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 :)

RED Color - A look at REDWideGamutRGB/Log3G10

I'm curious how these image pairs relate to one another. I don't see a perfect one-LUT-to-rule-them-all, but my preference is for D as the high contrast, high saturation look and H as the lower saturation look. I'm guessing that the lower color sat of H is creating the impression of slightly lower contrast.

In D and H, I'm not seeing any channel clipping like in C, blues are true and not skewing toward red as in C/E/G, skintones not skewing green as in B. Red are skewing slightly blue in H. So I guess F is ok as well. If I could only choose one right now, it would be D.
 
Funny, I really strongly prefer H on just about every shot! Weird to see no one else prefers it.

B and D are too saturated, C and E seem neutral but they really handle the color of the water differently in the shot of the surfer, as well as the surfer's skin tone. C just seems unremarkable until shot 18 of the flower, where its highlights on the yellow petal look weirdly yellow to me, and 13 of the wiskey bottle I see the same problem. In shot 12 only E handles the blown out lamp expertly, but that's the only time I didn't immediately think H looked best. A,E,C must be dragon color because they do look the least jaundiced in the shots of Phil. H looks like a slightly more contrasty F, and the darker warm tones on H look a little less yellowy.

edit: I actually wonder if H isn't dragon color with a small saturation bump...
 
Last edited:
Originally, I was thinking B & D (because they hold more colour in the highlights), but they also have that oversaturated yellowishness that I'm not really a fan of... I'd imagine saturation to taste could take care of that easily though. Then I started thinking G looks like a nice middle ground between B&D -- basically holding more colour in the high side without being too saturated/yellowish...

But I think I'm getting tunnel vision, as they're starting to look the same... (slight satch and colour temp differences.)

Phil, I'm presuming the curve/wb/sat you added was the same across all of them, correct? Like there were 8 different/new transform LUTs...
 
mostly what i'm seeing in c is that i like the skintones, but i like the tonality especially highlights of g and non-skintone color...
(like the wave shot which is a dealbreaker for c right now not having that strong color differentiation)

somewhere between c & g would be sweet.

01. C
02. C/G
03. A/C
04. C/G
05. C/E
06. G
07. C
08. C
09. C
10. C/G
11. C
12. C/G
13. G
14. C
15. C/G
16. G
17. C
18. G
19. C
20. G
 
I need to look closer but one thing is certain. I would scrap B. Thanks for posting this!
 
I actually see nothing that would rule out my using anyone of these shots. '-)
 
  • Thread starter
  • Moderator
  • #91
Mike P. said:
Originally, I was thinking B & D (because they hold more colour in the highlights), but they also have that oversaturated yellowishness that I'm not really a fan of... I'd imagine saturation to taste could take care of that easily though. Then I started thinking G looks like a nice middle ground between B&D -- basically holding more colour in the high side without being too saturated/yellowish...

But I think I'm getting tunnel vision, as they're starting to look the same... (slight satch and colour temp differences.)

Phil, I'm presuming the curve/wb/sat you added was the same across all of them, correct? Like there were 8 different/new transform LUTs...
Click to expand...

Some of the differences are certainly more subtle. I might actually upload 4K versions of these in a bit so you can really pixel peep. Additionally screen size of course does play a role in color perception, which is a battle we deal with commonly when going from the big screen to a smaller screen.

Generally these are all consistent processes of a straight RWGRGB/Log3G10 + Transform LUT, however a few have "options" and some of the samples themselves have received an additional curve on top just to see how they hold up. All of the individual letters however like A, B, C are the same overall process. So if you like G or whatever look closely at all the G samples, they are the same method.


Elsie N said:
I actually see nothing that would rule out my using anyone of these shots. '-)
Click to expand...

This points out some of the subjectivity that Graeme mentioned regarding color. I worked on a film that had 6 people approving color for a moment (Cat in the Hat) and through a simple process of getting them all in one place I was able to discern preference as well as strange things like "I don't the blue here" when there clearly is blue being seen and read. We switched the process over from there and had only one person responsible for color from there on out on a project due to chaos like that.


David Grigorian said:
If I had to pick an overall favorite for all of the images it would be a tossup between E and G. the only ones I don't like are B and D specially on skin tones. To be honest though I wouldn't mind having all of them as an option to play with.

Thank you Phil and Graeme for all your hard work, really appreciate it.
Click to expand...

Cheers David and that's good to hear. All of these looks in one way or another are achievable, also far more than this is of course achievable as well. REDCODE RAW can transform into whatever you want really, but this new approach is about maximizing exactly what you can possibly capture with all current and likely next generation RED cameras with a focus on the future and standards like Rec.709, Rec.2020, and a few other things. Thus far everything appears to be rather "gradeable" and I do have some VFX shots I'm pulling keys from and whatnot to test that sort of thing. It's working well as far as I can tell. And there's the additional benefit of seeing slightly less noise as the pixels are painted slightly differently. Very interesting stuff actually on that front however subtle it may be.
 
The differences are subtle but be wary of making a call on what you prefer from looking at a computer monitor. While some computer screens can be reasonable many are way off. To really make a proper call these need to be viewed on a broadcast aligned monitor.
 
Phil Holland said:
I do have some VFX shots I'm pulling keys from and whatnot to test that sort of thing. It's working well as far as I can tell. And there's the additional benefit of seeing slightly less noise as the pixels are painted slightly differently. Very interesting stuff actually on that front however subtle it may be.
Click to expand...

i wouldn't mind seeing some of those if they're available for viewing/testing.
 
Tom Gleeson said:
The differences are subtle but be wary of making a call on what you prefer from looking at a computer monitor. While some computer screens can be reasonable many are way off. To really make a proper call these need to be viewed on a broadcast aligned monitor.
Click to expand...
Very well said.
 
  • Thread starter
  • Moderator
  • #95
One thing I'll add. For me only 4 of these are interesting and there are merits to each of those.

Also, one of the options shown is straight DRAGONcolor2/REDgamma4 which we are all familiar with.
 
Are RWG & Log3G10 locked and can be relied upon as a basis or they are still worked on ?
 
When I take into account

a) what I've noticed people tend to do to REDcolor/REDgamma images here on REDuser to obtain a "final" image

b) what RED images tend to end up looking like on professionally graded TV/theatrical releases (with a neutral grade, not any stylized looks)

c) what skin has looked like on celluloid over the years

d) what I think Graeme is going for with the new approach...

I think G is consistently where it's at, followed closely by F. Although image 1 would be F or C, given its particular lighting mix. Also, an argument could be made for F nudging out G in image 3.

Since different web browsers can show colors and gamma differently on the same machine, I copied every image into Photoshop and viewed them in the space they were saved in.

My monitor is calibrated and has a Delta E of < 3 which, while it ain't spectacular, is respectable. I might feel differently about the images with a more accurate monitor.
 
This is fascinating! Thanks for sharing Phil! When I get in front of my grading monitor I'll list my answers.

But in the meantime, I think we can all appreciate how difficult Graem's job is. The amount of variance in these answers is staggering!
 
Brian Boyer said:
When I take into account

a) what I've noticed people tend to do to REDcolor/REDgamma images here on REDuser to obtain a "final" image

b) what RED images tend to end up looking like on professionally graded TV/theatrical releases (with a neutral grade, not any stylized looks)

c) what skin has looked like on celluloid over the years

d) what I think Graeme is going for with the new approach...

I think G is consistently where it's at, followed closely by F. Although image 1 would be F or C, given its particular lighting mix. Also, an argument could be made for F nudging out G in image 3.

Since different web browsers can show colors and gamma differently on the same machine, I copied every image into Photoshop and viewed them in the space they were saved in.

My monitor is calibrated and has a Delta E of < 3 which, while it ain't spectacular, is respectable. I might feel differently about the images with a more accurate monitor.
Click to expand...

Bingo. Back at the hut I test out things on a few displays, as does Graeme. Calibrated and "other" even. On the road I have 3 calibrated screens here and one extremely risky thing I do is slug this workstationtop thing into each hotel's TV via HDMI. Always exciting and infuriating.

It's a difficult thing to "define" what's truly right or wrong, but what's very clear here is there are improvements from where we were. That's the big thing. It does look different of course than the well traversed Dragon Color Spaces and Gamma Curves. In this case I think it's a good thing and it will plug into many standard workflows from there.

I personally love the flexibility of all of this. I remember a time when color wasn't this much fun and that wasn't too long ago really. Especially on cameras with far less captured/possible color or even some strange film scans for that matter.
 
You must log in or register to reply here.
Back
Top