HCompress Class Reference

#include <Compression.h>

Detailed Description

The H-compress algorithm.

This algorithm relies on some scaling parameter. When scaling is enabled, H-compress is lossy irrespective of quantization. In order to use H-compress losslessly, quantization and scaling must be zeroed, and the data values must be integers.

Analogously to quantization, the scaling can be set globally or tile-wise, which is recommended. In the latter case, it is generally provided as a multiplicative factor of the noise RMS:

HCompress algo;
algo.quantization(Tile::rms / 4);
algo.scaling(Tile::rms * 2.5);

Public Member Functions
	HCompress (Position<-1 > tiling=Tile::adaptive(), Quantization quantization=Quantization(), Scaling scaling=Scaling(0))
	Constructor.
bool	is_lossless () const override
	Check whether compression is lossless. More...
const Quantization &	quantization () const
	Get the quantization.
HCompress &	quantization (Quantization quantization) override
	Set the quantization. More...
const Scaling &	scaling () const
	Get the scaling parameter.
HCompress &	scaling (Scaling scale)
	Set the scaling parameter.
Public Member Functions inherited from AlgoMixin< HCompress >
const Quantization &	quantization () const
	Get the quantization.
virtual HCompress &	quantization (Quantization quantization)
	Set the quantization. More...
const Position<-1 > &	tiling () const
	Get the tiling.
virtual HCompress &	tiling (Position<-1 > shape)
	Set the tiling.
Public Member Functions inherited from Compression
virtual bool	is_lossless () const
	Check whether the compression is lossless for. More...
const Quantization &	quantization () const
	Get the quantization.
const Position<-1 > &	tiling () const
	Get the tiling.

---

The documentation for this class was generated from the following file:

• EleFitsData/EleFitsData/Compression.h