RED Helium whole workflow question


(Krzysztof Wlodarczyk) #1

Hello everybody,
I’ll just get straight to the point as I am preparing for a feature on which we wanted to apply ACES workflow to the whole process - from the set till color correction. In theory should be easy but I cannot find a simple example workflow, so I’ll try to create one here.

We have been working until now with RedLogFilm based workflow (create “look” LUTs after prep shooting, adjust them on-set, use for dailies in Resolve and finally use it like reference(after transfering to P3 for Nucoda) in grading). That was an easy way but since ACES can give some more benefits in transferring between systems/software using different monitoring it is time to try the new thing.

I would like to go point by point how to setup everything, so we can get through to final point. I am new to the ACES workflow and am still reading all the background, but it is easiest to understand when you do the actual thing in practice. Also when I am writing it I am trying to make all the steps clear to myself.

1.CAMERA - RED Helium 8K

We are going to work on RED Helium. The way to use ACES on-set with this camera is to set the “gamma” as ACESproxy as I understand. The color science stays as Dragoncolor2 if I remember correctly.
This is supposed to be related to ACEScc log curve as I found Graeme Nattress response on reduser forum.
This was an easy step.

2.ON-SET Grading - Livegrade Pro

We are using Livegrade Pro for on-set grading. Nucoda which is going to be the final grading software is using 1.0.2 ACES Transforms, so I set the same in Livegrade. Still easy.

Than I chose IDT. It should be “ACEScc v 1.0.2” and than “AMPAS - ACESproxy 1.0.1/1.0.2” as there are no IDTs for RED. This seems also clear.

I have my Livegrade setup as ACES CDL (or Advanced), ACES Freestyle - and this is where all the fun is happening on set.

Later I set my ODT. Since I am working on Flanders Scientific CM171 on-set in REC709 color space (could be P3, but let’s keep it 709 for now), gamma 2.4, 6500K I choose “AMPAS- Rec709 100 nits”. Not the “REC709 D60 100 nits”, right?

If everything above is correct I can go to the exporting…

  1. EXPORTING ON-SET LOOK

Here are 2 methods. If I chose grading mode ACES CDL (or Advanced) I can export CDL values that were aplied: on ACESproxy (using ACESproxy IDT) and before my certaing monitoring ODT. They should transfer without any problems and could be applied in any further step as CDLs.

But if I chose the second method ACES Freestyle - it allows me to export EVERYTHING as a 3D LUT. So the IDT and ODT I had set will be embedded. If I understand it correctly.
As I understand it, this method could be useful in the old workflow style like RedLogFilm. But I don’t exactly get it how to treat that LUT in further steps (because it has IDT and ODT embedded and all software will treat R3D files without IDT if I got RED workflow correctly). This is the more desired way of working, as I could make some “creative looks” imported as LUTs inside, do some HHS grading and basicaly much more than just simple primaries. For primaries I wouldn’t even need Livegrade - I could do it in camera in metadata using FoolControl - but that is for another topic.
So for now I am skipping that solution (but I would love to use it)- unless someone can give me a clue how to have a good use of the LUTs created in this way.

  1. CREATING DAILIES/EDITORIAL RUSHES USING RESOLVE

I setup my Master Project Setting first. I choose color science “ACEScc”, ACES version 1.0.2. And what goes next? I choose only ODT (REC709) if I want to stay in ACES RED workflow (according to RED there is no need to set IDT when using R3D files). Is it right or shall I choose something in ACES Input Transform? (While I am writing it I have only Resolve 12.5 on this computer so I cannot see any logical option corresponding to previous steps).
If I have everything set-up correctly I apply CDL and proceed to export as usual - DNx36 or Prores HD in Rec709 Video levels. Right?

  1. NUCODA - COLOR GRADING

Here I am just theoretical because I can only visit the post house. I assume if they are on ACES 1.0.2, the setup is similar to the one above in Resolve. I know thay can be ACEScc and outputing P3-DCI. The CDLs created in Resolve could also be aplied.

Is the above workflow correct? I plan to test it this week but was just curious if I get everything right.

The things I am not sure how everything will behave in Resolve and Nucoda. In the first 3 steps I work with ACESproxy and IDTs because I work with live feed from camera. But later in 4 and 5 - post soft steps - we deal with R3D files which skip those steps - ACESproxy + IDT, those programs are supposed to work straight from RAW in ACES. If I got everything right…

Also the route with “creative looks” in between is tempting. It would be nice to apply a small “creative” LUT in between to change some ugly RED color shifts.

PS. I was today in the grading suite to see the camera prep tests. What that camera does to yellows and REDs in ACES is really surprising for me(on set we were going the usual way like always - RedLogFilm + one over-the-years-developed-LUT. The ugly yellows and reds appeared while in ACES in post, which you couldn’t see by eye on set , or in the RedLogFilm workflow on monitors. Apart from that colors where OK). But please in comments let’s not concentrate on RED color science yet. I know it is scientifically great.


(Jim Houston) #2

I read this but there is a lot to digest. I have to find my copy of Resolve 12.5 to check on some things.

But a minor point, you can use the ODT_D60sim as it is actually for Rec709 g2.4 D65 monitors.
It shows the white point though of ACES directly and I think that white looks better in a theater.
D65 is when you want to convert to a video white point (and sometimes people want to do that)
it is a little bluer but the white point is all timetable by a colorist. The only important point is to keep
track of which one you used.

ACESProxy does not need an IDT. It represents the middle values of ACEScc directly with a little loss in prevision since it is only integer. But you can grade on it, and apply to an ACEScc setup in a color grading suite.

Jim


(Krzysztof Wlodarczyk) #3

Thanks,

the most important for me are the exact setups in practice. Theory follows. (maybe should be the opposite, but when dealing with images it is best to see the result I think)

I found Greame Nattress post from 2015 on reduser.net when they introduced ACESproxy mode in RED. I am refering to that one wherever I write ACESproxy. I want to stick to that specific workflow to understand all of it. Literally to digest ACES in the easiest possible way. So Graeme wrote:

"ACES Proxy is a way to map the ACES floating point data into a log format so you can work with it live. We implemented ACES Proxy for testing purposes last year, but since then the format and specification has been updated, and we’ll be putting out an update for that.

The ACES Proxy log curve is related to the ACEScc log curve, which means ASC CDL decisions on set can be applied in post."
http://www.reduser.net/forum/showthread.php?129811-ACES-Proxy-what-exactly-am-I-looking-at

Couple of years passed so I hope everything is up to date. If it was designed to work with grading on set than I hope that the option of “ACEScc 1.0.2 - IDT - AMPAS - ACESproxy 1.0.2” Is the proper option to choose (in Livegrade). Otherwise there is no option to work without IDT and with ODT only. What would that IDT for ACESproxy be anyway?

Here is the Livegrade workflow for ACES for reference:

Now to ODT REC709_D60. I think in my workflow I should stick to the regular one (so D65 I hope - no description in regular REC709 100 nits ODT), because most of people - editor,director,producers will look at the dailies output for that standard. Colorist can adjust the whitepoint for DP as he will be choosing his own ODT in Nucoda. Correct?

I hope that in the end ACES was designed for cinematographers not scientists to work with. Finding all the information is tricky, understanding a little more. But to introduce it to the artists (DPs I mean) we work with is a real challenge.