ACES IDT DCTL Generator

Tags: #<Tag:0x00007f58b941fd98> #<Tag:0x00007f58b941fcd0> #<Tag:0x00007f58b941fc08>

Hi, everyone

I developed an online ACES IDT DCTL generator. It is a tool that mainly provides a function of immediate building an IDT(Input Device Transform) into a format of DCTL (Davinci Resolve Color Transform Language) based on the guide of ACES documentation, camera vendors white papers and technical specification of some color science algorithm. Its usage includes:

  1. Generating DCTL for trending and more diverse camera types and a combination of gamut and OETF. I understand that the ACES official provides a series of public IDT, one purpose of this tool is to provides the solution for those camera models falls outside ACES IDT official support and provides a handy and relatively accurate IDT to those who facing various of camera types in this color grading project. I would like to keep update them based on the findable camera data.
  2. One function could be adding more parameters in IDT for parameter tuning and color science experiment. It allows getting a deeper understanding of the relationship between different parts of the ACES system and how it works. The data of color primaries, white point and log functions of different cameras can be visualized in this tool.
  3. Preview of results online in web browsers by uploading DPX format picture. The ACES workflow is developed and implemented in web GLSL and working in web rendering.
  4. Potential flexibility of adding upcoming algorithms designed in future ACES road map (e.g. gamut mapping) for fast prototyping and testing.

You can find the tool in the following URL
http://acesidtdctl.tcolorscience.com/

It opens publicly and hopes you enjoy it.

2 Likes

@tommyzenth this is very impressive work and I thank you for sharing. I am not yet sure how I can use it to my advantage since I haven’t had this particular need yet but I know that sure want to use it (because it is fun to watch the plots react to the setting changes :slightly_smiling_face:).

Already impressive and definitely a great showcase with potential extensions that could be very valuable to the ACES user community. I hope others take a few minutes to check out your work.

Thank you. I really feel motivated in this community and would like to keep enhancing it with more useful features in the future. :grinning:

1 Like

@tommyzenth Thank you for creating such a useful - and a great looking tool. I’m looking for some pointers as to how to create a DCTL that accurately brings Magic Lantern Raw into the ACEScc or cct space in Davinci Resolve. I think much transform happens under the hood in Resolve, so it’s not often easy to keep track of the effects of a transform. Any help gladly received! Many thanks.

This is very cool. I really like the UI and overall approach.

You can check out (if you haven’t already) the GLSL conversion of ACES 1.1

ACES 1.2 OFX Plugin also exports DCTLs, so that might also be worth checking out (for comparison’s sake)

I didn’t build a specific IDT DCTL for Magic Lantern Raw (which is automatically converted to ACES anyway) but I did build a set of tools for better working with the images, in particular the LMT_DNG_OFX DCTL

The images I worked on came from a hacked 5Dmk3 (some of them were dualiso).

@rivertim I am glad to help. I haven’t use the Magic Lantern Raw before during my work, but I did some investigation on this and you might try the setting shown in the attached image file.

Then use the Davinci ACES plug-in to convert from ACES to ACEScc. It is best if you can send me some dpx file in Magic Lantern Raw (just grab a still in Davinci and export this still in dpx format) via email or in this community so that I could figure out a clear workflow for Magic Lantern Raw.

Thanks for the response! Your plugins are awesome and used by many of my colleagues. I really love the LMT part ACES TOOLS you create that provides more precise controls of sat and hue compared with the tools in Davinci Resolve.

Hi Paul. Thank you for your toolset – they are amazing; I’m only just starting to get my head around them. :wink: I shall pay extra attention to the LMT_DNG_OFX DCTL now. [quote=“Paul_Dore, post:6, topic:2566, full:true”]
(which is automatically converted to ACES anyway)
[/quote]
I see… Do we knowing exactly what spectral data Resolve is using to do this? So does it not require any IDT at all? Have you seen @sdyer great tests gathering SS data here from amongst others 5DMk2, and Mk3? Could this be used to create an even more accurate IDT/DCTL?
PS You might also see a post from me in the ML forum on your Resolve thread. :slight_smile:

Thanks so much Tommy. I will send you the .dpx file. Should the Resolve Colour settings be anything special to grab the still from?

@sdyer Would it be ok if I sent @tommyzenth a still from your 5D3 tests? If so, which would you recommend?

Hi Tommy
You are really great
This tool is very helpful for colorists without programming knowledge
There is a doubt at the moment and I hope to get your answer
I tested IDT DCTL with ARRI and RED
Found that the conversion value
Deviation from the official PDF of ARRI and RED
I do not know whether it is caused by operational reasons
Below I provide screenshots generated by the IDT DCTL tool.
And the official PDF screenshot


ARRI_WGT_TO_ACES


RED_WG_TO_ACES AP0

@Dean thanks for the valuable response! I am to check it to see whether it is caused by program systemic error or other reasons.

Continuing the discussion from ACES IDT DCTL Generator:

If there was a guide of Davinci color management setting for Magic Lantern Raw, that would be the best. I think the default setting is sufficient as the Davinci will do the de-bayer process automatically. The question is that I am not sure what colorspace it cope with Magic Lantern Raw to my knowledge.

@tommyzenth Hey. ML Raw files from the camera (.mlv) are not read directly by Resolve. They have to be made into dng using a choice of apps/methods. So Resolve is just ‘seeing’ dng. Resolve does a pretty good job of debayer-ing into ACES without any specified IDT, but it would be nice to know exactly what is going on during the automatic process, and whether a custom IDT might improve the process?

@rivertim Got it. This may means the ML IDT process is done by the certain app that could be a black box. Or the colorspace infromation could be found in the documentation of that app.

@Dean I noticed that in the screenshots, the adaptation method used is not correct. The ARRI needs to use CAT02 adaptation, not Braford. You will get correct conversion with CAT02 method :slightly_smiling_face:. (There is a small info text line beneath the gamut select option UI, maybe I need to highlight them). However, I still see some deviation at .00005 to .0001 range. I would like to keep check it out.

@tommyzenth If I understand correctly maybe Resolve uses the standard transform for Adobe DNG and then into ACES AP0 > ACES cc/cct Maybe you could add this into your app as a source? I believe it is published data. https://wwwimages2.adobe.com/content/dam/acom/en/products/photoshop/pdfs/dng_spec_1.5.0.0.pdf

@rivertim thanks a lot. The document is quite informative and I think the gamut it uses is CIE XYZ and the gamma could be linear or canon c-log. The adaptation method is Braford. Given these cues, I am able to build options for Adobe DNG.

Also, I found a interesting post that validates this document about ML RAW IDT you’d like to see

https://www.magiclantern.fm/forum/index.php?topic=10626.0

@rivertim I added a testing CIE XYZ gamut and a linear option for gamma. Please refresh the tool page and feel free to play.