New rtcontrib program - mkillum conflict

Hi,

After installing mksource and rcontrib, mkillum does not finish any more. There seems to be a conflict. Mkillum without -ab 1 works fine, but anything else is now disfunctional, meaning, it just runs without end. Can others please check if they have similar experiences?

Thanks

Martin Moeck

···

-----Original Message-----
From: [email protected] on behalf of Greg Ward
Sent: Thu 5/26/2005 3:17 PM
To: Radiance general discussion
Cc:
Subject: [Radiance-general] New rtcontrib program

  Hi All,
  
  Somewhat related to the light pipe discussion, I have just completed
  work on a new utility called "rtcontrib," for computing arbitrary
  light contributions. This is a culmination of earlier work I did on
  rtrace to allow for the correct reporting of ray weights in trace
  output (the new -oTW option), which I mentioned on radiance-dev a
  little over a month ago. The benefits of this facility will not
  really be felt or known until a third (C-shell script) layer is added
  for calculating daylight coefficients and the like. Right now, the
  simplest thing you can do with rtcontrib is arrive at a set of images
  corresponding to the different light sources in your scene, such that
  you can recombine them with dimming multipliers to simulate lighting
  controls. An example of this is given in the man page, attached.
  
  Rtcontrib can also be used by the more adventurous among you to
  compute input/output relations for devices such as light pipes and
  shading systems, although I have yet to test such an approach,
  myself. I expect we'll be learning more about what one can do with
  this in the coming months, and I hope to work with Christoph Reinhart
  and John Mardaljevic on standardizing daylight coefficient
  calculations for starters.
  
  -Greg
  
  ---------
  
  RTCONTRIB(1)
  RTCONTRIB(1)
  
  NAME
          rtcontrib - compute contributions in a RADIANCE scene
  
  SYNOPSIS
          rtcontrib [ -n nprocs ][ -e expr ][ -f source ][ -o fspec ]
  [ -b binv ]
          -m mod .. [ $EVAR ] [ @file ] [ rtrace options ] octree
  
  DESCRIPTION
          Rtcontrib computes ray contributions (i.e., color
  coefficients) for
          objects whose modifiers are named in one or more -m
  settings. These
          modifiers are usually materials associated with light
  sources or sky
          domes, and must directly modify some geometric primitives to
  be consid-
          ered in the output. The output of rtcontrib has many
  potential uses.
          Source contributions can be used as components in linear
  combination to
          reproduce any desired variation, e.g., simulating lighting
  controls or
          changing sky conditions via daylight coefficients. More
  generally,
          rtcontrib can be used to compute arbitrary input-output
  relationships
          in optical systems, such as luminaires, light pipes,
  and shading
          devices.
  
          Rtcontrib calls rtrace(1) to calculate the contributions for
  each input
          ray, and the output tallies are sent to one or more files
  according to
          the -o specification. If an output specification contains a
  "%s" for-
          mat, this will be replaced by the modifier name. The -b
  option may be
          used to further define a "bin number" within each object if
  finer reso-
          lution is needed, and this will be applied to a "%d" format
  in the out-
          put file specification if present. The actual bin number
  is computed
          at run time based on ray direction and surface
  intersection, as
          described below. The most recent -b and -o options to the
  left of each
          -m setting affect only that modifier. (The ordering of
  other options
          is unimportant.)
  
          If a -b expression is defined for a particular modifier, the
  bin number
          will be evaluated at run time for each ray contribution
  from rtrace.
          Specifically, each ray's world intersection point will be
  assigned to
          the variables Px, Py, and Pz, and the normalized ray
  direction will be
          assigned to Dx, Dy, and Dz. These parameters may be combined
  with def-
          initions given in -e arguments and files read using the -f
  option. The
          computed bin value will be rounded to the nearest whole
  number. This
          mechanism allows the user to define precise regions or
  directions they
          wish to accumulate, such as the Tregenza sky
  discretization, which
          would be otherwise impossible to specify as a set of
  RADIANCE primi-
          tives. The rules and predefined functions available for
  these expres-
          sions are described in the rcalc(1) man page.
  
          If no -o specification is given, results are written on
  the standard
          output in order of modifier (as given on the command line)
  then bin
          number. Concatenated data is also sent to a lone output file
  (i.e., an
          initial -o specification without formatting strings). If a "%
  s" format
          appears but no "%d" in the -o specification, then each
  modifier will
          have its own output file, with multiple values per record in
  the case
          of a non-zero -b definition. If a "%d" format appears
  but no "%s",
          then each bin will get its own output file, with modifiers
  output in
          order in each record. For text output, each RGB coefficient
  triple is
          separated by a tab, with a newline at the end of each ray
  record. For
          binary output formats, there is no such delimiter to mark
  the end of
          each record.
  
          Input and output format defaults to plain text, where each
  ray's origin
          and direction (6 real values) are given on input, and one
  line is pro-
          duced per output file per ray. Alternative data
  representations may be
          specified by the -f[io] option, which is described in the
  rtrace man
          page along with the associated -x and -y resolution
  settings. In par-
          ticular, the color ('c') output data representation together
  with posi-
          tive dimensions for -x and -y will produce an
  uncompressed RADIANCE
          picture, suitable for manipulation with pcomb(1) and related
  tools.
  
          If the -n option is specified with a value greater than
  1, multiple
          rtrace processes will be used to accelerate computation on
  a shared
          memory machine. Note that there is no benefit to using more
  processes
          than there are local CPUs available to do the work, and the
  rtcontrib
          process itself may use a considerable amount of CPU time.
  
          Options may be given on the command line and/or read from
  the environ-
          ment and/or read from a file. A command argument beginning
  with a dol-
          lar sign ('$') is immediately replaced by the contents of
  the given
          environment variable. A command argument beginning with
  an at sign
          ('@') is immediately replaced by the contents of the given file.
  
  EXAMPLES
          To compute the proportional contributions from sources
  modified by
          "light1" vs. "light2" on a set of illuminance values:
  
            rtcontrib -I+ @render.opt -o c_%s.dat -m light1 -m light2 \
           scene.oct < test.dat
  
          To generate a pair of images corresponding to these two
  lights' contri-
          butions:
  
            vwrays -ff -x 1024 -y 1024 -vf best.vf | rtcontrib -ffc \
           `vwrays -d -x 1024 -y 1024 -vf best.vf` @render.opt \
            -o c_%s.pic -m light1 -m light2 scene.oct
  
          These images may then be recombined using the desired outputs
  of light1
          and light2:
  
            pcomb -c 100 90 75 c_light1.pic -c 50 55 57
  c_light2.pic \
           > combined.pic
  
          To compute an array of illuminance contributions according
  to a Tre-
          genza sky:
  
            rtcontrib -b tbin -o sky.dat -m skyglow -b 0 -o ground.dat \
            -m groundglow @render.opt -f tregenza.cal scene.oct < test.dat
  
  AUTHOR
          Greg Ward
  
  SEE ALSO
          cnt(1), getinfo(1), pcomb(1), pfilt(1), ra_rgbe(1), rcalc(1),
  rpict(1),
          rtrace(1), vwrays(1), ximage(1)
  
  RADIANCE 5/25/05
  RTCONTRIB(1)
  
  _______________________________________________
  Radiance-general mailing list
  [email protected]
  http://www.radiance-online.org/mailman/listinfo/radiance-general

Hi Martin,

Rtcontrib required some rather extensive changes to the diffuse interreflection methods, which definitely introduced one or two bugs that I have since fixed. You need to make sure you have the latest HEAD release, as I suspect you picked up a buggy version of src/rt/ambcomp.c or similar. If you recompile with last night's HEAD and still have problems, send me a list of the options you're using in mkillum and I'll see if I can figure it out. I did my own test runs this morning to check that it's behaving correctly, and I couldn't find anything wrong.

-Greg

···

From: "Martin Moeck" <[email protected]>
Date: June 5, 2005 8:08:24 AM PDT

Hi,

After installing mksource and rcontrib, mkillum does not finish any more. There seems to be a conflict. Mkillum without -ab 1 works fine, but anything else is now disfunctional, meaning, it just runs without end. Can others please check if they have similar experiences?

Thanks

Martin Moeck