Author Topic: Problem with Aces and Corona Tonemap Control Texmap  (Read 2023 times)

2022-09-12, 17:46:18

dragan_ilic

  • Users
  • *
  • Posts: 2
    • View Profile
I think there is a Bug with the Corona Tonemap Control Texmap, if i Put it with an CoronaColor completely "255 White" in the "Direct visibility orrerride" slot and turn on Aces in the render frame Buffer, the Background is not completely white it is 253.

2022-09-12, 18:18:56
Reply #1

TomG

  • Administrator
  • Active Users
  • *****
  • Posts: 5434
    • View Profile
The Tone Map Control notes in the text "Removing tonemapping does not work with non-monotonic operators - some LUTs, some values of filmic mapping or 0.0 contrast". A fair chance the ACES OT is one of those non-monotonic operators and cannot be "cancelled out" by the Tone Map Control.
Tom Grimes | chaos-corona.com
Product Manager | contact us

2022-09-14, 08:57:54
Reply #2

dj_buckley

  • Active Users
  • **
  • Posts: 871
    • View Profile
Sorry to piggy back on this thread, but keeping in mind what you've just said Tom, if we plan to capture and use backplates in render (rather than compositing in post) would it make sense to process our backplates with no tone curve applied, linear so to speak?  The way I see it, if we use backplates in render then they generally already have a tone curve of some sort applied either in camera or in lightroom etc.  ACES OT would then technically double up on this tone curve.  So if we can't cancel the ACES OT transform out, then perhaps starting without a base tone curve on the backplates would be the next best option?  Of course this is only possible if we're shooting our own backplates and/or have access to the raw files?

We often shoot 360 backplates (not HDRI's) on site and use them in render for both direct visiblity and reflections, but there is always a battle trying to get it to look right in render, without any additional colour transforms/brightess changes due to the issues you've described.

Disclaimer - I'm not technical, I could be way off here

2022-09-14, 13:31:21
Reply #3

TomG

  • Administrator
  • Active Users
  • *****
  • Posts: 5434
    • View Profile
I am not THAT technical when it comes to LUTs and curves, but what you say sounds right - something of a trial and error I'd say, since the ACES OT cancellation was close to back to pure white, it could be that in some scenarios the cancellation is enough to look fine in a render with a backplate, but if you find it is not, then if a Linear version of the backplate is available that would (I believe... and someone more tech could say) give better results as then the LUT and ACES OT would be the only curves being applied and should affect the backplate in a similar way to the rest of the render.
Tom Grimes | chaos-corona.com
Product Manager | contact us

2022-09-27, 10:27:50
Reply #4

NazarVitkovskyi

  • Corona Team
  • Active Users
  • ****
  • Posts: 84
    • View Profile
Hello, thanks for sharing the information. This is logged to the developer's team.
(Internal ID=970543269)

2023-03-29, 20:27:09
Reply #5

brr

  • Active Users
  • **
  • Posts: 96
    • View Profile
Hello, thanks for sharing the information. This is logged to the developer's team.
(Internal ID=970543269)

Hello Nazar,

Do you have any news about this issue?

Best regards

2023-03-29, 21:03:46
Reply #6

TomG

  • Administrator
  • Active Users
  • *****
  • Posts: 5434
    • View Profile
As far as I can tell this should be "Improved AcesOtOperator unmapping in case alpha is 1.0 - Fixes issue that CoronaTonemapControl did not negate the effects of the AcesOT operator." in https://forum.corona-renderer.com/index.php?topic=38385.msg206878#msg206878 , so is in the 10 daily builds since then.
Tom Grimes | chaos-corona.com
Product Manager | contact us