rsdec (Communications Toolbox)

Reed-Solomon decoder

Syntax

decoded = rsdec(code,n,k)
decoded = rsdec(code,n,k,genpoly)
decoded = rsdec(...,paritypos)
[decoded,cnumerr] = rsdec(...)
[decoded,cnumerr,ccode] = rsdec(...)

Description

decoded = rsdec(code,n,k) attempts to decode the received signal in code using an [n,k] Reed-Solomon decoding process with the narrow-sense generator polynomial. code is a Galois array of symbols having m bits each. Each n-element row of code represents a corrupted systematic codeword, where the parity symbols are at the end and the leftmost symbol is the most significant symbol. n is at most 2^m-1. If n is not exactly 2^m-1, then rsdec assumes that code is a corrupted version of a shortened code.

In the Galois array decoded, each row represents the attempt at decoding the corresponding row in code. A decoding failure occurs if a row of code contains more than (n-k)/2 errors. In this case, rsdec forms the corresponding row of decoded by merely removing n-k symbols from the end of the row of code.

decoded = rsdec(code,n,k,genpoly) is the same as the syntax above, except that a nonempty value of genpoly specifies the generator polynomial for the code. In this case, genpoly is a Galois row vector that lists the coefficients, in order of descending powers, of the generator polynomial. The generator polynomial must have degree n-k. To use the default narrow-sense generator polynomial, set genpoly to [].

decoded = rsdec(...,paritypos) specifies whether rsdec appends or prepends the parity symbols to the input message to form decoded. The string paritypos can be either 'end' or 'beginning'. The default is 'end'. If paritypos is 'beginning', then a decoding failure causes rsdec to remove n-k symbols from the beginning rather than the end of the row.

[decoded,cnumerr] = rsdec(...) returns a column vector cnumerr, each element of which is the number of corrected errors in the corresponding row of code. A value of -1 in cnumerr indicates a decoding failure in that row in code.

[decoded,cnumerr,ccode] = rsdec(...) returns ccode, the corrected version of code. The Galois array ccode has the same format as code. If a decoding failure occurs in a certain row of code, then the corresponding row in ccode contains that row unchanged.

Examples

The example below encodes three message words using a (7,3) Reed-Solomon encoder. It then corrupts the code by introducing one error in the first code word, two errors in the second code word, and three errors in the third code word. Then rsdec tries to decode the corrupted code.

m = 3; % Number of bits per symbol
n = 2^m-1; k = 3; % Word lengths for code
msg = gf([2 7 3; 4 0 6; 5 1 1],m); % Three rows of m-bit symbols
code = rsenc(msg,n,k);
errors = gf([2 0 0 0 0 0 0; 3 4 0 0 0 0 0; 5 6 7 0 0 0 0],m);
noisycode = code + errors;
[dec,cnumerr] = rsdec(noisycode,n,k)

dec = GF(2^3) array. Primitive polynomial = D^3+D+1 (11 decimal)
 
Array elements = 
 
     2     7     3
     4     0     6
     4     0     0


cnumerr =

     1
     2
    -1

The output shows that rsdec successfully corrects the errors in the first two code words and recovers the first two original message words. However, a (7,3) Reed-Solomon code can correct at most two errors in each word, so rsdec cannot recover the third message word. The elements of the vector cnumerr indicate the number of corrected errors in the first two words and also indicate the decoding failure in the third word.

For additional examples, see Creating and Decoding Reed-Solomon Codes.

Algorithm

rsdec uses the Berlekamp-Massey decoding algorithm. For information about this algorithm, see the works listed in References below.

Limitations

n and k must differ by an even integer. The maximum allowable value of n is 65535.

See Also

rsenc, gf, rsgenpoly

References

[1] Wicker, Stephen B., Error Control Systems for Digital Communication and Storage, Upper Saddle River, N.J., Prentice Hall, 1995.

[2] Berlekamp, Elwyn R., Algebraic Coding Theory, New York, McGraw-Hill, 1968.

rcosine rsdecof