--- ray/doc/notes/ReleaseNotes	2016/01/29 16:21:55	1.273
+++ ray/doc/notes/ReleaseNotes	2021/12/04 16:29:29	1.398
@@ -2058,7 +2058,7 @@ Fixed early exit of rcontrib with internal error (samp
 discovered by Rob G.
 
 =================================
-Version 5.1
+Version 5.1 (Released 8/17/2017)
 
 Fixed issue with vchars.mta and symbols.mta not being installed in
 the appropriate metafile library location, with help from Rob G.
@@ -2070,3 +2070,427 @@ Added ability to digest color (CIE-XYZ) input in pabop
 This change alters the .sir (interpolant) format, so these files
 will need to be regenerated.  Use the new "#colorimetry: CIE-XYZ" tag
 to enable on input.
+
+Rewrote bsdf2klems to support tristimulus color and call wrapBSDF.
+Did the same for bsdf2ttree.  Function specifications still gray only.
+
+Reduced a problem where rvu or rholo would be slow to quit when
+run with many processes on a large model or one with a large
+ambient file (taking lots of memory).  We were waiting for
+processes to quit one-by-one, which takes some time for each
+process as it disentangles its shared memory.  We now
+wait for all the processes together, so they can disentangle
+in parallel.  It can still take up to a minute in some cases,
+but not 10 minutes like it was taking before the fix.
+
+Enabled ambient cache value corral for all levels, not just final two.
+This may reduce errors in certain pathological scenes.
+
+Added -in option to rcalc, rlam, and total to limit the number of
+input records in hopes of circumventing Windows EOF issues on binary data.
+Also added -on option to rcalc and total for logical symmetry.
+
+Increased modifier limit to 10,000 in rcontrib.
+
+Fixed issues with glrad and added zoom response to mouse wheel input.
+
+Added missing definition of "ashik2" material type to reference manual.
+
+Changed genBSDF to calculate device thickness assuming Zmax==0 so
+pkgBSDF will generate correct proxy geometry.
+
+Created pabopto2xyz tool to combine different sensor measurements
+into a single CIE-XYZ measurement file for color BSDF interpolation.
+
+Jan Wienold updated evalglare to version 1.30
+
+Roland Schregle added out-of-core photon mapping routines, which
+may be compiled using the PMAP_OOC macro.
+
+Added -a option to getinfo for easy addition of header lines.
+
+Fixed issue with multiple records on picture output in rcontrib (and rfluxmtx).
+Credit to Sarith Subramaniam for pointing out the problem.
+
+Made mkillum more tolerant of non-planar polygons in its sampling routines.
+
+Added "-recover" option to genBSDF to pick up where it left off.
+
+Fixed genBSDF issue that was created incorrect normalization of
+Klems output.  Thanks to David G-M and Christian Knoflach for
+discovering the error.  Also improved sampling of indirect irradiance
+that showed some slight bias due to poor sample collision avoidance.
+
+Fixed long-standing issue with -aa 0 neglecting textures in calculating
+surface shading.
+
+Fixed bug in ranimove introduced during photon map integration.
+
+Added check for steep sections of BSDF in pabopto2bsdf, creating narrower
+peakes in those places and mitigating some normalization issues.
+
+Added view360stereo.cal by Mark Stock to handle 360-degree over/under
+stereo views.
+
+Added MAX2SHADE macro to place limit on number of light sources
+to precheck for obstructions.  (Currently set to 200.)  This
+avoids excessive start-up times for scenes with many light sources.
+Thanks to Axel Jacobs for pointing out the problem.
+
+Added -v option to obj2mesh to print out mesh statistics.
+
+Rewrote bsdf2rad from ground up and made it a standard part of build.
+
+Fixed bug in orientation of Reinhart/Tregenza sample generator in rfluxmtx.
+
+Added ability to perform shadow-testing through unproxied BSDFs with strong
+through (direct-direct) component.
+
+Added {+|-}a options to genBSDF and bsdf2ttree for switching off reciprocity
+averaging.
+
+Added bias to see through BTDFs with a strong "through" component
+rather than scattering view rays.  This is in lieu of peak extraction,
+which looked to be too difficult from an implementation standpoint.
+
+Added bsdfview tool analogous to objview but for BSDF files (XML and SIR),
+which calls bsdf2rad then rad with a few useful default views.
+
+=========================
+Version 5.2 (Release October 8, 2018)
+
+Created psketch program to apply non-photorealistic rendering to selected
+objects, identified by modifier name(s).
+
+Added '*' and '/' operations to rmtxop for element-wise multiplication and
+division of same-sized matrices.
+
+Changed falsecolor so "-s" option is adjusted to match legend if one.
+
+Falsecolor legend changes and improvements by David G-M and Stephen W.
+
+Fixed bug in peak extraction code that generated
+"Bad call to SDqueryTreProjSA" error, reported by Lars Grobe.
+
+Increased maximum luminance to 10^7 in pcond to fix issue noticed
+by Axel Jacobs where veiling glare not added for the sun under
+some conditions.
+
+Increased maximum number of open files in rcontrib to the system upper
+limit of 2048 rather than the default 512 under Windows.
+
+Added top-level resources directory with files from Rob Guglielmetti.
+
+Improved accuracy for peaky BSDFs and small light sources.
+
+Fixed issue with pmapdump noted by Axel Jacobs.
+
+Improved accuracy of bsdf2ttree when converting scattering interpolation
+representatons with strong peaks.
+
+Tweaked peak-extraction code to work more reliably.
+
+Georg Mischler updated SCons build system.
+
+Rob Guglielmetti and Alex Swindler updated CMake build system.
+
+Modified pcomb so the first view seen in the inputs is written into header.
+
+Wrote rtpict script to run rtrace in parallel mode and simulate rpict with
+-n option.
+
+Added ability to move forwards/backwards specified distance in rvu "origin"
+command.  Feature requested by John Mardaljevic, who essentially owns this
+command, now.
+
+Reduced bias in ambient super-sampling method.
+
+Fixed bugs associated with zero transmission in chromatic tensor trees.
+Thanks to Lars Grobe for pointing out issues.
+
+Added RAYPATH searching for -f options to genrev, gensurf, genworm,
+bsdf2klems, bsdf2ttree, and pcomb.  This makes behavior consistent
+with rcalc, calc, and rendering tools.
+
+Worked with Randolph Fritz to fix some issues with ies2rad and newer
+IESNA63 standards.
+
+Resurrected original bsdf2rad test program, renaming it bsdf2rado. Still an
+optional build in src/cv/.
+
+Fixed bug in rcontrib with virtual sources, where pretest was adding
+to modifier accumulation buffers.  Reported by Mostapha.
+
+Added an "aBSDF" material type that explicitly checks for peak extraction.
+Its arguments are the same as the original "BSDF" type, but without the
+thickness parameter, as proxy geometry is not supported by this new type.
+The old "BSDF" type no longer does peak extraction under any circumstances,
+since this behavior is superfluous for proxied geometry.
+
+Added -r option to bsdf2rad (and bsdfview) to fix BSDF min and max range
+for comparing BSDFs.
+
+Fixed issue with pfilt -m option and uneven image size changes pointed
+out by Rob Shakespeare.
+
+Updates to evalglare and gendaylit from Jan Wienold, including new
+colored sky option.
+
+=========================
+Version 5.3 (Released September 1, 2020)
+
+Created radcompare program to compare Radiance tool output to reference outputs
+during unit-testing.
+
+Fixed issue reading very large files (> 2 GBytes) in rcollate.
+
+Made picking function in rvu ignore transparent and void surfaces.
+
+Added -orRxX options to rtrace to enable output of mirrored and unmirrored
+contributions and distances to enable new types of reprojections.  This also
+cleaned up and unified handling of effective ray lengths throughout code.
+
+Bug fixes from Jan Wienold for evalglare.
+
+Changed indexed (numbered) output from dctimestep to start from 0 rather
+than 1, to make it more consistent with expected input naming.
+
+Increased default sampling in bsdf2klems (-n option) to 1024.
+
+Fixed an issue noticed by David G-M that was causing genBSDF to bias
+isotropic tensor tree samples resulting in subsequent failure of
+peak extraction with "aBSDF" type.
+
+Introduced max() and min() functions to .cal library.
+
+Added random incident patch sampling to bsdf2klems to improve accuracy
+with scattering interpolation representation input (e.g., from
+pabopto2bsdf).  The other input types already sampled incident patch area.
+
+Added gendaymtx -A option to compute average sky over all the input
+time steps.
+
+Added ability to output color image from rtrace with any of -ovrx options.
+Also fixed inconsistency of -oN output when ray casting versus full trace.
+Before, it would flip normal to front side unless ray casting.  Now, -oN
+always reports the unflipped normal.
+
+Changed rlam behavior so -iaN reads N lines from a stream rather than N
+characters.  This seems more useful, since -ibN is somewhat redundant with
+the old behavior.
+
+Created rsplit program as compliment to rlam and to handle more complex
+rtrace output into separate files or streams.
+
+Improved getinfo so that it can also hop over resolution string with -c
+and header elide mode.  Use +d to include resolution string in output,
+and -d to elide with - and -c options.
+
+Made pvalue -u option mean "uncompressed" with -r.  Also, allow skipping
+bytes on standard input.
+
+Created rcode_depth tool to convert to and from a 16-bit/pixel portable
+depth image encoding that uses a mix of linear and reciprocal distance
+with explicit representations of 0 and infinity.  Code for loading these
+depth maps is in src/common/depthcodec.{h,c}.
+
+Created rcode_norm tool to convert to and from 32-bit/pixel portable
+surface normal representation, matching the one in src/common/dircode.c.
+Code for loading normal maps is in src/common/normcodec.{h,c}.
+
+Created rcode_ident tool to create indexed identifier files.
+Code for loading these files is in src/common/idmap.{h,c}.
+
+Added -o* option to rtpict to produce layered images that handle most
+output types from rtrace (colors, surface normals, distances, IDs).
+
+Created rcode2bmp script to create images from new rcode_* output maps.
+
+Improved efficiency of rmtxop in cases where matrix multiplication is
+faster evaluated right-to-left, as when the final component is a
+column vector.  Also added ability to apply trailing unary operators
+(-t, and -s or -c) to resulting matrix.
+
+Added new depth, normal, and ID file types to radcompare.
+
+Added BigEndian= header line for automatic byte-swapping in rmtxop
+and radcompare.
+
+Added new "turbo" palette to falsecolor (thanks to Taoning).
+
+Nathaniel Jones created the "dcglare" program for annual glare
+simulations when the sun and sky are visible through the window.
+Tutorial is posted at
+"https://github.com/nljones/Accelerad/wiki/The-Imageless-Method-for-Spatial-and-Annual-Glare-Analysis".
+
+Added support for loading RGBE images into matrices in dctimestep.
+
+Added -y option to gensky and gendaylit to support more accurate Michalsky
+solar position calculation.
+
+Changed rsplit, so specifying a space as the separator parses words
+between whitespace.
+
+Added ability of vwrays, pinterp, and pmblur2 to read encoded depth files.
+
+Added -o RxCxR1xC1 facility to perform blocking in rcollate.
+
+Made it so number of segments in genworm, genrev, and gensurf can be
+given as expressions of previously defined variables, etc.
+
+Added -D and -n option to gendaymtx to output sun description and skip
+matrix output on request.  A -M option was also added to output solar
+modifiers for rcontrib.  Support also added for leap days in WEA input.
+Enhancements sponsored by Ladybug Tools.
+
+Updated eplus_uvf to work with IDF version 9.x, checking version number.
+
+Fix to glarendx issues with dgi calculation.
+
+Made it so rmtxop and dctimestep undo any exposure applied to Radiance
+pictures they may load.
+
+Added robjutil utility to manipulate Wavefront .OBJ files.
+
+Fixed a crash related to rtrace -om or -oM with mesh instances containing
+materials.
+
+Improved rtrace flushing function to handle any value for -x without
+deadlock on controlling process.
+
+Eliminated unnecessary start-up procedures for rtrace when only ray-casting
+is needed (i.e., no tracing or value output).
+
+Improved isotropic sampling method in bsdf2ttree using "zipper" approach
+applied in genBSDF.
+
+Added limit to memory used by high-resolution tensor tree BSDFs for
+cumulative cache during MC sampling (rendering).  Limit set to 250MBytes
+per BSDF for standard architectures.
+
+Added -n and -s options to bsdf2ttree to provide some user control
+over super-sampling.
+
+Improved surface normal codec used by ambient file to store orthonormal
+directions exactly.
+
+Improved peak extraction method, which was undervaluing transmission
+in many cases.
+
+New version of evalglare from Jan Wienold.
+
+Added -u option to gendaymtx to elide data columns (time steps) when
+the sun is not visible above the horizon.
+
+=========================
+Version 5.4
+
+Added rcollate -c option to force operation to continue even if it
+seems unnecessary.  Made tool more robust.
+
+Added -t option to rcontrib (and rfluxmtx) to report progress after the
+given number of seconds.
+
+Added "test" target to makeall to run unit tests.
+
+Modified bsdf2ttree to allow different parameters per input SIR file.
+
+Fixed issues with very large matrix files (> 2GB) in dctimestep, rmtxop,
+and rcollate.
+
+Added -rf and -rb options to rmtxop to load front or back reflectance
+matrix data from XML file rather than transmission.  (Lars Grobe suggested.)
+
+Improved behavior for coincident surfaces with different materials.
+Rather than returning one surface or the other based on floating-point
+vagaries, opaque materials are now preferred over transparent ones.
+If the materials are both transparent or both opaque, then whichever
+surface appeared later in the scene description will be selected.
+This came of a request for consistent behavior with legal CAD output
+from Peter Apian-Bennewitz.
+
+Added explicit tests for on-edge intersections with polygons.  This should
+eliminate stray rays and tiny light leaks that often appear even in
+well-constructed geometry.
+
+Added "-s up" option to pabopto2bsdf to allow for 180 degree symmetry,
+meaning that data repeats after 180 degree rotation (as distinguished
+from "-s bilateral" whose data is mirrored about a vertical plane).
+Thanks to Helen Rose Wilson for the suggestion and informing me that
+such a thing exists.
+
+Removed upper limit on number of modifiers in rcontrib.
+
+Put cap on maximum ambient gradient to avoid over-extrapolation of
+poorly computed irradiance cache values.
+
+Added robjutil +T option to triangulate all faces in .OBJ description.
+
+Altered BSDF library to have both front and back diffuse transmittance
+values, which can be different.  This was necessary to avoid issues
+caused by measurement and modeling discrepancies.
+
+Added pabopto2bsdf -g option to cull near-grazing scattering to avoid
+issues with noise and negative values.
+
+Increased default dynamic range from 32 to 100 in pcond and normtiff.
+
+Added -o and -n options to genbox to create .OBJ output and segment
+edges and corners for transparent materials.  Also added -s option
+to control smoothing for .OBJ and Radiance with segmented rounding.
+
+Made it so robjutil with the +r option for Radiance output now
+produces normal smoothing unless normals are removed with -n.
+
+Added exposure preservation to pcompos when possible (i.e., when
+there is only one input or inputs agree on exposure).
+
+Added -DSHARP_RGB compile-time option to use sharp RGB color space
+rather than historic default.  This can make for more accurate color
+calculations, but requires that materials be recalibrated.  The
+PRIMARIES= line is also produced on output from rpict or rvu with
+the correct set of primaries based on how the system was compiled.
+
+Added automatic flushing to the output of total, except when the -r
+(running subtotals) option is used.
+
+Improved behavior of little-used -aw option, so that it avoids sea-level
+rise caused by adding its own estimate back into the ambient level remainder.
+It ended up being simple enough to derate the final average by the factor
+corresponding to average surface absorpition, thus removing the average
+from the final bounce estimate.
+
+Introduced ies2rad upgrades by Randolph Fritz.  Randolph says it
+fixes the following problems:
+
+	1. Generates correct Radiance geometry for spheres in 1995, 2002,
+	and 2019 IES files.
+
+	2. Generates correct Radiance geometry for vertical cylinders that
+	are taller than they are wide.
+
+	3. Ignores the file source ("File Generation Type") field in the
+	2019 version of the file, which would otherwise be incorrectly used
+	as an output multiplier.
+
+    It also offers the following enhancements:
+
+	1. Shape information is included in the .rad file comments.
+
+	2. 1995, 2002, and 2019-version luminous opening geometry is
+	accounted for.
+
+	3. The 2002 and 2019 IES file versions are recognized and
+	processed appropriately.
+
+	4. I have made an attempt to do something intelligent with less
+	common "luminous opening" shapes in the 1995, 2002, and 2019 versions
+	of the standard, though I have not implemented any support for new
+	geometry; approximations are substituted and warning messages issued.
+	Unfortunately, this code is not tested, since I don't have any IES
+	files that use these shapes.
+
+Added support for depth-of-field blurring in vwrays using -pd option.
+Also, made -c count work with vwrays -i, producing multiple samples
+per indicated pixel position.  At the same time, support for the -pd
+option was added to rtpict, which calls on vwrays for its samples.