Frequently Asked Questions About Compact Discs

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

What do the tests referenced in CD standards really mean?

CD standards exist for the purpose of assuring interchange. This requires that every disc is readable in all drives, except those that are defective. Many specifications are physical, such as outer diameter, inner diameter, thickness, weight, unbalance, eccentricity, deviation, and deflection. Other requirements are optical such as index of refraction and birefringence. Measurements of these are either manual, or require expensive dedicated equipment, and are not discussed further in this FAQ.

Affordable, computerized equipment is available that evaluates many electrical requirments of CD standards by evaluating intensity variations of the laser beam that is reflected from the information layer of all CD discs. One beam that carries analog information is split both optically and electrically into various paths in the read drive. Standards assure interchange by specifying performance in each path.

The first path is total reflected beam intensity. Standards require minimum reflectance from a mirror region of the disc to assure sufficient signal strength for all paths. CD-R and CD-RW discs have an additional requirement, Rtop, that measures on-track total beam intensity. Rtop is lower than reflectivity because of diffraction loss from the pre-groove. Cross talk measures the ratio of off-track to on-track beam intensity, ensuring that signals from adjacent tracks do not interfere with data from the desired track.

The second drive path is radial tracking. This servo loop uses a radial error signal to center the focussed laser beam onto the track. Radial tracking, or push-pull, evaluates the sensitivity of the error signal to radial position, and must fall between upper and lower limits to assure proper servo loop operation. Radial eccentricity evaluates the radial runout of the tracks, assuring that variations remain within the range of the servo loop. Radial noise measures variations in radial track location that are at frequencies higher than the upper cutoff of the radial servo. Radial acceleration limits sudden radial track jumps to values within the capabilities of the electro-mechanical servo.

The third drive path is read data. Peak-to-peak signal strengths must be within limits at both the lowest, I11, and highest, I3, data frequencies. This analog information is then converted into binary data utilizing intervals between times when the analog signal crosses a d.c. decision level. Asymmetry assures that this decision level is within acceptable bounds. Effect length, or length deviation, verifies that averages of eighteen different time intervals are within tolerance. Jitter confirms that random variations of each time interval are not excessive. Out-of-tolerance read data usually results in read errors.

Very high data densities, noise, and physical defects can generate hundreds-of-thousands of read errors from even the best discs. Such errors are routinely detected and corrected using a Cross-Interleave Reed Solomon code (CIRC.) Before recording, data is organized into frames, each containing sync, subcode, 24 data bytes, and eight parity bytes. Four Q parity bytes are appended at the C2 level, frames are interleaved, then four P parity bytes are appended at the C1 level, and frames are again interleaved, but with a different pattern. C1 and C2 can each detect and correct two erroneous bytes in one frame. Various tests confirm that all errors are readily correctable.

De-interleaved read data is first sent to the C1 decoder for error detection and correction. BLER measures the rate (frames per second) of frames arriving with one or more errors. E11 represents the rate of correctable C1 frames having exactly one erroneous byte, E21 is the rate for frames with two bad bytes, and E31 is the rate for uncorrectable C1 frames having three or more errors. Good discs would have moderate E11 rates, low E21 rates, and very low E31 error rates.

Frames leaving the C1 decoder are de-interleaved again, distributing concentrated errors over many other error-free frames, and then go to the C2 decoder. E21 is the rate for correctable C2 frames having exactly one erroneous byte, while E22 and E32 are rates of uncorrectable frames having two and three-or-more errors respectively. E22 and E32 rates must be zero.

Burst is a different error test for scratches or other tangential defects. Standards require that no more than six successive C1 frames can each contain two or more erroneous bytes.

Forgiving read drives may tolerate an out-of-spec disc, but other drives will not. Discs that fail radial parameter tests can have acceptable error rates in one drive, but unacceptable rates in another because of servo loop differences. Drives may or may not be able to detect and correct E22 and E32 errors, but standards forbid them. Some players will ignore high radial acceleration while others will not. Every one of the above requirements must be confirmed by comprehensive testing in order to achieve confidence in predictable interchange.

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