Frequently Asked Questions About Compact Discs

Please contact Media Sciences if your questions are not answered on these pages.

Why do you insist that CD E22 and E32 errors must be zero? I cannot find this requirement in any of the standards. Only BLER and BURST are specified, and seem to be adequate quality indicators.

Both practical and technical justifications exist for requiring zero E22 and E32 error rates. Error rates are caused by two types of defects, "small" and "large". Noise or pit geometry variations create "small" defects that last for only a few microseconds. These are easily corrected at the C1 level, and usually determine BLER. Bubbles, black spots, and scratches are typical "large" physical defects that can have durations as long as a few milliseconds. They may be correctable at the C2 level, but severe defects generate many C2 uncorrectable frames that the third, sector level EDC/ECC will hopefully correct. These "large" defects generate E22, E32, and BURST errors, but may have only a small influence on BLER because of their very low density.

CD technology utilizes a cross-interleaved Reed-Solomon code (CIRC) for error detection and correction. Read data is de-interleaved and then analyzed by C1 CIRC circuitry that can locate and correct one or two error bytes in a 32 byte frame. If three or more error bytes are present, all bytes in the frame are flagged bad. Data is de-interleaved again and then analyzed by C2 CIRC circuitry that can correct up to four flagged bytes, or can locate and correct one or two error bytes. An E22 error indicates that the C2 circuit corrected two bytes in one frame. E32 indicates that three or more errors were present in the C2 frame, causing all bytes in the frame to be flagged bad, de-interleaved again, and sent to sector level RS-PC. A BURST error is defined as a region having seven or more contiguous C1 frames, each having two or more error bytes.

Practical justification for zero E22, E32, and BURST errors assumes that CD-ROM discs should not be marginal. Very "large" physical defects cause not only data errors but also the loss of all servo signals. Each servo must then reacquire lock when the read laser spot exits the defect region. Errors result both from the actual defect region as well as from the servo reacquisition region. Some drives quickly achieve servo lock while others require more time, because different drives can use dissimilar designs for radial, focus, and constant linear velocity servos. As a result, errors from a disc having a "large" defect might be correctable in one drive but not in another. Zero E22, E32, and BURST error requirements reject marginal discs and provide confidence that good discs will be readable in all drives.

Another practical justification arises from the deterioration of discs. New discs that pass BLER and BURST requirements but have E22 and E32 errors may initially be readable. These discs will soon become unusable when packaging, shipping, storage, and handling result in scratches, fingerprints, particulates, and other contamination of the readout surface. New discs that have zero E22 and E32 errors will have a much longer life under such conditions.

Technical justification is easily found in standards such as ISO/IEC 10149. Optical scanners are used by most manufacturers to 100% test all discs and to reject discs having "large" physical defects. Standards require air bubbles to be less than 0.1 mm in diameter, black spots surrounded by increased birefringence to be less than 0.2 mm, and black spots without birefringence to be less than 0.3 mm in diameter. Calculations show that such defects result in zero E22, E32, and BURST error rates that discs must satisfy if they are to meet the physical defect requirements of the standards.

Theory predicts that E22 errors occur for defects larger than 0.57 mm. In practice, loss of servo lock can generate E22 errors for 0.3 mm black spots. Theory also predicts that BURST errors result from defects larger than 0.95 mm. Servo limitations generate BURST errors from 0.5 mm black spots. E32 errors are predicted to occur when defect diameters exceed 1.3 mm, but E32's are actually observed from 0.8 mm black spots. The presence of E22, E32, or BURST errors indicate the presence of physical defects that violate the requirements of CD-DA, CD-ROM, and CD-R standards. Zero BURST alone is not sufficient. E22 is more sensitive, and certain unacceptable defects can generate E22 and E32 errors with no BURST errors.

A final technical consideration is the possibility of CIRC misdetects or miscorrects. Error detection and correction methods are not perfect. There is always a very small probability of a mistake. Data analysis routinely reveals C1 level misdetects or miscorrects that are subsequently detected and corrected at the C2 level. The presence of E22, E32, or BURST errors can saturate the capacity of C2 CIRC, leaving nothing in reserve for C1 misdetects or miscorrects.

The use of only BLER and BURST quality requirements does not assure high quality discs. Both technical and practical considerations support the requirement of zero E22, E32, and BURST errors. Such discs are readily available. Use them to realize predictable interchange and longer life of your CD-ROM and CD-R discs.

If it would help, Media Sciences will test one recorded sample at no charge. Please follow the free test instructions on our web site.