/*-----------------------------------------------------

/* SPU HLSW V6.0 - WHAT'S NEW !!

/*

/* Project: 		HSO/PACS, SPU HLSW

/* Originator: 		Ahmed Nabil Belbachir

/* Contact:		spuhlsw@prip.tuwien.ac.at 

/* User Manual issue:   6.1 

/* Date:		16-09-2003

/*-----------------------------------------------------



SPU HLSW 6.0 is positioned as an upgrade of SPU HLSW 5.8, containing the following 

new features:

 

  - Chance of Higher Compression Ratio using a bigger Range for RZIP

  - CPU Workload of Lossless Compression Improved (decreased)

  - Better Compression Ratio for Raw Channel data

  - New Fitting Algorithm in Case of 2 samples per sub-ramp





This will be explained in the specific sections that follow.



The required Decompression SW version is:

- Paladin6 (Standalone Decompression Software in ANSI C)

- Decompress6 (Ground Decompression Software in Java) 



The SPU HLSW Development Team.





Chance of Higher Compression Ratio using a bigger Range for RZIP

----------------------------------------------------------------



The Coding range of the RZip algorithm has been modified from 3 to 7 in order to have higher compression rate. RZIP algorithm encodes blocks of the data. Choosing the number 7, means that RZIP splits the data in blocks of 2power7 words (128 words) each. Bigger block size increases the chance to have higher compression ratio whenever redundancy found.   



CPU Workload of Lossless Compression Improved (decreased)

---------------------------------------------------------

The CPU workload of the Lossless Compression tasks has been decreased by:

1.) Decreasing the Iterations of RZIP algorithm from 3 times over the buffer to 1 time. The bigger range chosen (7) allows a higher compression ratio, therefore, there is no need for iterating the run of the algorithm. 

2.) Subtitute several functions by MACROs to shorten the execution time

3.) Two lookup tables have been added to speed up the bit coding. Setting a bit to one or to 0 has been performed in the previous SW version through a function. In this SW version this bit coding is performed via Look up Tables and MACROS.



Better Compression Ratio for Raw Channel data

---------------------------------------------

Raw channel data is compressed using 3 runs of RZIP algorithm. Ranges 7, 6 and 4 for respectively the 1st , the 2nd and the 3rd run of RZIP shows better compression ratio (based on the 6 Pack Data) 



New Fitting Algorithm in Case of 2 samples per sub-ramp:

--------------------------------------------------------

In case of 2 samples per sub-ramp, the sub-slope is calculated using the difference between two successive samples i.e. 32 sub-slopes are calculated out of a 64 samples ramp.



/* --- EOF ---