State-split based encoder and decoder with table compression
A data encoder includes an encoding logic circuit configured to encode input data to apply run length limiting using state splitting of a code trellis, and a lookup table configured to apply a block-enumerable map on an un-split version of the code trellis, wherein entries in the lookup table are compressed.
1. A data encoder comprising:
an encoding logic circuit configured to encode input data to apply run length limiting using state splitting of a code trellis; and
a lookup table configured to apply a block-enumerable map on an un-split version of the code trellis, wherein entries in the lookup table are compressed.
2. The data encoder of claim 1 , wherein the encoding logic circuit is configured to implement an E6SR cost-constrained run length limited code.
3. The data encoder of claim 1 , wherein the data encoder comprises a 96/97 rate.
4. The data encoder of claim 1 , wherein the encoding logic circuit is operable to encode the data based on a finite state transition diagram having a cost assigned to each transition to apply a constraint such that an average cost is less than a given value.
5. The data encoder of claim 1 , wherein the entries in the lookup table are compressed by a linear method.
6. The data encoder of claim 5 , wherein the entries in the lookup table are compressed using a first-order column difference to store column differences in place of original entries.
7. The data encoder of claim 5 , wherein each column in the lookup table approximates a skewed Gaussian curve and the entries in the lookup table are compressed using column averaging.
8. The data encoder of claim 5 , wherein each column in the lookup table are compressed using row and column differences.
9. The data encoder of claim 8 , further comprising a pair of adders operable to recover original entries for the lookup table based on the row and column differences.
10. The data encoder of claim 1 , wherein the entries in the lookup table are compressed by a polynomial method.
11. The data encoder of claim 10 , wherein each column in the lookup table approximates a skewed Gaussian curve approximated by a polynomial, and wherein the entries in the lookup table are compressed by storing differences between actual column values and a polynomial fit for the actual column values.
12. The data encoder of claim 10 , wherein each column in the lookup table approximates a skewed Gaussian curve approximated by a polynomial, and wherein a plurality of sections of each column are approximated by a different polynomial, and wherein the entries in the lookup table are compressed by storing differences between actual column values and the polynomials.
13. The data encoder of claim 10 , wherein each column in the lookup table approximates a skewed Gaussian curve approximated by polynomials of logarithms of the column values, and wherein the entries in the lookup table are compressed by storing differences between actual column values and the polynomials.
14. A method of encoding data, comprising:
receiving input data to be encoded;
performing state split-based encoding of the input data in an encoder circuit including retrieving compressed entries from a lookup table used in the state split-based encoding and reconstructing uncompressed entries based on the compressed entries to yield encoded data; and
outputting the encoded data from the encoder circuit.
15. The method of claim 14 , wherein the compressed entries in the lookup table are compressed with a linear method.
16. The method of claim 15 , wherein the compressed entries are compressed at least in part by calculating differences between columns of the uncompressed entries.
17. The method of claim 14 , wherein the compressed entries in the lookup table are compressed with a polynomial method.
18. The method of claim 17 , wherein at least some of the compressed entries are stored in the lookup table as differences between the uncompressed entries and polynomials.
19. The method of claim 18 , wherein at least some of the compressed entries are based on logarithms of the uncompressed entries.
20. A data processing system, comprising:
encoding means comprising:
an encoding logic circuit configured to encode input data to apply run length limiting using state splitting of a code trellis; and
a lookup table with compressed entries, configured to apply a block-enumerable map on an un-split version of the code trellis; and
decoding means comprising a second lookup table with compressed entries.