[Radiance-general] BSDF xml into Radiance

[Greg Ward]

Hi Lars,

Andy probably is the right person to respond to this, but as he's on a vacation until the end of the month, I thought I'd offer a couple of comments (inline).

Cheers,
-Greg

> From: "Lars O. Grobe" <grobe at gmx.net>
> Date: July 25, 2012 1:18:42 AM PDT
> 
> Hi Andy, hi list-subscribers,
> 
> I just came across this recent message about the usability of the bsdf
> material type with patch-based models of the sky including direct sun
> and complex fenestration. To avoid misunderstandings, I will try a short
> summary for others to comment on available options for annual
> simulations with complex glazing:
> 
> 
> 1) classic radiance tools (rpict, rtrace), complemented by mkillum to
> relax ambient setting.
> 
> Advantages: low noise, validated.
> 
> Disadvantages: very slow for annual simulations, no support when
> non-planar specular reflective surfaces are involved.

More specifically, non-planar, specular reflectors run into trouble for insolation.  Cloudy skies or sunless skies are no problem.

> 2) rtcontrib and patch-based model.
> 
> Advantages: faster for annual simulations.
> 
> Disadvantages: noise, nice images require high (slow) -ad and cannot be
> optimized using mkillum, limitations about specular non-planar
> reflectors apply.

The accuracy of non-planar, specular reflectors is actually better than #1, but the results are somewhat noisy.  A new -c option to vwrays (coupled with the rtcontrib -c option) is a good way to reduce noise that is available in the latest HEAD.  This is a better way to reduce noise than increasing -ad, and less costly.

> 3) rtcontrib and patch-based model, bsdf.
> 
> Advantages: support for non-planar reflectors, should be slightly faster
> than 2) as the fenestration system does not have to be traced internally
> - did anyone compare?
> 
> Disadvantages: still high -ad settings required leading to extended
> rendering times and still no way to get mkillum in, tends to
> underestimate direct sun (according Andy's message).
> 
> 
> 4) three-phase-method.
> 
> Advantages: very fast, can also be used with non-planar specular
> reflectors as bsdf data is supported.
> 
> Disadvantages: requires quite a lot of set-up work, e.g. subdivisions to
> reflect external obstructions. Patches visible in the results,
> fenestration geometry is not visible.

Andy has proposed an improved annual simulation method, which we hope to work on next year, to remedy the direct solar sampling difficulties in the 3-phase method.  It should also alleviate problems with external facade geometry and reduce the need to subdivide windows.

> 5) pmap.
> 
> Advantages: can be used with non-planar reflectors and multi-peak
> transmission.
> 
> Disadvantages: unknown status (any news?), not integrated with rtcontrib
> (contributions would need to be rendered manually).
> 
> 
> So if I need a way to generate images with visible fenestration
> geometry, the only reliable option would be 2), which requires very
> hight settings for -ad and thud will still be rather time-consuming, if
> noise is to be controlled.
> 
> Cheers, Lars.
> 
> On Tue, 2012-06-12 at 08:43 -0700, Andrew McNeil wrote:
> 
>> Though I've found that the BSDF material doesn't work well for
>> daylight coefficient based annual simulations (I'm assuming dds.bash
>> is a dynamic daylight simulation script).  Putting the solar radiance
>> into skypatches relies on probabilistic sampling to find patches
>> containing the sun, and if you don't have much direct transmission
>> from the direction of the sun, you aren't likely to find the sun.  Not
>> finding the sun causes big errors.
>