search

luox-app / luox

Code base for the luox platform
https://luox.app/
GNU General Public License v3.0
14 stars 11 forks source link

There are 3 wrong values in the CIE 026 dataset #96

Closed NilesDavis closed 2 years ago

NilesDavis commented 2 years ago

Hi! First at all: thank you for all the work you did! It is great to finally have some persons publishing light computations open source. This is what I was longing for since 2008, but I did not find any acceptance for it. I know it is a little nerdy, but I compared my implementation with yours and found 3 values in your dataset differing from those, given in the standard CIE S 026:2018. That are 3 values in the L-cone action spectrum at 661, 662 and 663nm each differing +1E-7 Can you confirm this? Hopefully I can contribute more. Still digging in the code ;)

spitschan commented 2 years ago

Thank you for flagging this. There is a discrepancy in the published tables in CIE S 026:2018 and the spectral sensitivities given in the toolbox.

CIE S 026:2018:

CIES026.png

Toolbox (see Sheet "Action spectra" in the spreadsheet):

CIES026Toolbox.png

In the implementation, we elected to use the version in the CIE Toolbox.

NilesDavis commented 2 years ago

Wow, I didn't check that! Of course this is just of minor impact but do you see a possibility to clarify this, which version is correct? I just saw that you were also involved in the toolbox work, so perhaps this is easier for you than for me? I just know Peter to get in touch with.

spitschan commented 2 years ago

I would expect this arises from a truncation / rounding convention.

If you want to get to the bottom of this, you could email Luc Schlangen or Luke Price.

CIE has validated the results, you can read the report here: https://raw.githubusercontent.com/luox-app/luox/master/docs/CIE_Software_Check_of_luox_app.pdf.