DVD BASICS

DFGDF


Figure 2: ECC block: Product Code RS(208,192,17)*RS(182,172,11)

ERROR CORRECTION (ECC BLOCK)
   Thanks to the development of new ECC decoding algorithms, DVD error correction coding, using a Reed-Solomon product code can be applied across a larger amount of data, with better correction capabili -ties. This reduces error correction redundancy to approximately 13%, or half of that of CD. After EDC calculation over the Data Sectors, scrambling data is added to the 2048 bytes of main data in the data sector. Then the error correction code is applied over 16 data sectors, or one ECC block.
   After putting the data bytes in a matrix of 192 rows * 172 columns, 16 bytes of PO-parity are added to each column. Next, 10 bytes of PI-parity are added to each of the 208 rows (192 data + 16 formed by the PO-bytes) to form a Reed-Solomon Product Code with 208 rows and 182 columns (172 data + 10 formed by the Pl-bytes).
   This code can correct at least 5 byte errors in each row and at least 8 byte errors in each column. By applying several alternating calculations over rows
and columns, much bigger error patterns can be solved.

Number are protected by an additional error detection code (IED). Furthermore the "header" contains copyright management information and region management information. The sync as used in CD-ROM is not needed, because DVD uses an advanced synchronisation method at the EFM level. The extra ECC layer as used in CD-ROM is superfluous, because the normal error correction incorporated in DVD is much more powerful than the CIRC in CD, and is able to give the high reliability needed for data applications. The EDC is maintained, because it gives a very easy and powerful error detection capability at sector level.

RECORDING SECTOR
   Finally the PO-rows are interleaved with the data rows in a regular order (12 data, 1 PO) and each interleaved ECC block is divided into 16 Recording Sectors. In this way each Recording Sector contains the original data from 1 Data Sector + 12 * 10 PI-bytes + 1 row of 182 PO-bytes, together 2366 bytes.
P
HYSICAL  SECTOR 
   A physical sector is formed by splitting each row of a recording sector into 2 parts of 91 bytes each. These bytes are converted to EFM+ and preceded by a special EFM+ sync pattern of 32 channel bits. By using alternating 8 


Figure 3 : Recording Sector

different sync patterns, decoders can easily synchronise (or resynchroniseafter clock loss) their timing. A physical sector has 38688 channel bits, which is equivalent to 2418 bytes before modulation. A Sync Frame has 1488 channel bits.
C
HANNEL MODULATION   
   DVD uses a modified version of the CD eight-to- fourteen channel modulation scheme (EFM), which is known as EFM+. This scheme converts 8-bit data directly into 16 channel bits.
   The 16-bit channel values are chosen from the EFM+ conversion table. There are 256 possible data symbol values and 4 possible run-length conditions, or States. For each data-symbol/state combination, the conversion table contains a 16-bit EFM+ codeword and the value of the run-length State for the next Data Symbol to be encoded. The coding not only allows unique mapping during demodulation, 


NOV - DEC   2002                                    OPTICAL DISC SYSTEMS                       Back to Content     Back to Magazine Cover