Consider candidates that are tied according to distance with the current
worst predictor in order to consistently use the lowest indexed candidate
and to prefer neighbours in coarser LoD to neighbours in current LoD.

This fixes an issue when during predictor construction, when
lodIndex == levelOfDetailCount causing the retained set to be empty.

In order to simplify the introduction of quantisation using the
multiplicative inverse, the existing quantiser state is refactored
into a Quantizer class that represents the current step size.

The quantization functions are applied to both fixed-point and unitary
precision data using the same QP.  This commit moves the implicit
conversion adjustment out of the quantiser and into the calling code.

Toshiyasu Sugio authored Jul 09, 2019 and David Flynn committed Aug 12, 2019

This functionality is now handled by the unified generation process.

Toshiyasu Sugio authored Jul 08, 2019 and David Flynn committed Aug 12, 2019

This adoption permits the unification of the single and multi
level-of-detail generation processes by always solving ties in sorting
predictors with identical weights according to relative buffer position.

A 64 bit integer can represent a three dimensional 21 bit position with
one spare bit.  In order te allow the computation of abs(a-b), a signed
representation should be used, making use of the full 64 bits.

Toshiyasu Sugio authored Jul 02, 2019 and David Flynn committed Aug 12, 2019

This commit makes all lod layers order points in a consistent,
ascending, morton order.

Toshiyasu Sugio authored May 10, 2019 and David Flynn committed Aug 05, 2019

This adoption fixes an issue whereby intra lod prediction is not applied
to the top layer.  By fixing this, intra lod prediction will be used for
the single-layer case.

Toshiyasu Sugio authored Mar 01, 2019 and David Flynn committed Apr 16, 2019

This commit introduces prediction by previously coded points within
the same level of detail using the predicting attribute coder.
The feature is controlled using --intraLodPredictionEnabled=0|1.

This method has been replaced by an LoD decimation method.

Valery Valentin authored Mar 05, 2019 and David Flynn committed Apr 16, 2019

This LoD generation (subsampling) method uses the Morton-based ordering
of points that is used to determine prediction neighbours.  Each LoD
level comprises 3/4 of the points of the previous.

Toshiyasu Sugio authored Mar 01, 2019 and David Flynn committed Apr 16, 2019

This commit introduces zero run length coding for attribute residual
value along with an eq1 flag which indicates if residual value is equal
to 1 based on the current residual coding specification. In zero run
length coding, the number of zeros prior to each residual value is
counted as zerorun, and then zerorun is encoded instead of encoding
sequence of 0s

Noritaka Iguchi authored Mar 27, 2019 and David Flynn committed Apr 16, 2019

This commit introduces control of the quantisation step size using
the familiar HEVC | AVC quantisation parameter.

This function has never been used.

This is part of a series attempting to remove unhelpful typedefs.

The old typedefs (PCCBox3D) is retained for compatibility, but marked as
deprecated.

This is part of a series attempting to remove unhelpful typedefs.

The old typedefs (PCCVector3D, PCCPoint3D, PCCColor3B) are retained for
compatibility, but marked as deprecated.

Vladyslav Zakharchenko authored Dec 14, 2018 and David Flynn committed Feb 06, 2019

This uses a binary-tree to generate the levels-of-details for the
lifting/predicting transforms as an alternative to Euclidean distance
thresholding methods.

Jungsun Kim authored Jan 30, 2019 and David Flynn committed Feb 05, 2019

This simplified attribute prediction scheme avoids computing the
LoD structure in order to reduce the computational complexity of
both the encoder and the decoder.  It is activated by setting
levelOfDetailCount=0.

NB: this is only configured for lossless/near-lossless predictive
attribute coding of cat3 sequences.

Khaled Mammou authored Sep 22, 2018 and David Flynn committed Nov 01, 2018

Replace the k-d tree based nearest neighbour search with an approximate
search based upon the points in Morton order.

Toshiyasu Sugio authored Sep 20, 2018 and David Flynn committed Nov 01, 2018

This commit provides an ability to vary the number of direct neighbours
used in the predicting transform.  Mode 0 corresponds to the existing
averaging, and mode n selects the n-th nearest neighbour for direct
prediction.

The provided code has been reworked to:
 - rewrite unary coding to be a little more obvious
 - tidy the mode decision path
 - add configuration for max_num_direct_predictors (renamed from
   MaxNumPredCand)
 - now counts the number of single predictors, rather than prediction
   modes.

This commit removes unnecessary include directives.

This commit moves various constants from PCCTMC3Common.h to a new
constants.h.  Hard coded values of constants have been replaced with
their symbolic name.

The geometry coder is quite large, especially with trisoup and has no
benefit to being a header only implementation.  This commit moves the
geometry octree coder out of the header files and into
geometry_octree_{en,de}coder.cpp.

Khaled Mammou authored Aug 23, 2018 and David Flynn committed Aug 29, 2018

This functionality is unnecessary given a uniform quant_step size.

Khaled Mammou authored Aug 23, 2018 and David Flynn committed Aug 29, 2018

This commit removes PCCBuildLevelOfDetail2()

Khaled Mammou authored Aug 22, 2018 and David Flynn committed Aug 29, 2018

This commit updates the hard coded limit and default values, and
updates the master configuration template.

Khaled Mammou authored Aug 27, 2018 and David Flynn committed Aug 29, 2018

This commit fixes an error in the hls: dist2 is capable of being a
64-bit value and must not be signalled as 32-bit.

Khaled Mammou authored Aug 22, 2018 and David Flynn committed Aug 29, 2018

This commit modifies PCCBuildLevelOfDetail to use the direct dist2
value, rather than attempting to derive dist2.  This unifies the
behaviour of the two PCCBuildLevelOfDetail*() implementations.

This commit rewrites the codec high-level syntax:

 - the bitstream is divided into "bricks" (akin to an AVC/HEVC
   slice/tile).

 - sequence, geometry and attribute parameter sets describe the coding
   parameters in use generally and for a specific brick.

 - marshalling the bitstream payloads to a file format is achieved using
   a type-length-value encoding scheme.

Additionally, the triSoup bitstream scale and translation values have
been unified with (replaced by) the octree counterparts.  For
compatibility, existing command line parameters continue to function
as before.

NB: this commit does not incorporate flexibility in the decoding order.
The decoder requires the bitstream to be presented in a fixed order.

Khaled Mammou authored Jun 30, 2018 and David Flynn committed Aug 20, 2018

This commit implements a lifting scheme optimized for lossy attribute
compression, consisting of predict/update operators, and an adaptive
quantization strategy.

Khaled Mammou authored Jun 30, 2018 and David Flynn committed Aug 20, 2018

Adaptively pick the number of neighbours considered during the
prediction based on analysis of the reconstructed attribute values of
the neighbours.

From m42642, this is performed by computing the maximum difference
between any two reconstructed attribute values of a points neighbours.
If the variability is higher than a threshold, then apply a
rate-distortion optimization procedure to choose between using either
a single neighbour or all neighbours. The index of the best prediction
strategy is arithmetically encoded.

This commit:
 - implements the m42642 method
 - moves buildPredictors into transform-specific attribute encoding/decoding
 - adds quantizationStepsChroma
 - renames quantizationSteps -> quantizationStepsLuma

df: fixed incorrect application of quantizationStepsLuma|Chroma
df: updated cat3 cfg to specify quantizationStepsChroma
df: fixed header encoding if quantizationStepsChroma missing

In the contextualisation of occupancy coding using neighbour patterns,
this commit replaces the case when neighbour contextualisation is
disabled with a constraint that restricts neighbour contextualisation
only examine the direct siblings of a node.

The restriction ensures that at most three neighbours may be present
for a node.

The configuration option "neighbourContextualisation=1" is replaced with
"neighbourContextRestriction=0".

This commit provides an implementation of m42522 to code occupancy
using a binary context model rather than a 255-ary model.  A total of
581 binary contexts are used compared to the previous 10 255-ary
contexts.

The original integration of IDCM implemented only the parent based
eligibility check of m42239 and erroneously omitted the six-neighbour
check.

This commit adds the six-neighbour check, making use of the neighbour
calculation from m42239 neighbour contextualised coding of occupancy.