ACESNext Shoot - Overview and goals

Hello everyone,

As mentioned during the last TAC meeting, I have volunteered to help lead a test shoot to collect footage that is needed for current (as well as any possible future) working groups. This is being done in collaboration with the ACES team. The test will be done in Rome, Italy, where the Covid-19 restrictions have just been lifted and it will happen across three days, May 18, 19 and 20. The crew will consist of a DP, a Gaffer, a DIT, a camera assistant and a couple of trainees, all of which offered to take part in this as volunteers. The camera and lighting equipment required for the shoot will be provided by Panalight Spa, a prestigious camera rental house in Italy. The principal location will be Cinecittà Studios and Panalight’s headquarters.

Here you can find a link to the list of cameras we will be shooting, along with some other data that we are collecting. It’s a living document, so it’s going to be updated during our shoot. If anyone needs editor access, please DM me.

As it stands now, we will be shooting with 13-15 cameras, all the usual suspects plus some less common. For some of the “black-box” cameras that offer different recording format/recording color-spaces, we will try to shoot as many combinations as possible, especially if there is a RAW vs RGB option. We will avoid this on the main cameras, such as Alexa, Red or Sony, and for these we will only shoot RAW, at minimum possible compression (if applicable).

At the moment, we are planning to dedicate the first shooting day to shoot a small montage. We were thinking at a sequence where we will be using as many cameras as possible to portrait a camera assistant walking in the camera rental house and set up a camera test (meta cinema, I know). Once cut together, we don’t expect this montage to be any longer than 4-6 minutes. We will use this shooting to do a proof of concept of AMF and CLF. We expect to put this footage through the whole production pipeline: we will live grade it on set using Pomfort Live Grade and we will process dailies using Colorfront OSD, both tools have beta integrations of CLF and AMF. We will then cut it together, conform it and post produce it (perhaps adding some basic comps/VFX to it). And we will finally archive it using Marquise and its current integration of AMF.

Through this effort we will create as many iterations and versions of AMF as required by a real production, to aim for:

  1. Stress test the AMF pipeline and make sure every major use-case is covered by the current specs
  2. Create good samples to help future implementers
  3. Have material to show and use for ACES demo/presentation/plugfest

The second and third days will be focused on creating some test footage for the IDT and the ODT working groups.

We are currently aiming to shoot, for each camera/format combination in the above list, where possible always using the same Zeiss CP3 lens, the following:

  • a Macbeth Chart (X-Rite Classic ColorChecker, 24 patches), along with a X-Rite 18% grey card, on a 45/0 setup, lit by Tungsten light and, if times permit, by a HMI light. For these we will follow the IDT check-list approach provided by Josh Pines and the IDT Implementation Group. For the cameras that have official ACES IDTs already, we will only bracket the exposure ±1 stops. For the cameras that have not yet an official IDT, we will bracket the exposure as much as possible, ideally ± 3/5 stops.
    I’ll take with me a CR-300 and collect spectral data of the lights, the lens, the light reflected on the 18% grey chart.
  • A full frame lens coverage clip for each camera, shot against a well maintained OLED screen displaying a white patch, covered by a frost gel. From this we will be able to analyse the lens shading and remove it for post image evaluation.
  • A scene set up with some known objects, skintone(s) and some extremely saturated colors. Potentially bracketing the aperture in the shot.

We will organise a side-meeting with the IDT Implementation group to discuss this further, trying to set priorities (we have two days, but there is a lot to shoot!) and exploring better what is required to do to make the most of these tests. We will share here the invite so that anyone that might have inputs could join the meeting.

Other than that, we will take lots of pictures, we will try to document everything as much as possible, and…have a good time.

I would invite @dtatut , @jzp , @KevinJW , @alexfry , @carolalynn , @matthias.scharfenber @nick to comment/add to the list anything that they feel might help their working groups and @Alexander_Forsythe and @stobenkin to comment if there is anything to say I might be missing here.

Let me know!

F

9 Likes

Great stuff! All looks pretty good to me. I’ll let you know if anything else occurs to me.

Nit picking, RAW files aren’t in any colour space, but I know what you mean :wink:

Is there a reason you have the C300 Mk II down as Canon Log 2 rather that 3? And why ??? for its IDT?

Do you have the custom CLF IDT needed for the Blackmagic Pocket 6K? If not, I can make one for you with the data I’ve extracted here.

That said, if the Pocket 6K is running the latest firmware, it will be using “Gen 5 colour science” and that uses a different curve, which is not yet available in Resolve, so I can’t extract it. Likewise the Gen 5 primaries are unknown.

Thanks Sir, the BM IDT would be great. I have shared with you what I had from my archives as well.

Good spot on the table wording, I just tried to make it less ambiguous.

Re log-2/log-3, I left it there for now as I’m not sure exactly which model I will have available and what firmware. Will aim for Log3 anyway.

F

One thing I would like to add, I’m going to try to get an SDI recorder (such as an ATOMOS or similar) to record the live SDI output as well, I guess it could be good to have that in the archives as well to compare sw vs in-camera hw debayer or any possible difference between live SDI and recorded RGB file.

F

Sounds fantastic @Giardiello!

@Giardiello
I’ve just checked the lists in the google sheets and it’s amazing! Will it be available for download one day?

Also I saw you’ve used V-Log and V-Gamut for GH5, but seems like Panasonic made a mistake because GH5 primaries aren’t actually V-gamut. And I can’t even generate any primaries that would be able to describe its colors. I was able to do it with 3d cube only.

Thanks Anton,
Footage will be available for everyone soon. If you need it for now, I can share it with you.

In re to the V-log thing, it’s a good call I should change the spreadsheet. We shall see what we can produce with the IDT group.