Color Space Name Improvements

  • Author
  • Message
Offline
User avatar

shebbe

  • Posts: 1004
  • Joined: Tue Mar 06, 2018 11:48 am
  • Location: Amsterdam
  • Real Name: Shebanjah Klaasen

Color Space Name Improvements

PostThu Nov 25, 2021 11:16 am

I would love to see a proper naming scheme for all the color spaces.
Preferably all the commonly used combined ones with a '/' to divide like most of them are.
For me the 'dummy proof' naming schemes create more confusion than clarity.

Project Settings Color Space List:
ACES -> AP0/Linear or ACES2065-1
ACEScc -> AP1/ACEScc
ACEScct -> AP1/ACEScct
ARRI LogC -> ARRI Wide Gamut/LogC or AWG/LogC

CST Color Space:
ACES -> ACES (AP0)
ARRI Alexa -> ARRI Wide Gamut

ACES Transform:
Alexa -> ARRI Wide Gamut/LogC or AWG/LogC
Add missing color space -> ACES2065-1

Remove all CSC labels on those color spaces and remove any duplicates.
RWG/Log3G10 is the same as RED Log3G10 REDWideGamutRGB - CSC
Initially the idea in the node was to only have Display targets as output spaces but in practice this is not the only task an ACES transform should do so CSC versions were added to both input and output lists but we really just need the actual to be the same on both sides. Make the input list equal to the output list!
This should also fix the sorting in alphabetical order so non display transforms (with - CSC at the end) are easily reachable like ACEScct which is probably used more often than the huge display list.

Alternatively you could split out these to ACES Display Transform and ACES Color Space Transform by having a setting inside the node to tell whether you want to go to display or not and thus changing the list accordingly to clean it up.

Curious how others feel about this.
Last edited by shebbe on Fri Nov 26, 2021 4:35 pm, edited 1 time in total.
Home System Resolve 18.6b9: Z790 / i9 13900K / 64GB DDR5 / RTX4090 / Win 11 / ASUS PA32UGC 1600 nits
Office System Resolve 18.6b9: X570 / Ryzen 9 5900X / 128GB DDR4 / RTX4090 / Win 11 / EIZO CG248-K
Offline

Hendrik Proosa

  • Posts: 3007
  • Joined: Wed Aug 22, 2012 6:53 am
  • Location: Estonia

Re: Color Space Name Improvements

PostThu Nov 25, 2021 11:34 am

+1

Current namings are ambiguous and don’t properly define colorspaces.
I do stuff.
Offline
User avatar

waltervolpatto

  • Posts: 10498
  • Joined: Thu Feb 07, 2013 5:07 pm
  • Location: 1146 North Las Palmas Ave. Hollywood, California 90038 USA

Re: Color Space Name Improvements

PostSat Nov 27, 2021 12:56 am

+1 for cleanup
W10-19043.1645- Supermicro MB C9X299-PGF - RAM 128GB CPU i9-10980XE 16c 4.3GHz (Oc) Water cooled
Decklink Studio 4K (12.3)
Resolve 18.5.1 / fusion studio 18
GPU 3090ti drivers 512.59 studio
Offline

twainrichardson

  • Posts: 268
  • Joined: Wed Mar 27, 2019 2:19 pm
  • Real Name: Twain Richardson

Re: Color Space Name Improvements

PostSun Nov 28, 2021 1:39 am

+1
2023 16' MBPro/M3 Max/48GB/macOS Sonoma
Output: UltraStudio 4K Mini, Desktop Video 12.4.2
Offline
User avatar

Olivier MATHIEU

  • Posts: 864
  • Joined: Thu Aug 04, 2016 1:55 pm
  • Location: Paris/Grenoble, FRANCE

Re: Color Space Name Improvements

PostWed Dec 01, 2021 7:17 pm

+1
Resolve Studio 18.6.x & Fusion Studio 18.6.x | MacOS 13.6.x | GUI : 3840 x 2160 | Ntw : 10Gb/s
MacbookPro M2 Max
Offline
User avatar

Chris Mierzwinski

  • Posts: 81
  • Joined: Wed Aug 16, 2017 5:58 pm

Re: Color Space Name Improvements

PostSat Dec 11, 2021 6:23 am

+1
Resolve 18.6.4 Studio
MICRO PANEL
ASUS ProArt PA329CV 32" Monitor
intel i9 7900x 10 core
Nvidia 1080ti
win 10 pro
64GB ram
m.2 512gb nvme scratch disk
Offline
User avatar

Olivier MATHIEU

  • Posts: 864
  • Joined: Thu Aug 04, 2016 1:55 pm
  • Location: Paris/Grenoble, FRANCE

Re: Color Space Name Improvements

PostSat Dec 18, 2021 11:55 am

Hello
I'm french and I teach Resolve
I use the word "ColorSpace/Espace Colorimétrique" only for referring a 3 dimensions "space" ...like ACES2065-1
I use the word "Gamut" only for referring a 2 dimensions ... like AP1 or AP0
I use the word "Gamma" only for referring a 1 dimensions ... like Linear
It's very simple for the the trainees to understand colorspace that way

But the OFX ColorSpaceTransform and other places in Resolve use the word ColorSpace" either for 3 & 2 dimensions .... witch is very confusing :?: don't you think so ?
Resolve Studio 18.6.x & Fusion Studio 18.6.x | MacOS 13.6.x | GUI : 3840 x 2160 | Ntw : 10Gb/s
MacbookPro M2 Max
Offline
User avatar

waltervolpatto

  • Posts: 10498
  • Joined: Thu Feb 07, 2013 5:07 pm
  • Location: 1146 North Las Palmas Ave. Hollywood, California 90038 USA

Re: Color Space Name Improvements

PostSat Dec 18, 2021 9:38 pm

Olivier MATHIEU wrote:Hello
I'm french and I teach Resolve
I use the word "ColorSpace/Espace Colorimétrique" only for referring a 3 dimensions "space" ...like ACES2065-1
I use the word "Gamut" only for referring a 2 dimensions ... like AP1 or AP0
I use the word "Gamma" only for referring a 1 dimensions ... like Linear
It's very simple for the the trainees to understand colorspace that way

But the OFX ColorSpaceTransform and other places in Resolve use the word ColorSpace" either for 3 & 2 dimensions .... witch is very confusing :?: don't you think so ?


... AP1 and AP0 still refer to a three primaries color space....

are you doing the distinction:
1) ACES2065-1 - that define both the volumetric primaries AND tonal range
2) AP1 - that only define primaries
3) "linear light" - that only define tonal range?
W10-19043.1645- Supermicro MB C9X299-PGF - RAM 128GB CPU i9-10980XE 16c 4.3GHz (Oc) Water cooled
Decklink Studio 4K (12.3)
Resolve 18.5.1 / fusion studio 18
GPU 3090ti drivers 512.59 studio
Offline
User avatar

Olivier MATHIEU

  • Posts: 864
  • Joined: Thu Aug 04, 2016 1:55 pm
  • Location: Paris/Grenoble, FRANCE

Re: Color Space Name Improvements

PostSat Dec 18, 2021 10:26 pm

waltervolpatto wrote:... AP1 and AP0 still refer to a three primaries color space....

For me AP1 & AP0 are "Gamuts". A gamut is defined by 3 primaries and a White point
Those 4 points are defined by only 2 coordinates each ➧ it is not 3 dimension

waltervolpatto wrote:are you doing the distinction:
1) ACES2065-1 - that define both the volumetric primaries AND tonal range
2) AP1 - that only define primaries
3) "linear light" - that only define tonal range?

Yes, exactly
For me
1) is a Color Space
2) is a Gamut
3) is a Gamma or "gradation" (in french" ) ... it's more like tonal range as you mentioned

e.g : The new resolve "ColorSpace" in Resolve 17 is made of
1) a Gamut called "Davinci wide Gamut"
2) a Tonal range called "Davinci intermediate"

Neither of "Davinci Wide Gamut" OR "Davinci Intermediate" IS a ColorSpace (3 dimensions)

it seems to me clearer. no ?

I think OFX ColorSpaceTransform confusing ... and may be to all people who want to learn more ....
EDIT
capture 2021-12-18 à 23.28.05.png
capture 2021-12-18 à 23.28.05.png (86.89 KiB) Viewed 2151 times

Look : about the mapping it's "Gamut mapping" (Correct for me) and not "ColorSpace Mapping" :D
Resolve Studio 18.6.x & Fusion Studio 18.6.x | MacOS 13.6.x | GUI : 3840 x 2160 | Ntw : 10Gb/s
MacbookPro M2 Max
Offline

Hendrik Proosa

  • Posts: 3007
  • Joined: Wed Aug 22, 2012 6:53 am
  • Location: Estonia

Re: Color Space Name Improvements

PostMon Dec 20, 2021 9:24 am

Olivier MATHIEU wrote:For me AP1 & AP0 are "Gamuts". A gamut is defined by 3 primaries and a White point
Those 4 points are defined by only 2 coordinates each ➧ it is not 3 dimension

Gamut is three-dimensional as Walter wrote. The fact that it is defined by tuples in xy space doesn't change that, because they are actually xyY values with Y == 1.0, so three dimensions with one conveyed implicitly.

Otherwise I agree, what needs clear separation in Resolve is gamut and transfer function (tonal range). Transfer function is more widely used term than tonal range I believe, because "range" means definition of endpoints, but we are also interested in what happens inbetween, the function of how input values are mapped to output values. Gamut and transfer function together form a colorspace. Neither of them separately does because one or another must be deduced somehow and it produces most of the current ambiguity.
I do stuff.
Offline
User avatar

Olivier MATHIEU

  • Posts: 864
  • Joined: Thu Aug 04, 2016 1:55 pm
  • Location: Paris/Grenoble, FRANCE

Re: Color Space Name Improvements

PostTue Dec 21, 2021 4:34 pm

Hendrik Proosa wrote:Gamut is three-dimensional as Walter wrote. The fact that it is defined by tuples in xy space doesn't change that, because they are actually xyY values with Y == 1.0, so three dimensions with one conveyed implicitly.

I'm confused : in xyY is Y the the "tonal range" ?
Is xyY a ColorSpace ?
Where Can I read more about that ?

Hendrik Proosa wrote:Otherwise I agree, what needs clear separation in Resolve is gamut and transfer function (tonal range). Transfer function is more widely used term than tonal range I believe, because "range" means definition of endpoints, but we are also interested in what happens inbetween, the function of how input values are mapped to output values. Gamut and transfer function together form a colorspace. Neither of them separately does because one or another must be deduced somehow and it produces most of the current ambiguity.

Agreed :)
Resolve Studio 18.6.x & Fusion Studio 18.6.x | MacOS 13.6.x | GUI : 3840 x 2160 | Ntw : 10Gb/s
MacbookPro M2 Max
Offline

Hendrik Proosa

  • Posts: 3007
  • Joined: Wed Aug 22, 2012 6:53 am
  • Location: Estonia

Re: Color Space Name Improvements

PostTue Dec 21, 2021 7:02 pm

Olivier MATHIEU wrote:I'm confused : in xyY is Y the the "tonal range" ?
Is xyY a ColorSpace ?
Where Can I read more about that ?

xyY is a derivate of XYZ colorspace, xy gamut plot is in essence a projection of xyY space which removes the third axis. Big Y component is the Y from XYZ and directly expresses luminance. Since x and y coordinates don’t depend on the absolute values of XYZ (XYZ values can be scaled with a scalar and derived xy won’t change), this projection is normalized by luminance. This stability of xy coords in relation to absolute power is also the reason why one can’t go out of gamut by simply changing the magnitude of values in linear RGB based colorspaces. Only way to go out of gamut (in its color science meaning) is to set one or more primaries negative.

It is covered a bit in wikipedia: https://en.wikipedia.org/wiki/CIE_1931_color_space
I do stuff.
Offline

twainrichardson

  • Posts: 268
  • Joined: Wed Mar 27, 2019 2:19 pm
  • Real Name: Twain Richardson

Re: Color Space Name Improvements

PostTue Dec 21, 2021 10:08 pm

+1
2023 16' MBPro/M3 Max/48GB/macOS Sonoma
Output: UltraStudio 4K Mini, Desktop Video 12.4.2
Offline
User avatar

Olivier MATHIEU

  • Posts: 864
  • Joined: Thu Aug 04, 2016 1:55 pm
  • Location: Paris/Grenoble, FRANCE

Re: Color Space Name Improvements

PostMon Dec 27, 2021 8:43 pm

Thanks
Hendrik Proosa wrote:xyY is a derivate of XYZ colorspace, xy gamut plot is in essence a projection of xyY space which removes the third axis.

So the gamut is 2D
I think we are agree but we express it with different way

Hendrik Proosa wrote:Big Y component is the Y from XYZ and directly expresses luminance. Since x and y coordinates don’t depend on the absolute values of XYZ (XYZ values can be scaled with a scalar and derived xy won’t change), this projection is normalized by luminance.

I learn something
Hendrik Proosa wrote:This stability of xy coords in relation to absolute power is also the reason why one can’t go out of gamut by simply changing the magnitude of values in linear RGB based colorspaces. Only way to go out of gamut (in its color science meaning) is to set one or more primaries negative.

It is covered a bit in wikipedia: https://en.wikipedia.org/wiki/CIE_1931_color_space

In this webpage it says "CIE xy chromaticity diagram". Xy is 2D and Gamuts are shown on chromaticity diagram....

Please enlighten me if I'm wrong .. :)
Resolve Studio 18.6.x & Fusion Studio 18.6.x | MacOS 13.6.x | GUI : 3840 x 2160 | Ntw : 10Gb/s
MacbookPro M2 Max
Offline
User avatar

latridutou

  • Posts: 36
  • Joined: Thu Oct 17, 2019 8:30 am
  • Real Name: Latri Dutou

Re: Color Space Name Improvements

PostMon Dec 27, 2021 8:51 pm

+1
Technically accurate naming is always preferable in professional environments.
Offline

Hendrik Proosa

  • Posts: 3007
  • Joined: Wed Aug 22, 2012 6:53 am
  • Location: Estonia

Re: Color Space Name Improvements

PostMon Dec 27, 2021 10:01 pm

Olivier MATHIEU wrote:In this webpage it says "CIE xy chromaticity diagram". Xy is 2D and Gamuts are shown on chromaticity diagram....

xy diagram is a 2d view of 3d space, thus it is a projection. To tell how long and wide a car is you need just the top view of it, but it doesn’t mean that car is flat 8-)
I do stuff.

Return to DaVinci Resolve Feature Requests

Who is online

Users browsing this forum: Stephen Dixon and 8 guests