Questions about Project Settings & Color Space Transform

Hello everyone,

i'm glad to post for the first time here, so If you have any feedback on how to raise my question, I would be grateful to read it.

I am advancing my knowledge in understanding the possibilities with nodes, and in parallel understanding what the color space of each camera is.

And after years of recording around, watching countless tutorials and courses, I see that there is no clear knowledge by my inner circle about the use of color space in the timeline and this honestly raises a lot of doubts for me, which is why this post will be huge!

If you can answer at least one of the questions accurately, it will be of great help.

I thank you in advance for all your attention.

-

What are we going to talk about:
- Part 1: I see that MANY people don't know and don't teach correctly when it comes to putting together a timeline in Da Vinci. Due to the configuration of the "Color Management" and "Camera Raw" tabs within the "File>Project Settings" path. So I want to understand what the "correct" configuration is
- Part 2: "How to deal" with this configuration. Mainly regarding the use of the "Color Space Transform" effect during my color grading.

So let's start with my case:
I'm recording with 3 different cameras, but only 2 different color spaces:
- Ursa Mini 12k = Generation 5 with "Film" as Dynamic Range, 4k
- Canon R3 in "Canon Log 3"; 4KD IPB; "Canon Cinema Gamut"
- Canon R5 in "Canon Log 3"; 4KD IPB; "Canon Cinema Gamut"

Question 1) As my main camera is the Ursa, I decided to configure the timeline as shown in the screenshots below. Is this setting in the "Color Management" and "Camera Raw" tabs correct?




Question 2) Regardless of what I have to adjust in the first configuration, how do I deal with this decision during my Ursa color grading? In other words, do I need to apply a node with "Color Space Transform" at the beginning to transform it into something else? Or even at the end to transform into "Rec709" and "Gamma 2.4"?

Question 3) In the case of correcting the Canon color, I need to insert 1 node at the beginning to transform the "Canon log 3" and "Canon Gamut" into "DaVinci Wide Gamut" and "Da Vinci Intermediate" and at the end insert one that transform from Da Vinci Wide Gamut" and "Da Vinci Intermediate" to "Rec709" and "Gamma 2.4"?

I believe that these doubts help to resolve A LOT of things
I really appreciate your patience in reading this far and making yourself available to help the community.

Yours sincerely,
Victor Hugo
instagram.com/heavy.fingers

1. If you're shooting BRAW on the Ursa, the Input Color Space will have no effect. Resolve gets that data from the file and ignores your setting. So it might help to set that for the Canon.

2. Typically you choose either Color Managed or CST's. They serve the same purpose, so it' unusual to use both.

3. See answers 1 and 2.

Bonus Answer: For now, try setting the Decode Using to Camera Metadata in the RAW settings. Change that only if you need to (and you know what the change is for).

What Jim wrote.

Color managed project will do colorspace transforms for you, based on project settings and selected input color space for the clips. If input color space isn't automatically detected correctly you can set it manually per clip. Raw files are automatically converted to working space as Jim wrote. Manually adding CST-s isn't necessary unless there is a very specific need (you want to apply an effect or correction in some other colorspace than working space etc).

If you don't want to use color managed system, selecting plain "Davinci YRGB" then you need to do all transforms manually using CST-s or just grade from whatever it looks like by turning the knobbyz the old way.

1. If you're shooting BRAW on the Ursa, the Input Color Space will have no effect. Resolve gets that data from the file and ignores your setting. So it might help to set that for the Canon.

Nice to know that! But even with the Ursa i need to insert CST with 709 and Gamma 2.4 at the end?

2. Typically you choose either Color Managed or CST's. They serve the same purpose, so it' unusual to use both.

But in case that i'm using Ursa and Canon in the same timeline, i was thinking to put Color Management for BlackMagic and then put CST in color nodes, that's right?

Bonus Answer: For now, try setting the Decode Using to Camera Metadata in the RAW settings. Change that only if you need to (and you know what the change is for).

Thanks for this last one!!

What Jim wrote.

Color managed project will do colorspace transforms for you, based on project settings and selected input color space for the clips. If input color space isn't automatically detected correctly you can set it manually per clip. Raw files are automatically converted to working space as Jim wrote. Manually adding CST-s isn't necessary unless there is a very specific need (you want to apply an effect or correction in some other colorspace than working space etc).

If you don't want to use color managed system, selecting plain "Davinci YRGB" then you need to do all transforms manually using CST-s or just grade from whatever it looks like by turning the knobbyz the old way.

Nice, Hendrik! Thanks for your response!

But answer me one thing: Theoretically, wouldn't I lose the file's color correction potential if I did it without the care of using CST and just doing it in the "old style" as you called it?

Color Management will handle both cameras, converting both to the desired Rec.709 at the end of the color processing chain. You don't need CST's at all.

Maybe have a look here. Darren usually explains things pretty well.

But answer me one thing: Theoretically, wouldn't I lose the file's color correction potential if I did it without the care of using CST and just doing it in the "old style" as you called it?

The only actual potential there is, is the data. It either has the "granularity" or it doesn't. But the point of color management isn't that. CM (automatic or manual, doesn't really matter) tries to fulfil some specific desires:
1) unify the initial baseline of data by moving it all to common working space;
2) apply all operations in that working space, producing expected response for all sources;
3) separate data from display purposes by applying operations in a data state useful for image manipulations while at the same time displaying the result in appropriate way (with tonemapping, targeting specific output device etc)

Sure you can turn the knobbyz manually, but you lose all the above. If that's okay, its okay.

The timeline color space has only two effects:

1. It limits the gamut to the selected color space so it is essential to take the same or a wider gamut as the delivery color space.

2. It defines how the controls behave.

If you want the controls to behave "natural" there is no reason not to take DaVinci Color Managed (with the wide gamut) or ACES (if the ACES control in DaVinci Resolve act naturally is yet another discussion). If you're a good old LGG fan then DaVinci YRGB might be your best choice.

But the most important thing is to remember that the timeline color space ain't gonna make an Arri out of your fill-in-the-blanks camera if you select some Arri timeline color space.

Darren Mostyn notes you can use NO color management, you can use Resolve color management (RCM) for the entire project, you can use ACES color management, or you can use CST node color management. You don't use any combinations of these -- you just pick one.

I generally use CSTs if we're doing a project shot with digital cameras, and I sometimes make little adjustments or changes where necessary (like if something was overexposed and the color management is freaking out). All of these methods can work, none of them are completely right, none of them are completely wrong, but I like the idea of sticking with the camera manufacturer's idea of what their color science was. (Though there are always situations where the client doesn't know/doesn't remember, and you have a flattened file of questionable origin, so you have to just experiment and find a way to normalize the pictures manually.

I think Cullen Kelly (who has done extensive classes and YouTube videos on color science and color management) has finally swung over to the side that says "use CST nodes" as well. To me, the only time this discussion has the potential to get really difficult is if you're delivering projects in different color spaces: like one to Rec709/2.4 broadcast/streaming, another to D-Cinema P3/2.6, and then a Dolby Vision ST2084/1000-nit version. That gets complicated and you will need to adjust the signal flow accordingly. DaVinci WideGamut and DaVinci Intermediate are both good "middle steps" to convert to anything, and I think it's a more flexible way to work than ACES.

1) Starting out, pick one path ... manual, Davinci color managed, or ACES. Doesn't matter so much which, but pick ONE and use it!

Stick with that one process until you understand what it does and how to bend it all over creation. How to break it, how to fix what you broke.

2) AFTER learning one path as in step 1, then test another process, to learn what is similar and what is different to the things you have learned to understand.


Because trying to sort all three out at once is needlessly complicating and confusing. Don't treat yourself so poorly!

Sent from my SM-S908U using Tapatalk