Hi,
As part of the ACES Virtual Working Group on Gamut Mapping, here is a computational Google Colab Notebook about spectral images processing.
CIE 1931 2 Degree Standard Observer - Rendered Spectral Image
Canon 5D Mark II - Rendered Spectral Image
CIE 1931 2 Degree Standard Observer - Chromaticities
Canon 5D Mark II - Chromaticities
I’m planning to refine the workflow in the coming days to be able to scale across many images.
Cheers,
Thomas
I’m currently pulling images from https://personalpages.manchester.ac.uk/staff/d.h.foster/Local_Illumination_HSIs/Local_Illumination_HSIs_2015.html and planning to download some more, given how slow their server is, I’m considering making some datasets for colour-datasets.
Cheers,
Thomas
Hi,
The notebook has been overhauled and likewise for the Spectral Images Generation thread scaled to mass-generate images and have them hosted on Fast.io: https://academy-vwg-gm-hyper-spectral-images.imfast.io/
Cheers,
Thomas
Hi,
I have almost all the pieces required to do that with Mitsuba 2. Another url hosting a few images: https://academy-vwg-gm-mitsuba-spectral-images.imfast.io/
It should be straightforward to generate images for the other cameras. The advantage with the synthetic renders is that clumping that was happening with the relight spectral images disappears.
Note that the Nuke screenshot is rendered using the ACES RRT while on the website, it is sRGB without proper View Transform, you can get the ARRI K1S1 by pressing t, ? shows all the shortcuts.
Cheers,
Thomas
Hi,
In continuation of the last post, I made two notebooks for Mitsuba 2:
Cheers,
Thomas
