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 :)

Signalpath when outputting to P3 and grading in rec

Gunleik Groven

Gunleik Groven

REDuser Sponsor
Joined
Dec 28, 2006
Messages
13,465
Reaction score
5
Points
38
Location
Norway
Website
quine.no
Hi.

I guess I will test for this, but still, here is the question.

If you are grading in a REC 709 environment and select jpg2k/P3 as output format, will the signal go through REC709 with a transform to P3, or is the signal set up to have a P3 output LUT instead of a REC709 LUT for the output?

In short, is this a viable temporary solution if one does not have the ability for a calibrated P3 environment?

Best

G
 
Resolve doesn't really know what you use for monitoring. So if you are set up for Rec709 and output as P3, colours will be wrong, if I remember correctly they will be more saturated as P3 is a larger colour space with more saturated colours.

Just monitor in Rec709 and let Resolve or your DCP program do a transform to XYZ colourspace. People do that all the time.
 
Now I am looking at it I know what you mean. There doesn't seem to be any choice available as to which color space to render to. Strange.
 
Ok, I did a little test and it seems there is no color transform at all going on, the images stay in the same color space you were grading in which makes me wonder what the jpeg2000 output is intended for. I had thought it to be so you can directly ingest them into your DCP package but in this case you can't.

Edit: looking at it a bit more it seems you can add a 3D Lut to the track and then you can find under 'CineSpace' a P3 to DCI XYZ and a Rec709 to DCI XYZ.

Edit 2: BTW to test you can load a jpeg 2000 image in Photoshop and assign it a DCDM XYZ profile, then the images should appear normal again
 
OK...

In that case it would probably be better to go for TIFFs or DPXs...

Thanks.

But I am thinking:

Going to EasyDCP, the transform has got to happen at a stage.

What I would like to see was that the REC preview/grading LUT was repøaced with an XYZ/P3 output LUT (not on top of the REC) on an jpg2k Render...

Maybe I have my thinking all wrong and thinking too close to ACES here...
 
Gunleik, the only really accurate way to monitor P3 is on a DCI Cinema projector ie Barco, Christie, NEC, Sony etc. or the Dolby Monitor. Most monitors aren't capable of truly hitting the spec accurately and hence shouldn't be trusted. Nor should you be looking for a LUT of some kind to show you a transform conversion of Rec 709 to P3 if your monitoring device isn't capable of fully showing you P3. It just doesn't make sense. If you can't see it , there's no point. Your better off doing a Rec709 grade, and letting EasyDCP do your XYZ transform for you. It does a really good job. But if you have a very saturated film and you want to see all the available hues available to you in P3 that Rec709 can't display then you need to work on a DCI projector. There are other differences to account for in a P3 world as well compared to Rec709. P3 has more of a green tint to the monitoring path natively, is 2.6 gamma etc. all of these you will have to account for in your grade.
Now if you are insistent in trying to see something on your non-capable monitor that will attempt to mimic P3 you can purchase Light Illusion or Cinespace (are they still around?) and create a P3 LUT to apply to your monitoring path which will emulate P3 as close as it can. Again, I wouldn't use these LUTs to grade a show in P3 space. I would work on a capable monitor/projector.
 
Mike. I know this. But thanks!

What I also have is DCP software which converts REC to 709 in a satisafctory way.

All good, but still hungry! ;)

What I want is to skip a generation.

So when rendering jpg2k for DCP, the output LUT of resolve is changed from REC to P3 without going through the REC 709 step.

A bit like ACES.

Grade for 1 target, render for different targets without a generation loss...

Should be possible with Resolve 10s structure...

Gunleik
 
I don't think there is any advantage of going from Rec709 to P3 if you grading is done with a Rec709 monitor path anyway. Going to P3 only makes sense if you are going to use those extra colors that it offers, otherwise you only risk colors going wrong somewhere. What you would end up with is a Rec709 colorspace image inside a P3 colorspace with no way to monitor it, which only sounds dangerous to me.

edit: Mike beat me to it and makes much more sense than I do :-)
 
sander kamp said:
I don't think there is any advantage of going from Rec709 to P3 if you grading is done with a Rec709 monitor path anyway. Going to P3 only makes sense if you are going to use those extra colors that it offers, otherwise you only risk colors going wrong somewhere. What you would end up with is a Rec709 colorspace image inside a P3 colorspace with no way to monitor it, which only sounds dangerous to me.
Click to expand...



On the contrary... That is done all the time.

Making a REC master and transforming into P3.

And the Idea behind ACES is to make a master grade with one preview LUT and being able to do transforms to "any" space from there.

This is doable and the precission should be better, even when grading for a REC space, given that the LUTs/transforms are accurate.
 
Personally, I try to avoid using transform LUT's when at all possible. If you monitors are calibrated correctly to the spec you are delivering to you are all good. If originally grade in P3 and then have to do a 709 trim pass which is most often times the case for Blu-Ray delivery etc then use a transform LUT and tweak the images in scenes you see fit to match as close as possible. Same goes for the other way around. If you originally graded in Rec709 and then want to redo the grade in P3 space, use a transform LUT to get you 75% of the way there and then tweak as necessary.
I think the problem you are running into is you want the off the shelf Resolve package to do all these transforms for you but in my personal experience Resolves P3 to XYZ and Rec709 to XYZ are not accurate. I've emailed them about this in the Beta multiple times with side by side image examples. Like i mentioned, you should buy LightSpace or purchase a single transform LUT that they can create for you from 709-P3 if you are that hell bent on doing this conversion a lot in your workflow. Of course, these LUT's also have to take your monitoring profile information into account to be fully accurate which is why you should just purchase the program and then you can create all sorts of transform LUT's and inverse LUT's etc to your hearts content. Again, I think you are slightly overcomplicating the workflow here but hey, if it makes you feel warm and fuzzy inside I say go for it.
 
Thanks Mike!

I will look into it.

I am into testing signalchains.

But if you cannot think it up and desire it, it will never happen... :)
 
Mike is correct. The 709 to XYZ lut in resolve doesn't adjust for the 2.2 to 2.6 gamma difference (at least not the one included with the last Beta release). It appears that the v10 beta uses the same transform for EasyDCP, so if your coloring in rec709 and then use the built in EasyDCP tool, your DCP will be incorrect.
 
You must log in or register to reply here.
Back
Top