Indigo Renderer Forum

We're in the process of moving the technical reference from a PDF file to a more accessible version:

http://indigorenderer.com/indigo-technical-reference

I have updated the ISL documentation a bit, there are some pretty pictures here:
http://indigorenderer.com/indigo-techni ... oise-funct

and here:

http://indigorenderer.com/indigo-techni ... oise-fun-0

A good move to put this into the web.
Example renders are also very welcome

Great idea!
Perhaps the document could also use an update on the content, not just here format and where it is located. The PDF is from 2011 so not exactly current. A good question to contemplate may be who is the target audience for the reference.
Most recently I've been struggling with two things and unfortunately the reference came up short in helping me answer my questions:

1. I wanted to know what technical unit is used for a tabulated emission spectrum for a light source. An example for such a light source would be the blue LED in the material database. I have a workaround in using normalized data for the tabulated values and then applying an emission scale, say in Lumen.
I still don't have an answer to that question.

2. This actually led me to investigate how lightlayers can be adjusted and exported correctly from Blendigo as that does not work when you have emission scale enabled in Blendigo. That question you was actually answered by Ono in an online chat.

Exporter writers are clearly in need of a good reference document.

Headroom wrote:Great idea!

1. I wanted to know what technical unit is used for a tabulated emission spectrum for a light source. An example for such a light source would be the blue LED in the material database. I have a workaround in using normalized data for the tabulated values and then applying an emission scale, say in Lumen.
I still don't have an answer to that question.

It's spectral radiance, Wsr^-1m^-3

Thank you very much!
I thought that's what it should be but was not sure.

Agreed, nice move

Ah meanCurvature(), I've been looking for this one !

Voronoï cells having a value of zero at contact edges would be a useful type.

CTZn wrote:Voronoï cells having a value of zero at contact edges would be a useful type.

You could detect this by taking a sample in each direction and checking the second derivative of distance to the nearest voronoi site with finite differences.

OnoSendai wrote:

CTZn wrote:Voronoï cells having a value of zero at contact edges would be a useful type.

You could detect this by taking a sample in each direction and checking the second derivative of distance to the nearest voronoi site with finite differences.

Exactly, that's why I can't (e: galinette taught us iirc)

I mentionned otherwhere that albeit I know how I would use them, I am not able to implement finite element methods myself :/

have a look at https://en.wikipedia.org/wiki/Finite_difference, "Higher-order differences"

so the code would look something like

second_deriv = voronoiDist(p + vec2(h, 0)) - 2*voronoiDist(p) + voronoiDist(p - vec2(h, 0))

where h is a small value (like 0.0001)

which would get the second derivative (curvature) in the x direction.

What it's doing is comparing the average of the two voronoiDist values to the left and right of our point p, with the value at p. If the average is bigger then it's positive curvature, if smaller, it's negative curvature, if it's the same, it's flat.

Well I will have to try this now that you made it seem affordable to me, thanks a lot Ono !