readme

This a framework to test new ideas in transmission technology. Actual development is a LDPC-coder in

convolution
===========


This module does a convolutional encoding/decoding.
The code can be chosen according to the following parameters
n Number of transmitted bits per k data bits
k Number of data bits per n transmitted bits
m Number of memory registers
polys Array of polynome definition. The number of polynomes must be equal to n

Any values of n and k are legal, but a useful code can only be achieved
if k>=n and n<=m

The transmitted code blocklength is >= n/k times the data blocklength. Thus
a large n/k reduces capacity but improves error resistance due to higher redundancy.

For implementation reasons, m is limited to 32 (number of bits in an int value).
Still, for memory and performance reasons only m << 32 are possible


convolution_rcv
---------------

For description also refer to convolution.c

This module is the receiver part (decoder) of the convolution code.

Besides the configuration parameters described in convolution.c, the
following decoder specific parameters must be set:

trunclength 	Indicates the length of the trellis which will never be exceeded.
  This Parameter must always be > 2, large values produce better error correction
  but reduce decoder speed performance and increase memory consumption

Inputs:
	U8

outputs:
	U8

Config:
	*polys<REG_STATUS>: 0x37,0x73
		array of polynome-definitions.
		The array has length n, and holds one REG_STATUS value for each polynome.
		The bits of a value define if this term should be added to the result or not.
		Of each value, only the m LSB are relevant. All others should be set to 0.
		Be aware that the LSB (bit 0) represents the term d^m,
		while the bit m represent the term d^0.
		E.g. A polynome which calculates the result as the latest, 4th latest and the 5th latest bits
		p=1+d^3+d^4 has value=0b00000000 00000000 00000000 00011001 = 0x000000019

	k<int>: 1
		The number of input bits (number of input bits to generate n output bits)

	m<int>: 8
		Number of memory registers used
		(limited to 32 because of use of unsigned int to define a state)

	n<int>: 2
		The number of output bits (output bits per input bit)

	truncblock<int>: 50
		indicates how many trellis-steps should be truncated in one truncate operation

	trunclength<int>: 100
		length of state-history of trellis. The trellis is truncated whenever it exceeds
		this size


Stats:
	time_calcsucc<double>:
		Stores the total time spent in calculating the trellis

	time_truncate<double>:
		Stores the total time spent in truncating the code

	total_time_decode<double>:
		Stores the total time spent in the decoding


convolution_send
----------------

For description also refer to convolution.c

This module is the sender part (encoder) of the convolution code.
Data blocks are coded independently, i.e. the registers are flushed after each block.

Besides the configuration parameters described in convolution.c, the
following decoder specific parameters must be set:
NO SPECIFIC CONFIG

Inputs:
	U8

outputs:
	U8

Config:
	*polys<REG_STATUS>: 0x37, 0x73
		

	k<int>: 1
		The number of input bits (number of input bits to generate n output bits)

	m<int>: 8
		Number of memory registers used
		(limited to 32 because of use of unsigned int to define a state)

	n<int>: 2
		The number of output bits (output bits per input bit)