CST Settings for Rec 709 In To DWG / Intermediate???

Are the "proper" settings for Tone Mapping, Custom Max Input, Custom Max Output, Gamut Mapping, Forward OOTF, Inverse OOTF, when transforming Rec 709 footage in to DaVinci Wide Gamut / Intermediate?

I have seen some colorists say to set Luminance Mapping and Gamut mapping to none, and to apply Inverse OOTF.

I have seen some colorists say to set Tone Mapping to Luminance, Gamut Mapping to None, leave the custom max input and output at default, and leave both forward and inverse OOTF UNchecked.

And other colorists have said to just "eyeball" the custom max input and the custom max output. (Meaning, adjust those values by hand to whatever looks good.)

I normally have my color management page settings set to:

Color science: DaVinci YRGB
Timeline color space: DaVinci WG/Intermdieate
Output color space: Rec 709 Gamma 2.2

Thanks in advance.

There is no right or wrong unless you want a perfect, untreated Rec709 going in and out.
By that I mean, if you care about the output (Davinci Intermediate going to R709) reflecting exactly what the input file was. Then you should of course either just bypass the transforms at all or at least recreate the output but inverse.

If you do any grading / creative decisions at all, there is no right or wrong. It's up to you.



A note about default values. Each transfer function in the CST has a certain default value assigned to it. Most of the times when people put in some custom values they put in the exact same values that are default anyway. So take their recommendations with a grain of salt. Just to give you an idea, Gamma 2.2, 2.4 R709 and R709-A have a value of 100, DaVinci Intermediate would be 10000. HLG is 1000. Most of the times people put those in manually for no particular reason.

Should you use Luminance Mapping or DaVinci mapping? You decide. I personally like DaVinci better. It's simpler and usually more pleasing to me, but others prefer Luminance. It's been in Resolve longer.

There is no right or wrong unless you want a perfect, untreated Rec709 going in and out.
By that I mean, if you care about the output (Davinci Intermediate going to R709) reflecting exactly what the input file was. Then you should of course either just bypass the transforms at all or at least recreate the output but inverse.

If you do any grading / creative decisions at all, there is no right or wrong. It's up to you.



A note about default values. Each transfer function in the CST has a certain default value assigned to it. Most of the times when people put in some custom values they put in the exact same values that are default anyway. So take their recommendations with a grain of salt. Just to give you an idea, Gamma 2.2, 2.4 R709 and R709-A have a value of 100, DaVinci Intermediate would be 10000. HLG is 1000. Most of the times people put those in manually for no particular reason.

Should you use Luminance Mapping or DaVinci mapping? You decide. I personally like DaVinci better. It's simpler and usually more pleasing to me, but others prefer Luminance. It's been in Resolve longer.

Thank you so much for the reply and detailed information. I really appreciate it.

Yes, I will be grading the Rec 709 footage, so I will not be concerned about it trying to reflect exactly what the camera recorded.

Thanks again.

I think it's important to know that if the desire is to have an 'untreated' image to start with you need to inverse your forward transform. So if you use a CST with DWG/I to Rec.709 with default DaVinci settings then the only way to invert it would be to also use DaVinci as tone mapper with default settings. That way the image will not have been altered yet before any user grading adjustments. But I would say this only matter most when some already graded Rec.709 shots should remain the same in the context of the product you are finishing.

If there is already a look applied to the whole timeline and you also need to mix/match different cameras then it's of course no big deal to tweak the 709->DWG settings if it gets a better result.

I think it's important to know that if the desire is to have an 'untreated' image to start with you need to inverse your forward transform. So if you use a CST with DWG/I to Rec.709 with default DaVinci settings then the only way to invert it would be to also use DaVinci as tone mapper with default settings. That way the image will not have been altered yet before any user grading adjustments. But I would say this only matter most when some already graded Rec.709 shots should remain the same in the context of the product you are finishing.

If there is already a look applied to the whole timeline and you also need to mix/match different cameras then it's of course no big deal to tweak the 709->DWG settings if it gets a better result.

Thanks for the elaboration That makes sense.

In my particular case, it is kind of a "flat" looking rec 709 source from aconsumer DJI drone shooting in the D-Cinelike profile.

I have noticed with my drone and phone Rec709 footage that clipped whites will have extents that are above 100% on scopes.

Put a clip on timeline with no color management. Drop gain a bit, you will see there is 'stuff' that was hiding above 100% on the scope (turn extents on in scope settings)

I have not seen anyone explain why this is. But I personally don't want to throw this away at the input CST. So I will use input tone mapping and set a custom max input to like 130, or whatever is the bare minimum that those extents come in.

Sadly this does tend to drop the overall exposure a bit, so you will need to correct downstream again.


Would love to hear someone smarter than me explain a better way.

Would love to hear someone smarter than me explain a better way.

Won't claim I'm much smarter, but having data outside 0-1 range from a typical 'movie file' which is integer 8 or 10 bits will likely mean that the stored data is intended for video range but the encoded range is full to preserve slightly more dynamic range for standard Rec.709 encodings. Also referred to as "super-white". I'm not sure if a CST clips that when converting from Gamma 2.4 to a logarithmic transfer function, never really tested this myself. But if you want to protect it you could also interpret your files as data levels instead, but it will lift your blacks slightly. But maybe that's not so much a bad thing if you intend to grade it anyway.

Values outside of 0 and 1 in video codecs are usually related to full/legal range settings. As shebbe said, go to the clip attributes and it should do the trick.

The CST does not clip values per se, but if the tone mapping method is set to anything other than None it can. It will for sure clamp at 0 and 1. Setting the custom input value to sth like 130 helps, but I'd try the full/legal settings first.

Values outside of 0 and 1 in video codecs are usually related to full/legal range settings. As shebbe said, go to the clip attributes and it should do the trick.

The CST does not clip values per se, but if the tone mapping method is set to anything other than None it can. It will for sure clamp at 0 and 1. Setting the custom input value to sth like 130 helps, but I'd try the full/legal settings first.

Good advice!!! Thanks for explainging this. Quite helpful.