Correction to Window 6 notes
Modified rtcontrib -c option to accumulate a specified number of rays
Added planisphere view type (-vts option) as requested by Axel Jacobs
Initial check-in of untested rsensor
Fixed bug in rotation computation when X' is aligned with Z-axis
Fixed reversed orientation for anisotropic distrubutions
Superficially working version of mkillum with BSDF input
Added -c option to rtcontrib (barely tested, and -r does not work)
Allow -P and -PP options to be set by user for rtcontrib
Corrected pixel alignment for downsampling & upsampling in pcomb
Partway addition of BSDF data in mkillum
Fixed bug in shadow cache related to antimatter holes
Decreased default -n setting in obj2mesh from 15 to 9
Corrected ray weighting for mixed materials
Fixed original octree bug that made it take 8 times as much memory as it should
Added checks for source surfaces, including zero-area polygons
Added byte-swapping i/o options to rcalc
Added -spec and "-s auto" options to falsecolor
Added -h option to pcompos and pcomb to avoid uncontrolled header growth
Fixed bug in virtual sources with intervening transmitters
Added -o option to ra_xyze for original pixel values.
Changed default turbidity from 2.75 to 2.45
Fixed bug in aliases of aliases
Fixed long-standing inaccuracy of +/-t threshold options
Fixed visual choosing problem for some SGI machines
Created optics2rad script for converting Optics materials to Radiance
Minor fixes to last change with IEEE float lookup function
Made missing calls to SET_FILE_BINARY() for Windows
Added credit for -u+ bug fix
Changed default in rtrace to -u+ (random sampling)
Made -u+ option truly random
Renamed librt.a to librtrad.a to avoid conflict on Solaris.
Fixed old bug in ranimate where NEXTANIM variable was not working
Added -V option to rtcontrib and -oV option to rtrace to report contributions
Fixed byte ordering for 16-bit/channel PPM files
Fixed problems with genblinds
Made scene i/o work for identifiers with spaces and quotes
Removed NICE macro settings in rpict and rtrace to improve performance
Added threshold to disable Fresnel approximation for specularities below 1.8%
Improved default ordinal color scale in falsecolor
Fixed rtcontrib flushing behavior
Added rtcontrib -fo option to override new reluctance to clobber existing files
Extensive changes to support rtcontrib -bn option and fix -r problems
Added -i option to allow rlam to work with binary input records
Rearranged some struct's for better efficiency on 64-bit architectures
Added -Dfseeko=fseek for Linux machines, as library call is currently broken
Fixed serious bug where rtcontrib would sometimes forget to write last file
Added -pj option to vwrays for ray jittering
Eliminated SPEED macro from makeall and source tree
Created ran2tiff script for batch animation exposure and conversion
Added preloading of mesh objects for parallel rendering
Added -e [default] and -o options for even/odd fields first
Created fieldcomb script to combine alternate fields in walk-through animations
Fixed another serious bug where raytrans() wasn't working with RR ray term.
Fixed bug in ambient calculation causing zero results for some scenes
Eliminated writing of EXPOSURE= when STONITS not set in input TIFF
Changes for 3.7.1 patch release
Changed to 98th percentile for default threshold
Fixed serious bug in new Russian roulette algorithm
Onward to 3.8a
Documented new behavior with blank input lines in total(1)
Added -z option to normtiff to output LZW-compressed TIFFs
Fixed ancient bug in cone bounding box calculation -- caused visibility errors
Changed wording in -defaults output (again)
Fixed bug in rcalc with -l option where it sometimes hung
Changed rlam -t option to accept strings as well as tab characters
Added -R option for pure Monte Carlo rendering, replacing -DMC compile option
Implementaed rtcontrib -r (recover) option, added MODIFIER= + BIN= to header
Changed rtrace to default to Russian roulette ray termination
Added binary i/o options to total and piped output to rtcontrib
Added RAYPATH directory searching to rtcontrib -f option
Added Russian roulette ray termination and fixed ambient weights & measures
Got rtcontrib working and wrote basic man page
Created rtcontrib program for computing ray contributions and coefficients
Added note on change to blackbody.cal
Created meta2bmp file to generate BMP output from metafile graphics
Added -T option to xshowtrace to trace rays to light sources
Extensive changes to enable rtrace -oTW option for tracking ray contributions
Added -oM option to rtrace to output material
Changed minimum number of ambient rays
Created mksource program
Fixed minor problems in 16-bit TIFF conversion
Made illum work properly with distant source objects
Switched from using "rsh" to "ssh" in ranimate
Fixed bug in rvu "move" command from last change and added "focus" command
Created pmdblur script and added DBLUR variable to ranimate
Added -pd option for depth-of-field sampling and -vd focal distance entry
Modernized tone-mapping routines with structure pointer r.t. stack
Fixed bug in calculation of diffuse transmission on polished "trans" surfaces
Fixed bug that caused inappropriate fatal error mixing meshes and instances
Added simple script to convert to/from Poskanzer Float Map image format
Fixed bug in acuity option for small view angles by skipping calculation
Removed unused programs and files from distribution (sources to CVS attic)
Changed /usr/tmp default temp file location to /tmp under Unix
Added comments on 3.6 bug fixes
Fixed bug in format parsing where extra EOL was sometimes added
Fixed minor spelling errors and typos
Altered ra_bmp so it doesn't reverse scanline order on grayscale output
Fixed bug where pvalue -o would report watts/sr/m^2 from XYZE input
Enabled -p option (display primaries) and XYZE reading to ra_bmp
Corrected Gaussian reflectance model normalization (cosine factor)
Added close-on-exec flag to pipes to prevent possible deadlocks
Fixed deadlock upon termination of mkillum with -n option
Added -n option to mkillum for multiprocessing on shared memory machine
Made it so -m option is ignored in gensky if time zone is given
Fixed bug spotted where spheres with sources inside were blocking themselves
Made it so rad accepts 'scene' variable with commands only
Added notes about recent fixes
Added -p option to rcalc for new passive mode
Fixed bug in pixel ordering out of protate
Added tone-mapping to ra_bmp and fixed but in RLE8 BMP encoding
Increased .OBJ face vertex limit to 512 (was 64)
Added -l option to search RAYPATH library locations for material file
Made oconv more robust to set overflow
Increased -aa default settings
Added support to ra_bmp for 16-bit images and RLE, plus created man page
Created ra_bmp program to convert between Radiance pictures and Windows BMP
Made it so pcompos does not cut off input command streams abruptly
Put in Rob Guglielmetti's falsecolor -lw and -lh options
Allow intervening modifiers and aliases in virtual source materials
Returned to 32-bit (x2) uv coordinates, since 16-bits was sometimes inadequate
Added -I option to replmarks to replace markers with mesh instances
Renamed rview, lam, calc, and neat to rvu, rlam, icalc, and neaten
Created a source occluder cache to accelerate shadow testing.
Added -p option to histo to report percentiles and changed output slightly
Fixed bug in glarendx pointed out by Phillip Greenup
Replaced all calls to vfork() with regular fork() calls
Added an in(n) column function to rcalc
Fixed old problem with drawsources and illum's
Fixed bug in otndx() that caused core dump with more than 2^16 modifiers
Increased default ambient options in rpict, rtrace, rview, rad & mkillum
Bug fix where ranimove was restarting processes unnecessarily
Improved rholo -o ogl interaction under Mac OS X
Made integer tone-mapping more robust for images with constant regions
Reduced texture coordinate size to 16-bits/sample
Changed meta2tga so it outputs correct colors
Added -dF and -dD options for byte-swapped float's and double's
Fixed bug in xform where it wasn't transforming mesh primitives
Allow source modifier to appear earlier in the chain
Added tone-mapping support for 16-bit/sample int and IEEE float TIFF images
Fixed serious bug in ranimate where mesh instances were not being freed
Added -N option to rad for parallel rendering (preliminary w/o using -PP)
Added -f and -w options to ra_tiff to produce float and 16-bit output
Fixed wrapping long in source hit testing code
Added mention of latest changes
Final updates for official 3.5 release
Removed duplicate copies of file and fixed comment in release notes
Added and updated documentation for 3.5 release
This form allows you to request diffs between any two revisions of this file. For each of the two "sides" of the diff, select a symbolic revision name using the selection box, or choose 'Use Text Field' and enter a numeric revision.