Howdy all!
In the next release of netcdf-c users will have access to the new quantize
feature. This allows users to specify a required number of significant
digits for a NC_FLOAT or NC_DOUBLE variable. Excess bits in each floating
point value are set to zero, and this makes compression work much better.
Today I gave a presentation of these new features (co-authored by Charlie
Zender), and we have an extended abstract that describes the features in
some detail.
The presentation:
https://www.researchgate.net/publication/360815823_EGU22-13259_Adding_Quantization_to_the_NetCDF_C_and_Fortran_Libraries_to_Enable_Lossy_Compression_1
The extended abstract:
https://www.researchgate.net/publication/360814804_Adding_Quantization_to_the_NetCDF_C_and_Fortran_Libraries_to_Enable_Lossy_Compression
Keep on NetCDFing!
Ed Hartnett
