Philips CDD522 Operating Instructions Manual page 23

The two sets of Parity Bits are used for error correction in two
layers, C1 and C2. The Error Corrections act on the 32 data
bytes, NOT on the Control Byte. That has been removed (see
figure 9 p 21 in Chapter 3: The Subcode Structure).
The C1 Error Handling is used for recovery from random errors,
originating from noise on the data. It restores missing databits,
using redundant information in the P-Parity Bits.
é
The C2 Layer is used for burst errors, like scratches and dirt-
spots on the disk that can disrupt whole strings of data. It
depends on the data being scrambled over many frames at
recording time. When read out, the data is descrambled, thus
scattering the burst error. The now dispersed errors can be
corrected frame by frame using the information in the Q-Parity
Bits.
After all these actions (retrieval of the Control Byte, C1-decoding
using P-parity bytes, C2-decoding using Q-parity bytes) frames of
24 bytes are obtained.
These 24-bytes-frames are the basic Audio Frames, coming at
7350 Hz rate. Every frame contains 6 Left Audio samples and 6
Right Audio samples, each 16 bits long (6 x 7350 =44100 Hz, the
familiar value of the CDDA Sampling Rate)
98 of these Audio Frames form an Audio Sector, containing 2352
bytes of audio-data. The Audio Sectors come at 75 Hz rate.