Getting To Know ACES Part 2 : Terminology and DaVinci Resolve

October 1, 2016

Start digging deeper into ACES terms and acronyms that will have you sounding like a pro. And you learn how to set up ACES in DaVinci Resolve.


Series

Update 1: This article was originally published in May 2016 and updated in October 2016 to include links to the new AcesCentral.com website and other minor edits

Update 2, Late October, 2016: Added mention of the ACEScct option included in the 1.03 specification

Update 3, Early November, 2016: Added mention of switching IDTs on timeline vs Media Page.


More ACES Terminology & Getting Setup To Work In DaVinci Resolve

NAB 2016 has come and gone (and life has kept moving), but I was reminded as I finished a feature in ACES this week that I needed to jump back on this series!

Speaking of NAB, huge thanks to DIT, technology guru and good friend Gary Adcock – we did a 3-hour session on ACES at Post | Production world that went great.

The Academy was also super helpful, providing Managing Director Of The Science Technology Council and ACES project lead Andy Maltz – Andy gave a 40 min ‘What Is ACES?’ presentation that was stellar.

In Part 1 of this series we got to know some terms associated with ACES including scene linear and display referred, as well as the different parts of ACES.

In this Insight, which is Part 2 of this series, I want to cover a few more ACES Terms and things that I realize I left out of Part 1 or wanted to go into a bit more detail. In the video at the end you’ll see how to set up ACES in DaVinci Resolve.

As a reminder, in Part 3 we’ll explore the essentials of a VFX pipeline using ACES – which I think is one of the main benefits of the ACES approach.

ACES 2065-1 AP0 Primaries, ACEScc AP1 Primaries & DaVinci ACES

One thing I want to spend a bit more time on and discuss in this Insight is the actual ACES color spaces.

Here are some of the things you’ll see (or hear) to describe ACES: 2065-1, AP0, AP1, Rec.2020+

What do these terms mean?

SMPTE ST 2065-1 is ACES standardized by SMPTE.

While this standard has many parts, in daily usage ACES 2065 has come to mean the full linear version of ACES that has a larger than the visual locus set of primaries (red, green & blue). This color space has a set of primaries known as AP0.

 

ACES-Primaries
In this graphic from Jeff Yurek at nanosysinc.com you can see the huge virtualized primaries of ACES 2065 AP0 primaries (left) compared to the smaller (still huge) ACES AP1 primaries also known as Rec. 2020 +

 

2065 -1 AP0 is mainly meant for archival and file exchange (more on that in Part 3) – in real-world usage –  grading, etc.,  that’s all happening in a smaller working space and that’s where AP1 primaries come into play.

Also known as Rec2020+,  AP1 primaries are slightly larger than Rec. 2020 and it’s the AP1 Primaries that are used in ACEScc, ACEScg and the Resolve specific ‘DaVinci ACES’.

Again, it’s with the smaller AP 1 primaries that most people work – both for vfx and grading.

In part one, I mentioned how ACEScc was log compared to ACES 2065 which is truly linear. Well, one thing that has caused a tremendous amount of confusion is the option of ‘DaVinci ACES’ in Resolve.

DaVinci ACES is not linear 2065 -1 AP0 ACES and it’s not log like ACEScc. While it uses AP 1 primaries, DaVinci ACES uses a 2.6 gamma modification.

Researching this more, the gamma modification is meant (I gather) to make the grading controls feel more ‘traditional’ and like ‘regular’ grading and I guess an attempt (prior to ACEScc being put into play) at a more traditional feel to ACES when grading.

In my own work with ACES projects, I’ve been avoiding DaVinci ACES and working with better-known quantity of ACEScc (EDIT 10/30/16: I plan to experiment with ACEScct  which is part of the newly released ACES 1.03 and compare it to ACEScc in a future Insight) .

In my opinion (and many others’), a special version of ACES by any particular manufacturer really defeats the purpose of ACES and a standardized color managed pipeline.  I don’t mean to pick on Blackmagic,  Filmlight does something similar with ACES Cineon Log, but again there are many that feel (myself included) this is a ‘hack’ of the ACES standard.

Version Number Of ACES & Why It Matters

AMPAS was very excited about (and promoted like crazy) the release of ACES 1.0 last year. But as you may know, ACES was around well before the official 1.0 release.

Resolve for example supports versions 0.1, 0.2, 0.7 and 1.0.

I don’t (and probably couldn’t) dive into all the specifics between these different versions, but it’s worth noting why older versions are still around in grading applications.

In Part 1 I mentioned the evergreen nature of ACES – well, support for older versions is part of that strategy.

Let’s say you were an early adopter of ACES and graded a film in ACES 0.2.  If you opened that project for a trim pass, years later, and used ACES 1.0 color science, things would look different.

In other words, ACES will (at least to my understanding) always be backward compatible – and that support is part of the deal that software developers agree to when implementing ACES.

Right now the official version of ACES is 1.03, but it’s worth mentioning that unless there are major future changes and/or needs for end users as determined by AMPAS, you will continue to see the major release versions in software and not the minor .0x releases even though software manufacturers behind the scenes may implement the .0X releases for performance and bug fixes.

Speaking of versions and different flavors of ACES, friend of MixingLight.com Nick Shaw has just released a set of CTL transforms to easily go between different flavors of ACES including ACES 2065, ACES CC

ACES Between Grading Platforms

One of the highly touted features of ACES is that when implemented correctly, shots should look the same no matter what tool is being used. As I’ll show you in Part 3 when we discuss a simple VFX pipeline, this is pretty much true.

However, after my NAB session on ACES this year, an attendee brought something to my attention about ACES and various grading platforms that I think is worth mentioning here.

While most of us will probably not worry about moving a project from Resolve > Baselight > Scratch > Pablo or some other combination of grading platforms this can be a major concern from an on-set platform i.e. LiveGrade, etc., to a finishing platform where the intent is to keep on-set grades ‘live’.

The thing is, even when these various platforms are setup to use the same flavor of ACES and version number, things might still not look the same – why?

Special sauce and the math of how the grading tools in one app work compared to another app, is why.

How one app applies Lift, Gamma, Gain, or Contrast & Pivot is going to be different from one app to another – even if just slightly This is issue is obviously not new to ACES, but there is some incorrect discussion that ACES solves this – it does not.

The tried and true method of moving corrections from place to place in a ‘live’ way is with an ASC CDL. This is true with ACES too.

Some apps, Baselight for example, have the ability to limit corrections to CDL compatible ones. Indeed, many on-set tools have this capability or really only allow you to make CDL compatible corrections (a good thing!).

While DaVinci Resolve has the ability to export CDLs, it doesn’t allow you to limit corrections for CDL compatibility – which if you’re not aware of what’s going on could be dangerous.

Again, this is probably not a huge deal for finishing colorists who don’t often move projects to different grading platforms, but could be an issue in on-set to post workflows.

ACES In Resolve?

Another attendee of my ACES session at NAB raised a very good question: ACES seems like an afterthought in Resolve, why wouldn’t I just use Resolve Color Management (RCM)?’

I think this is a fantastic question.

I agree ACES in Resolve doesn’t seem as developed compared to some other apps, and RCM in some ways (especially in Resolve 12.5) seems like a more full-featured tool-set and more controllable.

So then why do I continue to talk about ACES and use it on projects?

Pipeline.

While RCM is an amazing color management tool – it’s limited to Resolve.

ACES is all about pipeline from on-set to vfx, to final color and output.

RCM is really about Resolve, while you can make the argument that the standardization of ACES is not vital to many workflows (and I can understand that point) and that with some good workflow planning ACES doesn’t do anything ‘special’,  I think there is something to be said about the standardization of ACES and that’s why I’ve been using it.

Of course, confidence in something like ACES (or RCM for that matter) only happens with practice and implementing it in your own projects.

Video: Setting Up ACES in Resolve

With that background, in the video below I’ll discuss many of the above issues and explore how to setup ACES in Resolve.

I’ll also discuss using IDTs and ODTs, how the grading controls ‘feel’ and set the stage for the next part in the series.

UPDATE 11/16:  A few folks have noted that it’s not necessary to go back to the Media page to change an IDT.  That’s 100% true!  You can adjust a clip’s IDT right from the Resolve Timeline.  One advantage of going back to the Media page is changing IDTs for clips in bulk – but I didn’t show that in this video.  I went back to the Media page out of habit, not out of necessity.


Comments

22 thoughts on “Getting To Know ACES Part 2 : Terminology and DaVinci Resolve”

  1. Thanks a lot, Robbie! Seems like the Resolve team rejects to be standardised, which is not a good thing to me. Now they are putting time and effort into all 3: RCM, DaVinci ACES and real ACES.
    Looking forward to Pt 3, which you mentioned so often.

    1. ha! I did mention the next bit a lot – sorry about that. To be honest, I was planning on putting some stuff in this part about Nicks CTL scripts, what happens behind the scenes with no IDT or ODT applied etc but at the last minute thought that’d be better in the next part – so I think I over compensated with the call outs for part 3!

      To be fair, the Resolve team is doing what many color grading devs are doing – hedging their bets with additional color managed systems of their own. To be honest – when its just me, grading and finishing on my own system then I actually prefer the flexibility of RCM (especially in 12.5) but with more people in the pipeline ACES is compelling.

      I think a tell tale sign about ACES in Resolve will be if RCM continues to offer more input/output color spaces vs IDT/ODT for ACES.

      Time will tell….

  2. Great article, well done. Without a true linear ACES mode is there a way to send linear ACES plates to VFX or must these be in ACEScc? How can we render these without the RRT/ODT (seems like its all or nothing)? I’m guessing you will cover this in part 3. Just getting my questions in before its finalized.

    1. Thanks Steve – yes, covered in the next part. But the gist is – When no IDT applied footage is converted to ACES AP0 Linear – but then that is converted back your working space – ACEScc or DaVinci ACES. With no ODT same deal. So Resolve does have AP0 2065 just not a working space version of it! So to get a linear ACES 2065 AP0 EXR for VFX you’d turn off ODT. There are a few more nuances…but as I said I’ll cover that.

  3. In my own case, a major reason that I use ACES over RCM is that to my eye ACES gives a far better starting point when going out to REC709. If I bring the same footage in RCM versus ACES, the RCM shots all have very video/TV style oversaturated highlights. With ACES, the highlights don’t oversaturate and roll off to white very, very nicely. It’s my understanding that this is by design in ACES to mimic film response to light, and I have to say that it looks far better than RCM which seems to instead mimic the look of old school REC709 limited video cameras.

    1. I can see that. For me that type of thing is not totally consistent with RCM – depending on format etc. The RRT as you point out plays a big role as it has a bit of human touch and is one reason many love ACES but there are of course detractors that say they don’t like the ACES ‘look’. Can’t please everyone!

  4. I’ve been very curious about ACES and thank you so much for making this series of insights!
    My question though, still remains.

    If you are a small individual Colorist with your own “little! suite, working mainly on programs that only end in 709, there is really no benefit to using ACES, correct?

    ACES is only worth it, as soon as you have to to a multiple passes, like for P3 and Rec709, or have several VFX handouts, right?

    1. Jose –

      No, I don’t agree with that. ACES is a color management option that ‘might’ fit really with any project. While I’ve used it mainly on features, it could be right at home with TV, Corporate etc under the right circumstances.

      What I mean is there are a few factors to consider if ACES is right for you:

      1. You like how the controls ‘feel’ in ACES and you get the quality of images you want (really the most important thing)

      2. Do you know what camera(s) were used on a project? I think that’s vital in an ACES workflow.

      3. Do you like the ACES ‘look’. While AMPAS gets a little grumpy on that subject, ACES for sure has its own look & feel

      Technically there is a lot ACES offers – huge gamut – AP0, AP1, simple and precise processing. But its not the only game in town. Personally I’ve really come to love ACES. I wouldn’t think of it as ‘is there a beneift?’ but I’d encourage you to try it out as an approach to your projects and grading style.

      1. Thanks Robbie,

        So, I’m going to be coloring a 40 min short that’s supposed to be delivered in rec709 and DCIP3 so theoretically, using ACES, I can grade it in 709 and then simply change the ODT to P3 and make a trim pass right? ACES willl handle all the math automatically and I won’t need a conversion lut…

        1. Thru our own experimentation we noticed that this is not exactly true… Here is a response from BMD on that subject.

          “RCM simply converts accurately from one colour space to another, while ACES is more like an ‘artistic’ and dynamic transform.

          LUTs can be either colorimetric (like RCM), a mix of colorimetric and artistic (like ACES) or even completely artistic.

          Generally Resolve’s conversion LUTs are colorimetric and so should match more closely to RCM than to ACES.

          So basically we should not expect same result with ACES and LUTs, although with same colorspace name. The Results of RCM and LUTs in Resolve would be close.”

          ACES is always trying to give you the best looking image for a give ODT.

          And this is also why ACES is a great starting point for grading. Personally, I’ve noticed that I seem to have more range in colors when working with ACES. When starting from a LOG (or from the camera) it takes me more time to get image to the desired look VS using ACES. My looks seem richer than what they used to be… Maybe it’s placebo. But both myself and my colleague are feeling the same way.

          We we’re actualy able to prove this by using a standard test image from the CML image website. We created a grade (on our P3 projector) and probed the RGB colors to see where we were in terms of colors. our primaries were well inside 709. We did the same with the ACEScc base look (RRT/ODT combo). And low and behold, the primaries we’re already outside of 709. And this without having a super saturated base grade. So for us it was the only way of taking full advantage of our projector. And this will also be true for a 709 monitor!

          1. This is all great info, but not exactly what Jose was asking. He was saying when working in ACES (ACEScc/ACEScct) and he works on a project that is 709, can he switch the ODT to P3 and do a trim pass – yes that’s totally the idea!

            I agree that there is for sure a ‘feel’ difference between RCM and ACES but as you point out this is largely do the RRT that gives ACES its look/feel.

  5. So I’m just discovering ACES in post, and I love it. The color and feel is amazing. Definitely a tool I’ll be using from here on out when I grade. However, I also prefer to shoot all of my projects with a 3d lut loaded on to my monitor. One that shows me what I’ll be getting when I plug my footage into Resolve.
    What is the best way to monitor ACES on set? It’s not a 3d lut, so I can’t just load it on to my BM Video Assist monitor. So… I’m not quite sure how to go about that. Any tips?

  6. Thanks a lot for the guides! they are very helpful!

    I have just starting to colorgrade and I know that something like the oziris luts isn´t recommended for a true grading worflow but I need a little bit more knowledge and experience before I start doing my own looks. so my questions is simple. Most luts are made for use with rec.709 but now I wonder how ACES will react with luts? and if there are specific luts that are made for ACES?

    I´m also sorry if the my language is bad.

    thanks //Olle

  7. I’ve started using ACES and yes the control feel different. I use printer lights to grade the primary and finding that they behave differently. Each adjustment feels more dramatic. Without using ACES I would find that when I was using the + and the enter key to lighten or darken the image, everything would move up and down together. In ACES I find that the lift stays where it is and its the rest of the image thats adjusting and even when you lower the lift, its like ACES has determined where the lift even should be. Its only by using the LOG shadows that I can get the scopes to hit 0. Is that normal?

    1. Hi Peter – that dramatically bigger working space (AP1) def makes things feel a bit strange at first. I think the issue you’re reacting to is ACEScc vs ACEScct lift/shadow behavior. Check out Part 4 in the series (part 5 coming soon btw) where I discuss the difference between cc/cct.

      Sounds like you’re using ACEScc which you’re right it takes a tremendous amount of pull on the darker portions of the shot for them to move. Nick Shaw who is a contributor to this series describes the ACEScc behavior as like a piece of gum stuck to the floor you have to really stretch to get off the floor. While ACEScct (t stands for toe) offers a much more traditional lift approach.

      If you’re using cc you might try cct to see if you like it better. But the sticky shadow behavior is normal.

      Regarding the printer light issue you mention below I’ll test that when I’m back in front of a system. You’re using Resolve? What version? If in 12.5 what version of ACES?

      1. Thanks for the reply. I hadn’t got to part 4 but I’ll definitely check it out. Yes I’m using ACEScc. Yes I’m using Resolve 14 and the latest version of ACES (1.0.3). I just set the project to ACEScct and the printer lights behave in the same way.

      2. Heres what Blackmagic said about the Printer Lights
        “This is expected as there are cross-color channel transforms after your grade due to the ODT/RRT. If you use Resolve Color Management, and if you use a different output colorspace than the timeline colorspace, you should also see the same effect.”

Leave a Reply

Hundreds of Free Tutorials

Get full access to our entire library of 900+ color tutorials for an entire week!


Start Your Free Trial
Loading...