Locking down options for the DRT

In order to be able to deliver usable code to implementers, the unwieldy number of lines in the Blink development version needs to be significantly rationalised. Many options remain available which have not been used for some time. This thread is an opportunity for people to comment on which options people are happy to lock down, and note if there are any which they feel need to be kept available during final testing.

The currently selected options (v54) are:

  • Input AP1 clamp - ON in v54-pex1 (could be in an external default LMT)
  • HK mode - OFF (all stages)
  • Discount illuminant - ON (all stages)
  • Compress mode - ON in v54, OFF in v54-pex1
  • LMS matrix - From params (exact params vary between v54 and v54-pex1)
  • Input viewing conditions - Dim
  • Tone scale - Daniele Evo Curve
  • Reach mode clamp - OFF
  • Chroma compression reach - AP1
  • Eccentricity factor - NONE
  • Reach mode gamut compression - ON
  • Gamut compression reach - AP1
  • Boundary intersection method - “New from Nick”
  • “Nick method for reach” - ON
  • Upper hull gamma - Variable
  • Lower hull gamma - Constant
  • Focus distance gain - ON
  • Smooth cusps - ON
  • Six-axis compression - OFF
  • Output viewing conditions - Dim
  • Clamp output - ON
  • Soft clamp - OFF
1 Like

Are there any that should be kept for future debugging or testing?

As long as we have the final commit with the selected options and parameter values, we can always go back to that version for testing purposes if needed, after we have made a rationalised version that strips out all unnecessary code, but produces identical output to the previous version, can’t we?

I’ve started to look at some of these removals in this branch here:

Note this does contain an experiment using a tweak to the cone non-linearity that is not intended for use at this time. In particular it does not work with the current Y<->J functions we have.

Hi @KevinJW,

sorry for asking again about the “Input viewing conditions - Dim”.

From my understanding:
When I sit in a grading room for a TV commercial, the “right” selection for the Output viewing conditions should be “Dim”.

We also use the “dim” output for social media films, although a “Average” surround setting would be the “right” choice. The difference between dim and average seems to be a gamma value of 1.19 in nuke, as far as I can see.
The Output viewing conditions menu has still a tooltip that mentions ZCam (@alexfry) , but I guess this has no relevance anymore?

Does “Input viewing conditions - Dim” means I am sitting in a dim surround grading room and
“Input viewing conditions - Average” that I “grade” my footage in a brighter office surround?

And down in the menu " Output viewing conditions" I can still choose where my finished work is intended to be seen?

I hope this is not too confusing, but like this I can make sense of the menus provided in the Nuke DRT-v055.

To be honest, we have too many options exposed in the Blink, which are there for us to fiddle when testing, but don’t necessarily do what an end user might think they do!

Input viewing conditions as dim is just to match the output, so they null out, and the tone curve can do what it is expected to do. Having different viewing conditions for input and output has a much larger effect than we would want in reality. So probably best to leave both on dim.

1 Like

The following table enumerates different but common real-life viewing scenarios with surrounds ranging from dark to bright:

Viewing Scenario Ambient Illumination in Lx (or cd/m2) Scene or Device White Luminance in cd/m2 Adapting Field Luminance in cd/m2 Adopted whitepoint Surround
Viewing Slides in Dark Room 0 (0) 150 30 Projector Dark
Viewing Self-Luminous Display at Home 38 (12) 80 20 Display and Ambient Dim
Viewing Self-Luminous Display under Office Illumination 500 (159.2) 80 15 Display Average
Surface colour Evaluation in a Light Booth 1,000 (318.3) 318.3 60 Light Booth Average
Viewing a Smartphone in Outdoor Sunny Daylight 100,000 (31,830) 650 7,500 Outdoor Daylight Bright
1 Like