[Radiance-general] Re: Materials for scale model room

[Peter Apian-Bennewitz]

radiance-general-request at radiance-online.org wrote:

>Send Radiance-general mailing list submissions to
>	radiance-general at radiance-online.org
>
>To subscribe or unsubscribe via the World Wide Web, visit
>	http://www.radiance-online.org/mailman/listinfo/radiance-general
>or, via email, send a message with subject or body 'help' to
>	radiance-general-request at radiance-online.org
>
>You can reach the person managing the list at
>	radiance-general-owner at radiance-online.org
>
>When replying, please edit your Subject line so it is more specific
>than "Re: Contents of Radiance-general digest..."
>
>
>Today's Topics:
>
>   1. Materials for scale model room (Federico Giovannetti)
>   2. RE: Materials for scale model room (Francesco Anselmo)
>   3. Materials for scale model room (John Mardaljevic)
>   4. Re: Obtaining RGB output data (alira at gsd.harvard.edu)
>   5. Re: Obtaining RGB output data (Greg Ward)
>
>
>----------------------------------------------------------------------
>
>Message: 1
>Date: Thu, 8 Sep 2005 12:38:40 +0200
>From: "Federico Giovannetti" <f.giovannetti at isfh.de>
>Subject: [Radiance-general] Materials for scale model room
>To: <radiance-general at radiance-online.org>
>Message-ID: <002901c5b461$7a5cb2d0$e400a8c0 at PC501fg>
>Content-Type: text/plain; charset="iso-8859-1"
>
>Hallo Greg, hallo Radiance Community!
>
>I'm trying to compare some Radiance-simulation results with experimental measurements (i know it's not that easy, i'm just trying).
>As I would like to isolate (as good as possible) the effects of glazing system and skylight distribution, i'm looking for diffusing materials for my scale-room which can be  physically very good  modeled with Radiance. That is materials whose BRTDF is not more complex than a TRANS-type. So that i don't have to worry too much about them (??).
>Some textil materials (such as molleton for theater applications)  seem to be good diffusing ones, with an almost lambertian behaviour, but i'm still searching...
>Has anybody in the community a good tip (matt paint, etc.)?
>  
>
Hi Federico,

my two cents thoughts on validations in general: Rather simple ideas 
really. Sorry if they are a boring repetition of standard engineering 
procedures to many of you. Having watched and read some Phds on this 
subject, it seems worth mentioning so. Many other works on the subject 
don't need any of my humble comments. Nothing personal: "you" is just 
to-whome-it-may-concern.

    * Your results will most likely /not/ match you simulation. Be
      prepared to figure out why. Don't just say "oh, they are 15% off
      at point A und only 5% at point B, so we're better than 15%.
      Homework done, can I go play outside ?" . Check what's happening
      between point A und B.
      Compare whole curves rather than individual points to see whether
      there's an offset or a complete different shape. Both cases ask
      for an explanation.
      If they match perfectly, vary some parameter and check that the
      match results from you doing the right thing and not from sheer luck.
      If simulation and measurement differ by Monte Carlo and/or
      measurement noise, you're knighted and may retire to a splendid
      location on the planet.
      The aim of a validation is not so much an impressively small
      deviation between simulation and measurement, but a proven
      understanding of the mechanisms that lead to the deviations.
    * nature lies to you cold blooded at any moment you're not alert.
      Start with the simplest imaginable case and check that. Build
      trust in your results by adding complexity stepwise and in a
      controlled manner.
    * Keep track of your setup, materials and procedures. If you find
      out at a late stage that rho_dh of molton differs mysteriously
      from the value measured at the beginning (maybe because the large
      piece of molton that was cut down to smaller patches wasn't
      homogenous), you want to know which measurements might have been
      affected by this. Be prepared to recheck. That of course works
      more easily with automated measurements than with long runs of
      pure hand measurements.
      Unless you're taking data on Nessie or Mt. St. Helen, plan to
      repeat a measurement. Just for the fun of it and to see whether
      nature is still on your side.
    * get as much data points as you can. Timewise (some people are
      suprised to find that artifical lamps have a 50/60Hz modulation
      and that shows if values are not averaged over a longer period or
      if no special power source is used. output decrease with age, too)
      and spatial. IMHO the latter is higher valuable:
      Measurements at 3 points in a box are, at least to me, not enough
      to understand what's happening inside.
    * really diffuse reflecting materials are hard to find and it gets
      harder with higher incident angle (away from the surface normal)
    * "The systematic experiment", edited by J.C. Gibbings, Cambridge
      Univ. Press 1986 has some introduction to this art. Likely there
      are a piles of similar nice texts available.
    * Don't trust your superiors telling you that this bit of extra
      effort is not worth it.
    * To the superiors: Allocate enough money to do it right.
      Validations are like antibiotics: Do it fully or don't do it at all.

anyway, - enough nagging comments,
cheers
Peter

-- 
 pab-opto, Freiburg, Germany, http://www.pab-opto.de
 [see web page to check digital email signature]

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://radiance-online.org/pipermail/radiance-general/attachments/20050911/467c62ea/attachment.htm