflow your covered.
 
Ahhh yes, if there are overbrights the linear LUTs won't work. As noted above:
[quote="Gavin Greenwalt] RedLogFilm and Cineon/LOG are exactly the same. If your app (Nuke etc) has a built in filter/node/effect for converting between Log and Linear or Log and sRGB etc use that not the RLF_2_Lin or RLF_sRGB LUTs. The LIN 2 RLF and RLF 2 LIN or RLF 2 sRGB luts will be slower, less precise and generally worse than the mathematical cineon/log function built into the app. Some apps though I realize don't have any native log conversion so I included them for legacy purposes only.[/quote]

Maybe I should make that even more prominent. :D

Just an FYI there are RLF to RG luts in my LUT pack too. But it's nice to see there is another free package out there too.
 
Sorry Gavin, didn't notice you had the LOG LUTs in the package, there is a few :)...just had the others on hand, but they lacked any linear conversion.
Kudos for putting together a complete package.
 
Gavin,

I've been trying to use these LUTs to answer a question about noise using some GIO Scope LUTs of my own devising, but I'm stymied. My GIO Scope LUT tells me things about tones in ISO- and gamma-corrected spaces (such as ISO 800 / RG4), but it's giving me nonsense when I flip to Linear gamma. I'm trying to sort that out using LIN_2_RG4_64.cube, but I'm still coming up short. About 16x short. As in...if I set the source image to ISO 50(!) and look at the histograms in RG4, I can get a very blocky version of the same histograms using LIN_2_RG4_64.cube. But that's 4 stops from my concept of reality.

Moreover, I note that LIN_2_RG4_64.cube never gives an output above 0.833483 (would have expected 1.0), and looking at the math on the other side, RG4_2_LIN_64 gives a max output of 13.521684 (would have expected 1.0). Something tells me that scaling up the input by 16x will give me math that's in the ballpark, but it also means my 64x64x64 cube has the effective resolution of a 4x4x4 cube. Which is ugly, even if I'm just painting it with GIO Scope colors. How is it that your cubes can go beyond 1.0?
 
This is an older thread but just stumbled upon it today and is exactly what I've been after, thanks for the work Gavin!

Ive been using LUTs in grading for a year or so now but am still a noob on the technical aspects behind their exact make up. I stumbled upon this thread actually because I've been trying to hunt down a LUT package that can transform Sony's sLog2 and 3 into RedGamma 2 and 3. Essentially I'm trying to find the most efficient and consistent way to match my Epic with the a7s I I. Is this LUT package you've developed something that would work for this, and if not what would you recommend for matching these two cameras?
Thanks
Kris
 
Kris Tait said:
Essentially I'm trying to find the most efficient and consistent way to match my Epic with the a7s I I.
Click to expand...
Bear in mind that you're trying to match 12-bit images with 8-bit images, so you're going to be hamstrung right there.

Having said that, I'd advise have both cameras simultaneously shoot color charts on the same set under the same lighting conditions, like one of the DSC charts or the X-Rite Colorchecker Video. Pull them into Resolve and match them both with the scopes, using the available curves and primary controls (and possibly the Log controls for fine adjustments).

It is possible to use Resolve Color Management (and/or ACES) to put both cameras in essentially the same colorspace as a starting point. But I think it's equally valid to just find a look for the two cameras shooting the chart, save that look as a Gallery grade, and then apply that look as needed to each clip, then adjust them to suit the needs of the scene. While I don't think LUTs are the answer, I have no problem going with sLog or RedLogFilm and then using (say) a couple of nodes to create a starting point.

It's not easy to match two wildly-different cameras, especially if one is shot badly. Everybody has had to deal with this, sometimes on surprisingly big shows and features. When the differences are too great, sometimes you have to degrade the good camera to match the look of the bad, or else the transition between the two will be too jarring.
 
Hey Gavin, sorry for digging up the old thread and thank you for posting. I'll say I am a bit new to learning the technical side of coloring. I am currently coloring my own footage out of necessity. Anyways. I am using your luts to go from RLF to RG3 in Resolve. I am having trouble getting to true black though once the LUT is applied. I've tried using curves, offset, lift, etc. and they all basically crush the entire waveform into near black. In my experience, the offset usually shifts the entire waveform linearly. Am I oblivious to something? I don't know what linear gamma is, but I see lot's of talk about it here. Does this have anything to do with it?
 
Andrew Bissonnette said:
I am using your luts to go from RLF to RG3 in Resolve. I am having trouble getting to true black though once the LUT is applied. I've tried using curves, offset, lift, etc. and they all basically crush the entire waveform into near black.
Click to expand...

RG3's black point is above pure black. It's just a funky quirk of the curve. You should be able to use a curve though and move the bottom left curve point to the right to crush the blacks to pure black.
 
Kris Tait said:
I've been trying to hunt down a LUT package that can transform Sony's sLog2 and 3 into RedGamma 2 and 3.
Kris
Click to expand...
Here I made an slog2/3 to RLF. Apply that then apply a RLF -> RG3/RG2 lut from my pack. Marc's advice though is best. Go ACES if you want the easiest time matching footage. Although that will be a whole-nother bag of worms. Alternately it looks like RLF and sLog3 are nearly identical. You could just start in slog3.
 

Attachments

Gavin, Marc, and others! I'm awakening this post with some good news: I've managed to implement GIO Scope functionality using IPP2, a LUT, and a CDL.

GIO-LUT.jpg


Technical info here (plus link to a video demonstrating the functonality): https://manifoldrecording.com/2017/12/20/gio_lut/

Cross-posted to the other mighty LUT thread, in case that's actually the better one.

My own Christmas present to myself!
 
You must log in or register to reply here.
Back
Top